2‘3V1:|03 \q HGI AN STATEU :1W29l3 00611 lfluz‘lflmm Mica-IliaQRafiksIOte I University This is to certify that the dissertation entitled ANALYSIS OF ADHESIVE, CELL SURFACE, AND COLONIZING PROPERTIES OF CANDIDA ALBICANS presented by MICHAEL JAMES KENNEDY i has been accepted towards fulfillment of the requirements for Ph.D. degree in Botany and Plant Pathology ofcsso Date ”/0/98 / / MS U is an Affirmative Action/Equal Opportunity Institution 0-12771 PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before due due. DATE DUE DATE DUE DATE DUE F _J _ __W__ MSU It An Afflrmdlve AetlorVEquel Opportunity Institution | I flu. ‘ .I ANALYSIS OF ADHESIVE, CELL SURFACE, AND COLONIZING PROPERTIES OF QAHDIDA ALEIQAHE BY MICHAEL JAMES KENNEDY A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Botany and Plant Pathology 1988 ABSTRACT ANALYSIS OF ADHESIVE, CELL SURFACE, AND COLONIZING PROPERTIES 01“ 953.2195 M BY MICHAEL JAMES KENNEDY The ability of Candida M to attach to host mucosal surfaces and colonize the gastrointestinal tract may play an important role in pathogenesis. The present studies examined the adhesive properties of Q. albi cans and the interaction between this fungus and the intestinal microflora. Adhesion of Q. AIM to buccal epithelial cells (BECs) and plastic was significantly modified by altering growth and assay parameters, particularly growth medium. Significant differences in adhesion to BECs and plastic were noted when Q. M was grown in the same complex medium from different manufacturers and in different lots of medium from the same manufacturer. Electron microscopy also revealed significant differences in surface topography and cell wall ultrastructure of Q. albiggng. At least three different adhesins were observed to exist on the surface of Q. 112192.125 that mediate its adhesion to BECs: (i) a cell wall adhesin, (ii) an adhesin located in an outer floccular layer, and (iii) an adhesin expressed below the surface that becomes evident when the outer cell wall layer is thinned. The ability of white and opaque phenotypes of Q. albicans to attach to BECs and plastic were also examined. "White cells" were found to be significantly more adhesive to BECs than "opaque cells". No difference in adhesion to plastic between the two phenotypes was noted, although opaque cells were twice as hydrophobic. The adhesion of Q. albigans to plastic was further analyzed by binding isotherms, Langmuir isotherms, and Scatchard plots. Collectively, the results suggest that there are at least two distinct mechanisms of adhesion of Q. albigans to BECs and plastic. The interaction of Q. albicans with intestinal bacteria was examined in continuous-flow culture models. The suppression of Q. albicans by intestinal bacteria was collectively regulated by the depletion of nutrients and anaerobiosis, in addition to the production of inhibitory substances by intestinal anaerobes (e.g., short-chain fatty acids). The intestinal microflora apparently controlled Q. albicans collectively by several mechanisms, but Candida cells could become tolerant to some of these mechanisms and survive at a low levels in the gut probably by associating with GI mucosal surfaces. DEDICATION To my loving wife and daughters Laura Ann Kennedy and Alyssa Lauren Kennedy and Andrea Michelle Kennedy whose love and companionship have blessed my soul. An additional expression of love is extended to my parents and grandparents for their constant love, support, guidance and prayer. "For God so loved the world, that He gave His only begotten Son, that whosoever believeth in Him should not perish, but have everlasting life." (John 3:16) iv ACKNOWLEDGEMENTS I would like to express my deepest gratitude to my advisor, Dr. Alvin L. Rogers, for his friendship, support, and guidance during the years I have studied and trained in his laboratory. He has been an excellent mentor and has provoked me to expand my knowledge of medical mycology and mycology. It has been a privilege to have trained in his laboratory, learned mycology under his direction, and have him as a friend. I also respect the professionalism by which he has conducted his courses and laboratory. I would also like to thank Drs. Everett S. Beneke, Karen L. Klomparens, Ronald J. Patterson, and Paul A. Volz for serving on my guidance committee. I thank all of the above professors for their example of sound scientific concepts and principles, and for many helpful comments and stimulating discussions. I especially thank Dr. Paul A. Volz for stimulating my early interest in medical mycology, and for his friendship over the past ten years. My special thanks to Dr. Robert J. Yancey, Jr. for his encouragement and endorsement of the research described herein, and for his helpful ideas, comments, criticisms, and stimulating discussions. I also thank him for his friendship and encouragement. I thank my wife, Laura Ann Kennedy, for her love and support, her help in proofreading the manuscript and checking that all the references were in order. I also thank my wife and children for their patience as I was away from home, and especially for allowing me to postpone house duties during the final few months of my studies and research. Most of all I thank my wife for staying up with me when I worked through the night, and for sharing just the right verses of scripture (Isaiah 40:28-31) when I was very discouraged. Finally, I would like to thank all of my family for their constant love, encouragement, support, and prayers throughout all of my years of undergraduate and graduate study. I thank my mother for endeavoring to teach me professional qualities and for a godly heritage in Christ. This work was supported in full by The Upjohn Company. vi TABLE OF CONTENTS List of tables . . . . . . . . . . . . . . . . . . . xiii List of figures . . . . . . . . . . . . . . . . . . . xvi Introduction . . . . . . . . . . . . . . . . . . . . . . 1 Review of the literature . . . . . . . . . . . . . . . . 3 Qandidg and candidal infections . . . . . . . . . . 3 "Thrush" . . . . . . . . . . . . . . . . . . . 9 Gastrointestinal candidosis . . . . . . . . . 12 Systemic candidosis . . . . . . . . . . . . . 14 Candida albicans cell wall composition, ultrastructure and possible adhesins . . . . 15 Cell wall composition . . . . . . . . . . . . 16 Cell wall ultrastructure . . . . . . . . . . 23 Adhesins . . . . . . . . . . . . . . . . . . 24 Adhesion and association mechanisms . . . . . . . 34 Physicochemical considerations of adhesion . . . . . . . . . . . . . . 34 Mucosal surfaces . . . . . . . . . . . . . . 44 Nonbiologic surfaces . . . . . . . . . . . . 57 Microbial ecology of gastrointestinal ecosystems . . . . . . . . . . . . . . . . . 63 Importance of gastrointestinal colonization by andigg alpiggng . . . . . . 70 vii Article I: Influence of growth conditions on Qangiga albicans adhesion, hydrophobicity and cell wall ultrastructure . . . . . . . . . . . . . . . . . . 82 Abstract . . . . . . . . . . . . . . . . . . . . . 83 Introduction . . . . . . . . . . . . . . . . . . . 84 Methods . . . . . . . . . . . . . . . . . . . . . 86 Mucosal cells . . . . . . . . . . . . . . . . 86 Yeasts, culture conditions, and cell preparation . . . . . . . . . . . . 86 Adhesion assay . . . . . . . . . . . . . . . 88 Phase-partition test for hydrophobicity . . . 88 Electron microscopy . . . . . . . . . . . . . 89 Results . . . . . . . . . . . . . . . . . . . . . 91 Adhesion of Q. albigans to BECs after growth on different media . . . . . . . 91 Effect of growth temperature on Q. albicans adhesion . . . . . . . . . . . . . . . . 92 Effect of growth parameters on Q. alpigans cell surface hydrophobicity . . . . . . 93 Cell wall ultrastructure and surface topography . . . . . . . . . . . . . . . 94 Discussion . . . . . . . . . . . . . . . . . . . . 96 References . . . . . . . . . . . . . . . . . . . 102 Tables . . . . . . . . . . . . . . . . . . . . . 110 Figures 0 O O O O O O O O O O O O O O O O O O 0 O 114 viii Article II: Variation in adhesion and cell surface hydrophobicity in Qangiga albiggng white and opaque phenotypes . . . . . . . . . . . . . . . . 116 Abstract . . . . . . . . . . . . . . . . . . . . 117 Introduction . . . . . . . . . . . . . . . . . . 119 Methods . . . . . . . . . . . . . . . . . . . . . 120 Yeasts, culture conditions, and cell preparation . . . . . . . . . . . . . . 120 Mucosal cells . . . . . . . . . . . . . . . 122 Adhesion assays . . . . . . . . . . . . . . 122 Phase-partition test for hydrophobicity . . 124 Results . . . . . . . . . . . . . . . . . . . . . 125 Adhesion of Q. albicans phenotypes to BECS . . . . . . . . . . . . . . . . . 125 Adhesion of Q. albiggng phenotypes to plastic . . . . . . . . . . . . . . . . 126 Effect of phenotypic state on Q. albigang cell surface hydrophobicity . . . . . . 127 Discussion . . . . . . . . . . . . . . . . . . . 128 References . . . . . . . . . . . . . . . . . . . 134 Tables . . . . . . . . . . . . . . . . . . . . . 142 Article III: Environmental alteration and evidence for two distinct mechanisms of Qandiga albigans adhesion to plastic . . . . . . . . . . . . . . . 144 Abstract I O O O O O O O O O O O O O O O I O O O 145 ix Introduction . . . . Materials and methods Fungi . . . . . Growth conditions Cell preparation Adhesion assay . Phase-partition test hydrophobicity Results . . . . . . . cell surface Influence of growth medium on adsorption Influence of assay medium on adsorption Influence of ionic strength and ions on adsorption . Influence of detergents on adsorption Adsorption and cell surface hydrophobicity of white and opaque phenotypes Discussion . . . . . References . . . . . Tables . . . . . . . Figures . . . . . . . Article IV: An anaerobic continuous-flow culture model of interactions between intestinal microflora and NEW-- Summary . . . . . . . Introduction . . . . 147 148 148 149 149 150 151 151 151 153 154 155 155 156 161 165 168 174 175 177 HethOds O O O O O O O O O O O O O O O O Microorganisms . . . . . . . . . . Animals 0 O O O O O O O O O O 0 O Continuous-flow (CF) cultures . . Animal preparation, inoculation, and recolonization experiments . Enumeration of Q. albicans . . . . Volatile fatty acid (VFA) analyses Results . . . . . . . . . . . . . . . . Discussion . . . . . . . . . . . . . . References . . . . . . . . . . . . . . Tables . . . . . . . . . . . . . . . . Figures 0 O O O O O O O I O O I O C O 0 Article V: Mechanisms that control Qandida albigans populations in continuous-flow culture models of intestinal microflora and in the large intestine . . . . . . . . . . . . . . . Abstract . . . . . . . . . . . . . . . Introduction . . . . . . . . . . . . . Materials and methods . . . . . . . . . Microorganisms . . . . . . . . . . Continuous-flow (CF) cultures . . Growth studies . . . . . . . . . . Inhibitory activity of short-chain fatty acids (SCFA) . . . . . xi 179 179 179 180 182 183 185 185 189 193 201 203 206 207 209 211 211 212 212 213 Colonization of conventional and antibiotic-treated mice . . . . . . . . 214 Results . . . . . . . . . . . . . . . . . . . . . 215 Survival of Qandiga nlbicans in CF cultures . . . . . . . . . . . . . . 215 Inhibitory mechanisms . . . . . . . . . . . 217 Morphology of Q. albicang in CF cultures . . 219 Role of adhesion in colonization . . . . . . 219 Discussion . . . . . . . . . . . . . . . . . . . 221 References . . . . . . . . . . . . . . . . . . . 228 Tables . . . . . . . . . . . . . . . . . . . . . 234 Figures . . . . . . . . . . . . . . . . . . . . . 235 Discussion . . . . . . . . . . . . . . . . . . . . . . 238 Adhesion and association mechanisms of Candida albicans . . . . . . . . . . . . 238 Gastrointestinal colonization by Cnndida albicans . . . . . . . . . . . . . . 260 Literature cited . . . . . . . . . . . . . . . . . . . 276 xii LIST OF TABLES Table 1 Possible cell surface components suggested serve as Qandida adhesins . . . . . . . . . . . . 29 2 Adhesion and association mechanisms of Candidn nlbicans . . . . . . . . . . . . . . . . . 35 3 Major bacterial groups present in human fecal flora . . . . . . . . . . . . . . . . . . . 67 4 General classification of factors thought to predispose to colonization and dissemination from the gastrointestinal tract by angign . . . . 75 5 Isolation of Candida albicans and other yeasts in stool samples from healthy individuals on different diets . . . . . . . . . . . . . . . . . 76 Article I 1 Growth conditions used for studying Q. nlnicans adhesion, cell surface hydrophobicity, cell wall ultrastructure and various surface features . . . 110 2 Adhesive properties of Q. nlnignng grown on different media and under different environmental conditions . . . . . . . . . . . . . . . . . . . 111 3 Hydrophobic properties of Q. nlnignng grown on different media and under different environmental conditions . . . . . . . . . . . . . . . . . . . 112 4 Morphology and surface characteristics of Q. nlnignng grown on different media and under different environmental conditions Article II 1 Adhesion of Candign nlbicnns white and opaque phenotypes to buccal epithelial cells (BECs) Adhesion of angign nlnignng white and opaque phenotypes to plastic . . . . . Hydrophobic properties of Qandidn albignns white and opaque phenotypes . . . . . Article III 1 Comparison of binding affinities and adsorption 2 Effect of assay medium on the adsorption of Candidn albicnns to plastic . . . . . . . . . . 3 Adhesion of Qandidn nlnignng phenotypes to plastic . . . . . . . . . . . . . . . . . . . . Article IV 1 Population levels of intestinal bacteria in the ceca of test animals, human feces, and anaerobic continuous-flow (CF) cultures . . . . . . . . . 2 Concentrations of volatile fatty acids in test sites for angign albicans after growth in different media or with different assay media . animals, human feces, and in anaerobic continuous-flow (CF) cultures . xiv 113 142 142 143 165 166 167 201 201 3 Cecal population levels of Qandiga nlbignng in reconventionalized and control animals . . . . 4 Cecal characteristics of antibiotic-treated, control, and reconventionalized animals . . . . . Article V 1 Minimal inhibitory concentrations for short-chain fatty acids against angign nlnignng . . . . . . Table 6 Factors affecting fungal aggregation and adhesion to biological and nonbiological surfaces in yinnn . . . . . . . . . . . . . . . . 7 Factors influencing fungal aggregation and adhesion in 2119 . . . . . . . . . . . . . . . . 8 Experimental conditions for the beginnings of a standardized buccal epithelial cell adhesion assay for angign nlnignns . . . . . . . 9 Mechanisms by which the intestinal microflora inhibits Qandiga nlbicnns colonization and dissemination from the gastrointestinal tract . . 202 202 234 245 246 258 266 LIST OF FIGURES Article I 1 Electron micrographs of thin sections of Q. nlbigans AK785 stained by a modification of the "G-O-T-O-U" method. Cells harvested from (a) TSB+G25, showing an outer electron dense layer (arrow), (b) YNB+gal, showing an outer floccular layer (arrow), (c) TSB+G-S, showing an outer floccular layer (arrow), and (d) TSB-Z, showing a thin cell wall with no "extra" cell wall layers or obvious adhesive appendages . . . . . . . . . . . . . . . . . TEM of thin section of Q. nlnignng AK785 showing a tiny germtube-like structure . . . Article III 1 2 Adhesion assay . . . . . . . . . . . . . . . Adsorption of anQi n nlnignng to plastic after growth in different media . . . . . . . Scatchard analysis of angign nlnignng adsorption to plastic after growth in different media . . . . . . . . . . . . . . . Effect of KCl concentration on adsorption of angign nlnignng to plastic . . . . . . . xvi 114 115 168 169 170 171 5 Effect of Mg+2 concentration on adsorption of angign nlnicnns to plastic . . . . . . . 6 Influence of detergents on adsorption of Qandidn albicans to plastic . . . . . . . Article IV 1 The continuous-flow culture apparatus . . . . 2 A continuous-flow culture growth tube showing dense layers of bacterial growth formed on the wall that resemble bacterial populations that colonize intestinal mucosa . . . . . . . 3 Elimination of Q. albicans from continuous- flow culture of mouse cecal flora and mouse cecum. The values represent the means from 4 to 10 samples per data point . . . . . . . Article V 1 Elimination of Candida albicans from continuous—flow cultures of mouse cecal bacteria (a) and human fecal flora (b) . . . 2 Elimination of anginn albicans from an established ( ), newly colonized ( ), and sterile ( ) continuous-flow culture . . . . . 3 Growth of angign nlnignns in filtrates of continuous-flow cultures of mouse cecal bacteria (a) and human fecal flora (b) . . . xvii 172 173 203 204 205 235 236 237 INTRODUCTION Qandidn film is commonly found residing in the oral cavity, gastrointestinal (GI) tract, and female genital tract of humans (Rippon, 1988). The yeast is usually commensal, but in predisposed persons clinical infection can occur as a result of its opportunistic nature. In patients who are compromised immunologically and undergoing prolonged antimicrobic therapy, Q. nlnignng can reach high numbers in the GI tract and, subsequently, pass through the intestinal mucosa to initiate systemic infection by the hematogenous route (Myerowitz e1; 11. , 1977), particularly in patients with acute leukemia (Meunier-Carpenter e; Q1., 1981; Myerowitz 93; n1“, 1977; Ray, 1980). Similarly, colonization of the GI tract by Q. nlnignng may also lead to involvement in a number of other disease syndromes (Odds, 1988: Rippon, 1988). Although the determinants of colonization and dissemination of Q. M from the oral cavity, esophagus, and GI tract have not been completely defined, it is likely that adhesion and association of Qandjgg cells with host mucosal surfaces and prosthetic devices plays an important role (Kennedy, 1988; Rotrosen gt g1. , 1986). For instance, it is unlikely that colonization of various sites of the alimentary tract, such as the small intestine, could take place in the face of the rapid 1 2 passage of material and fluid through the tract due to peristalsis, unless these organisms could associate with the mucosa. Moreover, prevention of angign, mucosal association by bacterial antagonism has been shown to reduce oral and GI colonization and dissemination by this fungus (Kennedy and Volz, 1985a: Liljemark & Gibbons, 1973). Considering further the sloughing of epithelial cells and the bathing action of fluids over host mucosal surfaces and various tissues, and the fact that angign lesions may arise in virtually any tissue of the body, suggests that adhesion is also an important determinant of thrush and denture stomatitis, vaginitis, endocarditis, mucocutaneous candidosis, and renal Qangida infections. Because of the importance of QM adhesion and colonization to human health, then, there can be no reasonable doubt that it would be of great benefit to clinical medicine if adhesion and colonization by Q. glnignng could be regulated in such a way that the beneficial effects are retained, and the detrimental effects are eliminated. This will only be possible, however, when the adhesive properties of this organism are characterized at the molecular level and when the ecologic mechanisms that control Q. minnns in the gut are more thoroughly understood. To that end, the research described herein was undertaken to analyze adhesive properties of Q. 21215315 and to examine interactions between this fungus and the intestinal microflora. REVIEW OF THE LITERATURE a and'da i c 'o The genus Qandign is comprised of a heterogeneous collection of asporogenous yeasts that reproduce by budding and belong to the form-class Deuteromycetes and the form-family Cryptococcaceae (Rippon, 1988). Some members of this genus have sexual stages belonging to the genus mm in the Ascomycetes (Kreger-van Rij , 1984). There is also some evidence that Q. nlniggng has characteristics that are common to both Basidiomycetes and Ascomycetes, but classification in either of these ”perfect" classes will have to await the isolation of the teleomorph state (if there is one) of the organism. angign albicans is similar to other members of the genus in that nearly all species have a predominant morphologic form that is unicellular, with budding as the predominant form of reproduction. However, only Q. Mg produces germtubes when grown in serum at 37°C for one or two hours (Cooper & Silva-Hutner, 1985) and is the predominant species to produce chlamydospores when grown on chlamydospore agar or corn meal agar to which Tween-80 has been added (Beneke & Rogers, 1980: Kreger-van Rij, 1984). Assimilation and fermentation patterns for carbohydrates, biotype testing, and molecular and serological techniques have also been applied to the 3 taxonomy of andida species (Kreger-van Rij, 1984; Odds, 1985). Although some of these tests are still far from routine clinical use for yeast identification, they have provided a great deal of information regarding the antigenic similarities and relatedness of various angigg species. Jonsen (1955), for instance, first showed antigenic cross-reactivity among angigg species due to similarities in surface polysaccharides a little over thirty years ago. Studies .by others Ihave since confirmed this and. have classified QM species into six groups according to their complements of heat stable and heat sensitive antigenic components (Murray & Buckley, 1966: Segal gt g1., 1975: Sweet & Kaufman, 1970; Tsuchiya gt 91., 1961). The existence of several common antigenic components has likewise been observed with fluorescent antibody and agglutination reactions to both soluble and insoluble antigenic extracts (Dastidar, 1968: Geck & Novak, 1967: Gordon gt g1., 1967: Hasenclever, 1961; Kemp & Solotorvsky, 1964). Furthermore, all of the so-called pathogenic angigg species possess at least one common antigen, and most posses two (Tsuchiya gt al., 1961). It is apparent, then, that there are considerable similarities between antigens from the major opportunistic angigg pathogens. In fact, several investigators have shown that 9- means. s:- W and 9- mm possess as many as seven. common antigens, whereas (Q. 5 . lsnilliermondi. Q! krnsei. Q. untangilggig and Q. pggndgttgpigglig appear to be antigenically distinct (Tsuchiya gt g;., 1961). The existence of two antigenic types among strains of Q. Mpg, moreover, has been demonstrated by several investigators (Hasenclever, 1965: Hasenclever 8 Mitchell, 1961 a8b: Hasenclever gt n1., 1961: Summers gt g1., 1964). One antigenic type, serotype A, possess the same number of antigens as the other type, serotype B, plus at least two additional antigens (Hasenclever 8 Mitchell, 1961 a8b; Hasenclever gt nl., 1961). Antigenic similarities between type A Q. albicgns and Q. Mg, and type B Q. glbiggng and Q. gtgllgtgiggn_ have also been observed (Hasenclever 8 Mitchell, 1961a8b: Hasenclever (gt ‘11., 1961: Murray 8 Buckley, 1966: Stewart-Tull g 11., 1966: Sweet 8 Kaufman, 1970). Two-dimensional immunoelectrophoresis of soluble cytoplasmic extracts from Q. glbiggng and Q. ttopigglis, for instance, indicate that these two strains may share as much as 70% of common antigens (Odds, 1988). A third antigenic type of .Q. m, namely serotype C, has also been proposed by Muller and Kirchoff (1969) , however, there have been no other studies to date to either confirm or reject such a group for Q. nlniggng (Odds, 1988). One common antigenic factor for angign species, and other yeast (e-q-. fiassharsnzsss ssrsxisiae) appears to be cell wall polysaccharide, which is often found in cross- 6 reacting antigenic extracts (Hasenclever, 1965; Kemp 8 Solotorvsky, 1964: Stewart-Tull, 1966). As noted by Odds (1988) , there is good evidence to suggest that antigenic specificity in mannans and glucans (the primary make-up of yeast cell walls: see below) from QM species is predominately due to the nature of glycosidic linkages in side branches from the main polysaccharide chains (Suzuki 8 Sunayama, 1969: Yu gt nl., 1967 a8b), and that mannan structure is primarily responsible for the distinction between type A Q. aLb;cans/Q. ttopicaus and type B Q. gltiggng/Q. stellatoidea (Kolarova, gt gt” 1973; Odds, 1988) . Proton magnetic resonance (PMR) , furthermore, showed similar spectra for mannans isolated from serotype A 8 B Q. glbicans, Q. tzgnigglig, and Q. gtgllgtgifigg, while PMR spectra were quite distinct for mannans from Q. W911. 9.. 13131191. 9.. We as s and S;- pggnggttgpigglig (Spencer 8 Gorin, 1969). Electrophoresis of soluble extracts from broken andigg cells showed that there were at least 46 identifiable proteins from the extracts, and that there were characteristic similarities and differences between the protein patterns for several angigg species (Shecter gt g1., 1972). Such experiments, have confirmed the close similarities between Q. m and Q. gtgllgtgiggg, and that strains of Q. tunigalig appeared to exhibit extreme dissimilarities between this species and other angm species (Odds, 1988; Shecter gt g1“, 1972). Likewise, 7 others have also reported that Q. glniggng and Q. ttgpigglig showed quantitative differences with respect to phosphatase enzymes (Odds 8 Gonzales, 1974; Saugar gt n1.,1975) and several oxidases (Odds, 1988), but that both species uniquely produced an intracellular transglucosyl amylase (Sawai 8 Hehre, 1963). Measurements of the quanine plus cytosine (G + C) content of Qangigg DNA showed a very wide range of base compositions among members of the Qandida genus (from 35 to 50% G + C), but that the DNA base compositions of Q. glntggng, Q. ttgpigg11g_ and Q. gtgtlgtgiggg were very similar (Stenderup 8 Bak, 1968). In contrast, such similarities were not seen in DNA-DNA hybridization studies between Q. gltiggng and Q. tngpjgglig which showed a relative homology of only 8%, whereas Q. Mgng and Q. lgtgllntgiggn, exhibited. a Ihomology of over 80% (Bak 8 Stenderup, 1969). Finally, Q. gnLLLuamnnEUJ” Q. pgxgngilggig and Q. pseudotnopicglig have been shown to be capable of photoreactivation after ultraviolet irradiation, while strains of Q. glniggng, Q. ktnggi, Q. W and Q. trgniggljg were not (Miller 8 Sarachek, 1974). It appears, therefore, that although a wide variety of taxonomic studies have shown that there are a number of similarities between Q. glflgang, Q. gtgfltgidgg and Q. tzgpigglig, there is no tendency for the species of Qandidg that are pathogenic to form a taxonomic cluster 8 that would serve to distinguish them from the so-called nonpathogenic Qandtdg. One must assume, then, that angign species frequently found to cause disease in debilitated patients, and Q. glnignng in particular, possess one or more virulence mechanisms that render them pathogenic to their mammalian hosts. For instance, Q. nltjggng can produce a number of nonspecific hydrolytic enzymes, which are probably responsible for damage to host tissues in cases of candidosis. Kapica and Blank (1957) reported that Q. alntcgns could produce keratinase, an enzyme that allows fungi to attack keratinized tissues. The production of several peptidases, proteases and phospholipases have also been demonstrated and studied in detail (Odds, 1988). MacDonald and Odds (1983), moreover, recently showed a direct relationship between virulence for mice and protease production in Q. nlniggng. A correlation relationship between phospholipase and proteinase activity and the adhesion of Q gltiggng and other angigg species has also recently been made (Barrett-Bee gt g1., 1985: Ghannoum 8 Elteen, 1986). anQng infections have been reported in virtually every tissue of the human body (Odds, 1988: Rippon,1988). Different forms of candidosis, therefore, have been referred to in the past generally as either "superficial" or ”systemic" candidosis. This is primarily due to the fact that the oral cavity and GI tract have been considered to be continuations of the body surface. 9 However, since there are important differences in cutaneous and mucocutaneous involvement, both histologically and pathologically, infections of the alimentary tract will not be considered as superficial in this discussion. The various clinical forms of candidosis, then, may be listed according to the primary tissue system involved: cutaneous, mucocutaneous, and systemic (Rippon, 1988). The normal adult usually has a high resistance to infections by Q. glniggng (Rogers 8 Balish, 1980). angigg infections are, therefore, uncommon unless there are alterations in host defenses or predisposing environmental conditions, even though Q. glniggng can live as a commensal on several body surfaces (e.g. the oral cavity, GI tract and vagina). Factors which can predispose an individual to candidosis appear to be many and varied, an can be placed into four general groups: natural (e.ga, digressions from normal hormonal or physiologic status), dietary, mechanical (e.g., bowel resection), and iatrogenic medical factors (e.g., drug treatments) (Odds, 1988). From this list, as well as numerous publications on experimental animal infections and clinical observations, it appears that there are two major deterrents to infection by Q. gltigang that are disrupted by such factors: (i) suppression of 9mm; colonization by the indigenous host microflora and (ii) an intact immune system (Odds, 1988). 10 W. Oropharyngeal and esophageal thrush has been known since at least the time of Hippocrates, even though Q. altiggng was not discovered until about 1841 (Odds, 1988: Rippon, 1988). The disease, which is the most common form of candidosis, is most prevalent among infants, but may be seen in patients of all ages. Recently, chronic oropharyngeal and esophageal candidosis has been noted as a first sign of clinical AIDs (Rippon, 1988). Thrush is characterized by the presence of white patches that appear as discrete lesions on the surfaces of the buccal mucosa, throat, tongue, and gum linings. The lesions develop to become confluent pseudomembranes that resemble milk curds, which can be scraped off with difficulty to reveal a raw, erythematous base (Odds, 1988). The infection is not usually painful, although in severe cases erosion and ulceration of the mucosa may be seen (Jenkins, 1973: Ludlam 8 Henderson, 1942: Marples, 1968: Winner 8 Hurley, 1964). While oral thrush is the best-known and most common form of oral candidosis, Q. 8.12% also has an important role in the etiology of denture stomatitis, angular cheilitis, leukoplakia, and possibly other types of oral disease (Epstein gt 11., 1984: Odds, 1988: Rippon, 1988). It should be noted that oral thrush has seldom been attributed to yeast species other than Q. glbjggng (Epstein gt 11., 1984: Odds, 1988). Hewlett and Squier (1980) in an experimental model of anQidg-host interactions, and Wilborn and Montes (1980: 11 Montes 8 Wilborn, 1980) in patient sstudies, recently characterized the colonization and invasion of oral epithelium using scanning and transmission electron microscopy. From these studies, it was found that attachment of andidg cells to the superficial cells of the oral mucosa appeared to involve an intimate contact between the epithelial surface and layers of the fungal cell wall. Further examination showed the involvement of a number of layers of the cell wall of Q. glniggng within the epithelium, and. that changes 'were evident in. the organization and definition of the outer cell layers in budding yeast-phase cells and in hyphae and pseudohyphae participating in colonization and invasion of the epithelial cells. In other words, yeast-phase cells were closely associated with the epithelium, while hyphae and pseudohyphae invaded and penetrated into the epithelium. During hyphal penetration, a close seal was maintained between the invading hyphae and the surrounding epithelial cell envelope. No other damage to the epithelial cell surface was noted except at the site of entry. Within the epithelium there was only occasional loss of cytoplasmic components, these being in the vicinity of the invading hyphae. This would suggest that there was an enzymatic lysis associated with the invasive process, and that this enzymatic lysis was localized, possibly at the hyphal tip. angigg infections of oral mucosa, then, may be considered as occurring in three phases: (1) mucosal association and 12 (Liljemark 8 Gibbons, 1973): (ii) hyphal invasion of the epithelial cells and establishment of an intracellular habitat (Montes 8 Wilbern, 1968), which may be initiated by a tropism to the host tissue by Q. glniggng (Hewlett 8 Squier, 1980: Wilbern 8 Montes, 1980): and (iii) the provocation of an inflammatory response in underlying tissues (Hewlett 8 Squier, 1980: Montes 8 Wilbern, 1968: Wilbern 8 Montes, 1980). The major source of infections for infants has often thought to be the mother's vagina during birth, where Q. nlniggng can reside as a cemmensal (Rippon, 1988). Other sources have also been implicated, including infected feeding bottle nipples (Odds, 1988: Winner 8 Hurley, 1964), the skin of nurses and mothers (Jennison, 1966: Ludlam 8 Henderson, 1942: Odds, 1988: Winner 8 Hurley, 1964) and the air or dust of nurseries (Jennison, 1966: Odds, 1988) . Among adults, oropharyngeal and esophageal thrush may appear as a complication of infection or medical care resulting from a number of clinical situations, since Q. glniggng is known to reside unnoticed in the oral cavity in approximately 20 to 50% of healthy individuals (Odds, 1988: Winner 8 Hurley, 1964). W. Qandida infections of the stomach, small intestine and colon have been reported more frequently in recent years, but have been described far less often than cases of oropharyngeal and esophageal thrush or related diseases. Winner and Hurley (1964) have 13 reviewed the case histories prior to 1962, but there have been only a few descriptive reports since. Q. ngtiggng and other angigg species (as well as Iotnlopsig glgntgtg) have been isolated from samples of gastric ulcers, and have been suggested to interfere with the healing of the ulcers (Gotlieb-Jensen 8 Anderson, 1983: Thomas 8 Reddy, 1983). Likewise, Q. glniggng has been shown to invade the intestinal mucosa of the small intestine and colon (Joshi gt 91., 1981: Rippon, 1988). Endoscopic examination and biopsy specimens, for instance, have demonstrated the presence of andida organisms in ulcerative lesions and plaques (Joshi gt _a_;L., 1983: Kozin 8 Taschdjian, 1962). Macroscopically chronic gastrointestinal candidosis appears as two types: (i) the mucosa may be covered by a thick, shaggy, cream-colored confluent membrane (similar to that seen in oral thrush), which strips away easily to reveal grayish-brown plaques of confluent ulceration: or (ii) there may be more localized, discrete and scattered lesions, which may be few or many and may occur in any part of the gut (Rippon, 1988: Thomas 8 Reddy, 1983). The latter type of lesion or plaques may be grayish-brown or cream-colored raised plaques or areas of shallow ulceration, and may be covered by a rough friable brown exudate: thus making the second type of lesion very hard to observe endoscopically (Joshi gt g1., 1981). Biopsy specimens have shown that invasion of the mucosa can be accompanied by the downward, parallel growth of mycelial 14 elements (Koznin 8 Taschdjian, 1962: Rippon, 1988). In view of the frequent presence of andidg organisms in the gut, it has often been suggested that GI andidg infections are endogenous in nature (Odds, 1988). However, since Q. albigans frequently inhabits the gut of healthy, asymptomatic individuals as commensals (in low numbers), it follows that host resistance mechanisms must first be disrupted before Q. glniggng can initiate invasion and infection of host tissues. One of the most important factors that can predispose an individual to oral and/or GI candidosis and, subsequently, systemic candidosis, is a partial or complete disruption of the indigenous microflora (Odds, 1988). The oral and intestinal microflora protects the host against mucosal association, colonization and invasion of oral and GI mucosal surfaces (Liljemark 8 Gibbons, 1973: Kennedy 8 Volz, 1983: Kennedy 8 Volz, 1985 a8b). Sygtgnig_ggng1gggig. Systemic angigg infections are an important medical complication in seriously debilitated patients (Meunier-Carpenter gt gt. , 1981: Myerowitz gt g1., 1977: Ray, 1980), and are important causes of morbidity and mortality among patients who are compromised immunologically or undergoing prolonged antibiotic therapy (Odds, 1988: Ray, 1980: Rippon, 1988). Although angigg cells may be introduced into the bloodstream by a number of exogenous sources (including contaminated i.v. catheters and infected wounds), numerous authors consider 15 the direct passage of viable Q. glgiggng through the GI mucosa into the host bloodstream to be the most important mechanism leading to systemic candidosis (Krause it g1. , 1969: Myerowitz gt gl., 1977: Stone, 1974: Stone gt g1” 1973,1974). Factors which result in an overgrowth of Q. m in the gut (such as prolonged antibiotic therapy), then, may render the host more susceptible to systemic andiga infections. Such infections may arise as localized infections, with only one organ affected, or as disseminated infections as the result of mm cells being spread by the bloodstream to a number of organs (Odds, 1988: Rippon, 1988) . The most important factor determining the extent and severity of such infections are the underlying host immune system, the virulence of the organism and the amount of organisms entering the host bloodstream (Odds, 1988: White gt g1” 1979). Cancer patients who are being treated with immunosuppressive drugs, broad-spectrum antibiotics, and/or intensive radiation therapy are the most susceptible individuals to systemic candidosis (Bodey 8 Rodriquez, 1975: Meunier- Carpenter, 1981: Myerowitz gt g_., 1977: Young gt g;., 1974). 2.1-. 0.: ._b 318 ce_ .. _ 0011'? u a: 1 b ns The cell wall of Q. gltiggng is a multilayered structure approximately 100 - 300 nm thick whose 16 appearance varies with cell age, morphologic form and culture conditions (Cassone gt g1., 1979: Kulkarni gt g1., 1980: McCourtie 8 Douglas, 1981: Poulain gt g1., 1985). Because of the putative importance of cell wall ultrastructure and surface composition in anQng adhesion and pathogenesis, the chemical composition of the cell wall of Q. nlbtcans has been studied in great detail (Cabib gt g1., 1982: Chattaway gt gl., 1968: Gooday, 1977: Kessler 8 Nickerson, 1959: Poulain gt gt” 1985: Reiss, 1986). Biochemical studies of the cell wall of (Q. m indicate that it is composed mainly of glucan and mannan polysaccharides, with lesser amounts of chitin, protein and lipid (Chattaway' gt, g;., 1968: Kessler 8 Nickerson, 1959: Yu gt gl., 1967 a8b). The presence of glucose and mannose in Q. m polysaccharides has been a consistent finding (Chattaway gt g;., 1968: Jensen, 1955: Kemp 8 Solotorvsky, 1964: Odds, 1988: Stewart-Tull gt g1., 1966), and glucan and mannan have often been found combined with protein in cell wall extracts as glucoproteins, mannoproteins and glucomannoproteins (Kemp 8 Solotorvsky, 1964: Kessler 8 Nickerson, 1959). The finding that mannan and glucan represent about 75 - 85% of the dry matter of gagging cell walls, and that the remainder of the cell wall is composed of lesser amounts of chitin, protein, and lipid suggests that it is very similar to the cell wall of‘ Sgggngtgnyggg ‘ggzgyigigg (Ballou, 1976: Chattaway g g1., 1968: Kessler 8 17 Nickerson, 1959: Poulain gt g;., 1985: Reiss, 1986). lel ggll composgtign. Chemical analyses of the cell wall composition of Q. gltiggng have shown that mannan makes up nearly 35 - 40% of the total dry weight of the cell wall (Bartnicki-Garcia, 1973). Q. glttggng mannans contain mainly a-1-6- and a-1-2-linked mannopyranose residues and much less phosphorus than those reported for S. nggligtag (Bishop gt gl., 1960: Skil gt g1" 1964: Stewart 8 Ballou, 1968: Suzuki 8 Usuyama, 1971: Yu gt g1., 1967 a8b), while rare a-1-3-1inkages between mannose residues have also been reported (Odds, 1988: Stewart 8 Ballou, 1968: Suzuki 8 Usuyama, 1971). The chemical structure of this polysaccharide, then, is an a- 1-6-linked mannan backbone to which sidechains of up to six hexose units long are joined by a-1-2- and rare a- 1-3-linkages. The phosphate molecules associated with Q. nlniggng cell wall mannans appear to be linked together by phosphodiester bridges (Elorza gt g;., 1985). Analysis of mannans isolated from serotypes A and B of Q. glbiggng has revealed differences between the mannans of these serotypes. The differences between mannans of Q. gltiggng serotype A and B, as well as Q. W and Q. ttgpigglig, appear to be due to the extent of branching, the length of side chains (serotype A is thought to have longer side chains than serotype B [Okubo gt 11., 1979; San-Blas, 1982]), and the proportion and position of a- 1-3-bends in the side chains (Odds, 1988: Stewart 8 18 Ballou, 1968: Suzuki 8 Fukazawa, 1982). The impact these differences may have on anQng adhesion is presently unknown. Glucan appears to be the most abundant polymer in the cell wall of Q. albicgns (Bishop gt g1.,1960: Elorza gt 11., 1983: Gopal gt g1., 1984: Shephard gt g1., 1984: Sullivan gt, 11;, 1983), and is considered to Ibe the essential component of the "microfibrillar skeleton" (Gopal gt g1., 1984: Poulain gt g1., 1985). Glucan is a highly branched polysaccharide that consists of a backbone of fi-1,6-linked glucose residues to which side chains of fi-l,3-linked glucose residues are attached (Bishop gt g1., 1960: Yu gt g1., 1967). Recent studies suggest that there may actually be two types of glucan found in the cell wall of Q. g1nignng: either fi-1,3-linkages or fi-l,6-linkages, both of which are highly branched (Gopal gt g1., 1984). Gluc0protein, mannoprotein, and glucemanneprotein complexes have also been isolated from the cell wall of Q. finiggng (Kessler 8 Nickerson, 1959: Kolaraova gt g1., 1973). These molecules are all more or less tightly associated with one another and with glucan and mannan (Kolareva gt g1., 1973), and. may serve as necessary adhesive links between glucan and mannan (San-Blas, 1982). The proteins that are complexed with mannan and glucan, in yeast phase cells, are thought to be present as water- soluble proteins and water-insoluble glycopeptides (Kelarova gt g_., 1973) . Hannah-protein complexes are 19 thought to be linked through an N-acetylglucosamine residue and side chain amine groups (Odds, 1988). Glucan is also thought to be complexed to chitin (polymer of N- acetylglucosamine joined by B-1,4-linkages) adjacent to the plasma membrane (Gopal gt g1., 1984). Although chitin helps to anchor glucan to the cell membrane, it comprises only a small part (<1%) of the cell wall composition by dry weight (Chiew gt 11., 1980: Poulain gt g1., 1985: Sullivan gt g1., 1983). Most (90%) of the chitin has been shown to be associated with bud scars of yeast phase cells (Bacon gt a_1., 1966), and the remainder forms the link between the cell membrane and. the innermost insoluble glucan possibly by a mixed B-1,3-/fi-1,6- linkage (Gopal gt g1., 1984). All of these molecules are antigenic (Isenberg gt 31., 1963: Ponton 8 Jones, 1986b: Poulain gt 51., 1985: San-Blas, 1982: Smail 8 Jones, 1984: Sundstrom 8 Kenny, 1984), and some (e.g., mannoproteins) may serve as angigg adhesins (McCourtie 8 Douglas, 1985a: Sandin gt g1., 1982). The protein content of the cell wall of Q. g1niggng has been studied less extensively than the polysaccharide content. Protein may, in some instances, make up to about 30% of the cell wall content (Chaffin 8 Stocco, 1983: Chattaway gt 31., 1968: Kobayashi 8 Suginaka, 1984). It had generally been thought that nearly all of the proteins found in the outer layers of the cell wall of Q. glniggng were associated with large mannan molecules (Poulain gt 20 Q1. , 1985) . However, recent studies from several investigators do not support this view (Ponton 8 Jones, 1986a). Ponton and Jones (1986a) found that many proteins isolated from the cell wall of Q. fittggng had negligible mannose content. These authors further suggested that there may actually be a complex latticework of proteins situated in the outer layers of the cell wall of both yeast- and mycelial-phase organisms, and that inter- and intra-disulfide bonds in these proteins are important in maintaining the structure of this latticework. The ability of Q. glniggng to exclusively produce hyphae among angigg species has led to a number of studies on cell wall ultrastructure of yeast-mycelial conversion and its relationship to infection. A number of investigators, for instance, have shown that Q. gmiggng hyphae are: (i) usually seen in smears, scrapings or histological sections from lesions due to infection by Q. uniggng: (ii) often found in feces or from infected patients urine: and (iii) that Q. g1giggng yeast-phase cells produce hyphae in host tissues within a few hours of inoculation to experimental animals (Odds, 1988) . Thus, it has frequently been assumed, implied or suggested that hyphal formation by Q. miggng enhances its ability to penetrate and infect host tissues (Balish 8 Phillips, 1966 a8b: Blyth, 1959: Cantrell 8 Widra, 1964: Phillips 8 Balish, 1966). In fact, it has often been suggested that the hyphal form of Q. g1t1ggng alone is "pathogenic" or 21 "invasive" (reviewed in [Odds, 1988]). However, while the hypothesis that hyphal formation is the primary factor that confers an enhanced pathogenic ability on Q. n1bicang (or that such an ability is the only factor known to correlate with the fact that Q. Mgnng is the most common yeast pathogen) has considerable logical appeal, the fact remains that yeast-phase cells and hyphae are usually seen together in 35119. Furthermore, there have been relatively few instances in which the hyphal form of Q. Mng alone has been observed in infected tissues, whereas yeast-phase cells of Q. albicgng only are often observed in specimens from infected patients and animals. On the other hand, it is likely that the ability to form hyphae in infected tissues is probably influenced (and may be stimulated) by local environmental factors, and may allow Q. glnicans to avoid removal from these tissues. For instance, the ability to form germtubes affords Q. g1niggng an enhanced ability to attach to host epithelial cells, form emboli, and to resist ingestion by host phagocytes (Kennedy, 1988: Odds, 1988: Richardson 8 Smith, 1981) , and the ability to form hyphae has been shown to prolong survival within leukocytes and macrophages (Baine gt g1., 1974: Odds, 1988: Richardson 8 Smith, 1981). Nevertheless, although it has been shown that formation of hyphal forms is important in the pathogenesis of infections by Q. Mg (Sebel, 1984), the question of which form is more pathogenic may never be settled. 22 Studies related to the mechanism of morphogenesis have shown that numerous environmental factors can influence Q. g1t1ggng growth form, but the exact mechanism(s) responsible for Q. gltjggng morphogenesis has yet to be discovered (Odds, 1988) . Several studies, however, have shown dramatic changes in the cell wall composition and ultrastructure of Q. migang during the transformation from yeast to hypha (Chattaway gt g_., 1968: Odds, 1988: Yamaguchi a n1., 1971) . Such studies do suggest that the cell wall of Q. albicans is synthesized in two phases: (i) first an immature, soft, primary wall is laid down, and (ii) then it is subsequently rigidified (Odds, 1988). Analyses of the cell wall from Q. g1giggng yeast- phase cells and mycelium clearly indicate that there are differences in the cell wall composition of these two forms and that these differences are primarily quantitative and not qualitative (Bianchi, 1977: Braun 8 Calderone, 1978: Cassene gt g1., 1984: Chattaway gt g1., 19688 1976: Elorza gt L” 1985). Chattaway gt _1. (1968) , for instance, found that chitin was present at levels three times higher and protein at levels nearly three times lower in the mycelial form. Bianchi (1967) found twice as much lipid in hyphae as in blastoconidia, and three times as much carbohydrate has been found in hyphal membranes as in yeast membranes (Shwartz 8 Larsh, 1980). Recent studies confirm this and found that mycelial forms were more active than yeast forms in the 23 accumulation of steryl esters and triacylglycerols (Ghannoum gt g1., 1986b). Furthermore, it was also shown that yeast lipids contained much larger proportions of free sterels than the mycelial lipids, whereas the mycelial lipids contained several times more sterels than the yeast forms (Ghannoum gt g1., 1986b). Other differences in lipid composition from yeast and mycelial forms were also noted in that study. Chemical analyses indicate further' that. there (are quantitative differences in hexoses, hexosamines, and amino acids of purified cell wall preparations from blastespores and pseudohyphae (Chattaway gt g1. , 1968 8 1976), which may be related to changes in the activities of glycolytic enzymes at the branch points for polysaccharide synthesis from hexoses in the glycolytic metabolic pathway (Chattaway gt g1., 1968 8 1976). Quantitative differences of several hydrolytic and glycolytic enzymes between the two forms have also been reported (Odds, 1988: Shawartz 8 Larsh, 1980). W U st tu . The ultrastructure of the cell wall of Q. g1b_iggng has been examined by several investigators using a variety of procedures to reveal the precise location of the principle wall components (Cassone, 1973: Cassone gt g1., 1978: Djaczenko 8 Cassone, 1971: Poulain e_t g1., 1981). Several cell wall layers have been identified for Q. g1btggng (Poulain gt g1., 1985). These layers have been labeled L0 to L8 Poulain gt 24 Q1. (1985). Cytochemical staining (Djaczenko 8 Cassone, 1971: Poulain gt g1., 1981) revealed that each layer contains a dominant polysaccharide (Poulain gt g1., 1985). The gross appearance of the cell wall as viewed by electron microscopy is that of a sandwich structure with an outermost set of layers of high electron density (L0 to L4), an inner layer of high electron density (L8), and an intermediate set of layers of lower electron density (L5 to L7) (Poulain gt g1., 1978, 1981, 8 1985). It should be stressed that the gross appearance of the cell wall has been shown to be directly influenced by the growth parameters used for preparation of the test cells (Cassone gt g1., 1973: Cawson 8 Rajasingham, 1972: Djaczenko 8 Cassone, 1971: Garrison, 1981: Montplaisir Lt g1., 1976: Poulain gt g1., 1978, 1981, 8 1985). Poulain and co- workers (Poulain gt g1., 1978, 1981, 8 1985) found that both the growth medium and the age of the cells affected the number and appearance of layers in the cell wall. Recent studies have shown further that other growth parameters (e.g. temperature) can also have a significant effect on cell wall ultrastructure (McCourtie 8 Douglas, 1981) . Moreover, some of these changes, particularly in the outermost layers, have been shown to correlate with the adhesiveness of Q. g1tiggng to both biological and nonbiological surfaces (Douglas gt g1., 1981: McCourtie 8 Douglas, 1981 8 1984). This will be described in more detail below. 25 Agngging. Although there have been limited ultrastructural observations on the mechanisms of Q. g1biggng adhesion, such studies do suggest that at least two morphologic classes of anQth adhesins exist: floccular and fibrillar (Kennedy, 1988). The floccular adhesins are present as a thick (approximately 100 to 400 nm) "fuzzy" cell wall coat and appear to be somewhat amorphous (Tronchin gt g1., 1984). Occasionally, this material appeared to have an ordered alignment around the cell wall (Tronchin gt g1., 1984), whereas at other times it was unevenly distributed on the cell surface (McCourtie 8 Douglas, 1981) or localized only at an adhesive junction (Hewlett 8 Squier, 1980: Marrie 8 Costerton, 1981). Ultrastructural studies of the adhesion of Q. a1bicans to oral (Calderone gt g_., 1984: Hewlett 8 Squier, 1980: Marrie 8 Costerton, 1981: Montes 8 Wilborn, 1968: Tronchin gt 31., 1984), urinary (Marrie 8 Costerton, 1981), and vaginal epithelium (Calderone gt g1., 1984) showed the floccular structures to be present on yeast in infected tissues, and that they may indeed mediate angigg adhesion. Using tissue explants, for instance, Hewlett and Squier (1980) found that a floccular outer layer mediated the adhesion of Q. g1btgang to oral mucosal cells. angiga yeast phase cells harvested from media with high concentrations of certain carbohydrates have been found to be highly adhesive, and contained an outermost floccular layer that was absent from 26 blastespores that were significantly less adhesive which were cultivated in media without these carbohydrates (McCourtie 8 Douglas, 1981). The fibrillar material (called elsewhere fimbriae [Gardiner gt g1., 1982]), in contrast, can be seen on the cell surface as thin filamentous structures arranged perpendicularly to the cell surface and evenly distributed around the entire cell (Barnes gt g1., 1983: Lee 8 king, 1983a). Lee and King (1983a) viewed the fibrillar structures on the cell surface of Q. g1niggng mediating adhesion to renal endothelium. The diameters of these structures have not been measured, but appear to be within the size range (2 to 10 nm) of fimbrial adhesins of bacteria (Jones 8 Isaacson, 1983). Likewise, the length of the fibrillar adhesins has not been accurately measured, but these appear to be less than 0.5 pm long (Lee 8 King, 1983a), considerably shorter than bacterial fimbriae (Jones 8 Isaacson, 1983). Fibrillae also have been observed to mediate the adhesion of Q. gltiggng to other surfaces such as epithelial cells (Harrie 8 Costerton, 1981: Rotrosen gt g1., 1985: Tronchin gt g1., 1984) . Both types of adhesins stain with ruthenium red (Harrie 8 Costerton, 1981), which has an affinity for anionic polymers such as polysaccharides (Luft, 1961). The precise chemical and organizational nature of these two morphologic structures is not known. It is very likely, however, that these structures represent distinct 27 adhesive entities. There is some evidence to suggest that Q. g1niggng does produce more than one adhesin (Douglas, 1985a: Kennedy gt g1., 1987: Sandin, 1987a). It should be noted that the presence of one or both types of structures on the cell surface of Q. gltiggng does not, by itself, indicate that the structures play a role in adhesion (Kennedy, 1988). Cangiga cells devoid of floccular or fibrillar structures have been shown to be highly adhesive to buccal epithelial cells (BECs) (Kennedy, 1988). Furthermore, not all known microbial adhesins can be identifiable as discrete surface structures (Jones 8 Isaacson, 1983). Various studies have also suggested that two cell wall components primarily act as adhesins for Q. g1ntggng (Douglas, 1985 a8b: Rotrosen gt g1., 1986). It has been suggested, for instance, that a mannose containing moiety, probably a mannoprotein (McCourtie 8 Douglas, 1985a) , on the surface of Q. g1bicgns is the Cangiga adhesin (Sandin 8 Rogers, 1982: Sandin gt g1., 1982). However, analogous studies also have suggested a possible role for cell wall chitin as the adhesin (Segal gt g1., 1982). Lipids also have been suggested to play a role in adhesion to epithelial cells (Ghannoum gt g1., 1986a). Moreover, other studies suggest that there may be more than one adhesin (Douglas, 1985a: Kennedy gt 31., 1987: Sandin, 1987a) , and at least one report suggests that Qangigg adhesion may not be mediated by a specific adhesin- 28 receptor interaction but may instead be entirely nonspecific attachment (Reinhart gt g1., 1985). Possible cell wall components that serve as Candida adhesins are listed in Table 1. The majority of reports using lectin and carbohydrate blocking experiments have indicated that a mannoprotein on the surface of Q. g1niggng probably serves as the adhesin in a number of adhesive reactions (Douglas, 1985 a8b, Rotrosen gt 91., 1986). Rossano and Tufano (1975) and Sandin gt g1 (1982), for instance, reported that pretreatment of Q. a1b1cgns with Concanavalin A, a lectin which binds to a-D-mannosyl residues, inhibited adhesion of the yeast to human BECs. Adhesion could also be inhibited by adding mannose or a-D-methylmannopyranoside (a mannose derivative) to the adhesion assay medium during the incubation period (Sandin 8 Rogers, 1982: Sandin g g1., 1982). Addition of tunicamycin, an antibiotic which, in yeasts, specifically inhibits mannoprotein synthesis but not chitin or glucan synthesis (Kuo 8 Lampen, 1974), to cultures of Q. g1niggng inhibited the formation of the floccular surface layer of Q. glgiggng (Douglas gt g1., 1983). Concurrently, there was a significant decrease in adhesion of Q. gjtjtgng to BECs (Douglas gt g1., 1981 8 1983). Moreover, Lee and King (1983b) found that extracted mannan-rich cell wall fragments of Q. g1n1ggng attached to vaginal epithelial cells and that this attachment was abolished following treatment of the 29 Table 1. Possible cell surface components suggested to serve as anQng adhesins* Possible adhesin Examples of References moieties inhibitors Chitin Glucosamine Segal gt g1., 1982 Chitin-soluble extract " N-acetylglucosamine " Mannosamine " Mannan/ mannoprotein Conconavalin A Sandin gt g1., 1982 Mannose " a-mannosidase " D-methylmannopyranoside " a-methylmannosidase Lee 8 King, 1983 Tunicamycin Douglas 8 McCourtie, 1983 Glucose Glucose Glucan Protein Chymotrypsin Lee 8 King, 1983: Sobel gt g1., 1982 Trypsin " Papain Lee 8 King, 1982 Pepsin " Pronase ” Lipid Sterols Ghannoum gt g1., 1986 *Modified from Kennedy (1988). 30 fragments with a-mannosidase. This a-mannosidase selectively degraded mannan because it was free of proteases and did not contain glycosidic activity against p-nitrophenyl glucose or galactose substrate (Lee 8 King, 1983b). McCourtie and Douglas (19885a) showed that a crude mannoprotein preparation obtained from culture supernatants of Q. g1n1ggng also associated with BECs and inhibited subsequent yeast adhesion. This extracellular polymer of Q. albicans was found to contain a high mannose content (65 to 82%), and lesser amounts of protein (7%), phosphorus (0.5%) and glucosamine (1.5%) (McCourtie 8 Douglas, 1985a). The studies of Maisch and Calderone (1981) suggest a role of a-mannan in the adhesion of Q. albicang to fibrin- platelet clots formed in yittg (Maisch 8 Calderone, 1980). Sheep erythrocytes coated with an alkali-soluble cell wall extract attached to a fibrin-platelet matrix, whereas nonconj ugated sheep erythrocytes did not attach. The adhesion promoting effect was abolished by pretreating the alkali extract (72% polysaccharide and less than 1% protein) with a-mannosidase before conjugation to erythrocytes (Maisch 8 Calderone, 1981). Similarly, alkali extraction of ang1gg cells significantly inhibited Q. gm adhesion to human epithelial cells (Sandin, 1987a) . Likewise, a-mannan or a mannan-associated cell wall constituent may also serve as an adhesin in attachment to vascular endothelium. Rotrosen gt .31. 31 (1985) showed that anQng immune serum significantly blocked adhesion to endothelial cells, but that this activity was abolished by immuneprecipitation of immune serum with Q. g1niggng mannan but not by similar adsorption with particulate chitin. There is also evidence to suggest that protein alone may serve as an adhesin, or that the protein portion of the mannoprotein complex is more important than the carbohydrate moiety in mediating attachment. Adhesion of Q. g1bj_cgng to epithelial cells has been shown to be decreased after exposure to heat or various proteolytic enzymes (Lee 8 King, 1983b: Sobel g g1., 1981). Pretreatment of Q. g1bigans with trypsin, Chymotrypsin or pronase reduced adhesion to fibrin-platelet clots (Maisch 8 Calderone, 1981). Likewise, Critchley and Douglas (1985) found that pretreatment of an extracellular polymer (thought to serve as an adhesin) from Q. gjtigang with heat, dithiothreitol or certain proteases, but not sodium metaperiodate or’ a-mannosidase, either partially or completely destroyed its ability to block angtgg adhesion to BECs. Similarly, a protein component from Q. g1niggng pseudohyphae also reduced. the binding affinity of pseudohyphae for neutrophils (Diamond 8 Krzesicki, 1978). Studies on bacterial adhesion are consistent with this view as they have shown that most of the well characterized bacterial adhesins (with only two exceptions) are proteins (Jones, 1977: Jones 8 Isaacson, 32 1983). Other studies directed at identification of Gandhi; adhesins using selective inhibitors of adhesion have suggested that cell wall chitin may be involved in Q. g1niggng adhesion (Segal gt g1., 1982). Chitin, a hydrolysate derivative of chitin, and N-acetylglucosamine all significantly inhibited attachment of Q, glbicans to vaginal epithelial cells (Segal gt g1., 1982). Moreover, it was also observed that the latter two substances also reduced the infection rate in a rat model of vaginitis (Lehrer gt g_., 1983). Other investigators have not been able to reproduce these results (Sandin e_t g1., 1982). Although it seems unlikely that such a small portion of the cell wall, which is localized only to inner layers of the cell wall or bud scars, could serve as an adhesin, it should again be noted that those investigators (Sandin gt g1., 1982) did not use the same growth media as that used by Segal and co-workers (1982). Further blocking experiments with chitin or "chitin-soluble extract" should be performed after growth of several strains of Q. g1n1ggng on the yeast extract medium used by Segal gt Q1. (1982) , compared with that after growth on other media (Kennedy, 1988), to determine more conclusively whether or not chitin plays a role in adhesion. Another approach to determine whether chitin or other cell wall components serve as adhesins would be to isolate and characterize non-adhesive mutants of Q. g1niggng. It should be noted 33 also that the above experiments do not rule out the possibility that chitin or its derivatives block adhesion by means of steric hindrance or that chitin can bind to the same receptor sites that mannoprotein adhesins also bind. In another report, Sobel gt g1. (1981) showed that D- and L-fucose inhibited the attachment of Q. albigans to vaginal epithelial cells. However, in these experiments the fucose was allowed to react with the Cangidg cells first, suggesting not that a fucose-like component serves as an adhesin but that the receptor on the epithelial cell for Q. gmigng may also be a glycoprotein. Reinhart gt g1. (1985) have also suggested that QLndigg adhesion to epithelial cells may be entirely nonspecific, and that such adhesion is mediated solely by hydrophobic properties (Hazen gt g1., 1986). It seems unlikely, however, that a nonspecific phenomenon such as hydrophobicity could account for all of the distinct adhesive activities observed for Q. g1btggng (Kennedy gt g1., 1987). Studies from our laboratory suggest that specific adhesion predominates over nonspecific adhesion (Kennedy gt g1. , 1987). Therefore, the more likely explanation for Q. g1t1gang adhesion to various biological surfaces such as intestinal epithelial cells and other microorganisms is that adhesion is mediated by two or more distinct adhesive entities, in which nonspecific interactions (e.g. hydrophobicity) may or may not play a role. There is 34 already evidence to suggest that more than one adhesin may be involved in attachment to epithelial cells by (Q. g1b1ggng. However, the alternative proposition that many of the adhesive activities observed for Q. albtcans result fnmm a single mmltifactoral adhesive cell wall component cannot be ruled out from studies to date. d sso i 'o a The association of Q. g1t>1ggng with both biological and nonbiological surfaces has been suggested to play an important role in host colonization and pathogenesis. Data has been presented on the adhesion and association of Q. g1tiggng with nearly every tissue of the body, as well as to a number of "plastic" surfaces that are used to make dental protheses, catheters, and prosthetic cardiac valves (Kennedy, 1988). In this section, a detailed discussion of the physiochemical factors involved in Q. M adhesion to various surfaces is presented, and a number of adhesion and association mechanisms are described. These are summarized in Table 2. oc m' a Con e t'o s es' n. The adhesion of Q. AIM to any surface in vivg may vary considerably depending on host species, physiology, cell phenotype, and tissue involved (Anderson 8 Odds, 1985: Ryley, 1986: Segal gt g1., 1984: Sobel gt g1., 1981). Likewise, the attachment of Q. _a_l_n_igang to exfoliated 35 Table 2. Adhesion and association mechanisms of _C__:_daand' unifies" Nature of Mechanism Mechanism Active or Specific or Direct or Passive Nonspecific Indirect Adhesin-receptor A S D,I interaction Nonspecific A,P N D,I adhesion Coadhesion to A S,N I adherent organisms Entrapment in P N D mucus or tissue Germtube A N,S D,I penetration Enzymatic A N,S D,I digestion *Modified from Kennedy (1988). 36 epithelial cells, tissue explants, or various nonbiological surfaces, such as plastic or glass, can vary greatly in yitm depending on a number of experimental factors (e.g. growth medium: see the discussion on adhesion models below). Moreover, a number of long range, short range and hydrodynamic forces, only a few of which have been studied, may influence Q. a1bicgns adhesion ability (Loeb, 1985: Marshall, 1976 8 1984: Rutter 8 Vincent, 1980 8 1984: Skerl gt g1., 1984). Relatively little is known about the binding sites of Q. g1btggng adhesins and even less is known about andigg binding kinetics. Because a complete physiochemical description of the adhesion of Q. glnicans to any one surface is not yet possible, the following discussion will present the fundamental physiochemical principles involved in ang1gg- substratum interactions. In general, the cell surfaces to which Q. g1nigang attaches, as well as the surfaces of mtg; cells, have an overall negative surface charge (Jones, 1977) . These negative potentials result from the ionization of various chemical groups (e.g., sialic acid carboxyl groups of epithelial glycocalyces, acidic amino acid side chains, glycolipids and phospholipids, etc.) of the cell surface. Likewise, glass and plastic surfaces may also possess a net negative surface charge, depending, of course, on the type and ionic strength of the surrounding milieu (Marshall, 1976) . These charged surface groups loosely 37 attract oppositely charged ions (gegen- or counter-ions) from the surrounding solution to form a diffuse double layer of ions (Marshall, 1976). The resulting electric double layer is considered to be a part of the cell surface, in that as the cell is caused to move (e.g., due to Brownian motion or with an electric field) the thin layer of fluid and the ions of the double layer move with the cell (Jones, 1977). Two models of the diffuse double layer have been proposed (Loeb, 1985: Marshall, 1976). One concept, the Stern model, assumes that there is a one molecule thick layer of counter-ions held at the surface (the Stern layer) by electrostatic and London-van der Waals attraction forces which are sufficiently strong to overcome thermal agitation. The other concept, the Gouy- Champman model, describes that of a diffuse double layer without the Stern layer. These concepts have been described in detail elsewhere (Loeb, 1985: Marshall, 1976) and will not be discussed further here. However, it should be noted that the ionic strength of the surrounding environment can greatly influence the dimensions of the dielectric layer, and, therefore, can also modify attractive and repulsive forces (Jones, 1977) because repulsion will occur when both layers overlap (Paerl, 1985) . Not surprisingly, several studies have shown that Q. a1ttggng adhesion to epithelial cells varied according to selection of assay medium (Kimura 8 Pearsall, 1978 8 38 1980: McCourtie 8 Douglas, 1981: Persi gt g1., 1985). The most useful descriptions of adhesive interactions between two negatively charged surfaces, which theoretically should repel one another, are those provided by the new classical lyophobic colloid theory of Derjaguin and Landau (1941) and of Verwey and Overbeek (1948) (DLVO theory) and its many derivations (Jones 8 Isaacson, 1983). Briefly, the DLVO theory' proposes that the forces of repulsion (electrostatic interactions in the overlapping double layers) and attraction (London-van der Waals forces) between two similarly charged surfaces are additive, but vary independently with the distance of separation between the bodies. There are two separation distances between cells at which attraction is greater than repulsion. These are called the primary minimum (at small distances of separation of <1 nm) where attraction forces are strong, and the secondary minimum (at relatively large distances of separation of >10 mm) where attraction forces are weaker and are easily reversible, for example, by mild fluid shear (Jones 8 Isaacson, 1983). Interposed between these two positions is a point at which repulsive forces predominate and the potential energy is maximized (Jones 8 Isaacson, 1983) . The overall charge and shape of the bodies are important. and contribute significantly to the net interaction (Isaacson, 1985). For example, repulsion forces decrease with bodies of decreasing radius of curvature (Isaacson, 1985: Jones 8 39 Isaacson, 1983). Likewise, as ionic strength in an environment increases the energy maximum repulsion barrier decreases, and at high electrolyte concentrations the repulsion energy barrier may be eliminated altogether (Jones 8 Isaacson, 1983). The DLVO theory describes long-range adhesive interactions (i.e., those occurring at the secondary minimum) adequately where the bodies are held in a state of mutual attraction, but is inadequate in describing close-range interactions between adhesin and receptor, because bacteria and yeast do not possess sufficient kinetic energy to overcome the repulsion barrier (Jones 8 Isaacson, 1983: Rutter 8 Vincent, 1984). Nevertheless, the DLVO theory remains valid to describe certain adhesive interactions (Jones 8 Isaacson, 1983: Rutter 8 Vincent, 1980), and suggests that relatively irreversible adhesion (i.e., binding at the primary minimum) may require adhesive appendages (i.e. adhesins) that bridge the gap between cell surfaces to bind its complementary receptor (Jones 8 Isaacson, 1983, Marshall, 1984). Therefore, these interactions which occur at the secondary minimum are considered to be nonspecific and preliminary to irreversible adhesion (Jones, 1984a). Both types of interactions may be necessary for adhesion to and colonization of host tissues“ For’ instance, although curved bodies come closer together at the secondary minimum and require less kinetic energy to reach the 40 primary minimum, and adhesion of yeast cells to cells with microvilli is energetically more' favorable compared with other tissue of planar configuration, the effect of curvature may not decrease the repulsion sufficiently to permit Q. albicans to reach the primary minimum. Thus, in a system such as the small intestine, the effect of fluid shear due to peristalsis and the bathing actions of mucosal secretions should dislodge angigg cells resulting in their removal from the tissue surface. Therefore, as Q. glbicans is known to bind essentially irreversibly to small intestinal microvilli (Pope 8 Cole, 1981 8 1982), it is likely that specific adhesin-receptor binding is involved (Kennedy gt 911., 1987). However, without the weak attraction at the secondary minimum it seems unlikely that the presumed stereochemical fit of adhesin with receptor could take place (Fletcher, 1980: Jones, 1984a). Harrie and Costerten (1981) noted two types of andigg-epithelial cell interactions that preceded cell invasion. The first was a "loose" adhesion apparently mediated by a ruthenium red positive matrix, followed by a "tight" adhesion where no space could be seen between host and yeast cell. Furthermore, it seems likely that adhesive appendages of very small radii relative to the yeast cell would favor contact with a tissue surface. Individual fibrillar adhesins, as those described earlier, with considerably smaller radii than the yeast cell could overcome repulsion 41 due to the potential energy maximum required to bridge the gap between Q. Mng and epithelium. Lee and King (1983b) found that "fibrils" from Q. glbiggns yeast phase cells mediated adhesion to vaginal epithelial cells and endothelial cells lining renal peritubular capillaries. It is also interesting to note that cells of Q. g1tiggng which possessed germtubes were found to be significantly more adhesive to BECs than were blastespores which did not possess germ tubes (Kimura 8 Pearsall, 1980: Sandin gt n1., 1982). It follows from the physiochemical principles described that the smaller radii of curvature of tiny germ tubes and smaller yeast cells may have facilitated adhesion. The alternative proposition that the increased adhesive activities for cells with germtubes resulted from a concentration of adhesins, or the production of additional adhesins, cannot be ruled out at present. Nevertheless, the smaller radii of curvature of germtubes are more energetically favorable for adhesion. At short distances several other interactions are important in adhesion. Such interactions include: dipole-dipole (Keesom) interactions, dipole-induced dipole (Debye) interactions, ion-dipole interactions, chemical bonds (e.g. , electrostatic, covalent and hydrogen), and hydrophobic interactions (Rutter 8 Vincent, 1980 8 1984: Tadros, 1980) . A number of these types of interactions have been shown to be important in bacterial adhesion (Marshall, 1984) . It should be noted, however, that 42 short-range effects may be repulsive or attractive depending on the nature of the surfaces involved (Rutter 8 Vincent, 1980: Tadros, 1980), and are particularly important in aqueous systems (Marshall, 1976: Rutter 8 Vincent, 1984). The role most of these short-range forces may play in the adhesion of Q. g1bicgns to biologic or nonbiologic surfaces has not been examined. Another possibility that also has not been considered is that the variability of the surface of Q. albicans (e.g., in charge, shape, appendages, etc.) may lead to various types of synergistic interactions that take place simultaneously. In bacterial systems, hydrophobic interactions adjacent to ionic or hydrogen bonds have been shown to stabilize an otherwise energetically weak binding complex (Doyle _e_t 91., 1982). Synergistic interactions can easily be imagined for Qandidg adhesion and should be investigated. Likewise, the effect surface topography (Kulkarni gt g1., 1980) may have on anQng adhesion has been examined (Kennedy, 1988). Only a limited number of physiochemical parameters have been studied in relation to Q. miggng adhesion. Lee and King (1983b) examined the effect of certain physiochemical parameters on adhesion of Q. g1giggng to human vaginal epithelial cells in ytttg, and found that divalent cations, detergents, salts, and urea had no effect on adhesion. These findings suggest that hydrophobic, electrostatic or ion-bridging bonds may be of 43 little importance in adhesion to vaginal mucosal cells in this system. It should be emphasized that such results may be particular to the Q. glbicgns strain and growth medium (and other assay parameters) used in that study. On the other hand, reducing agents (e.g., fl- mercaptoethanol and dithiothreitol) significantly diminished Q. ggiggng adhesion, but did not effect viability (Kimura 8 Pearsall, 1980: Lee 8 King, 1983b), suggesting that the structural integrity of the adhesive factor(s) is important. In contrast to the study of Lee and King (1983b), Karaev gt Q1. (1986) showed that the introduction of Ca2+ and Mg2+ ions to in mg assays led to a significant increase in Q. g1ntcans adhesion to BECs. Likewise, divalent cations were shown to promote the adhesion of Q. gltigang to acrylic, and at high concentrations these cations caused extensive ceadhesion and aggregation of Q. Miggng (McCourtie 8 Douglas, 1981). These studies indicate that ion-bridging mechanisms and electrostatic forces may be of primary importance in the adhesion of Q. n1b_igang to BECs, at least under some conditions, and acrylic surfaces in m. Hydrophobic interactions also have been examined for their role in the adhesion of Q. M to various denture base resin materials, plastics, and BECs (Klotz gt 31., 1985: Minagi gt g1., 1985 & 1986: Miyake gt g1., 1986). Klotz gt Q1. (1985) found that the adhesion of Q. 44 n1b1ggng to plastic surfaces was predominantly controlled by hydrophobic forces, and that electrostatic forces also contributed to adhesion. This view is corroborated by the findings of Minagi gt g1. (1985 8 1986), who showed that Q. g1biggng could attach to denture materials in a similar manner. However, in the latter study Q. albicans was significantly’ less hydrophobic, indicating that the surface free energy of the denture material itself greatly influenced angigg adhesion. Whereas hydrophobicity appeared to contribute to the adhesion of Q. albicans to various plastics and denture materials, other nonspecific forces also probably contributed to the adhesion process. In contrast, recent studies from our laboratory showed that hydrophobicity may be of little importance in association with intestinal mucosa (Kennedy gt g1., 1987). However, hydrophobicity may have been important in yeast- te-yeast ceadhesion, and may have indirectly influenced the total number of attached yeast by promoting ceadhesion (Kennedy, 1988) . The role cell surface hydrophobicity plays in (Q. (n121ggng, adhesion i ‘vivo remains to be demonstrated. Mnggsn1 snrfgces. The adhesion and association of Q. miggng with oral, vaginal and GI mucosal surfaces has been studied in a number of in ytttg and in ytyg systems. Such studies have included adhesion to exfoliated epithelial cells, colonization and invasion of various 45 tissue explants, as well as the ultrastructural characterization of attachment and penetration of Q. g1n1ggng_ to oral epithelium obtained from infected patients or experimental animals. Some of these studies have focused on the nature of the adhesion, whereas others have focused primarily on tissue localization. Hewlett and Squire (1980) examined the colonization and invasion of oral epithelium using tissue explants. Tissues were obtained from Sprague-Dawley rats and 4 day- old or younger New Zealand white rabbits, maintained in yittg in a chemically defined medium, and inoculated with Q. g1ntggng. Infected explants were maintained in yittg for 12 to 30 hours, and harvested at regular intervals to examine the various features of adhesion and association. Three types of interactions were noted. Blastoconidia of Q. g1biggng were observed to be randomly adhering to the surface of the epithelium, and in many cases germtubes were observed extending from the parent blastespores and penetrating into the tissue. Adhesion of yeast cells to oral epithelium appeared to form an "intimate" contact between the cell surface and certain layers of the anQng cell wall. Five distinct layers in the cell wall of Q. Miggng could be observed in infected tissues, and the surface structure that appeared to mediate attachment was a floccular layer. Moreover, the cell wall of Q. g1b1ggng appeared to undergo ultrastructural changes during adhesion. This is similar to the observations of Tronchin 46 gt g1. (1984) who noted ultrastructural modifications of the fungal cell wall coat during the adhesion of Q. gltigm to BECs in ytttg. In that study, various cytochemical staining techniques were used to visualize the adhesion process, and showed that Q glbicgns developed a fibrogranular surface layer that appeared to mediate adhesion. Concanavalin A binding sites were increased after development of this material (Tronchin gt g1., 1984). The initiation of contact between Candida germtubes and the epithelium may represent a distinct adhesive mechanism because no alteration of the epithelial cell surface was noted either at the point of entry or around the vicinity of Candida microcolonies. This was true for all types of epithelia examined. Studies by Sandin and co-workers (1987a) suggested that germtubes may contain different adhesins, or concentrated amounts of adhesin (Sandin gt g1., 1982) . Others have also reported that germinated yeast cells were significantly more adhesive than their non-germinated counterparts (Kimura 8 Pearsall, 1980). Within the epithelial cells of infected explants occasional loss of cytoplasmic components in the vicinity of the invading hyphae was noted (Hewlett 8 Squier, 1980). These findings indicate that if enzymatic lysis was associated with the invasive process it was localized, probably to the hyphal tips. Mechanical support provided by the adhesion of Qangigg cells to the epithelium may 47 also facilitate growth and penetration into tissues (Hewlett 8 Squier, 1980). Other reports support the role of both hydrolytic enzyme activity (Barrett-Bee gt g1., 1985: Ghannoum 8 Elteen, 1986: Pugh 8 Cawson, 1975, 1977 a8b, 8 1978: Waid, 1978) and mechanical force (English, 1963) in the invasion of animal cells by this fungus. There was no evidence that non-germinated yeast cells invaded the epithelium directly. In addition to yeast adhesion and germtube penetration, it also was observed that long hyphae grew on and colonized the epithelial surface and penetrated deep into the tissue. Examination of the infected explants showed the hyphal form to predominate within the epithelium, especially at later times (Hewlett, 1976). Sequential characterization of hyphal penetration of the rat tongue showed further that penetration occurred in three stages: (i) penetration of the superficial keratin layer (at approximately 18 hours), (ii) penetration of basal epithelium and basement membranes (at approximately 28 hours), and (3) invasion of connective tissues (at approximately 35 hours) (Hewlett, 1976). A number of studies examining infected tissues from patients and experimental animals have shown results similar to those described above. Montes and Wilbern (1968 8 1985) examined scrappings of plaque from the tongue of patients with chronic mucocutaneous candidosis and found pseudohyphae of Q. g1n1ggng growing profusely 48 between epithelial cells and penetrating into them. In analogous studies, these authors found that Candtgg pseudohyphae grew on and colonized the surface of the buccal mucosa, and pseudohyphae jpenetrated keratinized cells (Wilbern 8 Montes, 1980). Pseudohyphal forms were predominant in patients with chronic mucocutaneous candidosis, although blastoconidia were often observed attached to and colonizing oral epithelium. In another study (Marrie 8 Costerton, 1981), examination of plaques from the tongue and buccal mucosa of patients with oral candidosis revealed similar interactions between Q. albicans and oral mucosa. These include a "loose" adhesion apparently mediated by a ruthenium red positive matrix (similar to floccular adhesins), a "tight" adhesion where no space could be seen between yeast and epithelium, and invasion of host cells by (Q. (g1niggng_ hyphae. Coadhesion of ang1gg blastoconidia and indigenous bacteria to hyphal elements were also frequently noted (Marrie 8 Costerton, 1981) . The adhesion of Q. g1ntcans to BECs 1n ytttg was characterized iby electron microscopy and appeared to proceed via similar mechanisms (Calderone gt g1., 1984). Again, yeast adhesion appeared to be of both a tight and loose nature. In addition, yeast cells were often observed in phagocytic vacuoles of epithelial cells. From these and similar studies it is not clear whether these two ultrastructural forms represent distinct adhesive 49 activities or phases of the same process. Although tight and loose adhesion of Q. albicans could represent two different mechanisms of adhesion, they probably represent different stages of the same adhesion mechanism (Sandin, 1987b) similar to phases of reversible adhesion that precede irreversible adhesion in bacteria (Jones 8 Richardson, 1981 a8b). Studies using 0.4% formalin saline to kill angtgg cells revealed that dead organisms formed a loose adhesion but that only viable cells bind irreversibly (Vudhichaminog gt g1., 1982). As noted earlier, a number of studies have tried to characterize the nature of Candidg adhesin(s) , and to a much lesser extent epithelial cell receptor(s), by introducing various substances into an assay that might block adhesion or by pre-treating host or yeast cells before preforming adhesion assays. Using the former approach it was found that Q. g1niggng may bind reversibly for as long as 20 minutes, but thereafter the cells bind irreversibly (Sandin, 1987b). This suggests ‘that ‘the initial yeast-epithelium contact may be due to non- specific adhesion, followed by specific adhesion, or that, if the initial attachment is the result of specific adhesion, for example, "loose" binding via floccular adhesins, additional adhesion or association mechanisms may be required for irreversible binding (Sandin, 1987b). Modification of both host and yeast cells have been observed. once (Q. (g1ntggng_ attaches (Hewlett: 8 Squier, 50 1980: Tronchin gt g1., 1984). It may be that as anQng cells "bump" into epithelial cells and bind reversibly, physiologic changes occur to both host and fungus that strengthen the adhesion. These changes could modify the epithelium to the extent that more or different receptors are exposed that could stabilize and strengthen adhesion. Likewise, modifications of the cell wall of Q. gltjtgng, possibly causing an increase or concentration of adhesins, or an unmasking of different or more adhesins, have been noted at yeast-epithelial binding sites. Alternatively, the deposition of new cell wall material could occur after initial attachment. Tronchin gt_(g1. (1984) have noted the reorganization and proliferation of an external cell wall layer of Q. g1ntggng during adhesion to BECs. Moreover, they noted an abundant extracellular material with numerous binding sites for Concanavalin A that appeared to be released from the yeast surface, leaving underlying cell wall layers exposed. Several descriptions of phases in anQth adhesion to epithelial cells exist (Ghannoum (gt (g1., 1986a), but further studies to characterize the affinity and number of binding sites will be necessary to present a more complete and accurate description. It also has been suggested that the adhesion of Q. g1b1gang to buccal mucosal cells might entail ionic interactions involving divalent cations (Karaev gt g1. , 1986). The adsorption of macromolecules to epithelial 51 cells is recognized to occur via electrostatic interactions involving calcium ions and other ionic groups. angigg cells could also attach by similar mechanisms. However, conflicting data do not permit definite conclusions to be drawn. For example, it was shown that at acidic and basic pHs angida adhesion decreases (Sobel gt 31., 1981). In another study (Samaranayake 8 McFarlane, 1982a) , the opposite results were observed. These results, however, may be strain dependent (Persi gt 31., 1985). Also, cell surface hydrophobicity has been implicated in mediating Q3n§1d3 adhesion (Hazen gt 31., 1986). One may speculate, therefore, that once long-range electrostatic ferces are overcome, specific adhesin-receptor interactions become the important determinants of adhesion. Because of the possible ecological and pathological consequences of mucosal association by Q3n3133 in the GI tract (Eras gt 31., 1972: Krause it 31., 1969), several methods have been used to examine the adhesion and association of Q. 3131' cans with gut mucosa. Most have involved removing infected tissues from animals and viewing the resulting association by scanning (SEM) or transmission (TEM) electron microscopy (Balish gt 31., 1984: Field gt 31., 1981: Kennedy 8 Velz, 1985a: Kennedy gt 31., 1987: Pope gt 31., 1979: Pope 8 Cole, 1981 8 1982). Others have included examination of stained histologic sections or quantitative cultures to determine 52 population levels of attached yeast in infected tissue (Pope 8 Cole, 1981 8 1982). Tissue slices or isolated mucus gel have also been used to study mucosal association 111 Mg (Kennedy 8 Volz, 1985a: Kennedy gt (a_l., 1987: Segal g 31., 1986). Animal models used to study colonization of the GI tract have included neonatal and adult conventional, specific pathogen-free, germ-free, antibiotic-treated, and/or athymic gnotobiotic mice, hamsters, chickens and rats (Guentzel gt 31., 1985), and have revealed mucosal association by Q. albicans in all regions of the GI tract (Balish gt 31., 1984: Field gt 31., 1981: Kennedy gt 31., 1987). Different preferential sites of colonization, however, have been noted between various animal models. Pope (e_t 31. (Field gt 31., 1981: Pope 8 Cole, 1981 8 1982) , for instance, found that the stomach was the primary site of colonization in infant mice, whereas studies in antibiotic-treated adult mice have revealed that the cecum was most heavily colonized by Q. 31313315 (Kennedy gt 31., 1987: Turner gt 31., 1976). It is also interesting to note that the GI tract of neonatal mice was not colonized as quickly with Q. a_l_b_ig3ng as were adult mice (Balish gt 31., 1984). Nevertheless, the aforementioned studies have demonstrated that Q. 31313315 can associate with GI mucosa presumably by several different (distinct) mechanisms. Examination of the stomach mucosa of infant mice 53 revealed that at early times after Q3n3133 inoculation, yeast cells were attached to both keratinized squamous epithelia and columnar secreting epithelial surfaces. At later times, yeast were also seen attached to the secreting epithelium surrounded by and embedded in mucus, and were observed to be associated with and attached to lactobacilli at the junction of the keratinized and secreting epithelium (Balish gt 31., 1984: Pope 8 Cole, 1981 8 1982). Hyphal invasion of the keratinized region of the stomach has also been observed in both infant and adult mice (Balish gt 31., 1984: Pope 8 Cole, 1981 8 1982) . The cardial-atrium section of the stomach appears to be the preferential location in the stomach in some experimental animals (Guentzel gt 31., 1985), and in some studies it was observed to be the sole colonization and invasion site by Q. a1b1cans (Balish e_t_ 31., 1984: Helstrom 8 Balish, 1979: Pope 8 Cole, 1981 8 1982). It may be that the cardial-atrium ridge contains an abundant number of receptors, which allow Q. 31b1'g3ns to colonize this region of the stomach preferentially. These findings suggest that Q. 31319313 can associate with stomach mucosa by a number of adhesion and association mechanisms. These include direct adhesion to the epithelium, indirect attachment to the epithelium by association with other microorganisms, and penetration of the epithelium. At present, no attempt has been made to identify Q3n3133 adhesins or mucosal receptors that might be involved in 54 gastric colonization by Q. 31b1cans. The association of Q3ndid3 with small intestinal mucosal surfaces has also been studied in infected animals. These studies have demonstrated that Q. 31nig3ng can associate with intestinal mucosa by several different and distinct mechanisms (Field gt 31., 1981: Pope 8 Cole, 1981 8 1982). Using an infant mouse model, Pope and Cole (Field gt 31., 1981: Pope 8 Cole, 1981 8 1982) reported that large numbers of Q. albicans were clearly visible on the surface of villi in the small intestine within a short time after inculation. Further examination revealed that yeast adhered to all areas of the epithelium. Many yeasts were also seen on the surface of the villi, frequently in association with mucus, and appeared to be attached to, embedded in, and covered by a layer of mucus. Histological sections at 3 and 6 hours post-challenge also revealed cells embedded in the villus surface, and that some yeasts were associating indirectly with the mucosa by attaching 'to adherent yeast. Ultrathin sections from small intestinal mucosa showed that yeast cells did not attach to microvilli, but instead they attached to overlying mucus. Examination of large intestinal mucosal surfaces from adult animals revealed that Q. Mug associated with these surfaces by similar mechanisms (Kennedy 8 Volz, 1985a: Kennedy, 1987). However, this was true only when the animals were given antibiotics to disrupt the ecology 55 of the indigenous bacterial flora (Kennedy 8 Volz, 1983 8 1985b). SEM studies revealed that large numbers of Q3n31d_a cells were present on the surface of the epithelium and mucus in antibiotic-treated animals challenged with Q. 31b133ng, whereas no yeast cells were observed associating with host mucosal surfaces of untreated animals similarly challenged. In the latter case, large numbers of indigenous bacteria were seen colonizing the mucosa. In mice given antibiotics, Q. 31m could associate with the mucosa of all areas of the GI tract, but the cecum had the highest population levels. Further examination of the cecal mucosa of antibiotic-treated animals challenged with Q. 3131<33ng revealed that Cangiga cells could attach to the epithelium, possibly by adhesion to the epithelial glycocalyx (Kennedy gt 31., 1987) . In addition, yeast cells were observed to attach directly to or were seen embedded in mucus material. It was also found that Q. 31319315 could associate with the cecal mucosal surface indirectly by attaching to other adherent organisms such as adherent yeast cells and bacteria. The possible importance of these interactions becomes apparent in the observation that microcolonies of 93113133 have been observed in the cecum at 72 hours after oral challenge: with some yeasts associated with the epithelium, and others attached to mucus material and 56 other yeast cells (Kennedy (gt (31., 1987). If these microcolonies are formed in intestinal mucosa in humans, this may explain the isolated plaques observed by certain workers on examination of intestinal mucosa infected with Q. 3131g3ng (Eras gt 31., 1972: Joshi gt 31., 1981). Microcolony formation has been suggested to be of ecologic and pathologic importance in colonization of mucosal surfaces by other pathogens (Cheng gt 31., 1981: Costerton gt 31., 1981 8 1985). Depressions in the epithelium also were observed under Qand133 cells, possibly due to enzymatic lysis which may expose underlying receptors and stabilize Q. 31b1cans to the epithelium after the initial adhesion. It is interesting to note that germtubes or hyphae were not observed to penetrate the mucosa of the small or large intestine in these studies (Kennedy 8 Volz, 1985a, Kennedy gt 31., 1987). Although Q. a1b1cans was observed to associate with, and appeared to attach to mucus in the specimens processed for electron microscopy described above, the mucus gel was not maintained intact and appeared to be somewhat amorphous. Therefore, to determine more conclusively if Q. 31n1g3ng could associate with the mucus gel proper, and to insure that the apparent attachment to mucus was not due to specimen. processing (which can cause :mucus to shrink), Q3n3133 adhesion to mucus gel was tested 1n yittg (Kennedy gt 31., 1987). It was found that large numbers of Q. 31b1c3ns associated, with intestinal mucus very 57 rapidly (within 5 minutes) when mixed together 1n vit o, and that Q3n3133 organisms were attached to and embedded in the gel. Quantitative cultures of Candig3 in the mucosa of infected animals showed that approximately 20% of the associated Q. 313133ng were present in the mucus gel (Kennedy gt 31., 1987). The nature of the adhesive events between Q. 31nig3ng and the intestinal mucosa, as well as between Qandida and other adherent microorganisms, remains to be determined. It seems unlikely that nonspecific interactions, such as cell surface hydrophobicity, played a significant role in the adhesion of Q3ndida to mucosa (Kennedy, 1987: Kennedy gt 31., 1987). Studies with bacteria are consistent with this view, and suggest that specific adhesion predominates over nonspecific adhesion (Jones, 1984b). Therefore, it seems likely that a complex adhesive system is involved in the attachment of Q. 31tig3ng to various mucosal surfaces and other microorganisms, and may be mediated by two or more distinct adhesive entities in which nonspecific interactions (e.g., hydrophobicity) may or may not play a role (Kennedy, 1987: Kennedy gt 3_., 1987). Nonnio1og1c Surfages. The adhesion and association of Q. 31tig3ng with a number of nonbiologic surfaces can create a wide range of health problems for humans. Studies on the oral distribution of @133 in chronic atrophic candidosis, a disease that is relatively common among elderly denture wearers (Odds, 1988), have shown that Q. 58 11.11% is recovered more often and in higher numbers from the acrylic denture surface than from the palate (Davenport, 1970). This suggests that colonization of the acrylic denture surface serves as a reservoir for infection, and that Q3n3133 adhesion to the denture surface may be prerequisite for colonization and infection of the palate (McCourtie 8 Douglas, 1981). Likewise, the adhesion of C3ndida cells to various plastic surfaces that are used to make catheters and prosthetic cardiac valves, may also cause serious medical complications in a variety of patients. Relatively few studies on the adhesion of Q. 31big3ng to these surfaces, however, have been conducted. The adhesion of Candida to acrylic has involved the majority of the studies (McCourtie 8 Douglas, 1981: McCourtie gt 31., 1985: Miyake gt 31., 1985: Samaranayake 8 McFarlane, 1980: Samaranayake gt 31., 1980), which have suggested that several factors are likely to be important in this association. Studies by McCourtie and Douglas (McCourtie 8 Douglas, 1981) have shown that growth of Q. 31big3ng in a chemically defined medium supplemented with a variety of carbohydrates can greatly modify adhesion to acrylic surfaces. Growth of Q3n3133 in medium supplemented with 500 mM galactose, for instance, showed a significant increase in the amount and rate of adhesion and the production of an extra floccular surface layer (McCourtie 8 Douglas, 1981). Production of this material, which may be a mannoprotein (McCourtie 8 Douglas, 1985a), 59 has also been found to cause an increase in adhesion to BECs (Douglas gt 31., 1981) and an increase in virulence in a systemic mouse model (McCourtie 8 Douglas, 1984) . These results have prompted some to suggest that dietary factors may have a profound influence on Q3n3133 adhesion 1n 2123 (Samaranayake 8 McFarlane, 1982b 8 1985). Nevertheless, studies aimed at characterizing the adhesive nature involved in the attachment of Q. 31ntg3ng to acrylic surfaces and denture base resins have suggested that cell surface hydrophobicity may mediate, or at least participate in adhesion (Klotz gt Q” 1985: Minagi gt 31., 1986 s. 1986: Miyake gt 31., 1986). Miyake gt 31. (1986) found a strong correlation existed between the adhesion capacities of Candid3 species to acrylic surfaces and their cell surface hydrophobicities. It is interesting to note that in these studies Q. 31m showed the lowest hydrophobicity (Minagi gt 31., 1986: Miyake gt 31., 1986), which was probably due to the growth medium and environmental parameters used (Kennedy, 1988). Growth of Q. 31b1c3ns in Sabouraud dextrose broth at 37'C produced cells that were only slightly hydrophobic (Miyake gt 31., 1986). Therefore, it would be of interest to examine cell surface hydrophobicity and adhesion to acrylic after growth of Q. 313133ng in a variety of media (Kennedy, 1988). It is worth noting that four different methods have been used to determine the cell surface hydrophobicity of 60 Q. 31%. These included contact angle measurements, separation in water-hydrocarbon two-phase systems, plastic adhesion, and hydrophobic interaction chromatography (Klotz gt 31., 1985: Miniagi gt 31., 1985 8 1986: Miyake, 1986: Reinhart gt 31., 1985). Comparison of the former two assays indicates that these methods gave similar results. Comparisons between hydrophobic interaction chromatography and the other two methods have not been conducted. All of these methods have certain drawbacks that may influence results. The currently described hydrophobic interaction chromatography method (Smyth gt 31., 1978) , for instance, has the drawback that 3333133 cells may be occluded in the gel, especially if they occur as pseudohyphae or aggregates. A modification of this procedure, where the sample is allowed to mix with the gel instead of passing through it, has been used to avoid this problem (Kennedy, unpublished data). Ion-bridging bonds have also been suggested to play a role in the adhesion of Q. 31ntg3ng to acrylic (McCourtie 8 Douglas, 1981). McCourtie and Douglas (1981) found that the addition of divalent cations to assay mixtures caused an increase in Q3n3133 adhesion to acrylic. The addition 2+ of Ca was shown to cause a greater increase in adhesion, 2+ than was the addition of Mn or "92+. Cations seemed to cause a slight decrease in adhesion, and EDTA had no effect on its own but abolished the adhesion promoting effect of Ca2+. Cells grown in medium containing glucose, 61 galactose, or sucrose were all found to have an increased adhesion due to Ca2+, but this effect appeared to be greater with the latter sugar. FeCl2 also promoted adhesion to acrylic, but at high concentrations it caused the yeast cells to aggregate and settle to the bottom of the assay wells (McCourtie 8 Douglas, 1981). Thus ion- bridging mechanisms, in addition to cell surface hydrophobicity and other unknown factors (Rotrosen gt 31., 1986), may play a role in the adhesion of Q. 31b1gans to acrylic surfaces. Other nonbiologic surfaces have been studied less extensively than acrylic in regard to Cangida adhesion. In one study, Q3ndida species were shown to attach to both polyvinyl chloride and Teflon catheters, although yeasts adhered more extensively to the former (Rotrosen gt 31., 1983). Furthermore, it was also found that Q. ttgn1g31tg adhered in greater numbers than Q. albigans. In another study, the adhesion of Candid3 species was examined microscopically and revealed adherent Q3n3133 enmeshed in fibrin-like strands (Locci gt 31., 1981). Reinhart gt 3_. (1985) studied the adhesion of Q. 323% to glass and concluded that adhesion was not due to cell surface hydrophobicity. Factors that may govern adhesion of Q3n3133 to a variety of plastic surfaces were investigated by Klotz gt 31. (1985) . These authors found that the adhesion of Q. 31tig3ng, to plastic was governed predominantely by hydrophobic properties of both the yeast 62 and plastic surface. Kinetic analyses of the interaction between Q. 31bicans and plastic also revealed negative cooperativity determined by Scatchard and Hill plots, due to electrostatic repulsion. This was determined by altering the surface charge of Cand1da cells by selectively blocking amino and carboxyl groups or increasing the ionic strength of the assay solution, and conducting adhesion assays. The more positively charged yeasts adhered in greater numbers, suggesting that the more negatively charged cells were more repulsive to the negatively charged plastic surface. Electrostatic forces were, therefore, deemed to be minor compared to the hydrophobic forces, as the adhesion of Q. a1b1c3ns to plastic without theses pretreatments occurred at high frequency. The investigations of the adhesion of Q. 31big3ng to acrylic and various plastics described above have provided some information that could be useful to prevent or reduce the adhesion of yeasts to these surfaces in yiyg. However, further studies will need to be carried out to determine whether or not these i_n vitro phenomena are applicable to the situation in tilg. In the latter situation, numerous substances are adsorbed to these materials that probably alter their surface properties, and, therefore, the adhesion of yeast and bacteria to them. Saliva, which contains numerous components that adsorb to plastics and acrylic, has been shown to reduce 63 Q3n3133 adhesion to acrylic when it was applied to acrylic before use in adhesion assays (McCourtie 8 Douglas, 1981: Miyake e_t 31., 1986a: Samaranayake gt L” 1980). The adsorption of these components may also suggest that specific adhesin-receptor interactions are involved in adhesion to molecules adsorbed to plastics in host tissue. Likewise, all of the previously described experiments have used different growth media and culture conditions, which have been shown to influence Candida adhesion and hydrophobicity 1n ytttg (Hazen gt 31., 1986). Although it is likely that these properties are important in the adhesion of Q. a1btgans to various nonbiological surfaces, further studies are needed to determine the relevance of each in vivo. 0 co 0 of as oi testina ecos stems Because the succession, development and composition of the microbial flora of the GI tract has been the subject of several excellent reviews and books (Drasar 8 Hill, 1974: Finegold gt 31., 1983 a8b: Freter, 1983 8 1984: Moore, 1977: Savage, 1977 8 1980), a complete overview of these topics will not be possible here. However, before a discussion on the factors or bacteria that control Q. 31313333 populations in the gut can be made, it will be necessary to at least list the major features of the GI microflora. These include: the complexity of the GI microflora, the diversity of the 64 microbial species found in various niches of the GI tract, and factors that might influence the composition of the microflora. By beginning with this subject, it is hoped that an appreciation can be gained for some of the stringent details that are involved with, and which can influence, the study of the survival, dissemination, and pathogenesis of Q. 313133ng and other fungi in the GI tract. The following paragraphs were modeled after an excellent review by Freter (1983) and were supplemented to include the following generalizations: 1. It is generally believed that the intraluminal contents of the normal human stomach is relatively sterile, with only low numbers (5103 colonies per' ml gastric contents) of a variety of organisms being present (Finegold gt 3_., 1983b: Franklin 8 Skoryna, 1966: Giannella gt 31., 1972). Studies indicate that these low counts may well represent oral and ingested microorganisms since counts tend to decrease to undetectable levels several hours after eating (Draser gt 31., 1969: Franklin 8 Skoryna, 1966), which is most likely due to changes in gastric pH (Giannella gt 31., 1972). Despite the apparent harshness of the stomach to colonization by microorganisms, some investigators have noted that certain microorganisms can be seen attached to the epithelium and deep in foveae in histologic preparations viewed by light or electron microscopy (Savage, 1970). A similar descripton of the stomach flora of laboratory rodents has 65 also been detailed, but the microbial species associated with the gastric mucosa are very different (Savage, 1969 8 1980). It should be noted that the indigenous gut microflora, for any portion of the GI tract, probably differs from host to host, even within the same species (Moore 8 Holdeman, 1974) . Studies have shown, however, that a-hemolytic streptococci, anaerobic cocci, lactobacilli, §t3pny1gggggng 33133331313, and Q. 31big3ng make up the vast majority of microorganisms that colonize the stomach (Franklin 8 Skoryna, 1966: Giannella gt 31., 1972). The microflora of the small intestine in the normal individual or laboratory animal, in contrast, contains numerous bacteria that are highly dependent on the location of sampling. Generally, the microbial counts range from being sterile (or very low) to very high as samples are taken from the duodenum and descending to the ileum, respectively (Finegold gt 3_., 1983b). Several studies have revealed that in the upper small intestine of normal healthy human subjects, low counts (0-104'5 colonies per ml) of both aerobic (streptococci, staphylococci, lactobacilli, and yeasts) and anaerobic (streptococci and lactobacilli) microorganisms, and almost a complete absence of coliforms and 33c_tflo_1d_eg species (Finegold gt 31., 1983). As samples are taken farther along the small intestine, at the distal ileum, ‘mean 3.5 6.5 bacterial counts increase: significantly (10 - 10 colonies per ml) and the microflora more closely resembled 66 colonic flora with higher counts of coliforms and fi3gtgtgiggg species (Finegold gt 31., 1983b: Plaut gt 31., 1967). The large intestinal flora of the healthy host consists of an average of at least 1011 colonies per gram which contains some 400-500 (mostly strictly oxygen- sensitive anaerobes) separate bacterial species from more than 45 genera (Draser 8 Hill, 1974: Holdeman gt 31., 1974: Moore, 1977: Savage, 1977). Finegold gt 31. (1983b), in their review of the "normal" indigenous microflora, have analyzed and summarized over 10 years of indepth studies characterizing the colonic microflora from several hundred individuals with various disease states and who were on various diets. These authors noted that there was a remarkable similarity between the major bacterial groups in general despite major differences in diet, disease state, and nationality. Table 3 summarizes the mean bacterial counts and percentages of specimens positive for the major bacterial groups present in the fecal flora as described in these studies. Several factors that may alter the composition or metabolic activity of the microbial flora of the gut include stress, diet, prolonged antibiotic and immunosuppressive therapy, surgery, the individuals genetic background, and a number of infections and debilitating diseases. All of these have been commented on elsewhere (Caugant gt 31. , 1981: Draser gt 31., 1969: Freter, 1976: Gerhardt 8 Iglewski, 1976: Gracey gt 3_., 1974: Hartley gt 31., 1978: Moore, 67 Table 3. Major bacterial groups present in human fecal flora % Loglo number of organisms per Bacterial of total gram (dry weight) group specimens positive Mean Range ngtinomyges 84.2 9.2 3.8 - 11.1 Anaerobic cocci 94 10.5 4.1 - 13.4 Arachmia- 9 9.0 4.3 - 12.0 Etopionobacterium Bacteroides 99 11.3 9.5 - 13.8 Bifidobacterium 74 10.2 5.0 - 13.6 0 'dium 100 9.8 3.8 - 13.1 Eubacterium 94 10.7 5.1 - 13.5 Fusobactgrium 18 8.5 5.1 - 11.0 Gram-negative 98 8.5 4.0 - 10.0 facultative anaerobes Other 93 6.5 1.0 - 11.2 facultative anaerobes §treptocgccus 99 8.9 4.0 - 12.8 *Adapted from Finegold gt 31. (1983b). 68 1977: Moore gt 31., 1981: Moore 8 Holdeman, 1974: Savage, 1977), but several of these may predispose individuals to colonization and infection with Q3n3133 (Odds, 1988). 2. The indigenous microflora of a given location within the GI tract, including various microsites within that location, is stable in its composition because the GI tract is an ecosystem in the climax stage (Zubrzycki 8 Spaulding, 1957 8 1962) . Thus "invading" microorganisms are usually prevented from colonizing the gut, unless, of course, they enter this site before the indigenous microflora has become established or after the microflora has somehow been disrupted (Dubos gt 31., 1963: Finegold gt 31., 1983a: Freter, 1983). This phenomenon has been described using several terms (e.g., bacterial antagonism [Freter, 1956], bacterial interference [Dubos, 1963] and colonization resistance [van der Waaij gt 31., 1971]), and constitutes a potent host defense mechanism (Freter, 1983). However, as noted by Freter (1983), although the gut microflora is stable in its function, the stability of individual bacterial species is not absolute. Consistent with this is the finding that the composition of the colonic microflora varies considerably in different hosts (Moore, 1977: Moore gt 31., 1981: Moore 8 Holdeman, 1974). Furthermore, it has been noted that, due to unknown factors, the population size of some indigenous bacterial species apparently may change from day to day within the same individual (Moore, 1977: Moore gt 31., 1981: Moore 8 69 Holdeman, 1974). One relevant example, as has been noted by Freter (1983), is that some individuals harbor no recognizable "resident" strains of Eggngtignig 3311 but only a frequent succession of different "invader" or "transient" E. 3311 strains (Anderson gt 31., 1973: Caugant gt 3_., 1981: Gerhardt 8 Iglewski, 1976: Sears 8 Brownlee, 1952: Sears gt 31., 1949: van der Waaij gt 31., 1971), and there are occasional individuals ‘who .allow "invader" strains to transiently coexist with "resident" strains (Anderson gt 31., 1973: Caugant gt 31., 1981: Gerhardt 8 Iglewski, 1976: Sears 8 Brownlee, 1952: Sears gt 31., 1949: van der Waaij gt 31., 1971). A similar situation may exist for Q. albicans (Abdelghaffar 8 Russell, 1979), but no definiative studies have been conducted to follow up this isolated observation. Nevertheless, one major feature of the gut microflora is its ability to resist implantation by allochtonous microbial species. 3. There are two general habitats of colonization in GI ecosytems, namely the lumen contents and the mucosa (Freter, 1983: Savage, 1977). This has been observed in several animal species and in man, and it has been found that both of these sites may contain common and/or unique microbial species (Freter, 1982 8 1983: Savage, 1970 8 1980) . Furthermore, within the mucosa there may be two micro-sites, which include those cells directly attached to the epithelium and cells entrapped in what is usually 70 described as mucus (Freter, 1982 8 1983: Kennedy gt 31., 1987: Savage, 1983). Recent studies indicate further that these micro-sites are closely interrelated in that a primary layer of bacteria "attaches" to the epithelium, and that several subsequent "layers" adhere to the primary one and to each other (Freter, 1983) . As will be noted below, for an area of the GI tract to function properly, at least in terms of resisting colonization to Qandida and other invaders, the mucosa-associated flora must remain intact. Inpgtt3nce of g3sttotntgst1nal colonizatign Iby Q3n3133 alhisans Colonization of the GI tract by Qandida species, and Q. 313133ng in particular, may play an important role in human health and disease since the GI tract has been implicated as an important source of infection in cases of napkin dermatitis (Rebora 8 Leyden, 1981), recurrent vaginitis (Miles gt 3_., 1977: Nystatin Multicenter Study Group, 1986) and systemic candidosis (Krause gt 31., 1969: Myerowitz gt 31., 1977: Stone, 1974: Stone gt 31., 1973 8 1974: Walsh 8 Merz, 1986). The over-growth of Q. 313133ng in the intestinal tract, and its subsequent passage through the gut mucosa into the host bloodstream, for instance, is believed to be the proximate mechanism leading to systemic candidosis (Stone gt (31., 1974), particularly in patients with acute leukemia (Meunier- 71 Carpenter gt 31., 1981: Myerowitz e_t 31., 1977: Ray, 1980). Similarly, colonization of the GI tract by Q. 313133ng and other 33133133 species may also lead to involvement in a number of other disease syndromes that include esophageal, gastric and intestinal thrush, gastric and intestinal ulceration, GI bleeding, diarrhea, peritonitis, perianal itch, napkin dermatitis, chronic "irritable bowel" syndrome, and auto-brewery syndrome (Odds, 1988). The importance of GI colonization by Q. 313133113, therefore, should not be underestimated, especially in view of the suggestion that "the most important source of Qandid3 species in human disease is endogenous" (Odds, 1988). On the other hand, GI colonization by Q. 31313333 may also have important beneficial effects on the host, in that it apparently results in a stimulation of the immune system leading to a protective response to systemic Candi d3 infection in the noncompromised host (Domer, 1988: Domer 8 Hector, 1987). The clinical recognition of candidosis of the GI tract has been traced by Odds (1988) to the late 1700's, which was long before Q. m was discovered (Odds, 1988: Rippon, 1988). Nearly a century elapsed from this date, when Q. 31b_i_g3ng was first isolated from a patient with candidosis of the stomach and colon (Odds, 1988: Rippon, 1988). The modern revival of the importance of GI tract colonization colonization by Q. 3112;33ng, however, had to wait nearly another century, and can probably be 72 credited, at least in part, to Krause and co-workers (1969) who demonstrated the rapid "persorption" of Q. 31313333 through the GI mucosa into the bloodstream when Krause drank a saline suspension containing 1010 viable Q. 313133ng. This was augmented by a number of studies that examined the role of the "normal" intestinal flora in inhibition of Q. alb133ng in experimental animals (Balish gt 31., 1974: Balish 8 Phillips, 1966 a8b: Clark, 1971: Hazen Q 31_., 1953: Huppert gt a_]._., 1955: Isenberg gt 31., 1960: Paine, 1958: Phillips 8 Balish, 1966) and to several clinical studies of the occurrence of 93n3133 gut colonization in various patient populations (Dewilde gt 31., 1982; Stone, 1974; Stone gt 31., 1973 8 1974). Widespread acceptance of this idea, however, had to wait a few more years until Myerowitz gt _1. (1977) showed a correlation between GI colonization by Q. 31b1c3ns and systemic infection in leukemic patients and to Miles gt 31. (1977) who showed a similar relationship to recurrent vaginitis. The latter findings, especially together with more recent studies (Balish gt 31., 1984: Ekenna 8 Sheretz, 1987: Field gt 3_., 1981: Hector 8 Domer, 1982 8 1983: Kennedy 8 Volz, 1983 8 1985 a8b: Kennedy gt 31., 1987: Lee 8 Balish, 1982: Pope gt 31., 1979: Umenai, 1978: Umenai gt 31., 1979 8 1980: Wingard gt 31., 1980 8 1982) , has new brought new prominence to the idea that colonization of the GI tract by Q. 3191mg may serve as an important reservoir for both superficial and systemic 73 candidosis (Odds, 1988). Recent data obtained from experimental animals reinforce this view and have shown that a number of fungi can spread systemically from the GI tract (Kennedy 8 Volz, 1983). Odds (1988) in his review on "Qandtda and candidosis" has examined in detail clinical and pathological features of GI candidosis, and has painstakingly summarized the incidence of GI carriage in "normal" and “patient" populations. Therefore, these topics will not be considered here. Nevertheless, two significant facts that are evident from this work should be noted before beginning a discussion on the survival and implantation of C 'da in the GI tract. (i) Not all healthy individuals, regardless of their age, sex, or the number of attempts (or methods) used to isolate "yeasts" from the GI tract, harbour a "resident" yeast flora. According to Odds (1988), about 40 to 50% of a given sample population can be shown to carry Candi d3 throughout most of the length of the GI tract. The incidence of Q3n3133 carriage, furthermore, has been shown to be highest in the oral cavity and lowest in the colon (Odds, 1988). Because Q. 31bi3ans can pass through the GI tract apparently without being "killed" (Kennedy 8 Volz, 1985a), the percentage of healthy individuals harbouring Q. 31t133ng in their gut is probably lower: since some positive fecal cultures may represent Q3n3133 removed from the oral cavity. Thus, common statements like "all of us acquire 74 Q3n3133 at birth", or "all of us carry Q3n3133 as a part of our normal flora", are unfounded. (ii) The incidence and population size of Q. 31t133ng in the gut, nonetheless, usually increases in individuals who are predisposed by illness, debility or a local reduction in host defense mechanisms (Odds, 1988: Rippon, 1988). Table 4 summarizes factors thought to remove, or at least lower, the "barrier" that inhibits the overgrowth of Q. 313133ng in the GI tract. Table 5 summarizes some recent data on the carriage of Candida species in the gut as related to diet. It is apparent from these brief comments that an overgrowth of Q. alnicans in the GI tract, and its subsequent pathologic consequences, could arise in a compromised individual either because of already harbouring a resident Q. a1bicans population or by acquiring a "hospital strain" because the barrier to implantation was reduced or eliminated altogether. There is good evidence for both conditions (Abdelghaffar 8 Russell, 1979: Carpenter, 1955: Cohen gt 31., 1969: Kennedy 8 Volz, 1985a: Pope gt 31., 1979: Shepherd gt 31., 1985) and it is not yet clear which is more relevant than the other. Nevertheless, both positions underscore that Q. 313133ng can become implanted in the gut either before the normal microflora has become established (e.g., during infancy) or after the ecological balance of the microflora has been disrupted (e.g., by antibiotic treatment). The 75 Table 4. General classification of factors thought to predispose to colonization and dissemination from the gastrointestinal tract by Q3n3133* Classification of predisposing Explanation Examples factors Natural 1. Infectious, Microbial idiopathic, infections, congenital or endocrine other debilitating dysfunctions, diseases and defects in disorders immunity 2. Digressions from Pregnancy, normal hormonal infancy or physiological status Dietary 1. Excess or deficiency Carbohydrate- of foodstuffs that rich diets, may alter the comp- vitamin osition of the micro- deficiencies biota or the metabolic activity of the flora 2. Intake of foodstuffs that may alter local immunity Mechanical 1. Trauma Stabbing 2. Prolonged stay in a Space flight closed environment Iatrogenic 1. Treatment with drugs Antibiotics, medical that alter the comp- cortico- osition of the micro- steroids biota or suppress host defense mechanims 2. Surgical procedures Bowel resections *Adapted and modified from Odds (1988). 76 Table 5. Isolation of Q3ndig3 31bicans and other yeasts in stool samples from healthy individuals on different diets % Loglo numbers of of total organisms per gram Diet Organism samples positive Mean Range Japanese Q. albicans 47 5.6 3.5 8.9 Candida spp. 13 6.3 3.7 8.8 Other yeasts 53 5.8 3.4 8.7 Strict Q. 31bicans 15 4.9 3.5 6.3 vegetarian Candida spp. 0 --- ------- Other yeasts 23 5.6 4.3 7.8 Vegetarian Q. albicans 0 --- ------- some meat Candida spp. 0 --- ------- Other yeasts 50 6.1 4.2 8.7 Western Q. 313133ng 14 5.4 3.6 9.4 Q3n§133 spp. 8 4.4 3.7 5.1 Other yeasts 31 5.2 3.6 8.1 *Data from Finegold gt _1. (1983b). 77 survival and implantation of Q3n3133 in the gut will, therefore, be considered from both views. Unfortunately, there is no study that has followed a large group of human subjects from birth to adult, for the specific purpose of monitering yeast flora population dynamics and/or the acquisition of new yeast strains. Therefore, the comments below will be derived almost entirely from data obtained from experimental animals. A number of animal models of GI colonization and dissemination by Q. albicans have been developed and used for studies of antifungal chemotherapy, colonization, mucosal association and invasion, and pathogenesis. These models have included several animal strains and species, and have included the use of both adult and neonatal conventional, specific pathogen-free, germfree, gnotobiotic, and athymic germfree or gnotobiotic animals (Guentzel gt 31., 1985). For obvious reasons, laboratory rodents, particularly mice, have been the most extensively used models described to date. No attempt will be made here to describe any of these models in any particular detail, but the interested reader is referred. to the review of Guentzel gt 31. (1985) who have described in some detail most of the currently available animal models for studying GI colonization and dissemination by (Q. 31t133ng. Certain salient features unique to a particular model that may aid in the identification and/or elucidation of the determinants of gut colonization or 78 invasion of the intestinal mucosa by Q. 313133ng, however, will be noted where appropriate. The following discussion focuses on the ability of Q. 31bicang to survive and implant in the gut and examines the effects various pre- or post-treatments have on colonization and systemic spread of C3nd1da from the GI tract. Several studies have been conducted over the past 35 years that make it impossible in the space given here to completely describe and compare the numerous data that has been generated on gut colonization by Q3n3133. Therefore, it became apparent at this writing that the only way to make such a comparison, especially when considering the wide variation in experimental parameters that has been used between laboratories, was to construct a table (not shown here) and compare directly all the available published data on the survival, implantation and dissemination of Q3n3133 species from the gut. The following generalizations regarding colonization and dissemination of Q. 313133ng and other Q3n3133 species from the GI tract are the result of this effort. The foremost generalization that is evident from a comparison of this data is that the indigenous microflora represents one of the most important (if not the most important) host defense mechanisms suppressing the colonization of Q3n3133 species in the GI tract. This is clearly demonstrated from several studies that showed conventional animals were significantly more resistant to GI colonization than were 79 germfree and antibiotic-treated animals (Andermont gt 31., 1983: Balish gt (31., 1984: Ekenna 8 Sherertz, 1987: Guentzel 8 Herrera, 1982: Helstrom 8 Balish, 1979: Huppert gt 31., 1955: Kennedy 8 Volz, 1983 8 1985 a8b). Another observation that is consistent with this view is that neonatal animals, which lack a complete indigenous microflora, are also readily colonized by Q. 313133ng (Cole gt Q“, 1988: Domer, 1988: Domer 8 Hector, 1987: Field gt 3_., 1981: Hector 8 Domer, 1982 8 1983: Pope gt 31., 1979). The few apparent contradictory results reported in the literature with regard to the effects of the normal intestinal microflora on GI colonization by Q3n3133, may be due to the animals selected for study or to other methodologic pitfalls that accompany the study of interactions between microogranisms in the GI tract. For instance, early findings of Balish and Phillips indicated that Q. a1bicans was established in "large numbers" in the gut of’ both. germfree and. conventional mice (Balish 8 Phillips, 1966: Phillips 8 Balish, 1966), may be explained by the finding that many strains of mice are unsuitable for research on intestinal floral interactions (Freter gt 31., 1979: Freter gt 31., 1983: Kennedy 8 Volz, 1985b). Such mice are characterized, as are most commercially available strains of mice, by a lack of strictly oxygen- sensitive anaerobes and therfore lack a true indigenous microflora (Freter gt 31., 1979: Freter gt 31., 1983: Phillips 8 Lee, 1984). Using a different strain of mice, 80 Balish and co-workers (Helstrom 8 Balish, 1979: Lee 8 Balish, 1981 8 1982) later reported that conventional animals were significantly more resistant to GI colonization by Q. M than were antibiotic-treated and germfree mice. Such data suggest further that the indigenous ‘microflora is also an important defense mechanism inhibiting the systemic spread of Q3n3133 from the lumen of the gut to other organs. Animals treated with certain antibiotics, for instance, showed that large populations of Q. albicans colonized their GI tracts, with a corresponding high incidence of dissemination to visceral organs (Kennedy 8 Volz, 1983 8 1985 a8b). Animals that posses a complex indigenous microflora, in contrast, had low numbers of Q3n3133 organisms residing in their GI tracts, with a very low percentage or none of the animals showing signs of Candida dissemination. The finding that not all antibiotics predisposed mice to GI colonization or dissemination, may suggest that only certain components of the microflora are primarily responsible for the suppression of Q3n3133 in the gut. This will be discussed in more detail below. It also appears that an intact immune system may also be involved in the suppression of both colonization and dissemination of Q3n3133 from the GI tract. This can be seen from the studies of Ekenna and Sherertz (1987) who showed that pretreatment with cyclophosphamide, or 81 cyclophosphamide and cortisone acetate in combination, predisposed mice to cecal colonization by Q. 313133ng. It should be noted, however, that it has also been shown that mice possessing a competitive bacterial flora are more effective than mice with an intact immune system in suppressing colonization with this fungus (Helstrom 8 Balish, 1979). Nonetheless, it may be that local antifungal immunity may act synergistically with bacterial antagonism in controlling Q. albicans populations in the intestine. This view is consistent with studies by Shedlofsky and Freter (1974) who showed synergism between ecologic and immunologic control mechanisms regulating bacterial populations in the intestine. Pretreatment of mice with a combination of antibiotics and immunosuppressive agents was found to be more effective than either antibiotics or immunosuppressive agents alone at predisposing the animals to cecal colonization by Q. M (DeMaria gt 31., 1976: Ekenna 8 Sherertz, 1987: Myerowitz, 1981). The possibility that the immunosuppressive agents used in these studies somehow modified the indigenous microflora is not known. ARTICLES INFLUENCE OF GROWTH CONDITIONS ON QANQIQA ALBICANfi ADHESION, HYDROPHOBICITY AND CELL WALL ULTRASTRUCTURE Michael J. Kennedy, and Ramon L. Sandin Published in: The J u na of d cal nd Ve co 0 26:79-92, 1988 Influence of growth conditions on 93313133 312133113 adhesion, hydrophobicity and cell wall ultrastructure Michael J. Kennedy1r2*, and Ramon L. Sandin2 1Microbiology and Nutrition Research. Unit, The Upjohn Company, Kalamazoo, Michigan, 49001, U.S.A., and 2Department of Botany and Plant Pathology, Michigan State University, East Lansing, Michigan, 48824, U.S.A. Running Title: FACTORS INFLUENCING Q. ALBIQANfi ADHESION *Correspondence address: Microbiology and Nutrition Research Unit, The Upjohn Company, Kalamazoo, Michigan, 49001, U.S.A. 82 83 ABSTRACT The effect of cultivation in 13 media (10 complex, and 3 synthetic), as well as altering growth conditions, on Q3n3133 albicans adhesion, cell surface hydrophobicity, and cell wall ultrastructure was studied. Adhesion of Q. 313133ng_ to buccal epithelial cells (BECs) was significantly modified by all of the factors tested, particularly growth medium. In general, optimal adhesive activity for Q. albicans was observed when the cells were grown in defined media (depending on the carbohydrate used) and/or at 25°C. Mereover, significant differences in adhesion to BECs were noted when Q. 31bicans was grown in the same complex medium from different manufacturers and in different batches of medium from the same manufacturer. Electron microscopy revealed significant differences in surface topography and cell wall ultrastructure of C_. alnicans grown in different media, but none of these differences, including presence or absence of an outer floccular layer, appeared to correlate with the adhesive changes noted, which raises questions regarding the location and nature of the Q3n3133 adhesin(s). Idkewise, cell surface hydrophobicity could not be correlated with adhesion to BECs, but may have influenced yeast ceadhesion. The results indicate that Q3n3133 adhesion is highly dependent upon the cultivation conditions of the yeast cells tested, and may explain discrepancies in the literature regarding the biochemical 84 nature of the surface component(s) responsible for Q. alhisans adhesion. INTRODUCTION The adhesion and association of Qandida 31bic3ns with various host tissues is considered to be an important event in colonization and pathogenesis (9,10,34). Consequently, numerous models (in yittg and in 2113) have been developed to study the adhesion of Q. 31313315 to mucosal surfaces, certain organs, and prosthetic devices (9). These models have led several authors to suggest that various surface components of this yeast mediate adhesion (9,34) . It has been suggested, for instance, that a mannose containing moiety, probably a mannoprotein (6,23), is the Candida adhesin (19,35-38), whereas analogous studies have also suggested a possible role for cell wall chitin as the adhesive component (39). Other studies suggest that there may be more than one adhesin (13,23,35), and at least one report suggests that Q3n3133 adhesion may not be mediated by a specific adhesin- receptor interaction, but. may be entirely nonspecific (31). One explanation for these different conclusions is the variations in experimental conditions used by different investigators: these include differences in the Q3n3133 strain or phenotype, growth and assay medium used, variation in growth conditions (e.g. , temperature), and 85 methods for isolation and preparation of test cells (9,12). However, while all of these factors may have influenced Candida adhesion data, growth conditions are probably the most important since the adhesiveness of microbial cells is directly dependent on the conditions under which the cells are propagated (9). Indeed, Douglas 8 McCourtie (7,21,22) have already reported that growth of Q. 31bigans in chemically defined media with different amounts and kinds of carbohydrates directly influenced Qandiga adhesion. While the importance of culture conditions, particularly growth media, on Can3id3 adhesion has been noted in the literature (7,8,14,16,28), it is somewhat surprising to note that over 15 different culture media have been used for propagation of Q. albigans for studies of epithelial cell adhesion (9). Of particular interest is the fact that very few investigators have used the same medium, and that there are major differences between media used by those investigators proposing that different cell surface components (adhesins) are responsible for Candida adhesion (9) . Consequently, we set out to examine the effects several media, as well as various growth conditions, had on Q3n3133 adhesion to buccal epithelial cells, cell surface hydrophobicity and cell wall ultrastructure. Furthermore, for reasons that will become apparent, we have discussed criteria and have suggested what, it is hoped, will be the initial selection of a defined basal medium for future development of a 86 standardized epithelial cell adhesion assay for Q. alhisans. METHODS Mucosal cells Buccal epithelial cells (BECs) were collected from 10 healthy adult volunteers by gently rubbing the inside of their cheeks with sterile cotton swabs and suspending the cells in 0.01M phosphate-buffered saline (PBS) at pH 7.2. The donors had no signs or symptoms of oral thrush and had not taken antibiotics for at least 12 months. Cells from the different donors were pooled, washed three times in PBS, resuspended to a concentration of 2 x 105 cells ml-1 of PBS, as determined by hemacytometer count and stored at 4'C until required (maximum 2 weeks). Preliminary studies showed that storage of BECs did not affect adhesion of Q3n3133, but controls were included throughout the 2 week experimental period to ensure that differences in adhesion were not due to the use of either fresh or stored BECs. Examination by light microscopy confirmed that none of the cell samples contained mucosal cells already colonized by yeasts. Yeasts, culture conditions, and cell preparation Q. 3131333g (AK785) , recovered from the oral cavity of a one week-old female with oral thrush, lyophilized and subsequently maintained on Sabouraud's dextrose agar (SDA: 87 Difco Laboratories, Detroit, M1) at 22°C to 25°C was used throughout the study: it was sub-cultured fewer than six times from the original isolation to minimize changes in adhesive or virulence properties. For all assays, Q. 313133ng was grown on SDA slants at room temperature for 4 days. A loopful of cells was then transferred to 100 m1 of test medium and the culture was incubated aerobically (temperature and other growth conditions as stipulated for each medium in Table 1) up to and past the early stationary phase (ESP). Stationary phase cells were used because they adhere more readily than logarithmic phase cells (14,16,38). One portion of the culture was then processed for electron microscopy as follows: Cells were harvested by centrifugation at 3,000 rev. min'1 for 10 min at 4°C, washed three times in PBS, and the pellet was gently resuspended in 3% glutaraldehyde in 0.1M cacodylate buffer (pH 7.2) and held at 4°C for 4 h to fix the cells. Following fixation, the cells were centrifuged, and the pellet was gently resuspended in the same buffer without glutaraldehyde and held at 4°C until they were processed for electron microscopy. Another portion of the culture was collected by centrifugation, washed three times in 7 PBS, and resuspended to 2 x 10 cells ml"1 so as to give a final yeast:buccal cell ratio of 100:1 for use in the adhesion assay. The remainder of the culture was then harvested by centrifugation and washed twice in PUM buffer (22.2 g KHPO4.3H o, 7.26 g KH 1.8 g urea, 0.2 g 2 4’ 2 2 88 MgSO4.7H20) (33) and used for hydrophobicity testing. Adhesion assay Adhesion of Q. alnic3ng to BECs was assayed as described previously (38). Briefly, 0.2 ml samples of BBC and yeast cells were placed into small test tubes (12 x 75 mm) and incubated on a rotary shaker (180 rev. min-1) for 1h at 37°C. Three tubes were used for each experiment. After incubation, BECs were collected and washed on polycarbonate filters (12 um pore size) (Nucleopore Corporation, Pleasanton, California). Cells were washed with approximately 100 ml PBS under continual (gentle) agitation. The filters were then stained with Gram crystal violet, and the number of Q. 313i33ng adhering to 200 BECs was determined by light microscopy (430x :magnification). Double-blind conditions were used. Results were read blind, without knowing which medium was being tested. C3ndida cells grown in T83 + G at 37°C (Table 1) were used as controls with each group of 4/5 media tested to compare adhesion values obtained with the same medium on different days. Phase-partition test for hydrophobicity A modification (13) of the phase-partition method of Rosenberg gt 31. (33), with hexadecane as the hydrocarbon phase (24), was employed to test for Q333133 hydrophobicity. The relative hydrophobicity was 89 determined by measuring the difference in absorbances of test and control cells, and determining the percentage of cells that entered the hydrocarbon phase. Electron microscopy For scanning electron microscopy (SEM), yeast cells fixed for 4 h in 3% glutaraldehyde in 0.1 M cacodylate buffer, were washed in three changes of buffer, and postfixed in 1% osmium tetroxide in the same buffer for 1 h. The cells were then washed with three changes of buffer, dehydrated by passage through an ethanol series, and critical point dried with a Ladd critical-point dryer. Specimens were coated under vacuum with gold:palladium (60:40) in a Hummer X sputter coater (Technics Inc., Alexandria, VA) and examined in a JEOL JSM-T300 scanning electron microscope at 15 kV. Variations in cell surface features of Q. albicans such as mode of budding, bud scar morphology, and surface topography, were examined with respect to growth conditions and compared to previously published reports (3,17). SEM was also used to confirm light microscopy observations on the predominant morphological cell type and the presence of short germ tube-like structures. The procedure used for transmission electron microscopy (TEM) was a modification of the "G-O-T-O-U" method described by Persi 8 Burnham (27). Briefly, Q3n3133 cells fixed in 3% glutaraldehyde in 0.1 M 90 cacodylate buffer were postfixed in 080 and tannic acid 4 and were washed three times in cacodylate buffer between each step in the schedule. After fixation, Q3n3133 cells were dehydrated by passage through an ethanol series, embedded in Polybed 812, sectioned with a Sorval MT 5000 ultramicrotome (DuPont Co., Wilmington, DE), and poststained with 2% aqueous uranyl acetate and Reynold's lead citrate (32). Sections were then examined in a JEOL- 1200 EX transmission electron microscope and photographs taken (at different magnifications) of several fields of cells from each culture. Average cell wall thickness was then determined from the micrographs by measuring the wall thickness at several different points for a number of cells. The presence. of a floccular cell wall layer (21,22), whether this was located uniformly or only amorphously around the cell, and the proportion of cells with such a layer was also determined from the micrographs as was the presence of an electron dense outer layer. Finally, the predominant morphological cellular forms, and whether the cells appeared to be germinating, were confirmed by TEM. All micrographs for SEM and TEM were coded, and were examined without knowledge of the growth conditions. 91 RESULTS Adhesion of Q. alnicang to BECs after growth on different media To determine the effect growth medium had on Qandig3 adhesion to BECs, Candida adhesion after growth in TSB+G was assigned a relative adhesion index of 1, and the adhesion of Q. albicans grown on the other 12 different media (Table 1) compared to it. The ability' of (Q. 31bi3ans to attach to BECs was found to be highly dependent on the growth medium used (Table 2). In general, Qangida cells grown in undefined liquid media at 37°C were usually less adhesive than those cells grown in defined media at the same temperature. However, for one defined medium (YNB) , the production of highly adhesive Q3n3133 cells was dependent upon the type of carbohydrate in the medium. For example, in assay 15, when glucose was substituted for the galactose in this medium (assay 16), adhesion was reduced. Moreover, significant differences in Q. 31bicans adhesion to BECs were noted when 33131133 cells were grown in the same complex medium (TSB) from different manufacturers (compare assays 1 8 12), and in different batches of medium from the same manufacturer (compare assays 11 8 12). The ability of Q. 31nig3ng to attach to BECs after growth in broth or on agar also was tested. It was found that Q. 31ti33ng grown on Sabouraud's agar at 37°C (assay 19) was significantly more adhesive than cells grown in Sabouraud's broth (assays 4 8 92 5). Similar results were obtained when the experiments were performed at 25°C, although the differences were not as marked (compare assays 6 and 20). A strong correlation was also found between Q. 31niQ3ng adhesiveness and the percentage of BECs to which Q. 313ic3ns had attached (Table 2) . As the number of 3333133 that attached to epithelial cells increased, so did the percentage of BECs which had attached yeasts. The percentage of BECs with one or more attached Q. 31nig3ng ranged from 28% to 94% after growth of Q. 313133313 in TSB+G-S (assay 2) and LBC-25 (assay 18), respectively. The ability of Q. albicans to associate with BECs indirectly, by attachment to already adherent yeasts, was also studied. It was found that yeast-to-yeast ceadhesion contributed to an increase in the total number of Q. 31313315, that attached to BECs after growth in 4 media, namely, LBC, BHI, SDA, and SDB (when grown at 25°C) (Table 2). Coadhesion was also noted. after’ growth in. TSB, however, as for adhesion, significant differences were noted, between medium from different manufacturers (compare assays 1 and 11) and between different batches of medium (compare assays 11 and 12). Effect of growth temperature on Q. 31bicans adhesion The adhesion of Q. 31bi3ang to BECs was significantly modified by the temperature at which the cells were grown. Growth at 25°C gave cells that were significantly more 93 adhesive to BECs than those grown in the same medium at 37°C (Table 2): compare assays 4 8 6, 17 8 18 and 19 8 20. Likewise, ceadhesion of Q. 313133ng cells was also found to be greater after growth at 25°C (e.g., assays 4 8 6). Thus, growth temperature, as well as medium, may influence Q. 31ni33ng ceadhesion. Effect of growth parameters on Q. 31m cell surface hydrophobicity To test the hypothesis that cell surface hydrophobicity may influence the adhesion of Q. 313133ng to BECs, the relative hydrophobic affinity of Q. 313133ng was determined. As is shown in Table 3, the strain of Q. 313i3ans used in this study proved to be relatively hydrophilic after growth in 12 of the 13 media tested (at either 37°C or 25°C), since on average only 510% of the cells entered the hydrocarbon phase. The only exception was Q. 31nig3ng grown in LBC at 25°C (assay 18) which was very hydrophobic with about 80% of the cells entering the hydrocarbon phase. Although these latter cells were also the most adhesive (Table 2), no direct correlation was observed between cell surface hydrophobicity and adhesion of Q. 31bic3ns to BECs. Indeed, at least one half of the total number of attached yeasts were attached indirectly due to yeast ceadhesion, with clumps of yeast cells attached to BECs gn mggg. Therefore, it seems likely that hydrophobic properties contribute little to adhesion 94 to BECs, but may instead be an important factor in promoting yeast ceadhesion: thereby contributing to the total number of Q. albicans attached to BECs by indirect adhesion. Effect of growth conditions on Q. 31m morphology, cell wall ultrastructure and surface topography The morphology and surface features of Q. 31213333 following growth on various test media were examined to determine if certain characteristics could be correlated with the ability of Q. albicans to adhere to BECs. The results are summarized in Table 4. Although a number of different. morphological characteristics ‘were associated with different experimental groups, only two factors appeared to correlate with increased yeast adhesiveness: namely, the presence of very small blastespores and yeast cells that appeared to be germinating and had short germ tube-like structures. To determine whether differences in cell wall ultrastructure or ‘various cell surface features could account for some of the differences in Q3n3133 adhesion to BECs, cells of Q. albicans were examined by SEM and TEM. Although differences were noted in cell wall thickness, number of distinct cell wall layers, and the presence of an extra electron dense cell wall layer by TEM (Fig. la), none of these differences appeared to account for either an increase (or decrease) in adhesion, ceadhesion, or cell 95 surface hydrophobicity. Furthermore, the presence or absence of a floccular material on the cell wall, which was previously shown to cause an increase in Q. 313133113 adhesion to BECs (7,22), did not appear to correlate with 513113133 adhesiveness. Such material was only found in large amounts on cells grown in YNB+gal (Fig. 1b), whereas smaller amounts were present on the cell surface of cells grown in LBC (at 37°C), YNB+G, and TSB+G (Fig. 1c). Growth of Q. albicans in the latter two media yielded cells that were relatively nonadhesive to BECs. Furthermore, this material occurred only amorphously on the cell surface and no other cells of Q. aibicang that were highly adhesive (e.g., assays 10, 12, 13, and 18) possessed such material. Indeed, some highly adhesive cells (e.g., assay 12) of Q. 313133ng appeared to have a relatively thin cell wall with no "extra" material or obvious adhesive appendages (Fig. 1d) . To ensure that differences were not due to extra cell wall layers being removed during the preparation of the cells for electron microscopy, two additional procedures (using different fixation methods or stains) were used to examine the cells (data not shown). The results were essentially the same as the results given in Table 4. Thus, the "floccular" material noted for cells grown in YNB+gal may contain only one type of adhesin, that is best expressed after growth in this medium. Examination of TEMs from cells with germ tubes or short germ tube-like structures revealed that 96 there were major differences in cell wall ultrastructure at the site of germination (Fig. 2). Finally, surface "smoothness" or "roughness," location of bud formation, and shape of the bud scar plug, as determined by SEM, did not appear to correlate with the level of adhesion, ceadhesion, or cell surface hydrophobicity observed for Q. 31ni33ng under the conditions tested (Table 4). DISCUSSION The findings presented in this paper clearly demonstrate the importance of growth conditions on the adhesiveness of Q. albicans in yittg. Growth of Q3n3133 at different temperatures, for instance, revealed that cells grown at 25°C were significantly more adhesive to BECs than cells grown at 37°C and that there was a higher percentage of BECs with attached Q3n3133. This was true for three of four media tested and is similar to the results of Lee 8 King (19) who showed that Q. 313i333g was significantly more adhesive to vaginal epithelial cells when Qandid3 cells were grown in PPG broth at 25°C compared to cells grown. in ‘the same ‘medium. at 37°C. Likewise, Segal gt 31. (39) also found that the percentage of vaginal epithelial cells with attached Q3n3133 was higher when Q. 31bi33ns was grown in YE broth at 28°C compared to cells grown in YE broth at 37°C. Other important variables that were observed to greatly influence Qandida adhesion in the present study included 97 differences in the type, batch, and manufacturer of the medium, whether growth was in liquid medium or on solid medium, and whether the medium was undefined or synthetic (the latter was also influenced by the type of carbohydrate used). One explanation for the effect of temperature on adhesion noted in this study may be the observation that at the lower temperature the cells with short germ tube- like structures, whereas at the higher temperature they did not (e.g., assays 4 and 6) or did so in lower numbers (e.g., assays 19 and 20). Indeed, this was true whenever Q. 31bi3ans was found to be highly adhesive, regardless of the growth temperature or medium used. The single exception to this pattern occurred when Q. 31big3ng was grown in LBC, in which highly adhesive yeast cells were produced that did not have germ tubes. The finding that most of the highly adhesive cells had germ tubes, therefore, is consistent with previous studies showing that germinated cells of Q. 3131331§ were significantly more adhesive than their non-germinated counterparts (15,38,40) . It is interesting to further note that for all media that yielded 38313133 cells that were highly adhesive to BECs, there was no other common morphological, ultrastructural, or physicochemical characteristic (of those parameters tested). Conceivably, then, the induction of germination, singly and collectively with other adhesive factors ( 9) , may have facilitated the 98 ability of Q. albiC3ng to attach to BECs. It also was found that the increased ability of Q. 31313333 to attach to BECs was accompanied by yeasts attaching to already adherent yeast cells. This ceadhesion occurred following growth in several media which yielded cells that were highly adhesive to BECs. Interestingly, yeast cells. grown in LBC at 25°C ‘were significantly more adhesive than cells grown in any other medium (at any temperature), manifested the largest amount of ceadhesion, and demonstrated significant cell surface hydrophobicity. Furthermore, although no direct correlation was observed between cell surface hydrophobicity and the adhesion of Q. 31315325 to BECs, the property appeared to contribute to the total number of bound yeast by promoting yeast ceadhesion. Coadhesion of Q3n3133 to yeasts and bacteria has been observed in 1113 (13,29), and should, therefore, be examined more closely to determine its role in colonization and pathogenesis of mucosal surfaces. The differences in adhesion of Q. 31313333 to BECs noted in this study raise questions regarding the location and biochemical nature of the Q3n3133 adhesin(s). For instance, at least two morphological classes of Q3n3133 adhesins have been noted in the literature, floccular and fibrillar (reviewed in [9]), both of which. have Ibeen observed to mediate the adhesion of Q. 313133ng in 2113 (2,4 ,25). Alternatively, it has also been shown that 99 adhesion of Q. ainicang to epithelial cells was mediated by cells that did not contain such outer surface layers (4) . Although it is not realistic to correlate directly the growth conditions of Q. 313i33ng E xitrg with those encountered in 2129. the present findings suggest that the presence of either floccular or fibrillar structures on the cell surface of Q. 31313315 does not, by itself, indicate that they play a role in adhesion to BECs. Cells grown in TSB+G (assay 2) , for instance, were found to contain an outer floccular cell wall layer and were relatively nonadhesive to BECs. It should also be noted that the chemical nature of the "floccular" layer noted between cells grown in various media may be very different. Therefore, some, but not all, such structures may contain Qandida adhesins. In addition, it is interesting to note that differences in cell wall ultrastructure (and presumably cell wall composition [9]) of Q. 313i33ng correlated with previous suggestions as to which cell wall component(s) may serve as the Qa_n_di_d3 adhesin(s) . As mentioned above, Segal and co-workers (39) have suggested that cell wall chitin, which is located at the innermost portion of the cell wall (30), may serve as the adhesive component. Cells of Q. 3&1;a_ns_ grown in the medium (YE broth) used by this group yielded cells that possessed the thinnest cell wall with the fewest apparent "layers." Growth of 3333133 in media (e.g. assay 16) that have been used in 100 studies suggesting that a mannoprotein on the surface of Q. 31bicans is the Q3n3133 adhesin (6,23), in contrast, yielded cells with thicker cell walls that possessed an outer floccular layer. It may be argued then, that at least three different adhesins exist on the surface of Q. 31bic3ns that mediate its adhesion to BECs: (i) a cell wall adhesin (demonstrated best when cells are grown in LBC), (ii) an adhesin in the floccular layer (YNB+galactose grown cells), and (iii) an adhesin expressed below the surface that becomes evident when the outer electron layer is thinned (YE grown cells). Consistent with the view that there is an adhesive component located below the outer layer(s) is the finding that adhesion of Q. 31bi3ans to BECs was inhibited by Concanavalin A (Con A), whereas after mannan extraction of Q3n3133 cells, Con A did not inhibit the residual, albeit lower, adhesion (35). Nevertheless, the variations in Q3 31ti33ng adhesion with changes in medium composition also highlight the potential of this organism for cell surface modification in vivo, and offers one explanation why Q. 31nig3ng has a greater ability to attach to tissues than other Candid3 species (5,16). In view of the findings presented in this paper, and the discrepancies regarding' which surface component(s) serves as the Q3ndid3 adhesin(s), the need for a reproducible and defined assay to study Q. 313133ng adhesion is very apparent. The most important variable to 101 be considered is that of the selection of growth medium. It is clear from this study that uniformity of adhesion results between batches of complex commercial media is unlikely. This is not surprising since peptone based media are highly complex and undefined, and consequently there are major differences in the chemical composition of peptones and "Sabouraud's medium" from laboratory to laboratory, from manufacturer to manufacturer, and from batch to batch from the same manufacturer (26). Therefore, it is proposed here that the medium of Lee gt 31. (18) or yeast nitrogen base (Difco Laboratories) supplemented with 500 mM galactose (21-23) be used because both media are chemically defined and produce cells of Q. 313icans that are highly adhesive. Several other factors which are important to the development of a standardized adhesion assay have been described in detail elsewhere (9,34). Nevertheless, further steps toward the standardization of Q3ndida adhesion assays, including determining the phenotypic state of the organism at the time of testing (12), may be necessary if the exact nature of the adhesion to BECs or other’ surfaces is to ‘be understood. Acknowledgments R.L. Sandin's work was supported by a post-doctoral Research Fellowship from Michigan State University. 102 REFERENCES 1. Anderson, M.L. 8 Odds, F.C. 1985. Adherence of Qa_n_31 8 £7: .. .2 R 8.5 82.. 0.1.2.. 88.... 2:. 8+8... 8.8.... 88> 2 an. R 8.5 ~28. m; 58.... 88> 3 a. i R d... “man. 0.... 8823255.... 2.8.... m. 2 4.... mos... Némh 58. 5... 88.5.; N. R .8. 3.09.... 38.. 58. .8 88.5.; : a 8.5 .852 E. .85.... :8. 88.. o. R 8.5 8.8. ms. 58. .8828... a 2. 8.... $22 2... 88.8... 8.8.5.8 83.. w a 8.5 8.8: m2. .5... 88...... 28.5.8 . a 8.5 $38 3-.-QO 58. 88...... 8.8.5.8 . R 8.5 282 2.08 ...8. 88...... 8.8.5.8 . 8.. .8 TN : a 8.5 9.88 7QO £8. 88...... 8.8.5.8 .. mm 8.5 .88. 20$... 88.... 81.5.. 5.. 8...»; n R 8.5 .88. 8-03.3 88...... £1.55 .8 8...; N a. .2. R 8.5 .88. o+mm~ 882.m $1.5... 5. 8...... . moon—0.5.3.— .AUoV deg noun—om .0: 53am :o_~a_>o.5Da 53002 .0: swan 2.030...— comuanzoc— 5:602 _ScoEtx—xm 8.33.. 88...... 32...? com u.=.oE.mE._= :95 :8 5.032383: 8&5... =8 628:9. .3333 0 9.3.3... .8 to... 283:...8 €320 ._ mam—«C. 111 TABLE 2. Adhesive properties of C. albicans grown on different media and under different environ- mental conditions. Mean no. % BECs with Assay Growth C. albicans Relative attached no. medium' 200 BEC : SD adhesion” Candida Coadhesion‘ l TSB+G 142: 32 1-0 36 - 2 TSB+G-S I73: 25 12 28 - 3 TSB+GZS 102: 38 0.7 32 - 4 SDBcl 152: 27 1-1 44 — 5 SDB—2 175: 22 12 34 — 6 SDB-l-25 622: 46 4-3 70 + 7 SMB 104: 36 0-7 36 - 8 FSM 1471: 36 IO 32 — 9 MB 156: 32 H 36 — l0 BR] 610: 14 4-3 82 + 11 TSB-l 260: 16 1-8 56 + 12 TSB-2 671:: 26 4-7 64 - 13 PPG 443: 53 3-1 70 + 14 YE 551-: 26 3-9 64 - lS YNB+G 156: 29 M 36 - l6 YNB+gal 759: 87 5-3 76 1 l7 LBC-37 579: 63 4-1 72 + 18 LBC-25 1233: 124 8-7 94 + 19 SDA-37 , 537: 60 3'8 70 + 20 SDA-25 752 1135 5-3 76 + 'Sec Table l for a complete description of growth medium and other experimental parameters. bAdhesion relative to that of cells grown in TSB+G (assay 1). Cindirect attachment to BEC mediated by yeast-to-yeast coadhesion. 112 TABLE 3. Hydrophobic properties of C. albicans grown on different media and under different environmental conditions Assay Growth Percentage of cells entering Relative no. medium' the alkane phase 2 SD hydrophobicity” l TSB+G 11-1: 0-1 1-0 2 TSB+G-S 7-1: 0-1 0-6 3 TSB+GZS 2-2: 3-0 0-2 4 SDB«1 7-2: 2-2 0-6 5 808-2 0 0 6 SDBol-ZS 4-2: 0-1 0-4 7 SMB O 0 8 FSM 8-8: 2-3 0-8 9 MB 0 O 10 EH] ”-2110! 13 11 TSB-l 12-01 0-1 1-1 12 TSB—Z 3-6: 5-0 0-4 13 PPG 11-1115-7 1-0 14 YE l3-7: 6-4 1-3 15 YNB+G 0 0 l6 YNB+gal 4-5: 0-1 0-4 17 LBC-37 116: 5-4 1.1 18 LBC-25 79-21: 6-5 7-2 19 SDA-37 12-0: 0-1 1-1 20 SDA-25 8-3: 0-1 07 'See Table l. l’Hydrophobicity relative to that of cells grown in TSB+G (assay 1). 113 TABLE 4. Morphology and surface characteristics of C. albicans grown on different media and under different environmental conditions Cell wall ultrastructure Surface topography‘ Average Electron cell wall Bud Assay Growth Floccular dense thickness Surface scar no. medium Morphology' layer layer (nm) detail Budding plug 1 TSB+G Sy, By - - 150 S MP Convex 2 TSB+G-S Sy, By + - 200 S MP Convex 3 TSB+st By, Sy — + 240 S MP Flat 4 SDB-l Sy, By - - 100 R MP Convex 5 508-2 Sy, By ND” ND ND ND ND ND 6 8084-25 Sy, Gt ND ND ND ND ND ND 7 SMB Sy, By - — 120 S MP Convex 8 FSM Sy, By - - 120 S MP+1 Convex 9 MB Sy, By - + 150 8 MP Flat 10 8H1 Sy, Gt - + 120 S MP Flat 11 TSB-l Sy, By - + 150 R MP Convex 12 TSB-Z Sy, By, Gt - - 100 S MP Convex 13 PPG Pm, Sy, Gt - - 180 S MP Flat 14 YE Sy, Gt - - 100 S MP+1 Convex 15 YNB+G By, Sy + - 300 S MP+1 Flat 16 YNB+ga1 Sy, Gt + + - 275 R MP Convex 17 LBC-37 Sy + - 140 S MP Flat 18 LBC-25 S - + 1 80 S MP Flat 19 SDA-37 VSy, Gt - - 120 S MP Convex 20 SDA-25 SB, VSy, Gt - - 100 S MP Convex +F1at 'Sy-single yeast cells; By-budding yeast cells; Gt-yeast cells with germ tubes or short germ tube-like structures; Pm-pseudomycelium; VSy-single yeast cells of various sizes and shapes; S-‘singlets’; SB-small (‘sticky’) blastospores. The predominate form is given first. ”Not determined. cS--smooth; R-rough, MP-multipolar, l-lateral. 114 .8: 8m I can ace—3523“ he _ 033. com £393 :8 8.. E8053 aqua—.253 30.8% .5 flflsfiézfifi 889.5“. :8: m2. min: 956.585? :2» =8 5 80:20.56 08 Z donut—.25“ 028:3 39:50 .8 E2?— =u3 :8. E36 0: £5 :9: =8 55 a nurses» Ndmh AB was 38.5 .5?— 31.68: .850 an 3505 m-0+flm._. A3 39:3 .5?— Hioooc .830 5 352: 3+? Any 39:3 .5»! 85v .8583 .850 an union—u nm0+mmh 3 80¢ Bantu: £00 :2 3:68 .2. On..- 0-0 2: Co aozwucfiofi a an 8:85. 352 3353 U .«o 8988 EE .8 Emacs—E ce_—Bom— ._ .OE _ c of o _ 115 .E: oomncam 659:3 2; 35 z... 2303.53.53 =9: =8 2: no .5388? 95 202 .8303; 3_.—.32 ES» 3: m an??? 352 $3.53 U Co :28qu 55 .8 2m... .N .OE VARIATION IN ADHESION AND CELL SURFACE HYDROPHOBICITY IN CANDlQA ALQLQAflfi WHITE AND OPAQUE PHENOTYPES Michael J. Kennedy, Alvin L. Rogers, Laurey R. Hanselman, David R. 8011, and Robert J. Yancey, Jr. Published in: co 0 102: 149-156, 1988 VARIATION IN ADHESION AND CELL SURFACE HYDROPHOBICITY IN CANDIDA ALBICANS WHITE AND OPAQUE PHENOTYPES O * 0 Michael J. Kennedyl's’ , Alv1n L. Rogersl’2'3, Laurey R. Hanselmen115I+, David R. 80114, and Robert J. Yancey, Jr.5 Departments of 1Botany and Plant Pathology, 2Microbiology and Public Health, and 3Medical Technology, Michigan State 4 University, East Lansing, Michigan 48824, Department of Biology, University of Iowa, Iowa City, Iowa 52242, and 5Microbiology and Nutrition Research Unit *(address for correspondence), The Upjohn Company, Kalamazoo, Michigan 49001 USA Running title: Adhesion of C. albicans phenotypes *Corresponding author +Present address: College of Osteopathic Medicine Michigan State University, East Lansing, Michigan 48824 116 117 ABSTRACT A previous study had established that a select group of pathogenic isolates of Candida albigggg was capable of switching heritably, reversibly and at a high frequency (10"2 to 10-3) between two phenotypes ("white" or "opaque") readily distinguishable by the size, shape, and color of colonies formed on agar at 25°C. This paper describes experiments designed to determine the ability of these two phenotypes to attach to buccal epithelial cells (BECs) and plastic, and to compare the cell surface hydrophobicities of white and opaque phenotypes from three clinical isolates. "White cells" were found to be significantly more adhesive to BECs, and a strong correlation was also found between phenotype adhesiveness and the percentage of BECs to which C. albicans had attached. The percentage of BECs with one or’ more attached C. albicans was approximately 90% for the white phenotype and approximately 50% for the opaque phenotype. "Opaque cells", in contrast, were twice as hydrophobic as white cells, and the percentage of opaque cells bound to BECs by coadhesion was also double that of white cells. The differences in adhesion to plastic between the two phenotypes were not statistically significant and there was no distinct trend to suggest which phenotype might be more adhesive to plastic. These results indicate that several factors are involved in the adhesion of _C. 118 albicans to plastic, and confirm the hypothesis that cell surface hydrophobicity is of minor importance in direct adhesion to epithelial cells but that it may contribute to indirect attachment to epithelial cells by promoting yeast coadhesion. Moreover, the data presented in this paper also revealed that under identical growth conditions, adhesion of C. albicans. was significantly altered depending on the phenotypic state of the organism tested. Therefore, because C. albicans can switch at a high frequency to various phenotypes in yitrg, it may be that in future adhesion studies involving Qandida the phenotypic state of the organism at the time of testing will have to be determined. Otherwise, the results, even within the same laboratory, may be difficult to interpret. 119 INTRODUCTION Recently, it was reported that a select group of pathogenic isolates of Candi g; M was capable of switching heritably, reversibly and at a high frequency -2 to 10-3) between two general phenotypes ("white" or (10 "opaque") readily distinguishable by the size, shape, and color of colonies formed on agar at 25'C (38). The differences noted in colony characteristics apparently were due to the dramatic difference in cell size, shape, and budding pattern. "White cells" are round to ellipsoidal and exhibit a budding pattern similar to most strains of Q. albicans; in contrast, "opaque cells" are elongate, or bean shaped, and exhibit a different budding pattern. While white cells undergo the bud to hypha transition under standard laboratory conditions, opaque cells do not (38). In addition, opaque cells exhibit an unusual pimpled pattern on the cell surface and express an opaque-specific antigen (J .M. Anderson and D.R. Soll, manuscript submitted for publication and unpublished observations). Because the white-opaque transition has now been observed in a number of pathogenic isolates obtained from several body sites (41), it may be that the ability of C. albicans to switch between various phenotypes (37) in 1119 may play a role in pathogenicity (38) . However, it is unclear what role the opaque phenotype might play since opaque cells are differentially 120 sensitive to the normal temperature of the human body, at least under laboratory conditions (38). It is possible that the opaque phenotype plays only a transient role in m, for instance, if it were more resistant to host defense mechanisms. Alternatively, it may be that the opaque phenotype plays a role in another habitat such as water or in the hospital environment. Such environments would foster a temperature range that would allow for maximal growth of opaque cells. Consequently, we examined cells in the white and opaque phenotypes of several independently isolated strains of C. albicans for adhesion to both buccal epithelial cells and plastic surfaces. Significant differences in adhesion. characteristics as well as cell surface properties were observed between these two phenotypes. This report decribes these differences, considers them in relation to recently observed differences in cell surface architecture (J .M. Anderson and D.R. Soll, manuscript submitted for publication and unpublished observations), and discusses the putative role white and opaque phenotypes may play in pathogenesis or alternative habitats. METHODS easts cu ture condit ons l e r 0 Three isolates of C. albicans exhibiting a white- opaque transition were recovered from patients with 121 systemic or vaginal candidiasis. These isolates were used throughout the study and were passed less than six times from the original isolation to minimize changes in adhesive properties. For all assays, Q. albigans cells were inoculated onto a modification (37) of the medium of Lee _e_t_ _a_]_.. (22) , designated here MLBC agar, and grown aerobically at 24°C for four to five days. A colony of cells was transferred to 100 ml of MLBC broth, which was then incubated aerobically with shaking (180 rpm) at 24°C. Cells were grown to stationary phase, and were selected for study because they have been shown to adhere more readily than logarithmic phase cells (18,19,35). One portion of the culture was then processed for adhesion to buccal epithelial cells (BECs). Yeast cells were collected by centrifugation at 3,000 rpm for 20 min at 4°C, washed three times in buffered KCl, and resuspended 7 to 2 x 10 cells/ml to give a final yeast:buccal cell ratio of 100:1 for use in the epithelial cell adhesion assay. Another portion of the culture was collected by centrifugation, washed three times in PUM buffer, which consisted of 22.2 g KZHPO4-3H20, 7.26 g KHZ 4, urea, 0.2 g 119804.711 0 and distilled water to 1000 ml 2 (29) , and resuspended to 5 x 106 1.8 9 cells/ml for use in an assay to measure adhesion to plastic. The remainder of the culture was then harvested by centrifugation, washed three times in PUM buffer, and used for hydrophobicity testing at a concentration of 1 x 108 cells/ml (17). All 122 cultures were examined by light microscopy prior to testing to confirm the phenotypic state of the organism. Mucosa; cells Buccal epithelial cells were collected by gently rubbing the inside of the cheeks of ten healthy adult volunteers with sterile cotton swabs and suspending the cells in 0.05 M KCl containing 1 mM phosphate, 1 mM CaClZ, and 0.1 mM MgCl2 ("buffered KCl": 4) at pH 7.2. This assay medium was chosen because it mimics the ionic composition of saliva (10). The donors were not suffering from signs or symptoms of oral thrush and had not taken antibiotics for at least 12 months prior to the present study. Cells were washed three times in buffered KCl and resuspended to concentrations of 2 x 105 cells/m1 of assay medium, as determined by hemacytometer count. This single, large batch of cells was used throughout the study to assure that differences observed in adhesion could be attributed to variables of yeast phenotypic parameters and not to changes or differences in the BECs (16) . When examined by light microscopy prior to experimentation, none of the cell samples collected from the different donors contained mucosal cells already colonized by yeasts. d e 'o ssa s The adhesion of C. albicans white and opaque 123 phenotypes to BECs was studied using a previously described assay (16). Briefly, 0.2 ml samples of BECs and yeast cells were placed into small test tubes (12 x 75 mm) and incubated on a rotary shaker (180 rpm) for 1h at 37'C. Three tubes were used for each experiment. After the incubation period BECs were collected and washed on polycarbonate filters (12 um pore size) (Nucleopore Corporation, Pleasanton, California). Cells were washed with approximately 100 ml of PBS under continual (gentle) agitation. The filters were then stained 'with Gram crystal violet and the number of C. albicans adhering to 200 BECs was determined. by light. microscopy’ at 430x. Double-blind conditions were used. The adhesion of C. albicans cells to plastic was studied using polystyrene microtiter trays containing 24, 16 mm-diameter wells (Costar, Cambridge, MA) as follows. A 0.5-ml sample of a suspension of white or opaque Qandida cells was placed in each well, the tray incubated for 1.5 h at 37°C without shaking, and the wells then washed three times with 1.0 ml of assay medium. The trays were inverted and allowed to dry overnight at room temperature. Adherent Qandida cells were counted by light microscopy at 100x. Five to ten 1 mm2 fields were counted per well, and the assay was performed in quadruplicate. Only the center area of each well was included in the counting procedure because several nonadherent Qandida cells were deposited at the outer edge of the wells during the drying process. 124 This procedure provided an efficient and reproducible method of quantitating the adhesion of C. albicans to plastic surfaces (M.J. Kennedy, A.L. Rogers, R.V. Thomas, P.A. Volz, and R.J. Yancey, Jr., manuscript submitted for publication). Phase-partition test f9; cell surface hydrophobicity A modification (17) of the phase-partition method of Rosenberg et a1. (29), with hexadecane as the hydrocarbon phase, was employed. to test for' Candida cell surface hydrophobicity. Briefly, yeast were grown as described, washed twice in PUM buffer, and resuspended to a final concentration of 1 x 108 yeast/ml. To round-bottom test tubes (12 mm-diameter), containing 2.5 ml of washed cells in PUM buffer, 0.5 m1 of hexadecane (Sigma Chemical Co., St. Louis, MO) was added. The suspension was preincubated at 37 ‘C for 10 min, then mixed on a vortex mixer for 2 min. After separation of the aqueous and hydrocarbon phases, the aqueous phase was measured at 400 nm, using a Spectronic 20 spectrophotometer (Bausch & Lomb, Inc., Rochester, NY). Control cells were put through the identical procedure without the addition of hexadecane to the system. The relative cell surface hydrophobicity was then determined by measuring the change in absorbance between test and control cells, and determining the percentage of cells that entered the hydrocarbon phase. This method was selected because of its ease, and because 125 it has previously been shown to give results similar to all other current methods of testing cell surface hydrophobicity (20,26). RESULTS Adhesion of C. albieans phenotypes to BECs To determine the effect the phenotypic state of Q. elbieans had on adhesion to BECs, three individual isolates of Q. albicans (WG-1, BL6, and 6a) were grown under identical growth conditions in either the white or opaque phenotype (38) and tested by the E gigs; assay described in Materials and Methods. To assure that differences were due to phenotype induced changes and not to differences in epithelial cells (7,13,19,33,34), a constant and standardized pool of BECs was used throughout the study (16). The ability of Q. elbicegs to attach to BECs was found to be highly dependent on the phenotype used. These results are summarized in Table 1. For all three isolates, the white phenotype was significantly (p<0.001) more adhesive to BECs than was the opaque phenotype. A correlation was also found between phenotype adhesiveness and the percentage of BECs to which Qengige cells had attached (Table 1). The percentage of BECs with one or more attached Qengige cells was approximately 90% for the white phenotype and approximately 50% for the opaque phenotype. In general, as the mean number of 126 attached Qandida cells increased, so did the percentage of BECs with attached Qandide cells. Nevertheless, the white phenotype attached to significantly more BECs than did the opaque phenotype. The ability of Candide cells to associate with BECs indirectly, i.e. by coadhesion to adherent Qendide cells, was also studied. It was found that yeast-to-yeast coadhesion contributed to an increase in the total number of Candida cells that attached to BECs for both phenotypes (Table 1). However, although the percentage of the total coadherent yeast population was low for both phenotypes, opaque phenotypes had approximately twice as many cells attaching to BECs in this manner. Adhe 'on of C albicans heno e to ast The adhesion of white and opaque Qengige cells to plastic was studied using a modification of a microtiter tray assay described by Klotz et a1. (20). In preliminary studies (M.J. Kennedy, A.lu Rogers, R.V. Thomas, P.A. Volz, and R.J. Yancey, Jr., manuscript submitted for publication) it *was found. that. two types of adhesion occurred in this system. The first was the direct adhesion of Candida cells to plastic, and the second was adhesion to the liquid-plastic interface. Consequently, as plates dry, cells floating freeLy in a thin layer of fluid are deposited at the outer edge of the wells. Therefore, to accurately determine the relative number of 127 Candida cells adherent to the plastic surface, only the center area of the wells were counted. Using this procedure, the adhesion of (Q. ‘eibieene phenotypes to plastic was determined. The results are summarized in Table 2. The differences in adhesion to plastic between the two phenotypes were not statistically significant for any of the isolates tested, and there was no distinct trend to suggest which phenotype was more adhesive to plastic. For instance, for the WO-l-isolate, the white phenotype was more adhesive to plastic than the opaque. In contrast, this trend was reversed for isolate BL6. Furthermore, for isolate 6a the cells of both the white and opaque phenotype attached in similar numbers. This suggests that several factors may be involved in adhesion of Q. elbicans to plastic surfaces. Effect of the phenotypic state on C. albicans ceii enzfeee nydzonhebieity To test the hypothesis that cell surface hydrophobicity may influence the adhesion of Q. einieene to BECs or plastic (11,20), the relative hydrophobic affinity of Q. elbicans phenotypes was determined. As is shown in Table 3, both the white and opaque phenotypes of the three isolates of Q. eibicans used in this study proved to be relatively hydrophobic after growth in MLBC at 24'C since about 50% of the cells entered the hydrocarbon phase. Furthermore, it was also noted that 128 the opaque cells were significantly more hydrophobic than the white cells, with at least twice as many cells adhering to the hydrocarbon phase. DISCUSSION It was recently demonstrated that most strains of Q. elbicens switch heritably and reversibly at high frequency between a number of general phenotypes distinguishable by colony morphology (37-39,41). There are at least three switching systems which are strain specific (40) . The "white-opaque transition" is perhaps the most interesting system since it appears to represent a phase transition and involves a dramatic effect on cell shape, cell size, gene expression, growth dynamics, budding dynamics, actin localization and cell wall morphology (38,39: J.M. Anderson and D.R. Soll, manuscript submitted for publication and unpublished observations). Strains which possess a white—opaque transition system are capable of switching back and fOrth indefinitely between two general 2 to 10-3 phenotypes at frequencies of 10- (39). The results presented in this report demonstrate that switching between the white and opaque phenotype dramatically influences the adhesive and cell surface properties of individual cells. White cells of three independently' isolated strains of Q. elnieene were significantly more adhesive to BECs than their opaque cell 129 counterparts. In addition, there was a higher percentage of BECs with attached Qengige cells of the white phenotype. This was true for all the isolates tested, and was similar to previous reports that showed a strong correlation to exist between Q. mile adhesiveness and the percentage of BECs to which Q. m had attached (16,36). In contrast, the opaque cells of each Q. albi cans isolate tested were far more hydrophobic than their white cell counterparts. The present data, therefore, clearly indicates that under identical growth conditions, adhesiveness, and other cell surface properties, can be significantly different depending on the phenotypic state of the organism tested. One factor which may be involved in the apparently superior adhesion of white cells to BECs is the differences in size and shape of the phenotypes. White cells are relatively round, whereas opaque cells are elongate, or bean shaped. The mean volume of a population of Q. albicans WG-1 white cells was shown to be 33 nm3, 3 (38). whereas the mean volume of opaque cells was 114 um Because curved bodies require less kinetic energy to overcome repulsive interactions during adhesion (12,14), such as between Q. 6.121% and epithelial cells (14,30), the smaller, rounder white cells may more effectively reduce the yeast-epithelial gap to allow binding to take place (14) . Indeed, it was previously shown that small blastospores of Q. el_b_i_cene were significantly more 130 adhesive to BECs than larger yeast cells taken from the same culture (16). However, because Q. elnieene is known to bind essentially irreversibly to BECs (14,15,30), and nonspecific binding is not strong enough to account for this type of adhesion (12), it is likely that specific adhesin-receptor binding was also involved. Although both factors are important in adhesion to epithelial cells, the latter is probably more important since even mild fluid shear can prevent adhesion by dislodging Qengjnie cells that are "attached" only by nonspecific mechanisms (14). While the data presented here do not allow definitive conclusions to be drawn regarding the increased adhesiveness of white cells to BECs compared to opaque cells, there are, nevertheless, at least two possible mechanisms which could account for this difference. (1) There may be major changes in the organization of the cell wall of these phenotypes (8,14) that allows cells of the white phenotype to produce more or different adhesins that increase their adhesiveness. There is some evidence to suggest that Q. albicans does produce more than one type of adhesin (17,24,32), but it is not clear if these types are phenotypically regulated (25) or whether a cell can produce two distinct types of adhesin simultaneously. Although there is no evidence indicating that differences in the composition of the wall exist between white and opaque Candide cells, both scanning and transmission electron microscopy have revealed dramatic differences in 131 wall morphology (J.M. Anderson and D.R. Soll, manuscript submitted for publication and unpublished observations). The function of "pimples" on opaque cells is still unknown, but such dramatic differences in wall morphology may indeed be related to the difference in adhesion delineated by the BEC adhesion experiments and the hydrophobicity measurements. (2) Because ultrastructural and surface characteristics of white and opaque cells are different, other' surface features (e.g., charge) ‘might also be different. Such differences may allow white cells to favor adhesion to BECs by further reducing repulsive interactions that occur during adhesion between cells (14,30). Moreover, such changes may also allow surface molecules adjacent to adhesins to bind nonspecifically to BECs and thereby facilitate and strengthen the adhesion (14) . Factors involved in the adhesion of Q. elbjeane to plastic surfaces were previously investigated by Klotz e; ei, (20), who found that hydrophobic ,properties predominately governed Candide-plastic interactions. However, kinetic analyses also revealed negative cooperativity due to electrostatic repulsion (20). Thus, if cell surface hydrophobicity was the only factor involved in the adhesion of Qengjge to plastic, opaque cells should have attached in higher numbers. The present study, therefore, corroborates the view that other factors are involved in the adhesion of Q. e_Lkflgene to plastic. 132 Factors other than cell surface hydrophobicity that may have influenced the attachment of Q. eifiieene phenotypes to plastic include cell size and surface charge (14,20). Previous studies have also shown that Q. EEQEIEQLLE was more hydrophobic and attached in greater numbers to these surfaces than did Q. gigging (20,26-28,31). In those studies SDB (at 37'C) was used for cultivation of Qendide cells, whereas in this study MLBC (at 24’C) was used. We have found that growth of Q. einieene under the latter conditions produced cells that were significantly more hydrophobic (16) . The observations of the present study, therefore, were also consistent with previous findings that Q. albicans is not very hydrophobic after growth in SDB (16,17,20,26-28), and confirms the suggestion that growth in the medium of Lee e; 31. (22) at 25°C produces cells that would attach to plastic in the greater numbers (16) . Furthermore, the differences in adhesion and coadhesion to BECs noted here for white and opaque phenotypes, corroborates the hypothesis that cell surface hydrophobicity plays only a small role in direct adhesion to BECs but that this property may be important in promoting yeast coadhesion. Opaque cells were generally twice as hydrophobic as white cells, and the percentage of opaque cells bound by coadhesion was also double that of white cells. In contrast, white cells were significantly more adhesive to BECs. Finally, it is worth considering whether the white- 133 opaque transition, the resulting differences in adhesion, or both, play a role in colonization and pathogenesis by Q. eibicans. To begin with, it should be noted that strains exhibiting a white-opaque transition have been isolated from systemic infections, vaginal infections and oral lesions (39) . The transition between white and opaque occurs frequently and spontaneously, and appears to involve the activation of opaque specific genes (J.M. Anderson and D.R. Soll, manuscript submitted for publication and unpublished observations). The transition also involves very refined changes in nearly every aspect of growth and cell architecture (38,39) , suggesting that the transition is a highly evolved form of hertitable, high frequency variability. It seems likely, therefore, that this transition does play some role in the pathogenic success of this organism. However, because white cells are significantly more adhesive to epithelial cells than opaque cells, and because white cells have selective growth advantage at 37°C (38), it seems likely that of the two phenotypes the white phenotype would predominate in yiye. Nevertheless, it is possible that. the opaque jphenotype ‘may' play a transient role during the initial phases of colonization, or that the opaque phenotype may be more resistant to irradication by antifungal chemotherapy (B. Slutsky, Ph.D. thesis, University of Iowa; D.R. 8011, M. Staebell, S. Eisely, and B. Slutsky, manuscript in preparation). It is 134 equally plausable that the dramatic alteration of cell surface properties may offer the opaque phenotype an increased resistance to host defense mechanisms similar to that of encapsulated forms of ‘szn§eeeeene, neefeznene (2,21). Alternatively, the opaque phenotype may predominate in some other environment. Indeed, Q. eibicans has been isolated from IV lines, hospital linens, soil, water, and toothbrushes (3,5,9) . Further studies will be necessary, therefore, to characterize the specific role(s) of the white-opaque transition in pathogenesis, and to determine which phenotypes are present in the host during colonization, invasion, and infection of various body sites. Differences in adhesive properties, nevertheless, may be a major factor in the alternative roles of white and opaque switch phenotypes. REFERENCES 1. Anderson ML, Odds FC: Adherence of Qeneide elnieene to vaginal epithelia: Significance of morpological form and effect of ketoconazole. Mykosen 28: 531-540, 1985. 2. Breen JF, Lee IC, Vogel FR, Friedman H: Cryptococcal capsular polysaccharide-induced modulation of murine immune response. Infect Immun 36:47-51, 1982. 3. Buck JS: Candide Me. In: AW Hoadley and BJ 135 Dutka (eds) a te i 'c t s s Asso witn Watez. ASTM STP 635. American Society for Testing and Materials, Philadelphia, 1977, pp 139-147. 4. Clark WB, Bammann LL, Gibbons RJ: Comparative estimates of bacterial affinities and adsorption sites on hydroxyapatite surfaces. Infect Immun 19:846-853, 1978. 5. Cook WL, Schlitzer RL: Isolation of Qeneide eieieene from fresh. water' and sewage. .Appl Environ. Microbiol 41:840-842, 1981. 6. Critchley IA, Douglas LJ: Differential adhesion of pathogenic Candida species to epithelial and inert surfaces. FEMS Microbiology Lett 28:199-203, 1985. 7. Davidson S, Brish M, Rubinstein E: Adherence of Candida albicans to buccal epithelial cells of neonates. Mycopathologia 85:171-713, 1984. 8. Douglas IJ: Adhesion of pathogenic Qeneiee species to host surfaces. Microbiol Sci 2:243-247, 1985. 9. Feo M: Supervivencia y desinfection de Qengiee eibieens en el cepillo de dientes. Mycopathologia 74:129- 134, 1981. 136 10. Gibbons RJ, Etherden I, Peros W: Aspects of the attachment of oral streptococci to experimental pellicles. In: SE Mergenhagen and B Rosen (eds) W anl Microbiai Aeheeien. American Society for Microbiology, Washington, 1985, pp 77-84. 11. Hazen KC, Plotkin BJ, Klimas DM: Influence of growth conditions on cell surface hydrophobicity of Qeneiee eibicans and Candida giabrata. Infect Immun 54:269-271, 1986. 12. Jones GW, Isaacson RE: Proteinaceous bacterial adhesins and their receptors. CRC Crit Rev Microbiol 10: 229-260, 1983 . 13. Kearns MJ, Davies P, Smith H: Variability of the adherence of Candida albicane strains to human buccal epithelial cells: Inconsistency of differences between strains related to virulence. Sabouraudia 21:93-98, 1983. 14. Kennedy MJ: Adhesion and association mechanisms of _aaL_C ida alums. In MR McGinnis (ed) WM Medical Myeolegy, vol 2, Springer-Verlag, New York, 1987, pp 73-169. 15. Kennedy MJ: Role of motility, chemotaxis, and adhesion in microbial ecology. Annal New York Acad Sci 137 506:260-273, 1987. 16. Kennedy MJ, Sandin RL: Influence of growth parameters on Candide EM adhesion, hydrophobicity, and cell wall ultrastructure. J Med Vet Mycol 26:79-92, 1988. 17. Kennedy MJ, Volz PA, Edwards CA, Yancey RJ: Mechanisms of association of Qeneige elbicane with intestinal mucosa. J Med Microbiol 24: 333-341, 1987. 18. Kimura LH, Pearsall NN: Adherence of Qeneiee eibieans to human buccal epithelial cells. Infect Immun 21:64-68, 1978. 19. King RD, Lee JC, Morris AL: Adherence of Qengige eibieans and other Qeneiee species to mucosal epithelial cells. Infect Immun 27:667-674, 1980. 20. Klotz SA, Drutz DJ, Zajic FE: Factors governing adherence of Qendide species to plastic surfaces. Infect Immun 50:97-101, 1985. 21. Kozel TR: Dissociation of a hydrophobic surface from phagocytosis of encapsulated and non-encapsulated Q;yp§ocoecus neofognans. Infect Immun 39:1214-1219, 1983. 138 22. Lee KL, Buckley HR, Campbell CC: An amino acid liquid synthetic medium for development of mycelial and yeast forms of Candida einieene. Sabouraudia 13:148-153, 1975. 23. McCourtie J, Douglas 1.1: Relationship between cell surface compositian of Qeneiee elnieene and adherence to acrylic after growth on different carbon sources. Infect Immun 32: 1234-1241, 1981. 24. McCourtie J, Douglas LJ: Extracellular polymer of Candida einicans: Isolation, analysis and role in adhesion. J Gen Microbiol 131:495-503, 1985. 25. McCourtie J, Douglas LJ: Unequal distribution of adhesins within populations of QeLdiee eibieens. FEMS Microbiol Lett 27:111-115, 1985. 26. Minagi S, Miyaka Y, Fugioka Y, Tsuru H, Sunginaka H: Cell-surface hydrophobicity of Qeneiee species as determined by the contact-angle and hydrocarbon-adherence methods. J Gen Microbiol 132:1111-1115, 1986. 27. Minagi S, Miyake Y, Inagaki K, Tsuru H, Suginaka H: Hydrophobic interaction in Qeneige emieene and Qeneige tgepiealis adherence to various denture base resin materials. Infect Immun 47:11-14, 1985. 139 28. Miyake Y, Fuj ita Y, Minagi S, Suginaka H: Surface hydrophobicity and adherence of Candida to acrylic surfaces. Microbios 46:7-14, 1986. 29. Rosenberg M, Gutnick D, Rosenberg E: Adherence of bacteria to hydrocarbons: A.simple method for measuring cell-surface hydrophobicity. FEMS Microbiol Lett 9:29-33, 1980. 30. Rotrosen D, Calderone RA, Edwards JE: Adherence of Candida species to host tissues and plastic surfaces. Rev Infect Dis 8:73-85, 1986. 31. Rotrosen D, Gibson TR, Edwards JE: Adherence of Candida species to intravenous catheters. J Infect Dis 147:594, 1983 32. Sandin RL: Studies on cell adhesion and concanavalin A induced agglutination of Qandida albicans after mannan extraction. J Med Microbiol (in press), 1987. 33. Sandin RL, IRogers AL, Beneke ES, Fernandez MI: Influence of mucosal cell origin on the in xitro adherence of Candida albicans: Are mucosal cells from different sources equivalent? Mycopathologia 98:111-119, 1987. 34. Sandin RL, IRogers .AL, Fernandez MI, Beneks ES: 140 Variations in affiity to Candida albicans in 11:19 among human buccal epithelial cells. J Med. Microbiol (in press), 1987. 35. Sandin RL, iRogers .AL, IPatterson. RJ, Beneke ES: Evidence for mannose-mediated adherence of Qandida m to human buccal cells in 1113;9- Infect Immun 35:79-85, 1982. 36. Segal E, Lehrer N, Ofek I: Adherence of Candida albicans to human vaginal epithelial cells: Inhibition by amino sugars. Exp Cell Biol 50:13-17, 1982. 37. Slutsky B, Buffo J, Soll DR: High frequency switching of colony morphology in Qandida albicans. Science 230:666-669, 1985. 38. Slutsky B, Staebell M, Anderson J. Risen L, Pfaller M, Soll DR: "White-opaque transition“: A second high- frequency switching system in Candida albicans. J Bacteriol 169:189-197, 1987. 39. Soll DR: High frequency switching in Candida. In DE Berg and MM Howe (eds) Mobile DNA. American Society for Microbiology, Washington, 1988, in press. 40. Soll DR, Langtimm CJ, McDowell J, Hicks J, Galask R: 141 High-frequency switching in Candida strains isolated from vaginitis patients. J Clin Microbiol 25:1611-1622, 1987. 41. Soll DR, Slutsky B, Mackenzie S, Langtimm C, Staebell M: Two newly discovered switching systems in Candida albicans and their possible roles in oral candidiasis. In IC Mackenzie and CA Squier (eds) EMILE—9.2a]. Mucosa. Published for Dows Institute for Dental Research by Laegeforeningens Folarg, Denmark, 1987, in press. 142 Table 1. Adhesion of Qandida albicans white and opaque phenotypes to buccal epithelial cells (BECs). Strain Phenotype Mean no. % of yeast % of BECs yeast co-attached with yeast per BEC attached * SD WG-1 White 14.9 5.8 6.1 92.0 Opaque 2.1 3.3 12.2 50.0 BL-6 White 13.5 3.4 12.8 88.7 Opaque 1.9 4.7 27.5 31.2 6a White 18.2 5.8 13.8 96.0 Opaque 3.0 5.3 23.2 44.0 Table 2. Adhesion of Candida albicans white and opaque phenotypes to plastic. Strain Phenotype Mean no. yeast per mm2 * SD WO-l White 154.2 * 31.8 Opaque 94.8 i 41.3 BL-6 White 76.2 i 12.4 Opaque 110.8 * 26.0 6a White 62.7 i 30.1 Opaque 76.1 t 33.6 143 Table 3. Hydrophobic properties of Qandida albicans white and opaque phenotypes. Strain Phenotype % change in absorbance * SD WO-l White 45.2 i 1.1 Opaque 82.7 i 3.8 Opaque 80.0 * 0.5 6a White 40.3 * 1.6 Opaque 96.8 * 0.5 ENVIRONMENTAL ALTERATION AND EVIDENCE FOR TWO DISTINCT MECHANISMS OF m MS ADHESION TO PLASTIC Michael J. Kennedy, Alvin L. Rogers, and Robert J. Yancey, Jr. Submitted for publication to: Infection and Innnnity October 10, 1988 Environmental alteration and evidence for two distict mechanisms of Candida nininnng adhesion to plastic Michael J. Kennedy114*, Alvin L. Rogerslr2r3, and Robert J. Yancey, Jr.4 Departments of 1Botany and Plant Pathology, 2Microbiology and Public Health, and 3Medical Technology, Michigan State University, East Lansing, Michigan 48824, and 4The Upjohn Company, Kalamazoo, Michigan 49001 Running title: *Correspondence address: The Upjohn Company Microbiology and Nutrition Research 7922-190-MR Kalamazoo, Michigan 49001 144 145 ABSTRACT Few studies on the adhesion of Qandida mm to nonbiologic surfaces have been conducted in standardized systems. In this study, the adsorption of Q. ninignns to plastic was examined after growth in two chemically defined media, Lee-Buckley-Campbell (LBC) and yeast nitrogen base (YNB), using a standardized microtiter tray assay. Two types of adhesion were found to occur in this system: (i) direct adsorption to plastic and (ii) adsorption to the plastic-liquid interface. The adhesion of Q. albicans to plastic only was subsequently analyzed by binding isotherms, Langmuir isotherms, and Scatchard plots. The number of binding sites (N) and the affinity constants (K) were calculated. The K and N were twofold and fourfold higher, respectively, after growth in LBC compared to YNB. Scatchard curves for both LBC and YNB grown cells had negative slopes, which is supportive evidence for the view' that negative cooperativity is involved in the binding process. A comparison of different assay solutions gave similar results, with the solution given to dehydrated patients (D5.45) allowing for the highest K and the largest N. Additional experiments to examine the role of cell surface hydrophobicity (CSH) in adsorption to plastic were conducted using white and opaque phenotypes of Q. ninigans. These studies showed that there was no significant difference in the adsorption of these phenotypes to plastic, although the opaque 146 phenotype was significantly more hydrophobic. Adsorption but not CSH of both phenotypes was significantly greater in D5.45. Moreover, relatively hydrophilic mycelial forms of C. albicans were found to attach only when D5.45 was used as the assay medium. In addition, yeast phase cells, but not mycelial forms, were sensitive to reduction in adsorption by non-ionic detergents. These results suggest that there may be at least two distinct mechanisms of adsorption of Q. albicans to plastic, and that growth and environmental parameters can have a profound effect on this process. 147 INTRODUCTION The "attachment" of angign nininnng to various biomaterials and host tissues has been deemed an important step in the initiation of both superficial and deep seated candidosis (5). Adhesion of (Q. nlninnng, to various plastic surfaces that comprise catheters, prosthetic cardiac ‘valves, and other' protheses, for' example, ‘may allow this opportunistic fungal pathogen: (i) a direct entrance into the vascular system through a breach in vascular' patency, such. as is caused. by“ an indwelling catheter; (ii) to establish a reservoir for chronic inoculation and thereby dissemination of fungal cells; (iii) to damage tissue cells adjacent to colonized biomaterials; and (iv) to become more resistant to host defenses and antifungal therapy (5). The ability of Q. aibiQans to adhere to intravenous catheters, various plastics and acrylic surfaces has been documented (6,10, 13—15). Although biomaterials are being used with increasing frequency for tissue substitution (3), relatively few studies on the adhesion of Q. ninicans and other angidn species to plastic surfaces have been reported (5) . In one study, the adhesion of 931nm; species was examined microscopically after "colonized" catheters were removed from patients in an attempt to obtain information on the morphologic nature of the attached fungi (12). In another study, attached Candid; species were examined 148 quantitatively and were shown to adhere to polyvinyl chloride and Teflon catheters in large numbers in Mm ( 17) . Factors that might govern the adhesion of Qangign cells to plastic were also examined by Klotz g; 31. (10) and it was suggested that hydrophobic properties of the yeast and/or the plastic surface predominately governed the adhesion of Q. M to plastic. Other factors were also examined, but two different, nonstandardized assay methods were used. Furthermore, Sabouraud dextrose broth, which is an undefined peptone based medium that produces cells of relatively low cell surface hydrophobicity (CSH) (4,7,15), was used for cell growth (10). Therefore, in the present study the adhesion of Q. ainicnns to plastic was examined after growth in two synthetic media; one of which produces cells of high CSH (LBC) and one that produces cells of relatively low CSH (YNB). Using this approach it was found that two distinct mechanism are involved in the adhesion of Q. M to plastic. MATERIALS AND METHODS Fungi. Q. nibicans AK785, recovered from the oral cavity of a 1 week-old female with oral thrush, 1yophilized and subsequently stored in Sabouraud dextrose broth (Difco) supplemented with glycerol (30% final concentration) at -74°C, was used throughout the study 149 except where noted. The organism was subcultured fewer than five times from the original isolation to minimize changes in adhesive or virulence properties. In addition, five isolates recovered from the blood of patients with systemic candidosis were also used (BL-6, CA30, CA34M, WO- 1, and 6a). Growth conditions. Unless otherwise noted, yeast phase cells of Q. albicans were grown aerobically at 25°C in the medium of Lee _e_t; gi. (LBC; 11) or yeast nitrogen base (YNB; Difco) (13) supplemented with either 50mM glucose or 500mM galactose. Cells were grown to stationary phase (48 h), and were selected for study because they have been shown to adhere more readily than logarithmic phase cells (9). For growth of white and opaque cells, Q. ainignns cells were inoculated onto a modification (18,19) of LBC, designated here MLBC agar, and grown aerobically at 24°C for four to five days. A colony of cells was transferred to 100 ml of MLBC broth, which was then incubated aerobically with shaking (180 rpm) at 24°C. Cells of white and opaque phenotypes were grown to stationary phase. In addition, a mycelial form of isolate 6a arose spontaneously and was also tested. Cell preparation. After cell growth, yeasts were collected by centrifugation, and the pellet was gently washed twice in buffered KCl (1,2,5), phosphate buffer (pH 150 7.0), or a variation of one of these buffers. The cells were suspended to give a final concentration of 5 x 106 Candida per ml. A.range of cell concentrations (1 x 105 to 1 x 108 cells per ml) was also tested to study angidn binding kinetics. Adhesion assay. The adhesion of Q. ninians cells to plastic was studied using polystyrene microtiter trays (Costar, Cambridge, MA) containing 48, 11 mm-diameter wells as outlined in Figure 1. Briefly, a 0.5 m1 sample of a suspension of Candida cells was placed in each well, the tray incubated for 1 h at 25°C without shaking, and the wells then washed three times with 1.0 ml of assay medium. After incubation of the assay mixture and washing, the trays were inverted and allowed to dry overnight at room temperature. Adherent Q. ainiang were counted by light microscopy, without staining, at 100x. Five to ten 1 mm2 fields were counted per well, and the assay was performed at least six times. Only the center area of each well was included in the counting procedure because several nonadherent Q. nininnng cells were deposited at the outer edge of the wells during the drying process. Occasionally, clear patches with few or no yeast cells were observed around or near the center of the well, presumably due to static, and were not included in the areas to be counted. This method provided an efficient and reproducible assay for quantitating the adhesion of Q. 151 albicans to plastic (6). Phase-partition test for cell surface hydrophobicity. A modification of the phase-partition test of Rosenberg n; ni. (16) was used as described previously (8). RESULTS Influence of growth and assay media on adsorption. To determine the effect growth or assay medium had on the adsorption kinetics of Q. 113.1% to plastic, a wide range of yeast cells were examined for their ability to adsorb to plastic after growth in LBC or YNB and the cells were washed in one of several buffers. The number of Q. (nininnng, cells which adsorbed to the plastic when suspended in D5.45 varied significantly according to the growth medium used, although the concentration of cells at which saturation of the available binding sites occurred was the same (Figure 2). The concentration of Q. nininnng which saturated the available binding sites in this assay medium was 5 x 107. The number of bound yeast cells increased in a curviliner fashion, but a dip in an otherwise smooth curve became evident at concentrations beyond the equilibrium point. When these experiments were repeated in another buffer, namely PBS, the concentration at which the available binding sites became available was 1 x 107 cells per ml (data not shown). Nevertheless, the 152 adsorption isotherms followed the same curviliner pattern, dipping at concentrations above the equilibrium point. This appeared to be due to yeast cells aggregating during the assay. In all assays various degrees of yeast aggregation became evident at concentrations above 5 x 106 cells/ml, but the degree of yeast aggregation was dependent upon the growth and assay medium used (data not shown). Because of this, a cell concentration of 5 x 106 cells/ml was chosen as the standard cell concentration for the remainder of the studies. The number of binding sites (N) and the affinity constant (K) were calculated for each growth and assay medium (Table 1). The values for these adsorption parameters varied considerably depending upon the growth and assay medium used. Yeast cells grown in YNB were usually less adhesive than those grown in LBC, regardless of medium carbohydrate or assay medium. The K and N were two-fold and four-fold higher, respectively, after growth in LBC compared to YNB. A comparison of different assay solutions gave varying results, with the solution given to dehydrated patients (D5.45) allowing for the highest K and the largest N (Table 1). The product of the K and N for each growth and assay medium was also calculated and the KN for D5.45 was on average two-fold higher than for PBS. Similarly, the KN for Q. nlnignns grown in LBC supplemented with glucose or galactose were three- to ten- fold higher than those cells grown in YNB, respectively 153 (Table 1). These data were also prepared as a Scatchard plot to determine if positive or negative cooperativity influenced binding to plastic (Figure 3). It was noted that a dip in the curves for cells grown in LBC and YNB occurred regardless of the carbohydrate used. The bowing of each line in the Scatchard plot indicated that adsorption to plastic under the conditions used in the present study displayed negative cooperativity. However, at cell concentrations between 5 x 107 and 1 x 108 cells per m1, positive cooperativity was noted (data not shown). Influence of assay medium on adsorption. The effect of various assay media on adsorption of Q. nibiggng to plastic was determined and the results are shown in Table 2. The optimal assay medium was D5.45, and this was true regardless of the medium in which the cells were grown. Comparison of several assay media suggested that adsorption of Q. aibicans to plastic is a multifactoral process that can be manipulated by several factors. For instance, when LBC supplemented with 50 mM glucose was used as the growth medium, yeast cells adhered in significantly greated numbers when Mg+2 ions were added to D5.45. This trend held true when other assay media were employed in the adsorption assay (data not shown). A comparison of adsorption ability was also made by comparing the number of adherent yeast cells when 154 variations of a particular assay medium was used in the adsorption assay. D5.45, for instance, was split into its component parts (i.e., 5% glucose and .45% NaCl) and it was observed that adhesion of Q. alums AK785 was significantly lower when DS.45 was used as the assay medium. The ionic strength of the medium as well as the relative adhesion strength and average cell size are also listed in Table 2. From these data it appeared that as the ionic strength of the assay buffer increased, the adsorption of Candida cells to plastic decreased. However, because several factors can apparently influence the adsorption of Q. albicans to plastic, and because of the large differences between the assay media used, the influence ionic strength had on adhesion was examined more closely. Influence of ionic strength and ions on adsorption. The effect of KCl concentration on adsorption to plastic is shown in Figure 4. The buffer was 2 mM phosphate, pH 7.0. As is shown in the figure, adsorption of Q. Mn; to plastic decreased as the ionic strength of the assay buffer increased. The effect various concentrations of ions (Ca+2 and Mg+2) had on the attachment of Q. nihidnng to plastic is shown in Figure 5. All concentrations of calcium and magnesium ions increased adsorption to plastic. 155 Influence of detergents on adsorption. Adsorption of Q. niniddns CA34M to plastic was reduced by 90% and 60%, respectively, when Tween 80 (TW) or Triton X-100 (TX) were included in the assay suspension, whereas the inclusion of either sodium dodecyl sulfate (SDS) or cetyltrimethylammonium chloride (CTMAC) to similar preparations had no effect (Figure 6). A similar pattern was observed for Q. nlniddng CA30 in that the inclusion of TW or TX reduced adsorption to plastic, but SDS or CTMAC did not (Figure 6). In contrast, no reduction in adsorption was noted when mycelial cells of Q. m CA30 were used in the assay, regardless of the detergent used (Figure 6). Adsorption and cell surface hydrophobicity of white and opaque phenotypes. The adsorption of white and opaque phenotypes to plastic was determined in two different buffers. The results are summarized in Table 3. The differences in adsorption to plastic between the two phenotypes were statistically significant, but there was no trend to suggest which phenotype was more adhesive to plastic. For instance, for the WO-1 isolate, the white phenotype was more adhesive to plastic than was the opaque. In contrast, this trend was reversed for isolate BL6, whereas for isolate 6a the cells of both the white and opaque phenotypes attached in similar numbers. Furthermore, a mycelial form of isolate 6a attached to 156 plastic only when D5.45 was used as the assay medium. These results suggest that several factors may be involved in the adsorption of Q. ninidnng to plastic surfaces. To test the hypothesis that CSH is the primary mechanism of adhesion of mm cells to plastic, the relative cell surface hydrophobicities of Q. .alhisans white, opaque, and mycelial phenotypes were determined. As is shown in Table 3, both the white and opaque phenotypes of three isolates of Q. nlnidnns used in this study proved to be relatively hydrophobic after growth in MLBC at 24°C since on average 40% or more of the cells entered the hydrocarbon phase. The only exception to this ‘was for mycelial cells of isolate 6a, which were relatively hydrophilic. Furthermore, it was also noted that the opaque cells were significantly more hydrophobic than the white cells, with at least twice as many cells adhering to the hydrocarbon phase. In addition, there was no difference in the relative CSHs of the phenotypes when the assay was performed in PUM versus D5.45, with the single exception of white cells of isolate WO-l which were slightly more hydrophobic in D5.45 (Table 3). DISCUSSION Adsorption of microorganisms to nonbiological surfaces has been proposed to occur in two phases (5) . The first is the nonspecific, reversible phase, which is 157 achieved by reaching an equilibrium between the attraction of London-Van der Waals forces and the repulsion of electrostatic forces. During this initial phase, microorganisms are susceptible to environmental alterations which influence electrostatic and other nonspecific forces (5). The second or specific phase of adsorption involves the bridging of cells to the surface through the interaction of macromolecules and leads to firm adhesion (5). In the current report, as well as other studies on microbial adsorption, it is probably more appropriate to describe adsorption as occurring in three phases: (i) adsorption to the liquid-plastic interface, (ii) initial or reversible adsorption to plastic, and (iii) irreversible or "permenant" adhesion. The adsorption properties of Q. nlniddns to plastic in the present report were found to be very different under several environmental conditions tested as well as due to the phenotypic state of the organism at the time of testing. Collectively, the data presented here suggest that: (i) the reversible phase of attachment greatly influences the final adsorption of Q. ninidnng to plastic, since those environmental parameters which altered electrostatic forces interfered with irreversible attachment, (ii) several factors can modify adsorption of Q. niniddng to plastic, and (iii) Q. niniddng can attach to plastic by more than one mechanism. Repulsive electrostatic forces between Q. M 158 and plastic surfaces have been proposed to be minor in comparison with hydrophobic forces (10), suggesting that the adsorption of yeast cells to plastic should either be altered little by increasing the ionic strength or that this change should result in a greater adsorption by "salting-out" hydrophobic surface moieties (20). When the ionic strength of the assay buffer was increased, however, a reduction in the number of adsorbed yeasts was noted. In contrast, the addition of Ca+2 or Mg+2 ions to the assay buffer increased the adsorption of Q. dim-gs to plastic. This is in contrast to the findings of Klotz gt 51. (10) who found that increasing the ionic strength of the assay buffer or altering the yeast surface charge to be more positive led to an increase in the number of yeast cells that adsorbed to plastic. It is not clear, however, whether the differences noted from the present study and that of Klotz §§.§l- (10) are due to differences in the ”strains" of Q. 2.1m used or to differnces in experimental conditions. In the present study, the medium of Lee gt d1. (11) was used for cell growth, whereas in the study reported by Klotz and co-workers yeast cells were grown in SDB. Yeast phase cells of Q. Mans grown in these two media have been shown to very greatly in both their CSH and cell wall ultrastructure (7) . This alone could account for the differences noted between the results of experiments examining the influence of ionic strength on the 159 adsorption to plastic, in that cells grown in LBC are very hydrophobic whereas those grown in SDB are relatively hydrophilic (7,15). Indicating that there were fewer hydrophobic moieties to be salted-out in LBC-grown cells, that cells of Q. albidang that are very hydrophobic are more influenced by electrostatic interactions during binding, or both. The finding that there were little difference in the CSHs between yeast cells washed and suspended in PUM, PUM + NaCl, D5.45, 0.45% NaCl, and distilled water, but that there were differences in adsorption to plastic in D5.45 is certainly in agreement with this. Thus while these data tend to confirm the importance of hydrophobic properties of the yeast cell surface in adsorption to nonbiological surfaces (5,10), these data also suggest that other factors may be of equal or greater importance in adsorption depending upon the prevailing environmental parameter. The use of detergents in adsorption assays produced additional data concerning view and the role electrostatic and hydrophobic interactions may play in the adsorption of Q. dibicans with plastic. For instance, the adsorption of yeasts to plastic was not influenced by the presence of ionic detergents (i.e. , CTMAC or SDS) , but was reduced significantly by the presence of the non-ionic detergents Tween-80 and Triton X-100. The latter two detergents interfer with hydrophobic binding. Thus suggesting that CSH is the predominant force governing the adsorption of 160 yeast phase cells of Q. mm to plastic. However, while CSH is certainly a major adhesive force involved in the attachment of yeasts to plastic, it is not the only one. Studies with white and opaque phenotypes of Q. ninignns are consistent with this view. Thus if increased CSH alone were responsible for adsorption to plastic, opaque cells should have been found to attach in greater numbers. The finding that there was no distinct pattern to suggest which phenotype is more adhesive to plastic, although opaque cells were significantly more hydrophobic, confirms the suggestion that several factors collectively are responsible for adsorption of Q. dinidnns to nonbiological surfaces (6). One factor which may be involved in the superior adsorption of yeasts in certain buffers is the differences in size and shape of the cells. Cells that are relatively round in shape and smaller in size, for instance, may require less kinetic energy to overcome repulsive interactions during adsorption (5). In addition, it appears that there is more than one mechanism by which Q. aibicdns binds to plastic. A mycelial form of Q. dinigdnfi 6a, for instance, was found to be relatively hydrophobic in both D5.45 and PUM buffers (which should excentuate CSH) , but was found to attach in significantly higher numbers in D5.45. In addition, the adsorption. of a mycelial form of strain CA30 was not influenced by the the incorporation of either ionic or non-ionic detergents into 161 the assay. Collectively, these results suggest that there may be at least two distinct. mechanisms of adsorption. of ‘Q. dibicans to plastic, one of which appears to be hydrophobic in nature and another which may be mediated by an ion-bridge type of mechanism, and that growth and environmental parameters can have a profound effect on these processes. These data also underscore the great variability that can arise from the use of "simple" adhesion assays, and suggest that standardization of such assays as well as determining the phenotypic state of the organism at the time of testing may be required if meaningful results are to be obtained between and within laboratories (5,6). REFERENCES 1. Clark WB, Bammann LL, Gibbons RJ: Comparative estimates of bacterial affinities and adsorption sites on hydroxyapatite surfaces. Infect Immun 19:846-853, 1978. 2. Gibbons RJ, Etherden I, Peros W: Aspects of the attachment of oral streptococci to experimental pellicles. In: SE Mergenhagen and B Rosen (eds) W Qral Microbial Adhesion. American Society for Microbiology, Washington, 1985, pp 77-84. 3. Gristina AG: Biomaterial—centered infection: microbial 162 adhesion versus tissue integration. Science 237:1588- 1595, 1987. 4. Hazen KC, Plotkin BJ, Klimas DM: Influence of growth conditions on cell surface hydrophobicity of Qandida albicans and Qandida ginnzdta. Infect Immun 54:269-271, 1986. 5. Kennedy MJ: Adhesion and association mechanisms of Qandida albicans. In MR McGinnis (ed) W Medical Mycology, vol 2, Springer-Verlag, New York, 1987, pp 73-169. 6. Kennedy MJ, Rogers AL, Hanselman LR, Soll DR, Yancey RJ: Variation in adhesion and cell surface hydrophobicity in andidd_ dibicans white and opaque phenotypes. Mycopathologia 102: 149-156, 1988. 7. Kennedy MJ, Sandin RL: Influence of growth parameters on Qandida albicnns adhesion, hydrophobicity, and cell wall ultrastructure. J Med Vet Mycol 26:79-92, 1988. 8. Kennedy MJ, Volz PA, Edwards CA, Yancey RJ: Mechanisms of association of andidd dinidnng with intestinal mucosa. J Med Microbiol 24:333-341, 1987. 9. Kimura LH, Pearsall NN: Adherence of Qandida ditigdng 163 to human buccal epithelial cells. Infect Immun 21:64-68, 1978. 10. Klotz SA, Drutz DJ, Zajic FE: Factors governing adherence of Qandida species to plastic surfaces. Infect Immun 50:97-101, 1985. 11. Lee KL, Buckley HR, Campbell CC: An amino acid liquid synthetic medium for development of mycelial and yeast forms of Candida aibicans. Sabouraudia 13:148-153, 1975. 12. Locci R, Peters G, Pulverer G: Microbial colonization of prosthetic devices. IV. Scanning electron microscopy of intravenous catheters invaded by yeasts. Zentralbl Bakteriol Hyg (B) 173:419-424, 1981. 13. McCourtie J, Douglas LJ: Relationship between cell surface composition of Candida niniddng and adherence to acrylic after growth on different carbon sources. Infect Immun 32:1234-1241, 1981. 14. Minagi S. Miyake Y, Inagaki K, Tsuru H, Suginaka H: Hydrophobic interaction in Candida ninitdnn and Qandidn .LLQEigdlié. adherence to various denture base resin materials. Infect Immun 47:11-14, 1985. 15. Miyake Y, Fujita Y, Minagi S, Suginaka H: Surface 164 hydrophobicity and adherence of Qandida to acrylic surfaces. Microbios 46:7-14, 1986. 16. Rosenberg M, Gutnick D. Rosenberg E: Adherence of bacteria to hydrocarbons: A simple method for measuring cell-surface hydrophobicity. FEMS Microbiol Lett 9:29-33, 1980. 17. Rotrosen D, Gibson TR, Edwards JE: Adherence of Candida species to intravenous catheters. J Infect Dis 147 : 594 , 1983 18. Slutsky B, Buffo J, Soll DR: High frequency switching of colony morphology in Candida ainiQana. Science 230:666-669 , 1985 . 19. Slutsky B, Staebell M, Anderson J. Risen L, Pfaller M, Soll DR: "White-opaque transition": A second high- frequency switching system in Qandida altiQana. J Bacteriol 169:189-197, 1987. 20. Smyth CJ, Jonsson P, Olsson E, Soderlind O, Rosengren J, Hjerten S, Wadstrom T: Differences in hydrophobic surface characteristics of porcine enteropathogenic Eagnatidnia Qnii with or without K88 antigen as revealed by hydrophobic interaction chromatography. Infect Immun 22 : 462-472 , 1978 . 165 Table 1. Comparison of binding affinities and adsorption sites for Candida albicans after growth in different media or with different assay media Affinity Constant No. of binding Growth Assay . KN . . (K) SlteS (N) per well -2 Medium Medium (X105) (X103) (X10 ) LBC + glucose PBS 2.0 5.5 11.0 05.45 3.0 5.4 16.2 LBC + galactose PBS 2.5 6.2 15.5 05.45 4.1 9.0 36.9 YNB + glucose PBS 1.2 2.5 3.0 05.45 1.8 3.3 5.9 YNB + galactose P85 2.4 2.0 4.8 05.45 1.0 3.6 3.6~ 166 Table 2. Effect of assay medium on the adsorption of Candida albicans to plastic Assay Ionic Mean No. Candida / mmZ t 50 after growth in: Madium Strength LBC + 50 mM glucose YNB + 50 mM glucose Distilled dionized H2O 0 409.6 f 37.0 388.4 1’ 27.8 PBS 0.49 25893270 316.0: 17.8 PBS + Urea 0.52 238.0 1 34.4 257.6 t 21.4 PUM 0.38 418.0 f 27.6 399.6 1 27.8‘ 05.45 0.08 495.4 f 34.2 405.0 3 17.8 05.45 + Mg + 2 0.08 - 569.6 3 30.3 342.6 f 19.8 05 (5% glucose) 0 224.0 1' 25.8 459.6 1 29.6 .45 (45% NaCl) 0.08 309.6 1- 11.8 344.4 f 21.2 PUM + NaCI 0.52 193.0 1 38.4 178.4 1 24.6 P8 + Urea 0.38 230.4 2 27.4 216.6 1 16.8 Table 3. 167 Adhesion of Candida albicans phenotypes to plastic Mean No. Candida/mmZ i 50in: Strain Phenotype PUM 05.45 WO-l white 156.0: 10.9 395.8: 31.0 opaque 108.8:89 210.5:230 8L6 white 104.5 f 30.6 274.9 f 44.3 opaque 196.0 f 28.0 314.0 1’ 39.8 6a white 72.5 i 24.6 543.3 1: 39.2 opaque 88.0 i 29.5 N0 mycelial 10.3 i' 5.2 384.5 :1: 41.4 168 Figure 1. ADHESION ASSAY Candida growth 1. Incubation: 48h at 25°C. 100 rpm 2. Media: LBC and YNB supplemented with SOmM glucose or 500mM galactose lWash X3 and adjust cell concentration to 5 X 106/ml Assay solutions . B-KCI D5.45 H20 PBS PUM LD-hCaJN‘ Assay conditions 1. Incubation: 1h at 25°C, n0 shaking 2. Wash X3 48-Well 0 Plate 3. Invert and dry overnight lExamine by light microscopy Microscopy 1. Invert plates and examine under high power 2. Several fields at the center of each well are counted 169 Figure 2. Adsorption of Candida albicans to plastic after growth ln dlfferent media 0’. 6 - ,I \ .’ \. LBC + glucose Assay medium = P85 l.’ \ ‘ ' ‘ . . LBC 4- galactose YNB + glucose ‘ YNB + galactose Number of Candida albicans bound per well (X103) 1 t L l 1 l I 1x105 5x105 1x10‘ 5x10‘ 1x107 5x107 1x108 Cell Concentration 170 Figure 3 . SCHATCHARD ANALYSIS OF CANDIDA ALBICANS ADSORPTION TO PLASTIC AFTER GROWTH IN DIFFERENT NIEDIA 250 I _. Assay Medium = PBS I = LBC + glucose CI = LBC + galactose 200 q o = YNB + glucose 0 = YNB + galactose Bound/ 150 —1 Unbound (x104) 100 -I 50- Number of Candida albicans bound per well (x103) 171 Figure 4. Effect of KCl concentration on adsorption of Candida albicans to plastic Mean no. adherent cells/mm2 .01 .1 1 10 Concentration of KCl (mM) 172 Figure 5. Effect of Mg+2 concentration on adsorption Mean no. adherent cells/mm2 (x 102) of Candida albicana to plastic .1 1 10 100 Concentration of Mg+2 (mM) 173 Figure 6 . Influence of detergents on adsorption of Candida albicans to plastic 100 F .m— 90 L 80 —- 70 l— 7'. 60 1'? I , Percent of .j, i g , adhesion to 50 - ‘ .17.?" plastic i ,7; ,; 40 — 7x2 . _ “Ii/j . 30 - ', 20 — lo — E o - SDS CTMAC C M TX SDS CTMAC C TIN TX 505 CTMAC CA34M CA30 CA30 - PM AN ANAEROBIC CONTINUOUS-FLOW CULTURE MODEL OF INTERACTIONS BETWEEN INTESTINAL MICROFLORA AND QANDIDA ALELQANS Michael J. Kennedy, Alvin L. Rogers, and Robert J. Yancey, Jr. Published in: Wis 103: 125-134, 1988 An anaerobic continuous-flow culture model of interactions between intestinal microflora and Qandida ainicana Michael J. Kennedy,1’ Alvin L. Rogers, and Robert J. Yancey, Jr.4 Departments of 1Botany and Plant Pathology, 2 3 Microbiology and Public Health, and Medical Technology, Michigan State University, East Lansing, Michigan 48824 4 U.S.A., and Microbiology and Nutrition Research, The Upjohn Company, Kalamazoo, Michigan 49001 U.S.A. Running title: Models of interactions between intestinal bacteria and Qandida *Corresponding address: Microbiology and Nutrition Research, The Upjohn Company, 7922-190-MR, Kalamazoo, Michigan 49001 U.S.A. 174 175 SUMMARY The finding by earlier workers that Eadngzidnia lei suppressed the growth of Qandida ainidana in yittd or in gnotobiotic mice has led to numerous, erroneous conclusions regarding the identity of the organisms and mechanisms responsible for the suppression of Qandida in the gut. This is due, in part, to the fact that nearly all studies to date have not reflected interactions as they occur in the intestinal tract. This paper describes a series of experiments that establish that an anaerobic continuous-flow (CF) culture model of the ecology of the large intestinal flora reproduces interactions between bacteria and Candida as they occur in the large intestine. This was determined in the following ways. (i) Bacterial counts in CF cultures of conventional mouse cecal flora or human fecal flora closely resembled that found in the mouse intestine and human feces. (ii) Dense layers of bacterial growth that formed on the glass walls of the CF culture vessels resembled bacterial populations that colonize intestinal mucosa. (iii) Total and individual levels of certain metabolic end-products of the predominant anaerobic bacterial flora present in CF cultures coincided with those found in the large intestine of conventional mice or human feces used to establish the CF cultures. (iv) Q. ainidans was eliminated from CF cultures of mouse cecal flora at a rate similar to that of 176 untreated experimental animals. (v) Contents of CF cultures fed to antibiotic-treated mice redressed several cecal abnormalities, and suppressed Qandida populations to levels found in conventional animals. Thus, a number of complex ecological mechanisms were maintained in CF cultures which normally control Qandida populations in the large intestine. It is suggested, therefore, that the CF culture model should help to further define the mechanisms which control Q. albicans and other fungi in the intestinal tract, as well as define which components of the indigenous microflora are responsible for suppression of Candida in the gut. 177 INTRODUCTION Historically, the study of interactions between the indigenous microflora and Qandida altidana in the intestinal tract has been hampered by numerous problems. As one example, in an individual experimental animal, one is limited. to studying such. interactions at one time point, since, in most instances, the animal must be sacrificed to remove the specimen of interest. Furthermore, there is a limited range of experimental manipulations that are possible in the human or animal gut. To overcome these and other problems that come from using experimental animals, or' human volunteers (23), there have been numerous attempts to study Qandida- loacterial interactions in, yittd (16,17,29,30). ‘Unfortunately, however, these studies have also been beset with difficulties, and it is significant that none of these reflect interactions as they occur in the intestinal tract (9,10,22). Virtually all in yitrn studies have demonstrated suppression of Q. ainidana by a single bacterial species (22,27), which is not representative of the 400 to 500 different bacterial species that normally inhabit the intestinal tract (7,25,33,37). Moreover, suppression of Q. ainidana using the former method has been shown to be dependent on culture conditions (29) . Likewise, it has also been shown that certain intestinal bacterial species inhibitory to the growth of Q. ainigang 178 in yittd (16,17) were not inhibitory for Q. alnidana under normal in yiyd conditions (1,5,22). One may conclude with some certainty, therefore, that there is little similarity between the inhibitory mechanisms for Q. ainidang in these simple in yittd systems and those found in the intestine, and that suppression of Q. ainidana by a single bacterial species cannot be expected to reflect interactions of a complex indigenous microflora (22). Recently, an anaerobic continuous-flow (CF) culture model of the ecology of the large intestinal flora was developed that reproduces a number of bacterial interactions that occur in the large intestine of mice (14). For instance, it was found that mixed populations of mouse cecal bacteria in CF cultures were able to suppress "invader" bacterial populations (e.g., Qiostridium difficila or Eschetidnia lei) to levels similar to those found after being fed to conventional mice (12,14, K.H. Wilson, J.N. Sheagren & R. Freter, 1984, Abstracts 84th Annual Meeting of the American Society for Microbiology), and that contents of CF cultures fed to germfree mice redressed several germfree abnormalities (e.g., cecal size and mucosal histology) (14). The finding that this model system can reproduce a variety of bacterial interactions that occur in the murine large intestine makes it likely that the underlying mechanisms that control the bacterial populations in anaerobic CF cultures are similar to those operating in 1139 (12,14). 179 Because of this, and due to the shortcomings of the simple in xittd systems mentioned above, the present studies were designed to examine the use of an anaerobic CF culture model to study interactions between Q. ainidana and intestinal bacteria. METHODS Microorganisms Bacteria for the inoculation of CF cultures or for recolonization of antibiotic-decontaminated mice were obtained from the cecal contents of normal mice, or the feces of a healthy volunteer who had not taken antibiotics for at least 24 months prior to the study. The bacterial suspensions were prepared in a Coy anaerobic chamber (Coy Laboratory Products, Ann Arbor, MI) as described previously (22). Filtrates of cecal or CF culture contents were also prepared by a modification (22) of a previously' described. method (32). The isolate. of ‘Q. ai_l_o_i9_an_§ used was recovered from the blood of a patient with systemic candidosis (CA34). Qandida cells were grown aerobically (without shaking) in Sabouraud dextrose broth (SDB; Difco Laboratories, Detroit, MI) for 18 h at 37'C. Animals Swiss Flow DUB/kr mice, 2 to 3 months old, were used in all experiments. These mice were originally obtained 180 from a closed colony at Eastern Michigan University, and a colony was subsequently established at The Upjohn Company. It is important to note the source of the mice used for the present studies because all major breeders supply "specific pathogen-free" (SPF) mice which originated from germfree stock and are unsuitable, therefore, for research on intestinal floral interactions (21). Such SPF animals lack strictly oxygen-sensitive anaerobes (13,14) and intestinal mucosa-associated populations (31), which predominate in the intestinal tracts of conventional animals (3,7,25,33,37), and are usually characterized by a flora containing environmental and handler microorganisms ( 13,14) . To avoid these problems, mice were selected which contain a conventional intestinal microflora, and were not started from a germfree or SPF commercial stock. Continuous-flow (CF) cultures The CF culture apparatus described by Freter at al. (14) was used (Fig. 1). The growth medium was veal infusion broth (Difco) supplemented with 5 g l.1 yeast 1 hemin (ICN Nutritional 1 extract (Difco), 100 mg l- Biochemicals, Cleveland, OH), and 0.5 mg l- menadione (ICN Nutritional Biochemicals), and pumped into a growth tube containing a 7-ml culture at the rate of 1.17 ml h-1 by means of a Manostat Cassette Pump (Manostat, New York, NY). This flow rate allowed fresh broth to be added to the 181 culture at a rate equivalent. to» displace 1/6 of the culture volume per h, and was chosen because previous studies had established that flow rates between 1/3 and 1/6 h-1 reacted similarly; but the slower flow rate conserved media, thereby facilitating long term experiments (14). Moreover, this flow rate simulated very closely the flow of intestinal contents through the mouse cecum (14) . The 7-ml culture was contained in a round glass tube (approximately 12-mm inner diameter), and was stirred at 500 rpm by means of a Teflon-covered magnetic stirring "flea" (lo-mm length, 3-mm diameter). Because of the small size of the stirring bar, this stirring action was adequate to cause a barely noticeable movement of the medium at the culture surface. The entire apparatus was maintained in an anaerobic chamber at 37‘C. The atmosphere in the chamber consisted of 5% CO 10% H2, and 2' 85% N2, and was maintained at less than 5 ppm 0 by means 2 of a coated palladium catalyst (2). The oxygen level was monitored with a Trace Oxygen Analyzer ( 2) . Inoculation of CF cultures was carried out in the following manner. A freshly killed mouse or human feces was placed immediately into the anaerobic chamber. Dissection of the mouse was performed by reflecting the skin coverings and abdominal walls in layers to avoid contamination of the viscera. The cecum was then aseptically removed, and part of the cecal contents and a small piece of the cecal wall were introduced into the growth tube. For human feces, samples 182 were well mixed according to previously described methods (15), and part of the feces was similarly introduced into a separate CF culture growth tube. CF cultures were allowed to equilibrate for a three-week period before being used in experiments. Animal preparation, inoculation, and recolonization experiments Mice were housed in sterile, covered plastic cages and supplied with sterile Purina mouse chow and sterile 1). 1) (VAG) ad iibitun for 5 days (19,20,22). Control animals distilled water containing vancomycin (500 ug ml- ampicillin (1 mg ml-l), and gentamicin (100 pg ml- were housed identically but without the addition of VAG to the drinking water. After decontamination, mice were transferred to sterile cages as described previously (19,20). To test the ability of indigenous cecal bacteria to protect against Q. aibicans colonization, cecal or CF culture homogenates (or filtrates as controls) were administered to antibiotic-decontaminated mice Iby Iboth oral and rectal routes once daily for 5 consecutive days, beginning 5 days after stopping antibiotic treatment. For oral administration, mice were lightly anesthetized with methoxyflurane (Pittman-Moore Co., ‘Washington Crossing, NJ), and cecal homogenates or filtrates were administered by carefully inserting a 5-cm 18-gauge plastic catheter 183 through the mouth into the stomach and injecting 1 ml of the material. For rectal administration, the same catheter was placed approximately 5 mm into the rectum, the skin was pinched tightly around the catheter, and 2 m1 of the material was injected. The latter dose has been shown to be an amount adequate to reach the cecum (40). On the day after the last administration of cecal homogenates or filtrates, the animals were challenged intragastrically (as above) with 0.5 ml of Sabouraud dextrose broth containing 107 CFU of Q. ainidana. Enumeration of Q. albicans Counts of viable Q. ainidana were performed on cecal or CF culture contents at various intervals after administration of Q. aininana. Mice were killed by cervical dislocation, their abdomens were soaked with 70% ethanol, and skin coverings and abdominal walls were reflected in layers to avoid contamination of the viscera. Ceca were aseptically excised and placed into separate glass tissue homogenizers, each containing 2 ml of sterile SDB. Organs were homogenized, and serial 10-fold dilutions of organ homogenates or CF cultures were plated on Sabouraud dextrose agar supplemented with vancomycin (7.5 pg ml-l) and ampicillin (100 ug' ml-1 ) for the enumeration of Q. ainidana. Plates were incubated aerobically for 48 h at '37‘C to determine the number of colonies per organ. 184 Enumeration of indigenous cecal bacteria Population levels of indigenous bacteria colonizing the CF cultures were determined in the anaerobic chamber. Samples were removed aseptically from the CF culture, and serial 10-fold dilutions were made in tryptic soy broth (Difco), supplemented with 0.04% Na CO 2 3' for the 5% CO2 in the anaerobic chamber (39), and 0.1 ml was plated on blood agar, MacConkey agar (Difco) , and to compensate modified AII agar (21,22) supplemented with 0.05% maltose (Rolf Freter, personal communication) without palladium chloride (28). Mammary powder (ICN Nutritional Biochemicals) was also substituted for placenta powder in the A11 medium (R. Freter, personal communication). The resulting medium yielded colony counts of strictly anaerobic bacteria equal to the "old" AII agar (21). All manipulations were performed in the anaerobic chamber, and the media and other materials were placed in the chamber at least 48 h prior to use. Modified AII agar plates were incubated at 37°C in the anaerobic chamber for at least 5 days. Numbers of facultative anaerobic bacteria and Enterobacteriaceae were determined by incubating for 48 h at 37'C one set each of MacConkey and blood agar plates aerobically and an identical set of plates anaerobically for 48 h at 37'C. VOlatile fatty acid (VFA) analyses The concentrations of VFAs in the cecal contents, 185 feces, or CF cultures were determined chromatographically (6), with 2-methylpentanoic acid as an internal standard (24), as described previously (22). RESULTS In the first study, the similarity of the bacterial flora from CF cultures and mouse cecal contents or human feces was determined by performing selective plate counts. These results are summarized in Table 1. Facultative and strictly anaerobic bacteria were found in almost identical numbers in CF cultures and in the ceca of mice or human feces. Likewise, the total numbers of Enterobacteriaceae found in CF cultures of human fecal flora and in human feces were also nearly identical, while the numbers of Enterobacteriaceae were found to be slightly higher in CF cultures of mouse cecal flora compared to those found in the mouse cecum. Therefore, both the total number of bacteria and the populations of various types of intestinal bacteria were very similar in CF cultures and their in yixn counterparts. Furthermore, gram stains of cecal contents from CF cultures, mouse ceca, and human feces were also remarkably similar. In all instances, the flora was characterized by a predominance of gram- negative rods. Another similarity between these systems was that dense layers of bacterial growth formed on the glass walls of the CF culture vessels, analogous to 186 bacterial populations that colonize intestinal mucosa. Figure 2 shows a bacterial layer that formed on the CF culture vessel containing mouse cecal flora, and is typical of the growth on the glass wall after equilibration with the entire complex intestinal flora. Table 2 summarizes the results of a series of experiments in which VFAs present in CF cultures, mouse ceca, and human feces were determined. The VFAs are regulated by the predominate anaerobic flora present in an ecosystem [15] . Both the total and individual levels of VFAs present in CF cultures coincide with the quantity of VFAs found in the ceca of the conventional mice used, as well as with those found in the feces used to establish the CF cultures of human fecal flora. The only significant differences found were for acetic and isovaleric acid (P <0.05), which were lower and higher, respectively, in CF cultures of mouse cecal flora. The patterns of VFA levels were, nevertheless, consistent for all experiments, and strongly suggest that the bacterial floras of the CF cultures were very similar to those found Figure 3 illustrates another critical test to determine the similarity of interactions between Q. aibicana and the intestinal microflora in these systems, and whether CF cultures would serve as good models of such interactions. QaLdida cells were inoculated into mice or CF cultures of mouse cecal flora, and the numbers of 187 viable Q. ainidana were determined at various intervals thereafter. The data points represent the logarithmic means of three separate experiments. As may be seen, the elimination of Q. ainidana frmm the mouse cecum occurred at a rate similar to that observed when Q. aim was inoculated into a CF culture of mouse cecal flora, thus indicating that the CF cultures reliably reproduced the populations of Q. ainidana as they occurred in mice under normal in yiyd conditions. As a final means of testing the ability of CF cultures to maintain the indigenous bacterial flora, and, therefore, a number of interactions as they occur in yiyd, the material from CF cultures of mouse cecal flora was fed to antibiotic-decontaminated mice to determine if they could redress several abnormalities that occur when the bacterial flora is eliminated (11,35,36), including the ability to resist colonization by Q. ainidana. For these studies (Table 3), mice were given a combination of antibiotics (VAG) designed to eliminate most of their intestinal flora, and subsequently they were given oral and rectal injections of either CF culture material, cecal homogenates, or filtrates of cecal homogenates. These animals, as well as appropriate control animals, were then inoculated with 107 Q. ainigana, and cultured 24 h postinoculation to determine the levels of Qandida in their ceca. Antibiotic-treated animals challenged with Q. ainidana had high levels of Q. ainidana colonizing their 188 ceca, as did animals given antibiotics followed by cecal filtrates. Control animals not given antibiotics, and antibiotic-treated mice recolonized with the indigenous cecal microflora, in contrast, had significantly lower cecal populations of Q. albicana (P <0.001). Likewise, animals that had been given CF culture material also had levels of Candida colonizing their ceca that were significantly lower than mice given antibiotics only (P <0.001). The effects of antibiotics and recolonization on the indigenous cecal microflora and certain cecal characteristics are summarized in Table 4. Animals treated with antibiotics and. animals treated with antibiotics followed by cecal filtrates had several cecal abnormalities. Gram stains, for instance, showed these animals to contain few’ bacteria (mostly gram-positive cocci), with a complete disappearance of the predominant gram-negative and fusiform-shaped rods. In contrast, antibiotic-treated mice, which had been recolonized with cecal bacteria isolated from the ceca of mice or with CF cultures of mouse cecal flora by oral and rectal injections, had bacterial floras that were similar to those found in conventional animals. Furthermore, dissection of test and control animals. revealed. other differences in cecal characteristics according to the experimental group examined. Animals that possessed a complex bacterial flora (i.e., normal mice, or mice 189 recolonized with mouse cecal contents or CF cultures of mouse cecal bacteria) had small ceca (approximately 1% to 2% of their total body weight) which contained thick, pasty contents. Mice given antibiotics to eliminate the indigenous microflora, in contrast, had enlarged ceca (approximately 5% to 8% of their total body weight) with very watery contents, a trait attributable to mice lacking a complex ("normal") intestinal microflora (8,11,35,36,38). DISCUSSION The results presented in this paper strongly support the hypothesis that the major ecological mechanisms that control and maintain a balance among bacterial populations of’ the large intestine are reproduced in. CF cultures (12,14). The ability of CF cultures to maintain. a bacterial flora that closely resembled that found in the mouse large intestine, for instance, is probably the most convincing of these data. Furthermore, thick layers of bacteria formed on the walls of the CF cultures, and these resembled mucosa-associated populations found in the intestine (14,34). It is known that indigenous bacteria of the large intestine occur in thick layers in the mucosa (34), presumably because the bacteria adhere to each other rather than to the epithelium itself (14). Scrapings of the wall layer of CF cultures revealed that the bacteria 190 were adhering to each other. This similarity, therefore, constitutes additional supporting evidence that CF cultures reproduce a number of bacterial interactions that occur in the large intestine (12,14) , and is consistent for both mouse and human intestinal flora. If the mucosal flora is disrupted, as with antibiotic therapy, the intestinal microflora undergoes a significant metabolic change (8,35). Likewise, if the formation of these layers is prevented in the CF culture, the resulting flora no longer resembles the intestinal flora in its function (14). As noted by Freter e_t _L. (14), one or a few bacterial species usually will overgrow all others when a clinical specimen is removed from the body and allowed to multiply in artificial nonselective culture. Because the major bacterial groups were present in similar numbers in CF cultures and the ecosystems from ‘which they were obtained, it is likely that the ecological control mechanisms operating in yittd were the same, or at least very similar to, those controlling the microflora in 2129. This interpretation ‘was further supported by the findings that the qualitative and quantitative composition of various VFAs present in CF cultures closely resembled that in the mouse cecum and human feces. Since these acids are characteristic metabolic end products of the predominant anaerobic bacteria that colonize the large intestine (15), it may be concluded that the CF cultures were able to support the major metabolic activities of the 191 predominate intestinal anaerobes. Therefore, because it is unlikely that two different sets of mechanisms would bring about similar equilibria among a complex microflora ( 14) , such as those studied here, one may conclude with some certainty that the predominant microbial interactions were reproduced in CF cultures as they occur in the large intestine. Indeed, when the survival and passage of Q. ainidana through CF cultures of mouse cecal flora and the mouse cecum proper were compared, Qandida cells were found to be eliminated at a similar rate. A final stringent test of the ability of CF cultures to propagate all of the numerous bacterial species of the cecal flora came from experiments showing that CF culture material could redress several cecal abnormalities of antibiotic-treated mice (11,35,36,38). For example, when CF culture material was fed to antibiotic-treated mice, a complete reduction of the enlarged cecum to the normal size took place. likewise, a bacterial flora resembling that of conventional mice not given antibiotics was confirmed by observing gram-stained smears of the cecal contents from normal and antibiotic-treated mice reconventionalized with CF culture suspensions. Moreover, such reconventionalized mice were able to resist colonization by Q. aini_can_s_, again suggesting that at least those organisms responsible for the suppression of Qandida in the gut were maintained in this culture system. It is significant to note that a complete reduction 192 of the enlarged cecmm to normal size and a reduction of the intestinal population of Qandida to levels found in conventional animals have never been fully achieved in germfree or antibiotic-treated animals following the implantation of one or a few bacterial species (4,5,16,17,26). Clark (5), for instance, showed that Q, albicana grew unchecked for several weeks in the GI tracts of‘ gnotobiotic ‘mice containing an. intestinal flora lof fiacteroides sp., Ladtobaciiina sp., fitxaptnddddna faecalis, Streptococcus laQtis, and Eadnatidnia 9911. Nevertheless, studies by Freter and coworkers (11,38) indicated that a complex intestinal flora, comprised predominantly of strict oxygen-sensitive anaerobes, was required to redress several ceca abnormalities of germfree mice. Furthermore, it was also found that a complex intestinal flora was required to maintain a balance among several bacterial populations similar to that found in conventional animals (1A0. In fact, close approximation of these parameters to normal values required the implantation of whole cecal contents or no less than 95 metabolically distinct strict anaerobes (11). The present findings that whole cecal contents or CF cultures of the cecal microflora were able to convert antibiotic-treated mice to the normal state and to reduce Qandida populations to levels similar to those found in conventional animals, then, are consistent with the view that a complex and diverse series of interactions is responsible for the 193 control of Qandida in the gut. Consequently, the CF culture model presented in this paper offers an excellent and accessible way to study Qandida-bacterial interactions in yittd. This model is currently being used to address the question of the nature of the mechanisms that control Q. alnidans in CF cultures of intestinal microflora. REFERENCES 1. Andremont, A., Raibaud, P. 8 Trancrede, C. 1983. Effect of erythromycin on microbial antagonism: a study in gnotobiotic mice associated with a human fecal flora. Journal of Intactious Qiseases, 148, 579-587. 2. Aranki, A. & Freter, R. 1972. Use of anaerobic glove boxes for the cultivation of strictly anaerobic bacteria. Anenigan Journal of Ciinidai Muttitign, 25, 1329-1334. 3. Aranki, A., Syed, S.A., Kenney, E.B. & Freter, R. 1969. Isolation of anaerobic bacteria from human gingiva and mouse cecum by means of a simplified glove box procedure. Appiied Midrobioldgy, 27, 568-576. 4. Balish, E. & Phillips, A.W. 1966. Growth and virulence of Qandida aibidans after oral inoculation in the chick with a monoflora of either Eadnatidnia Qdii or Strentocoddus gaecaiis. ddnznai__dt__fiadtatidlngy, 91, 194 1744-1749. 5. Clark, J.D. 1971. Influence of antibiotics or certain intestinal bacteria on orally administered Qandida aibicana in germ-free and conventional mice. mm and Immnnity, 4, 731-737. 6. Cottyn, B.G. & Boucque, C.F. 1968. Rapid method for the gas-chromatographic determination of volatile fatty acids in rumen fluid. ddnIna1__df__Agziintn:al__Eggd Chemistry. 16, 105-107. 7. Dubos, R., Schaedler, R.W., Costello, R. & Hoet, P. 1965. Indigenous, normal, and autochothonous flora of the gastrointestinal tract. 0 e d , 122, 67-76. 8. Dubos, R., Schaedler, R.W. & Stevens, M. 1963. The effect of antibacterial drugs on the fecal flora of mice. dournal of Ennanimentai Madidine, 117, 231-243. 9. Freter, R. 1976. Factors controlling the composition of the intestinal microflora. In: H.W. Stiles, W.J. Loesche, and T.C. O'Brien (Ed.), ' ° cr 8 s o e 'es (a special supplement to Microbiology Abstracts), vol. 1, pp 109-120. Information Retrieval Inc., Washington, D.C. 195 10. Freter, R. 1983. Mechanisms that control the microflora in the large intestine. In: D.J. Hentges (Ed-)1 Human_iatestinal_mi2r9fl9ra.in.health_and_disea§e. p. 33-54. Academic Press, Inc., New York. 11. Freter, R. & Abrams, G.D. 1972. Function of various intestinal bacteria in converting germ-free mice to the normal state. Infestign_and_lmmunitx. 6. 119-126- 12. Freter, R., Brickner, H., Botney, M., Cleven, D. & Aranki A. 1983. Mechanisms that control bacterial populations in continuous-flow culture models of mouse large intestinal flora. nf t 0 nd u ' , 39, 676- 685. 13. Freter, R., Brickner, H., Fekete, J., O'Brien, P.C.M. & Vickerman, M.M. 1979. Testing of host-vector systems in mice. Racombinant DNA Tecnninal finilatin, 2, 68-76. 14. Freter, R., Stauffer, E., Cleven, D., Holdeman, L.V. & Moore, W.E.C. 1983. Continuous-flow cultures as in vitto models of the ecology of large intestinal flora. Infgction and immunity, 39, 666-675. 15. Holdeman, L.V., Cato, E.P. & Moore, W.E.C. , eds. 1977. aero e La orato anu , 4th ed. Virgina Polytechnic Institute and State University, Blacksburg, 196 Virginia. 16. Hummel, R.P., Maley, M.P., Miskell, P.W. 8 Altemeier, W.A. 1975. Suppression of Qandida ainidana by Ari—HS e 'chi 29.1.1. W. 15. 413-418- 17. Hummel, R.P., Oestreicher, E.J., Meley, M.P. 8 MacMillan, B.G. 1973. Inhibition of Qandida alnidang by Escherchia coli in vitzd and in the germfree mouse. doutnal 9f Sntgidai Reseancn, 15, 53-58. 18. Iwai, H., Ishihara, Y., Yamanaka, J. 8 Ito, T. 1973. Effects of bacteria on cecal size and transit rate of intestinal contents in mice. WM Ennetinantal Medicina, 43, 297-305. 19. Kennedy, M.J., Bajwa, P.S. 8 ‘Volz, P.As 1982. Gastrointestinal inoculation of andrgtntix Mi in mice. M o atholo ia, 78, 141-143. 20. Kennedy, M.J. 8 Volz, P.A. 1983. Dissemination of yeasts after gastrointestinal inoculation in antibiotic- treated mice. fiannntandia, 21, 27-33. 21. Kennedy, M.J. 8 Volz, P.A. 1985. Effect of various antibiotics on gastrointestinal colonization and dissemination by Candida aibidans. Sabouraudia1__geurnal 197 2f_Medi2a1_and_Yeterinarx_stolesY. 23. 265-273- 22. Kennedy, M.J. & Volz, P.A. 1985. Ecology of Qandida ainidana, gut colonization: Inhibition of Qandida adhesion, colonization, and dissemination from the gastrointestinal tract by bacterial antagonism. Intadtipn and immunity, 49, 654-663. 23. Krause, W., Matheis, H. 8 wulf, K. 1969. Fungaemia and funguria after oral administration of Qandida aibicans. Lancet, 1, 598-599. 24. Mayhew, J.W. 8 Gorbach, 8.1” 1977. Internal standards for the gas chromatographic analysis of metabolic end products from anaerobic bacteria. Applied an v'ronmental icrob o , 33, 1002-1003. 25. Moore, W.E.C. 8 Holdeman, L.V. 1974. The normal flora of 20 Japanese-Hawaiians. Appiiad_MiQtppidlpgy, 27, 961-979. 26. Nishikawa, T., Hatano, H., Ohnishi, N., Susaki, S. 8 Omura, T. 1969. Establishment of Qandida alpiQana in the alimentary tract of the germfree mice and antagonism with Eadnatidnia coli after oral inoculation. Japan HiQIQQLQlQQYi 13: 253-275- 198 27. Odds, F-C- 1979. Qandida_and_§andidgsi§. University Park Press, Baltimore. 28. Owens, D.R., Rolfe, R.D. 81 Hentges, D.J. 1976. Effectiveness of palladium chloride for the isolation of anaerobes. J2uraal_2f_£linisal_uisrohiglegx. 3. 854-861. 29. Ozawa, K., Jabu-Uchi, K., Yamanaka, K., Jamachita, Y., Ueba, K. 8 Micwatani, T. 1979. Antagonistic effects of Bacillus nato and Sttaptpnpndna fiadaalta on growth of Candida albicans. Microbioiogy and Immunoiogy, 23, 1147- 1156. 30. Paine, T.F., Jr. 1958. The inhibitory activity of bacteria on Candida growth. Antibiotics and Qnanothenapy, 8, 273-281. 31. Phillips, M.W. 8 Lee, A. 1984. Microbial colonization of rat colonic mucosa following intestinal pertubation. Microbial Edology, 10, 79-88. 32. Rifkin, G.D., Fekety, F.R., Silva, J. 8 Sack, R.B. 1977. Antibiotic-induced colitis: implication of a toxin neutralized by Qlosttidinn aptdallii anti-toxin. Landat, ii , 1103-1106 . 33. Savage, D.C. 1977. Microbial ecology' of the 199 gastrointestinal tract. MW. 31, 107-133. 34. Savage, D.C. 1980. Adherence of normal flora to mucosal surfaces. in: E.H. Beachey (Ed) Bantaziai Adnazence, pp. 33-59. Chapman 8 Hall, London. 35. Savage, D.C. 8 Dubos, R. 1968. Alterations in the mouse cecum and its flora produced by antibacterial drugs. Journal of Enperimental Medicina, 128, 97-110. 36. Savage, D.C. 8 McAllister, J.S. 1971. Cecal enlargement and microbial flora of suckling mice given antibacterial drugs. intantipn_and_innnnity, 3, 342-349. 37. Schaedler, R.W., Dubos, R. 8 Costello, R. 1965. The development of the bacterial flora in the gastrointestinal tract of mice. Wins. 122. 59- 65. 38. Syed, S.A., Abrams, G.D. 8 Freter, R. 1970. Efficiency of various intestinal bacteria in assuming normal functions of enteric flora after association with germfree mice. inflection and innnnity, 2, 376-386. 39. Wilson, K.H., Kennedy, M.J. 8 Fekety, F.R. 1982. Use of sodium taurocholate to enhance spore recovery on a 200 selective medium for 2126;115:1911 911119.119- 1213111214: CiiniQai Mictobioiogy, 15, 443-446. 40. Wilson, K.H., Silva, J. 8 Fekety, F.R. 1981. Suppression of Qinstpidinn diflidiia by normal hamster cecal flora and prevention of antibiotic-associated colitis. WW. 34. 626-628- Table l. Population levels of intestinal bacteria in the ceca of test animals. 201 human feces. and anaerobic continuous-flow (CF) cultures. Experimental group tn) Logl0 mean no ( x SD) per g (wet Wt.) or ml Facultative Enterobacteriaceae Strictly anaerobic bacteria bacteria Mouse cecal flora (10) 8.4 t 0.4 4.9 1: 0.2 3.: z 3; CF culture of mouse cecal flora (4) 8.3 r 0.3 6.3 x 0.3 . z . Human fecal flora (2) 8.5 x 0.9 6.5 t 0.8 1;}: i 3.: CF culture of human fecal flora (2) 8.1 t 0.8 6.8 t 0.7 . - . Table 2. Concentrations of volatile fatty acids in test animals. human feces. and in anaerobic continuous-flow (CF) cultures. Experimental Group (n) Volatile fatty acid concentration (mM) of contents 1 SD Acetic Propionic lsobutyric Butyric lsovaleric Valerie Mousececalcontents (5) 116.1 x 19.2 20.4 t 4.5 4.5 t 0.8 20.8 s: 5.7 3.8 :t 0.8 1.3 t 0.6 CF culture ofmousececal flora (5) 79.3 3: 2.9 19.1 t 6.0 3.9 t 1.2 21.8 3: 3.4 7.4 :e 3.1 0.6 t 0.8 Human feces (2) 77.7 x 12.6 29.9 t 4.6 4.1 1: 0.8 30.4 t 6.3 9.2 :1: 0.7 1.5 1 0.1 CFculture of human fecal flora (2) 70.5 3: 15.1 21.3 :1: 4.9 5.4 t 1.1 29.5 3: 16.4 8.0 t 1.7 1.7 t 0.6 202 Table 3. Cecal population levels of Candida albicans (CA34) in reconventionalized and control animals. Antibiotic treatment Log“, mean no. :1: SD of Candida per g" Antibiotics only (VAG)- 5.7 :1: 1.2 Antibiotics. followed by cecal filtrates 5.8 t 0.9 Antibiotics. followed by cecal homogenates 1.8 :1: 0.9 Antibiotics. followed by continuous-flow culture material 2.3 :t 1.5 None 2.6 :1; 1.6 ‘ Animals received vancomycin. ampicillin. and gentamicin (VAG) in the drinking water as described in the text. 5 The values represent the means and standard deviations from 5 to 10 mice per group cultured for C. albicans at 24 h postinoculation. Table 4. Cecal characteristics of antibiotic-treated, control. and reconventionalized animals. Animal treatmenta Gram stainb Predominant Consistency Cecal sized bacterial of cecal organisms“ contents Antibiotics only (VAG) - Gm— coal. and Soft 5 - 8% small Gm- rods Antibiorics. followed by - Gm- cocci. and Soft 5 — 8% cecal filtrates small Gm- rods Antibiotics. followed by + Large, Thick and I — 2% cecal homogenates fusifonneshaped pasty Gm- rods Antibiotics. followed by + Large, Thick and 1 - 2% continuous-flow culture fusiform-shaped pasty material Gm- rods None + Large, Thick and 1 - 2% fusiform-shaped pasty Gm- rods ' Animals received vancomycin. ampicillin, and gentamicin (VAG) in the drinking water was described in the text. " Gram strins: - = abnormal; + = normal compared to normal. untreated animals. 1* Predominant organisms seen on gram stained smears. d Percentage of total body weight. 203 Figure 1. The continuous-flow culture apparatus. 204 Figure 2. A continuous-flow culture growth tube showing dense layers of bacterial growth formed on the wall that resemble bacterial populations that colonize intestinal mucosa . 205 Figure 3. Elimination of Q. alpidang CA34 from continuous-flow culture of mouse cecal flora and mouse cecum. The values represent the means from 4 to 10 samples per data point. o - Mouse («um O I CFC Logic C. albicans per ml S 10 15 20 2 5 30 nouns POST mocuumon MECHANISMS THAT CONTROL QAEDIQA ALELQAES POPULATIONS IN CONTINUOUS-FLOW CULTURE MODELS OF INTESTINAL MICROFLORA AND IN THE LARGE INTESTINE Michael J. Kennedy, Alvin L. Rogers, and Robert J. Yancey, Jr. Submitted for publication to: W October 10, 1988 Mechanisms that control Qandida aipidana populations in continuous-flow culture models of intestinal microflora and in the large intestine Michael J. Kennedylr4*, Alvin L. Rogerslrzr3, and Robert J. Yancey, Jr.4 Departments of 1Botany and Plant Pathology, 2Microbiology and Public Health, and 3Medical Technology, Michigan State University, East Lansing, Michigan 48824, and 4The Upjohn Company, Kalamazoo, Michigan 49001 Running title: Control of Qandida by intestinal microflora *Correspondence address: The Upjohn Company Microbiology and Nutrition Research 7922-190-MR Kalamazoo, Michigan 49001 206 207 ABSTRACT It has been well established that the intestinal microflora controls Qandida aipidana populations in the gastrointestinal (GI) tract. However, because most early studies did not reflect interactions as they occur n _v_i_v_d, several erroneous conclusions have been made regarding the identity of the organisms and mechanisms responsible for the suppression of Qandida in the gut. This paper describes the suppression of Q. ainipana in a continuous-flow (CF) culture model of the large intestinal microflora and in the mouse large intestine by a complex indigenous microbiota. When inoculated into CF cultures of established mouse cecal flora, Q. aidigna was rapidly eliminated at a rate greater than that of the system. In contrast, when Q. aipidana was inoculated into CF cultures of established human fecal flora or simultaneously or before establishment of mouse cecal flora, Q. aibiQana was removed with the flow rate of the system. These results suggest that Q. aipidana can develop resistance/tolerance to some or all inhibitory mechanisms in the CF culture. Growth rates of Q. albicans in CF cultures or in sterile filtrates of CF cultures incubated anaerobically were considerably slower than in CF culture medium incubated aerobically. Anaerobic condition in CF culture medium, however, was only partially inhibitory to growth of Q. aibicans, indicating that other mechanisms were also operating to suppress the growth of Q. ainicang. The 208 inhibitory effect on ‘Qandida, moreover, could not be abolished by adding carbon or nitrogen sources, vitamins, or all three to CF culture filtrates and incubating the system (under varying pHs) aerobically. Apparently, suppression of Q. albicana was not due solely to depletion of nutrients and anaerobiosis, but may instead be due to a combination of these two factors in addition to the production of inhibitory substances by intestinal bacteria. The production of the short-chain fatty acids acetic acid and butyric acid by intestinal anaerobes were shown to be inhibitory to Q. ainidana growth. In addition, mucosal association was found to be important for long term colonization of the cecum of antibiotic- treated animals as the flora reestablished. It is suggested that the indigenous intestinal microflora controls Q. alnicans populations collectiveLy by several mechanisms, but that Candida cells can become tolerant to these mechanisms and can survive at a low levels in the gut probably by associating with GI mucosal surfaces. 209 INTRODUCTION Colonization of the gastrointestinal (GI) tract by Candida aibidans may play an important role in diseases of humans since the GI tract is considered the proximate source of infection in recurrent vaginitis (23,26) and systemic candidosis (22,24,30-32). Although the determinants of GI colonization by Q. aip'mana are not completely defined (20,21), it is known that certain members of the indigenous microflora suppress the growth of Q. albicans within the gut (4-7,11-13,18-21,25,28-32). Nevertheless, because the intestinal microflora is able to regulate Candida populations in the gut and control the spread of Q. albicans to other body sites (18-20), elucidation of the inhibitory mechanisms should lead to a better understanding of the ecology of Qandida gut colonization. In addition, such knowledge may also lead to ways to manipulate the GI environment for the specific purpose of inhibiting fungal growth (15). Previous attempts to determine the mechanisms by whidh intestinal bacteria inhibit Q. aitidana (ll-13,29) have been beset with methodologic difficulties (see 17). As commented on by Freter (8-10), there are several reasons why such studies were "doomed to failure". First among these is that most in yittd studies have relied on simple static cultures of a diflora of Q. ainidana and a single bacterial species (4,11-13,29), which do not reflect interactions as they normally occur in the 210 intestinal tract by a complex indigenous microflora (8,20). Moreover, simple dicultures can be manipulated to give varying results depending on, for example, the culture medium used (28) . Moreover, several of these studies have shown that Q. aipinana was strongly inhibited by certain bacterial species in yitzd (4,11,12,25), whereas others have shown that these same organisms had no effect on Q. ainidana growth in the mouse intestine (6,19). It is clear from the above list that in yitrn model systems will have to closely simulate the processes occurring in conventional animals or humans if meaningful results are to be gained from studies of bacterial-Q_ndida interactions (10). An anaerobic continuous-flow (CF) culture model of the ecology of the mouse large intestinal flora was recently developed which meets these stringent requirements, since it reproduces a number of microbial interactions and maintains the indigenous microflora in yitrn as it occurs in mt) (9,10). In a previous paper (17) evidence was presented which supports the findings that populations of intestinal bacteria in CF cultures of mouse cecal flora resemble those in the mouse intestine, and offers additional evidence that ecological mechanisms and environmental conditions prevailing in both systems are very similar. Furthermore, it was shown that interactions between Q. ainidana and the intestinal microflora in CF cultures and in the mouse large intestine 211 were very similar, since population levels of Q. ainidang and elimination rates were almost indistinguishable. That study, however, did not address the question of what mechanisms were responsible for the suppression of Q. ainigana by intestinal bacteria. This paper, therefore, describes the mechanisms which regulate Qandida populations in this culture system, and describes the suppression of Q. aininanp not by a single bacterial species but by a complex intestinal microflora. MATERIAIS AND METHODS Mi Qrooganisms. Bacteria for the inoculation of CF cultures were obtained from the cecal contents of Swiss Flow DUB/kr mice (19) or the feces of a healthy male volunteer on a Western diet who had not taken antibiotics for at least 24 months prior to the study ( 17) . The bacterial suspensions were prepared in a Coy Anaerobic chamber (Coy laboratory Products, Ann Arbor, MI) as described previously (17). The isolates of Q. aLQiQana used were recovered from the blood of patients with systemic candidosis (CA30 and CA34) , or from the oral cavity of an infant with thrush (AK785). Unless otherwise noted, Qandida cells were grown aerobically (without shaking) in Sabouraud dextrose broth (SDB; Difco Laboratories, Detroit, MI) for 18 h at 37°C. 212 out nu us-fl CF . The CF cultures were maintained in an anaerobic chamber essentially as described by Freter at a]; (9,10) . The growth medium, flow rate, inoculation, and operation of the CF culture is described in detail in a previous paper ( 17) . To study the passage of Q. ainidana through CF cultures, Qandida cells were inoculated and viable counts were made by plating samples of serial 10-fold dilutions onto Sabouraud dextrose agar (SDA: Difco) supplemented with chloramphenicol (100 pg ml-l) and streptomycin (100 pg ml- 1). The plates were incubated at 37'C and observed for up to 5 days to determine the number of viable Qandida. The effluent from CF cultures of intestinal microflora for use in subsequent experiments were collected in a flask kept at 54°C, to prevent changes due to bacterial growth occurring after the material had left the growth tube (10). Filtrates from this material were prepared by filtration through 0.22 pm Millipore filters inside the anaerobe chamber. In other experiments, filtrates were prepared from CF culture effluent, or from culture fluid removed directly from the tube as described previously (17). Grdwth studies. Growth curves of Q. ainidana in filtrates of the effluent from CF cultures and in sterile CF culture broth medium, were performed under aerobic and anaerobic conditions at 37’C. For some experiments, 213 nutrients were added to filtrates of CF culture effluent alone or in combination at various pHs, and in other experiments substances were added to sterile CF culture medium to determine what factors might suppress the growth of Qandida in the CF culture. 1 4)2 so4 (5.0 g 1 1 ), and biotin (0.001 g 1’ The nutrients added were (NH 1 )1 glucose (12.5 g l- ). Possible inhibitory factors included the short-chain fatty acids (SCFAs): acetic acid (83.8 mM), propionic acid (24.0 mM), isobutyric acid (4.7 mM), butyric acid (25.0 mM), isovaleric acid (8.9 mM), and. valeric acid (1.8 mM), neutralized to pH 6.8 with NaOH. Five-ml cultures were incubated aerobically and anaerobically at 37'C in 18 x 150 mm tubes and stirred continuously with the same magnetic stirring fleas used in the CF culture apparatus. Growth was assessed using an LC-55 spectrophotometer (Perkin-Elmer, Coleman Instrument Division, Maywood, Ill.) at 600 nm. In some instances, samples of the material were spread onto SDA to determine if Qandida remained viable in tubes that showed no visible growth. nhib to ct'vit o o - ' c (SQFA). Each SCFA (acetic, propionic, isobutyric, butyric, isovaleric and valeric acids) was prepared in CF culture medium at various concentrations starting at double its CF concentration. The pH of the broth was adjusted to 6.8 with NaOH, and each broth was filter 214 sterilized through 0.22 p Millipore filters. Serial 2- fold dilutions in CF broth were made, 5-ml broths were then inoculated with 0.1 ml of a Q. ainidana suspension to 5 cells ml-l, and give a final concentration of 10 incubated aerobically at 37'C for up to 5 days to determine the minimal inhibitory concentration (MIC). The cell suspension was prepared by diluting an 18 h SDB culture of Q. aibicans to the appropriate concentration in CF broth. Colonization of conventional and antinidtiQ-tpgtg: nida. To assess the role of mucosal association in long- term implantation of the large intestine, conventional and antibiotic-treated mice were inoculated with Q. ainigana CA34 and lumen contents and mucosal walls were cultured for Q. alnicans as described previously (18,19). Animal preparation and inoculation has been described in detail elsewhere (16,18-20). Animals receiving antibiotics were given vancomycin (500 pg ml-l), ampicillin (1 mg ml-l) and 1) (VAG) ad iipitnn for 5 days in the drinking water prior to Qandida inoculation (20). gentamicin (100 pg ml- 215 RESUBTS WWW- As reported in a previous paper (17) , Q. aipiaana is eliminated from the CF culture of mouse cecal flora at a rate indistinguishable to the elimination rate from that of the mouse cecum. By plotting the elimination of Q. aibicana from various CF cultures against the dilution rate of the system, it was determined whether Qandida organisms were being killed, were multiplying, or were simply being "washed-out" with the flow of material through the system. Figure 1a shows the passage of CA30 and CA34 from CF cultures of mouse cecal flora. Qandida cells for both isolates were removed/eliminated from CF cultures at rates faster than the dilution rate. When inoculated into CF cultures of human fecal flora, in contrast, these same strains were eliminated at apparently the same flow rate (Figure 1b) . Thus in both systems Q. aibidans failed to implant and was eliminated, but only in CF cultures of mouse cecal flora did "killing" of Qandida appear to be occurring. The ability of Q. aipigana (CA34) to implant and grow in a sterile anaerobic CF culture is shown in Figure 2. In this instance, Q. alpigana was initially removed (for 20 h) from the system with the flow rate, but thereafter Qandida cells began to multiply and maintained a steady population of 104 cells ml-l. This steady state condition was maintained for longer than 2 216 weeks. It is well known that Q. aipigana cannot easily be implanted into an established intestinal microflora, but that Qandida populations can be maintained if given prior to or during the succession and development of the flora (eg, in infant mice). To study this phenomenon, Q. ainicans CA34 was inoculated into sterile CF cultures simultaneously with 1.0 ml of an established CF culture of mouse cecal flora that had been shown to rapidly eliminate Q. albidans CA34. In this experiment, Q. ainLcana was removed from the system with the flow rate for about the first 24 h, but thereafter was removed at a much slower rate (Figure 2). This indicates that Q. aibi cans had become implanted and was multiplying (albeit slowly) in the CF culture in the presence of a functional indigenous microflora. It should be noted, however, that the Qandida population was eventually eliminated as the microflora reached equilibrium. To confirm that the bacterial flora was functioning properly, a sample was drawn from the CF culture to determine whether characteristic SCFAs were being produced. It was found that the predominant acids were present, however, both the total and individual levels of SCFAs were present in lower quantities than that found in the donor CF culture. This was not surprising since the microflora had not yet reached equilibrium. Furthermore, phase-contrast microscopy revealed that the predominant bacterial flora was indeed maintained. 217 Wm. To study the mechanisms that were responsible for causing the inhibition of Q. aininana in CF cultures of conventional cecal microflora or human fecal microflora, the effluent from these cultures were sterilized by filtration, inoculated with Q. ainigana 5 1), incubated as static (final concentration 10 cells ml- cultures aerobically and anaerobically at 37'C. The growth was monitered for up to 5 days compared to that in sterile CF culture medium. As is shown in Figures 3 a8b, Q. aipidans was unable to multiply in the culture filtrates from CF cultures of mouse or human intestinal microflora. This was true whether the cultures were incubated aerobically or anaerobically. To determine if the inability of Q. aininana to multiply was simply due to the depletion of nutrients, carbon and nitrogen sources, vitamins and trace elements were added to the effluent (Figures 3 a8b). It was found that the inability of Qandida to multiply in CF culture filtrates could not be reversed by adding any of these nutrients, either individually or collectively, to the cultures. Furthermore, this trend was not altered by changing the culture pH (range of 6.0 to 8.0) , incubating the cultures aerobically or anaerobically, or both. However, the growth of Q. aipidana in sterile CF culture medium was reduced when the cultures were incubated anaerobically. Nevertheless, the reduction in growth was not as severe as that in CF culture filtrates. This indicated that 218 something other than anaerobic condition or substrate depletion, i.e. the production of an inhibitory substance(s), was also involved with controlling the growth of Q. aipidana in CF cultures. Figures 3 a8b illustrate the results of experiments in which all six SCFAs (at concentrations resembling those in CF cultures) were added to static cultures of IQ. albicans. These data indicate that the SCFAs added to fresh medium were able to inhibit the growth of (Q. albicans, or at least cause a pronounced lag phase to be imparted. By plating samples of these cultures at the end of the experiments, it was determined that Qandida organisms remained viable. It was also found that the inhibition of Q. alnidana was more pronounced when the pH of the medium *was lowered. to $6.0 (data not shown). Furthermore, the inhibitory effect of SCFAs for Qandida was found to decrease if the cultures were not stirred. Nevertheless, from the above data it could not be concluded whether one short-chain acid was responsible for the observed suppression, or the mixture acted collectively to inhibit the growth of Q. aipi_an§. Therefore, to test this the MIC of each SCFA was determined using the SCFA mixture as a positive control. Table 1 summarizes the results of these experiments. It was found that all of the short-chain fatty acids inhibited the growth of Q. ainidana, but that some of these did so at levels exceeding those found in the mouse 219 cecum, human feces, or CF cultures. 0 he of c. mmmm. C_.Jdaand cells were inoculated into CF cultures in the yeast phase, and were not observed to form germ tubes or "switch" to pseudomycelial or other forms in any of the CF cultures inoculated. This was true whether Qandida cells were inoculated alone, simultaneously with intestinal bacteria, or after the intestinal microflora had become established. Role 0 a sion ' o i a . As was described above, Q. albiQans CA34 became implanted (for a short time) in CF cultures of mouse cecal flora only when it was inoculated into the system before or simultaneously with intestinal bacteria. Moreover, it was also noted that while this strain was eliminated from CF cultures of established mouse cecal flora faster than the flow rate (i.e., CA34 was apparently being "killed"). However, when it was inoculated simultaneously with intestinal bacteria Q. albicans CA34 was, after an initial period of implantation, removed from the system with the flow rate indicating that it had somehow become "adapted" to the environment. Theoretically, then, Q. ainiQana should be able to remain implanted for longer periods of time if it were able to attach to the wall of the vessel. To test this, the walls of several established CF cultures were cultured after inoculation of Q. aipidana CA34. Also, 220 other cultures that were inoculated with Q. alpidana CA34 simultaneously with intestinal bacteria from an established CF culture were similarly cultured for Qandida. Finally, the wall of an established CF culture was cultured after draining its contents, washing it three times and inoculating it with Q. aipidana CA34. In no instance were viable Qandida recovered from the CF culture walls. Because a glass culture vessel wall is very different from intestinal mucosa, which contain thick layers of mucus that "entrap" Qandida (14,20,21), this same type of experiment was repeated using an adult mouse model. For these studies, mice were given a combination of antibiotics (VAG) designed to eliminate most of their intestinal microflora, and subsequently they were challenged with Q. ainiQana. A nonantibiotic-treated control group was also included for comparison. Cecal populations of Candida in the mucosa and lumen contents were then monitered. In the control animals, Qandida populations were eliminated rapidly, wheras in the VAG- treated animals Q. aipidana populations were eliminated at a much slower rate. Moreover, it was observed that in the latter group of animals lumen populations of Qandida predominated initially, but as the microflora reestablished mucosal populations predominated. 221 DISCUSSION Survival and implantation of Q. aipinana in the GI tract may be influenced by a number of factors (15), but the presence of an intact indigenous microflora is probably the most important deterrent to whether C_andida can colonize the gut (27). Several patient and animal studies support this hypothesis (7,18-21,23-26), and other studies have shown that certain intestinal bacteria are inhibitory to both in vitro growth and GI colonization and dissemination by Q. aibicana (7,17-21,28,29). However, while the importance of the intestinal microflora has been noted for a number of years, the mechanisms which control Candida in the gut have remained somewhat obscure. This is due, for the most part, to the fact that most studies have reported on Candida suppression by a single bacterial species (4,5,11-13,25,29), which cannot be expected to reflect interactions of a complex indigenous microflora (8). The present studies, therefore, examined the mechanisms by which a complex indigenous microflora inhibits the colonization of Q. albicans. The results presented in this paper strongly support the hypothesis that the indigenous intestinal microflora suppresses Q. ainicans and suggest that the regulation of Qandida populations in the GI tract is the result of a complex series of interactions. The present findings also indicate the importance of using a model that simulates, 222 as close as possible, the ecosystem under study. This may be inferred from the finding that more than one mechanism was found to suppress Q. ainidana. Thus, studies in which a monoflora of Escherichia 9311 (or any other single bacterial species) antagonized Q. ainiaana in the gut of gnotobiotic animals (4,5,12,25), or similar studies performed in yittd (12,13,29), should be interpreted cautiously. Nevertheless, the data presented here, together with the findings by earlier workers, indicate further that the indigenous microflora inhibits QM colonization by two general mechanisms: (1) antagonizing and suppressing the growth of Q. aipidana and (ii) inhibiting the association of Qandida with the intestinal mucosa. The data concerning the passage of Q. ainidana through CF cultures, and the growth of Candida in CF culture effluent, show ‘that the intestinal microflora caused Q. aibidans to remain in a prolonged lag phase (apparently up to several days) and to have a long doubling time. For instance, when Q. aipidana was inoculated into CF cultures of human fecal flora (Figure 2) yeast cells were removed with the flow rate, whereas in sterile controls Qandida cells maintained a steady population of 104 cells ml-l. Discussions in the literature of the possible inhibitory mechanisms involved with the suppression of Qandida have suggested that anaerobiosis may be important (13,29). However, anaerobic 223 condition was only partially inhibitory to growth of Q. albidans and did not prolong the lag phase as did CF culture filtrates, indicating that other inhibitory mechanisms were also operating to suppress the growth of Q. aibicans in this system. The inhibitory effect of Qandida growth in CF culture filtrates, moreover, could not. be reversed. by adding' carbon or’ nitrogen souces, vitamins, or trace nutrients (individually or collectively) to CF culture filtrates and incubating the cultures (under varying pHs) aerobically. Thus the suppression of Q. aibicans in CF cultures was not due solely to depletion of nutrients and anaerobiosis, but appeared to be due to a combination of these factors in addition to the production of inhibitory substances by intestinal bacteria. To begin to examine other factors which might be involved in the suppression of Qandida growth, certain metabolic end-products from the predominant anaerobic microflora were introduced into sterile CF culture medium and the growth of ‘Q. albicans was :monitored, In la previous publication (17 ) , it was found that both the total and individual levels of short-chain fatty acids present in CF cultures coincided with the quantity of acids found in the large intestine of conventional mice or human feces used to establish the CF cultures. Furthermore, short-chain fatty acids have been shown to inhibit the adhesion of Qandida to intestinal mucosa (20). 224 Because of this, because much has been speculated about the role of short-chain fatty acids in the control of the intestinal microflora (1,2), and since short-chain fatty acids have been shown to be inhibitory to Sandpandnynaa detayisiae (2,3) , these acids were thought to be likely candidates for inhibitory substances. It was found that short-chain fatty acids were inhibitory to Qandida growth. Further comparison between ”MICs" for individual acids and "gut levels" or CF culture levels were performed to determine if the entire mixture collectively was necessary for suppression, or whether one or more acids individually could antagonize Qandida growth. The results indicate that acetic and butyric acid antagonized Qandida growth at levels that are present in ‘the lgut and. CF cultures. Although it is not known whether the in yittd results of the present study adequately reflect the sensitivity of Q. albinana to these acids in the GI tract or CF cultures, it is not likely' that inhibitory’ concentrations, of 'these short-chain fatty acids would be higher. Therefore, these acids are likely to play a role in the suppression of Qandida in the gut. A similar argument could be made to suggest that secondary bile acids also play a role in the suppression of Q. ainidana in the GI tract. For instance, secondary bile acids have been shown to inhibit both the in_ yittp growth and the adhesion of‘ Q. aipidana, to intestinal mucosa (20). According to Freter at al. (8,9) , however, it is 225 theoretically impossible to account for the total suppression of a sensitive microorganism (e.g., Q. aipidana) in the GI tract on the basis of the production of growth inhibitors. If this were the case, constant populations of the sensitive species could only be maintained at precise inhibitor concentration (8) . Thus if too much inhibitor were present Q. aibidana would be eliminated, or if too little were present the Qandida population would increase until it became limited by some other mechanism. Although m population levels in the GI tract do show some increases or decreases from day to day (7), it may be argued that other inhibitory mechanisms are also involved with the suppression of Q. ainidana in the gut. Considering that there must be fluctuations in the production of growth inhibitors, and considering that short-chain fatty acids appear only to be fungistatic, this is likely to be the case. It should also be noted that there may be an enormous supply of growth inhibitors that can be produced by the indigenous intestinal microflora, which would be subject to fluctuation. For instance, several breakdown products from the large number of primary dietary or host-derived substrates in the gut may antagonize Q. ainidana, as might a number of substances synthesized by the microflora itself. Although Q. ainidana was apparently being "killed" when inoculated into established CF cultures (only mouse cecal flora), it was not "killed" when it was 226 inoculated before or simultaneously with the microflora. Thus it may be that Qandida can adapt itself to these "toxic” substances, which, nevertheless, are probably still inhibitory to Qandida growth. This view is consistent with the finding that Q. aipidana was able to implant itself (albeit only for’ a short time) in CF cultures of mouse cecal flora. The reason Q. aipidana could not permanently colonize CF cultures in large numbers was probably due to a lack of adhesion sites or the inability to attach to the glass vessel in the presence of, for example, short-chain fatty acids (14,15,20). This suggests that adhesion to the gut wall may be an important determinant to colonization for Q. aibicans, since association with the intestinal mucosa would probably allow Candida to remain in the gut even at a severely depressed growth rate. Mbreover, it may also be that colonization of the mucosa would provide a more sheltered site from growth inhibitors. The finding that mucosa-associated populations of Qandida predominated after the intestinal microflora reestablished is certainly consistent with this view. Therefore, the ability of the intestinal microflora to inhibit the association of Q. aipigana with the intestinal mucosa also represents an important defense mechanism that inhibits Qandida colonization (and systemic spread from) of the GI tract. The mechanisms by which intestinal bacteria do this have been commented on elsewhere (20). 227 Our current hypothesis holds that Qandida populations are controlled collectively by several factors. First, Q. aipidana is bombarded by several growth inhibitors (e.g., short-chain fatty acids, secondary bile acids, etc.) that cause Qandida cells to remain in a prolonged lag phase and have a long doubling time when entering the GI tract. This alone may cause Qandida organisms to be removed from the GI tract due to peristalsis and the flow of mucus; unless, of course, the Qandida are able to associate with the mucosa. Second, there are several mechanisms which are operating in concert to inhibit and suppress Qandida cells from attaching to the gut wall, including competition for adhesion sites, production of inhibitor substances, prevention of penetration into the mucus gel, and an increased disassociation rate (20). Finally, superimposed upon and modifying this type of regulation is the metabolic competition for limiting nutrients, which probably becomes the predominant control mechanism if Qandida cells become implanted in the gut. There is good evidence to suggest that the presence of certain metabolic end-products produced by the predominant anaerobes restricts the range of substrates which a given organism can efficiently utilize for anaerobic growth (9). Thus Q. aipima is probably collectively controlled by growth inhibitors and anaerobiosis (which severely depresses its growth rate), competition for mucosal sites, and substrate competition. It is not known to what extent these 228 mechanisms individually control the growth of Qandida in the gut, but the range of factors which can cause and "overgrowth" of Qandida in the GI tract suggest that no one single mechanism is responsible for the regulation of Q. albicans populations. This may help to explain discrepancies in the literature regarding which organisms of the intestinal microflora are responsible for controlling Qandida in the gut. REFERENCES 1. Argenzio, R.A. 1981. Short-chain fatty acids and the colon. Digest Dis SQi 26:97-99. 2. Bergeim, O. 1940. Toxicity of intestinal volatile fatty acids for yeast and Escherichia ccli- J.Infcct.Dic 66:222-234. 3. Bergeim, 0., A.H. Hanszen, L. Pincussen, and E. Weiss 1941. Relation of volatile fatty acids and hydrogen sulphide to the intestinal flora. d intact Dig 69:155-166. 4. Balish, E., and A.W. Phillips. 1966. Growth and virulence of Qandida albicana after oral inoculation in the chick with a monoflora of either Escnetichia coii or Streptococcus fcccclic. _J_Bcctcricl 91: 1744-1749- 229 5. Balish, E., and .A.W. Phillips. 1966. Growth, morphogenesis and virulence of Qeggige elbieene after oral inoculation in the germ-free and conventional chick. g gecterio; 91:1736-1743. 6. Clark, J.D. 1971. Influence of antibiotics or certain intestinal bacteria on orally administered Qeggige elbicens in germ-free and conventional mice. Infee; Immun 4:731-737. 7 . Ekenna, o. , and R.J . Sherertz . 1987 . Factors affecting colonization and dissemination of gengige albicans from the gastrointestinal tract of mice. igjeet Immun 55:1558-1563. 8. Freter, R. 1983. Mechanisms that control the microflora in the large intestine. In: D.J. Hentges (Ed.), Hugen intestinal microfloge in health end giseese, p. 33-54. Academic Press, Inc., New York. 9. Freter, R., H. Brickner, M. Botney, D. Cleven, and A. .Aranki. 1983. mechanisms that control bacterial populations in continuous-f low culture models of mouse large intestinal flora. Ingeeg leggn 39:676-685. 10. Freter, R., E. Stauffer, D. Cleven, L.V. Holdeman, and W.E.C. Moore. 1983. Continuous-flow cultures as 1Q 230 £1332 models of the ecology of large intestinal flora. Ihjeeh Immhn 39:666-675. 11. Hummel, R.P., M.P. Maley, P.W. Miskell, and W.A. Altemeier. 1975. Suppression of Qandida elhieene by Eschegichia eeli. g Igehme 15:413-418. 12. Hummel, R.P., E.J. Oestreicher, M.P. Meley and B.G. MacMillan. 1973. Inhibition, of ‘thgige,lelhieehe_ by Eschegchie coli ih vitro and in the germfree mouse. Journal ef Surgieei Reeeereh 15: 53-58. 13. Isenberg, R.D., M.A. Pisano, S.L. Carito, and J.I. Berkman. 1960. Factors leading to overt monilial disease. I. Preliminary studies of the ecological relationship between gengige elhieehe and intestinal bacteria. Ahgihie§_ghehe 10:353-363. 14. Kennedy, M.J. 1988. Adhesion and association mechanisms of Qandide M. In H.R. McGinnis (ed): Charreht [topics in hediee]. Mycolmy, 19]. a. New York, Springer-Verlag, pp 73-169. 15. Kennedy, M.J. 1987. Role of motility, chemotaxis and adhesion in microbial ecology. WW geiehee 506:260-273, 1987. 231 16. Kennedy, M.J., P.S. Bajwa, and P.A. Volz. 1982. Gastrointestinal inoculation of mm eehenehii in mice. Mycopathelegie 78:141-143. 17. Kennedy, M.J., A.L. Rogers, and R.J. Yancey. 1988. An anaerobic continuous-flow culture model of interactions between intestinal microflora and Qandid3 e_l_b_i_<_:_ahe. Myeopathologia 103:125-134. 18. Kennedy, M.J., and P.A. Volz. 1983. Dissemination of yeasts after gastrointestinal inoculation in antibiotic-treated mice. Sabohraudia 21:27-33. 19. Kennedy, M.J., and P.A. Volz. 1985. Effect of various antibiotics on gastrointestinal colonization and dissemination by candida M. W let M 23:265-273. 20. Kennedy, M.J., and P.A. Volz. 1985. Ecology of Qandide aibicans gut colonization: Inhibition of gehgige adhesion, colonization, and dissemination from the gastrointestinal tract by bacterial antagonism. infee; Ihhhh 49:654-663. 21. Kennedy, M.J., P.A. Volz, C.A. Edwards, and R.J. Yancey. 1987. Mechanisms of association of mg; eibieans with intestinal mucosa. W 24:333- 232 341. 22. Krause, W., H. Matheis, and K. Wulf. 1969. Fungaemia and funguria after oral administration of Candiee albicehs. LQDQQE. 1, 598-599. 23. Miles, M.R., L. Olsen, A.L. Rogers. 1977. Recurrent vaginal candidiasis: Importance of an intestinal reservoir. J Am Med ASSOC 238:1836-1837. 24. Myerowitz, R.L., G.J. Pazin, and C.M. Allen. 1977. Disseminated Candidiasis: Changes in incidence, underlying diseases, and pathology. Ah e Ciin Beth 68:29-38. 25. Nishikawa, T., H. Hatano, N. Ohnishi, S. Susaki, S. and T. Omura. 1969. Establishment of theiee aibicehs in the alimentary tract of the germfree mice and antagonism with Eseheriehia coii after oral inoculation. gee J Micgobioi 13:263-276. 26. Nystatin Multicenter Study Group. 1986. Therapy of candidal vaginitis: The effect of eliminating intestinal Cendida. Am J Obstet gyneeol 155:651-655. 27. Odds, F.C. 1988. Caheide end geneieesie, University Park Press, Baltimore. 233 28. Ozawa, K., K. Jabu-Uchi, K. Yamanaka, Y. Jamachita, K. Ueba, and T. Micwatani. 1979. Antagonistic effects of LiL_aC' lus e_cat and Sircpcccccccc fcccelic on growth of Qaneida albicehs. hierehiei_lmhehei 23:1147-1156. 29. Paine, T.F., Jr. 1958. The inhibitory activity of bacteria on Candida growth. We 8:273- 281. 30. Stone, H.R. 1974. Studies in the pathogenesis, diagnosis and treatment of Candida sepsis in children. g Pediatr Surg 9:127-133. 31. Stone E.H., C.E. Geheber, L.D. Kolb, and W.R. Kitchens. 1973. Alimentary tract colonization by Caheiea eihieehe. J §urg Res 14:273-276. 32. Stone, K.H., L.D. Kolb, C.H. Currie, C.E. Gebeher, and J.F. Cuzzell. 1974. Candida sepsis: pathogenesis and principles of treatment. Ann Surg 179:697-711. 234 Table 1. Minimal inhibitory concentrations for short-chain fatty acids against Caheida elbicens. ma. MIc identification Acetic Propionic lsobutryic Butyric Isovaleric Valeric CA30 SDB grown 40 40 20 1O 20 10 CAD30 LBC yeast 40 40 20 20 10 S grovvn CA30 "singlets" 40 40 >20 40 20 10 CA34 SDB grown 40 20 20 10 10 2.5 "Gut levels" (mM) 116.1 20.4 4.5 20.8 3.8 1.3 Figure 1 a&b. 235 Passage of Q. elhieehe through continuous- flow cultures of mouse (a) and human (b) intestinal flora. L0910 C. albicans per ml Lo910 C. albicans per ml Dilution Rate 1 l l 20 30 40 50 60 70 80 nouns POST Inocuumou ‘- Dilation .. In 1 20 30 40 50 60 70 80 nouns POST mocuumou 236 Figure 2. Survival and implantation of g. elhieehe through continuous-flow cultures of established mouse cecal bacteria ([3), freshly inoculated cultures ((3) , or sterile anaerobic control cultures (0). ..... n... e.- ......... ..... "n- on. .. L0910 4 C. albicans 3 per ml 2 <— Dilution""-.... Rate A l L h l 10 20 30 40 50 60 70 80 HOURS POST mocuumou Figure 3 a&b. 237 Growth of Cahdiee aibicahs in filtrates of continuous-flow cultures of mouse cecal bacteria (a) and human fecal flora (b). Optical Denna! at Opfiafl Denna; at 2.0 1.5 1.0 0.5 2.0 1.5 1.0 0.5 O - CFC Medium (aerobic) o - CFC Modiumhnmobic) . - Effluent (aerobic) - f" . aniig'iio'coiiiéim' (NH4)2504. . - CFC medium 9 VFA (aerobic) O - CFC Mediumhuobic) O I CFC Medium (anaerobic) . a Effluent (aerobic) a I , A $3333.13“: W1”:- - - CFC medium 9 VFAiaetobic) DISCUSSION DISCUSSION i.!_ O ‘ SOC .X2!Q1-i- g- 0 ‘;_IP:_;=CAN The "attachment" of Q. elhieehe to both biological and nonbiological surfaces is considered an important event in colonization and pathogenesis of humans (Kennedy, 1987 & 1988). Moreover, adhesion and aggregation may also be an important virulence factor that allows this fungus to resist host defenses and antifungal therapy. Because of the obvious importance of theiee adhesion and aggregation to human health, then, numerous ih 213:9 and ih vivo models have been developed to study the adhesion of Q. elbicans to various nonbiological surfaces and host tissues (Kennedy, 1988). Several models, for instance, have been developed to study the adhesion and association of Q. elbieans to mucosal surfaces, certain organs, and prosthetic devices. These models have been used with some success to characterize Q; eihieehe adhesion and association mechanisms, and the role attachment plays in colonization and pathogenesis of this important pathogen (Kennedy, 1988). However, the use of many of these same models has also led to several discrepancies in the literature as to the cell wall component(s) of Q. eihieehe that mediates its attachment to, for example, epithelial cells as well ast to ‘the :mechanism(s) of invasion of mucosal and endothelial surfaces (Kennedy, 1988). It has 238 239 been suggested, for instance, that a mannose containing moiety, probably a mannoprotein (Critchley & Douglas, 1987; McCourtie & Douglas, 1985a), is the cell wall adhesin of Q. eihieeee (Lee & King, 1983; Sandin, 1987 a&b: Sandin & Rogers, 1982: Sandin e3 e1., 1982), whereas analogous studies have suggested that chitin is the adhesin (Segal eh ei., 1982). Other studies suggest that there may be more than one adhesin (Kennedy e; ei., 1987: McCourtie & Douglas, 1985a: Sandin, 1987a) and at least one report suggests that the adhesion of Q. eihieehe may not be mediated by a specific adhesin-receptor interaction but may be entirely nonspecific (Reinhart e; el., 1985). The obvious means of resolving such extreme positions lies first in the development of models that closely mimic the infection or adhesive process as it occurs in the natural host or’ environment and second in ‘the proper critical interpretations and caution that must be applied to the data obtained with adhesion models. As has been described in detail elsewhere (Kennedy, 1988) , there are numerous methodologic pitfalls that accompany the use of adhesion models. A. survey of the literature on the adhesion of Q. elhigehe to epithelial cells showed that this fact has not generally been appreciated (Kennedy, 1988). The reasons for this lie not only in the fact that few critical studies have examined experimental factors that influence adhesion data but also because adhesion models to study fungal attachment mechanisms have been in 240 use for a relatively short period of time. That is to say, knowledge of fungal adhesion and aggregation mechanisms and the factors that influence fungal attachment is still in an early state. Thus at present, one can still only speculate at the actual contribution of many of the numerous attachment mechanisms that have been identified using various adhesion models. Although any adhesive reaction observed for Q. albicans in an adhesion model must be viewed as a real event, rigorous proof that adhesion is necessary for colonization, pathogenesis, or both for a number of adhesion mechanisms that have been identified to date must still be proven. For instance, if an adhesion mechanism is observed in an experimental animal model, it cannot epgiegi be assumed to be important in human disease (Kennedy, 1988). Likewise, any adhesive reaction observed for Q. efligehe in a simple adhesion assay performed in vitro should not automatically be assumed to represent adhesion in humans. This is true whether the adhesion is in the oral cavity, or to the surface of a prosthetic device. Moreover, at present there seems to be little appreciation for the complexities of host-fungal interactions including the interdependence of factors that regulate colonization, the various factors that can regulate adhesion and aggregation, and the redundancy of some of these mechanisms. For example, as will be discussed below, there is more than one mechanism by which 241 Q. eibicans can interact with a surface so as to reduce its rate of removal. Although numerous other examples of the complexities of host-fungal interactions could be given, all of these point to the fact that a satisfactory understanding of the role any one mechanism of adhesion has in colonization or pathogenesis comes from: (i) an elucidation of the biochemical and molecular basis of that mechanism and (ii) an understanding of its precise interaction with all the other mechanisms that collectively regulate the sequential processes of colonization and pathogenesis. Consequently, it is suggested that the study of fungal attachment should include: (i) the development of model systems that closely mimic the physicochemical environment of the host and (ii) models that address the interaction of adhesion mechanisms with other factors that are equally, and in some instances more, important in colonization and pathogenesis. As noted above, efforts to study Q. eihieehe adhesion and association mechanisms have led to the development of numerous models. These models have included both ih xihre and ih 1113 models, and have included models to study interactions between Q, .eihieehe, and. host tissues or nonbiological surfaces as well as intrageneric and intergeneric microbial aggregate formation. In addition, several models have been developed to study various cell surface properties of both yeast and hyphal forms of Q. eibicans that may play a role in adhesion and aggregation. 242 Although the use of adhesion models, and simple in yihze assays in particular, has led to several erroneous conclusions regarding the nature of how Q. eihieehe "attaches" to host tissues or the role certain adhesive reactions play in colonization and pathogenesis, there can be no reasonable doubt that very little progress in the understanding of theiee attachment could be made without their use (Kennedy, 1988). Nevertheless, it is important to note that there are also a number of limitations that accompany the use of adhesion models. For instance, the pathogenesis of any fungal infection, whether it involves Q. eihieehe or some other fungal pathogen, is the result of a complex series of separate but interdependent interactions that cannot all be reproduced ih vitro. Consequently, simple ih yihze assays may yield little information to the more important question of the role of Candide attachment in colonization and pathogenesis. Moreover, adhesion may only represent one step in the sequential process of colonization and in this process one type of adhesion (i.e. initial or reversible adhesion) may yield itself to yet another type of adhesion (i.e., secondary or permanent adhesion). Thus in some instances adhesion may only represent a transient step in colonization or pathogenesis. Furthermore, in the host, Q. elbicans may attach to surfaces by more than one attachment mechanism. As one example, there appears to be at least five distinct mechanisms by which Q. elhieehe can 243 associate with intestinal mucosa (Kennedy e; e1., 1987). The complexity of theiee-mucosal association, therefore, makes it difficult to devise a single model system (e.g., adhesion to isolated intestinal epithelial cells) which can be relied on to duplicate in yitige the process as it occurs in 21:39 (Kennedy, 1988). Therefore, while ih Lim systems remain valid tools to identify, differentiate and study various adhesion and aggregation mechanisms of theiee at the molecular level, proper critical interpretation as well as caution must be applied to the data when extrapolating to the situation found ih yige or in the natural environment. It may be, then, that the continued use of any in 113132 model of adhesion may require demonstration of the similarity of the adhesion or aggregation mechanism(s) gm yihge to the process as it occurs in humans (Kennedy, 1988). A complete understanding of the physicochemical or molecular basis of Q. eihieehe adhesion or aggregation, and the role that a particular mechanism plays in colonization or pathogenesis, may only be gained by studying each aspect of adhesion (e.g., the mechanism(s) of adhesion versus the role of attachment in colonization and pathogenesis) separately in different adhesion models. For instance, it may be necessary to initially study the adhesion of Q. eihieehe to intestinal mucus gel and epithelium seperately ih yiege, but thereafter to proceed to more complex ih vihgo models (e.g., intestinal slices), 244 followed finally by studies in experimental animals. Only by following this type of strategy can one be sure of the relevence of the adhesive reaction studied [in jyihxe. Furthermore, only by using several models to study adhesion and the role adhesion plays in colonization and pathogenesis can important information regarding the regulation and synthesis of adhesins as well as factors that might influence (either positively or negatively) adhesion be gained. An examination of factors present in the gut of conventional mice but absent in the gut of antibiotic-treated mice, for instance, has revealed that several bacterially-derived substances (produced by the metabolic activity of the indigenous microflora) may inhibit the attachment of Q. elhieene to intestinal epithelium in the noncompromised host (Kennedy & Volz, 1985a). Thus factors present in a given ecosystem may be as important as the presence of an adhesin itself in controlling adhesion. Finally, this type of strategy has previously' been ‘used. to show' that not all attachment mechanisms observed in gigge models of adhesion (Freter & Jones, 1976; Jones eh ei., 1976: Jone & Freter, 1976) are involved with the association of microorganisms with host tissue ih yiye (Freter eh ei., 1981 a&b). There are a number of experimental factors that can influence the adhesion or aggregation of Q. M in yitre and ih gilt; (Tables 6 & 7). Although a number of factors have been shown in the present studies to 245 Table 6. Factors affecting fungal aggregation and adhesion to biological and nonbiological surfaces in vihzo* Factor Examples Environmental factors Epithelial factors Fungal factors Growth factors Agitation speed Antibiotics Assay medium Assay temperature Incubation time pH of assay medium Ionic strength Yeast cell concentration Body site of origin Cell type Donor Indigenous microflora Viability "Cell-type" or "phenotype" Coadhesion Morphology Strain Viability Growth phase Medium composition Medium viscosity Temperature *Modified from Kennedy (1988). 246 Table 7. Factors influencing fungal aggregation and adhesion in vive Factor Examples Host factors Antimicrobic therapy Body site Diet Environmental factors (eg, pH) Hormonal status Immunologic status Number and type of receptors available Physiologic status Indigenous microflora Alteration of physiochemical nature of environment Colonization of adhesion sites Modification of substrate Production of inhibitors Fungal factors "Cell-type" or "phenotype” Concentration and type of adhesins synthesized Enzyme production (eg, proteinase(s) Germination Morphology Surface properties *Modified from Kennedy (1988). 247 influence the adhesion of Q. eihieehe to BECs and plastic (Kennedy, 1988) , growth conditions are probably the most important (Kennedy, 1988: McCourtie & Douglas, 1981 & 1984). As can be noted from Table 6, not only can the phase of growth modify adhesion but so can the growth temperature and growth medium composition. Of the factors that can influence Cehdide adhesion and aggregation, growth medium has been suggested to be the most important (Kennedy, 1988). While this may seem intuitively obvious, it is somewhat surprising to note that over 15 different media have been used in the literature for studies on Q. alhicans adhesion (Kennedy, 1988). McCourtie and Douglas (1981 & 1984) noted in studies in the early 1980's that growth of Q. elbicans in media that enhanced adhesion to BECs and acrylic correlated with changes in yeast cell surface composition. Yet a recent survey of the published data on the adheion of Q. albicahs to BECs showed that although these findings are often cited they have not generally been appreciated (Kennedy, 1988). Of particular interest is the fact that very few investigators have used the same growth medium, and that there are major differences between media used by those investigators proposing that different cell surface moieties are responsible for ‘Q. eihieehe adhesion (Kennedy, 1988). Because of this, because there was no way of comparing previously published adhesion data directly, and since there has been little or no agreement as to the selection 248 of the assay conditions for a standardized buccal epithelial cell adhesion assay for W, the effects several growth media (as well as various growth conditions) had on Q. elhieehe adhesion to BECs and plastic were studied. To determine the effect growth medium had on the adhesion of Qandide to BECs, a virulent and adhesive strain of Q. elbieens (AK785) isolated from the mouth of a 1 week-old female with oral thrush was grown in several media and examined for its adhesiveness to BECs. The ability of Q. albicens to adhere to BECs was found to be highly dependent on the growth medium used for cell growth. As can be noted from the Table 2 from article 1, theiee cells grown in undefined liquid media at 37°C were usually less adhesive than those cells grown in defined media at the same temperature. However, for one defined medium (YNB) the production of highly adhesive cells was dependent upon the type of carbohydrate present in the medium during cell growth. This finding confirms the earlier results that cells of Q. eihieehe grown in YNB + 500 mM galactose are significantly more adhesive than cells grown in YNB + 50 mM glucose (McCourtie & Douglas, 1981 & 1984). Moreover, significant differences in adhesion of Q. elbicahs to BECs were noted when the cells were grown in the same complex medium from different manufactures as well as from different lots of medium from the same manufacture. Other differences in cell growth 249 parameters, such as growth temperature and whether the cells were grown on agar or in broth, also had a significant effect on Q. elhieehe adhesion data. All of these factors also modified Qeneiee cell surface hydrophobicity and the ability of Q. elhieehe to attach to BECs indirectly by yeast-to-yeast coadhesion. To determine whether differences in morphology, cell wall ultrastructure or various cell surface features could account for some of the differences in Q. eihieene adhesion and coadhesion noted, thdide cells were also examined by SEM and TEM. Although several morphologic characteristics were noted, only the presence of tiny germtube-like structures, which were almost imperceptible by light microscopy, and very small blastospores appeared to correlate with increased yeast adhesiveness. Of all the cell types observed in that study, small blastospores were the most adhesive followed by yeast cells with short germtube-like structures, yeast cells with longer germtubes, and pseudomycelia and yeast cells, respectively. Furthermore, although differences were noted in cell wall thickness, the number of distinct cell wall layers, and the presence of an extra electron dense cell wall layer, none of these differences appeared to account for either an increase or decrease in adhesion, coadhesion or cell surface hydrophobicity. Similarly, the presence of an outer cell wall floccular material, which previously was shown to confer an increase in Q. eihieene 250 adhesion to BECs (McCourtie & Douglas, 1984), did not appear to correlate with m adhesiveness or cell surface hydrophobicity. These findings clearly demononstrate the importance of growth parameters on the morphology, budding pattern, cell wall ultrastructure and various surface properties of Q. alhicans. These include not only cell surface hydrophobicity but also cell surface topography and the ability of Qandiea cells to attach to BECs either directly (adhesion) or indirectly (coadhesion) . Although these data suggest that growth conditions can significantly influence Q. albicans adhesion data, they do not allow definitive conclusions to be drawn regarding the increased adhesiveness of small blaStoconidia or cells with germtubes. Nevertheless, there are at least three possible mechanisms which could account for this. (i) It is apparent that there are major changes in the organization of the cell wall at the site of germtube formation (Cassone eh ei., 1973; Schertwitz eh el., 1978: Tronchin eh e_., 1984), which has been shown to occur in infected tissue (Montes & Wilborn, 1968 8 1985). These changes may allow for a concentration of adhesins or the production of different adhesins, both of which may increase the adhesiveness of Q. eihieehe (Kennedy, 1988). There is some evidence that Q. elhieehe does produce more than one adhesin (Kennedy e§_‘e1., 1987: McCourtie & Douglas, 1985a: Sandin, 1987b). Furthermore, recent 251 studies have shown that germtubes of Q. elhieehe possess cell surface antigens which are distinct from those found on the surface of budding yeast cells (Brawner & Cutler, 1985 & 1986: Ponton & Jones, 1986: Smail & Jones, 1984: Sundston & Kenny, 1984: Sundstrom e_t; ei., 1987). It should be noted, however, that it is not known whether these antigens serve as adhesins or not, but these findings again emphasize that there are major differences between cell wall composition of yeast cells and germtubes (Cassone e; e_., 1973). This fact becomes very important, then, when selecting a growth medium for Q. e_lhieehe adhesion studies. (ii) There is evidence to suggest that production of certain enzymes (e.g., proteinases) in concert with germtube formation may facilitate adhesion (Ghannoum & Elteen, 1986). (iii) It is likely that the germtubes, which have very small radii of curvature compared to the yeast cell, would favor adhesion by reducing repulsive interactions that occur during adhesion between cells (Kennedy, 1988). The latter factor may also help to explain why very small blastospores of Q. eihieehe were found to be significantly more adhesive to BECs than were "normal" size yeast cells that did not possess germtubes. The differences noted in Q. eihieehe cell wall ultrastructure and adhesion to BECs with respect to various growth parameters, especially growth medium, raise questions regarding the location and biochemical nature of 252 the theiee adhesin(s). For instance, at least two morphologic classes of theicie adhesins (floccular and fibrillar [reviewed in Kennedy, 1988]) have been noted in the literature, both of which. have been observed to mediate Q. eihieehe adhesion in yiye (Barnes e; e1., 1983: Calderone e_t; e]._,,, 1984: Montes & Wilborn, 1968 & 1985). In contrast, it has also been shown that the adhesion of Q. elbieane to epithelial cells was mediated by cells that did not contain such outer surface layers (Calderone e_t ei., 1984). Although cells of Q. elhieehe with a floccular or fibrillar cell wall layer have been observed to apparently mediate adhesion to various tissues, as have Cahdida cells with no apparent adhesive appendages, no study to date has determined the similarity of adhesins produced in yihpe and in yiye. These will be key future experiments for the selection of growth conditions and assay parameters for the design and standardization of relevent and useful adhesion models. This seems especially true in view of the finding that the presence of either floccular or fibrillar structures on the cell surface of Q. eihieehe does not, by itself, indicate that these structures play a role in either adhesion. theiee cells grown in TSB, for instance, were found to contain an outer floccular cell wall layer, and were relatively nonadhesive to BECs. In contrast, growth of Q. eihicehe in YNB + gal produced cells that contained an outer floccular layer which were highly adhesive to BECs. 253 Furthermore, growth of Q. eihieehe in other media produced cells that were highly adhesive to BECs that had relatively thin cell walls with no "extra” outer layers. Thus, these findings might be interpreted to suggest that neither of the aforementioned structures is required for adhesion of Q. eihieehe to BECs, although they may mediate attachment. It may also be that after the initial binding, Cahdide cells modify their cell walls, and floccular or fibrillar structures are synthesized which stabilize and strengthen adhesion (Kennedy, 1988) . This is supported by the data of Tronchin e1; e1. (1984) who noted the reorganization and proliferation of an external cell wall layer of Q. eihieehe during adhesion to BECs. The observation that the floccular layer was produced only amorphously around the cell in some instances, and that not all cells in a population had the material, however, may also correlate with the suggestion that adhesins are distributed unequally in populations of Q. eihieehe (McCourtie & Douglas, 1985b) . This suggests that the biochemical nature of the "floccular" layer noted between cells grown in various media may be very different, and some Ibut not all such. structures ‘may' contain .anéiflc adhesins (Kennedy, 1988). The findings that growth of Q. eihieehe in various media produced significant differences in cell wall ultrastructure, and presumably cell wall composition, may also clear up previous contradictory results in the 254 literature regarding the nature of the cell wall component that serves as the theige adhesin (Kennedy, 1988). For instance, Segal and co-workers (Lehrer e; e1., 1983; Segal eh ei., 1982) have suggested that cell wall chitin, which is located at the innermost portion of the cell wall or at bud scars (Poulain e; ei., 1985: Tronchin e; e1., 1981), may serve as the adhesive component. In contrast, Douglas and co-workers have suggested that mannoprotein is the Candida adhesin (Critchley & Douglas, 1987: McCourtie & Douglas, 1985). It is important to note that the growth media used by these two groups is extremely different and yields cells that have a significantly different cell wall ultrastructure. Growth of Q. elbicens in the former medium (yeast extract medium) yielded cells of Q. eihieene that possessed the thinnest cell wall with the fewest apparent layers, whereas the latter medium (yeast nitrogen base supplemented with 500 mM galactose) yielded cells with thicker cell walls that possessed an outer floccular layer. From these data, and the data presented above, it is suggested that at least three different adhesins exist on the surface of Q. elhieehe that mediate its attachment to BECs: (i) a cell wall adhesin (demonstrated best when cells are grown in LBC broth), (ii) an adhesin in the floccular layer (YNB + 500 mM galactose), and (iii) an adhesin expressed below the surface that becomes evident when the outer electron layer is thinned (yeast extract medium). The variations in cell wall ultrastructure and 255 adhesion noted from these studies, then, underscore the capacity of this organism for cell surface modification in yihge and the importance of standardizing fungal adhesion assays. Moreover, one is left to wonder which, if any, of the numerous studies reported in the literature mimic adhesion to buccal epithelial cells as it occurs in yiye. Another important factor that may significantly alter fungal adhesion and aggregation data comes from the finding that Q. albicans is capable of switching, reversibly and at a high frequency between several phenotypes that are readily distinguishable by the size, shape, and color of’ colonies formed on agar’ at 25°C (Slutzsky eh el., 1985 8 1987). The differences noted in colony characteristics apparently were due to the dramatic difference in cell size, shape, and budding pattern (Slutzsky eh 1., 1985 & 1987). Two of these phenotypes (”white” and "opaque") were examined for their ability to attach to BECs and plastic, and to compare their cell surface Ihydrophobicities. ”White cells" are :round 'to ellipsoidal and exhibit a budding pattern similar to most strains of Q. aibieens, whereas ”opaque cells" are elongate, or bean shaped, and exhibit a different budding pattern. As was shown in the studies described above, white cells were found to be significantly more adhesive to BECs than were opaque cells. In contrast, opaque cells were found to be twice as hydrophobic as white cells as was the percentage of opaque cells that attached to BECs 1:1 256 by coadhesion. The differences in adhesion to plastic between the two phenotypes were not statistically significant and there was no distinct trend to suggest which phenotype might be more adhesive to plastic. These results indicate that several factors are involved in the adhesion of Q. eihieehe to plastic, and confirm the hypothesis that cell surface hydrophobicity is of minor importance in direct adhesion to epithelial cells but that it may contribute to indirect attachment to epithelial cells by promoting yeast coadhesion. Moreover, these findings revealed that under identical growth conditions, cell surface topography, cell surface hydrophobicity, and adhesion of Q. elbieehs was significantly altered depending on the phenotypic state of the organism tested. The above findings further stress the importance of culture conditions in the study of fungal adhesion and aggregation as well as maintaining a proper perspective on the data obtained from simple 3 mm adhesion assays. For instance, while the opaque phenotype of Q. _a_ihieehe may predominate in some environment (e.g., soil or water) other than the host, differences in the adhesive and cell surface properties between white and opaque phenotypes may yield important information about how theiee cells attach to biological and nonbiological surfaces. Therefore, although the importance of the white-opaque transition in colonization and infection has not been established, the white-opaque system lends itself as a useful tool to aid 257 in the characterization of Q. elhieehe adhesion mechanisms. In view of the findings described above, and the descrepancies regarding which surface component(s) serves as the Candida adhesin(s) , the need for a defined and reproducible assay to study adhesion of Q. eihieege to BECs is very apparent. The fact that over 15 different growth media have been used to cultivate theiee cells for adhesion studies, coupled with the numerous other variations in experimental factors used by different investigators, further stresses this point but also suggests that at present there is little, if any, agreement on the selection of criteria required for the standardization of a buccal cell adhesion assay (Kennedy, 1988). Because such an assay is important to describe Q. aibieahs adhesion at the molecular level and to confirm the role an adhesion mechanisms might play in colonization or pathogenesis, the beginnings of a standardized assay are outlined in Table 8 and the paragraphs below. The most important factor to be considered, and the only one that will be discussed here, is the selection of a culture medium and growth conditions (Kennedy, 1988) . It is clear from the data discussed above, that uniformity of adhesion results between lots of complex commercial media (e.g. , Sabouraud dextrose broth, which apparently has been the medium of choice [see Kennedy, 1988]) is unlikely. This is not too surprising since peptone based 258 Table 8. Experimental conditions for the beginnings of a standardized buccal epithelial cell adhesion assay for Qandide clbiccns* Assay parameter Specific condition Assay parameters 1. Temperature: 37°C 2. Yeast: Buccal ratio (100: 1) 3. Reaction time: 1 hour 4. Assay buffer: Buffered KCl at pH 6.81 Epithelial cells 1. Source 5 to 10 healthy adults (male & female) 2. Buffered KCl at pH 6.8 Yeast cells 1. theiee eihieehe AK785 (as the reference strain) 2. Growth media: LBC + 50 mM glucose YNB + 500 mM galactose 3. Growth phase: early stationary 4. Growth temperature: 37°C *Designed from criteria described by Kennedy (1988). 1Buffered KCl is a solution that mimics the ionic composition of saliva (Gibbons e; e_., 1985). 259 media are highly complex and undefined. Major differences in the chemical composition of peptones and "Sabouraud's medium” from laboratory to laboratory, from manufacturer to manufacturer, and from lot to lot of medium from the same manufacturer has been noted by Odds and collegues (Odds, 1985: Odds eh e1., 1978). Moreover, the use of complex and undefined media can modify several cell surface properties (e.g., cell surface .hydrophobicity) which may directly or indirectly modify adhesion of Q. elbicans to epithelial and plastic surfaces. Therefore, it is suggested that the medium of Lee e; ei. (Lee eh e1., 1975) and yeast nitrogen base (Difco Laboratories) supplemented with 500 mM galactose (McCourtie & Douglas, 1981) be used because both media are chemically defined and produce cells of Q. elhieehe that are highly adhesive. Furthermore, both media apparently yield 2mm cells that have different adhesins and different cell surface characteristics. Early stationary phase cells are also suggested because stationary phase cells have been shown to adhere more readily than logarithmic phase cells (Kimura & Pearsall, 1978: King e; ei., 1980: McCourtie & Douglas, 1981 & 1984: Sandin e_t e1., 1982) and growth should be at 37°C (shaken at 100 rpm) (Kennedy, 1988). Several other factors which are important to the development of a standardized buccal cell adhesion assay have been described in detail elsewhere (Kennedy, 1988). Nevertheless, further steps toward the standardization of 260 gm adhesion assays will be necessary if the exact nature of Q. eihieehe adhesion to BECs and other surfaces is to be understood. Because there are few similarities between the results of different investigators, the assay described in Table 8 was designed to incorporate what is presently known regarding Q. elhieehe adhesion to BECs and to similate the environment of the oral cavity (e.g., the assay medium was designed to mimic the ionic composition and pH of saliva [Gibbons e; ei., 1985]). Given that this adhesion assay is probably incorrect in some respects, its use could lead to the formulation of incisive questions to be addressed in the future as well as to help determine which experimental parameters best mimic the interaction of Q. eihieehe with BECs as it occurs in yiye. Therefore, it is urged that this assay be adopted for use with Q. aibicans, even if only to determine which experimental conditions are to be kept and which are to be changed. G', '0_N. STIN L CO,0 I_;T O ; xi? 9; ;_=, ; The survival, implantation, and dissemination of Qenfiiee QM and other QeDQiee species from the GI tract appear to be influenced by a number of factors. The presence of an intact and functioning indigenous microflora, however, is probably the most important deterrent to whether Q. eihieehe can persist in the gut and at what population levels. Several studies, in fact, have shown that intestinal bacteria. are inhibitory' to 261 Mich ih lien; or in the gut of experimental animals containing a diflora of Q;_ e1hieene_ and some single bacterial species (e.g., Eeeheziehie eeii) (Balish 8 Phillips, 1966b: Hummel 9.1: a]... 1973 s 1975: Isenberg e_t LL, 1960: Nishikawa e3; ei., 1969: Ozawa e3; e1., 1979: Paine, 1958). Nevertheless, although these early studies suggested that intestinal bacteria play an important role in controlling Q. eihieehe in the gut, they have led to several erroneous conclusions regarding the identity of the organisms and mechanisms responsible for the suppression of theiee in the GI tract. This is probably due to the use of inappropriate experimental models that do not reflect interactions between Qeneige and intestinal bacteria as they occur i_n M (Kennedy 8 Volz, 1985a). Therefore, the present studies examined the mechanisms which regulate Q. eibieans populations in yiye, and describes the suppression. of .theige. not. by' a single bacterial species but by an entire complex intestinal microflora. In addition, the beginnings of a general theory to explain the regulation of We population levels in the gut is put forth. As was described above, Q. elhieene and other fungi are able to readily colonize the GI tract of germfree and antibiotic-treated animals, whereas similar colonization of conventional animals or man is difficult to achieve. This in addition to numerous reports in the literature showing a high correlation between thgige infections and 262 antibiotic therapy (Odds, 1988: Rippon, 1988), has led to the now commonplace assumption that the incidence and severity of alimentary tract colonization by Q. eihieehe, as well as pathologic consequences that can be associated with this colonization, is somehow controlled by antagonistic interactions exerted by the numerous microbial species that. normally inhabit ‘the gut. .As mentioned already, several studies support this hypothesis and have shown that certain intestinal bacteria are inhibitory to both ih yihge growth and GI colonization by Qandida. and other' fungi (Odds, 1988). UnfOrtunately, however, most of the studies to date, although supporting this hypothesis, cannot be relied upon to reflect interactions as they’ normally occur in ‘the. GI tract. Nevertheless, although there may be little similarity between the inhibitory mechanisms observed in simple ih yihze studies and those found in the intestine, these and the data described in this dissertation do suggest that the suppression of Candida i_h y_i_\ge is likely to be the result of the interdependence of a complex set of interactions. Moreover, these data further suggest that these interactions are probably regulated, collectively, by a complex indigenous microflora, and not to one or even a few bacterial species as has been suggested previously (see Kennedy 8 Volz, 1985a). While it seems obvious that the normal flora is clearly important in the suppression of gut colonization 263 and dissemination by Q. eihieehe, the data presented here also help to clear up previous contradictory results reported in the literature regarding the effects of the "normal” intestinal microflora on GI colonization by this fungus. Balish 8 Phillips (1966 a8b: Phillips 8 Balish, 1966) in their early studies, for instance, had reported that Q. eihieehe established in "large numbers" in the gut of both germfree and conventional animals, whereas DeMaria et ei. (1976) and Helstrom 8 Balish (1979) in later experiments found that conventional animals were significantly more resistant to intestinal colonization than antibiotic-treated and germfree animals, respectively. Such discrepancies, however, may be due to the animals selected for study, since many strains of mice are unsuitable for research on intestinal floral interactions because they lack a true indigenous microflora (see Kennedy 8 Volz, 1985b). Historically, the study of interactions between theiee and the indigenous intestinal microflora have included, with rare exceptions, two general approaches. The first type of approach is to estimate theiee population levels in the presence of intestinal bacteria (both ih yiye and ih yihge). The second includes determining how well thd__ide grows in a milieu after intestinal bacteria have grown in it first and then were removed or measuring thgiee, growth after certain chemicals have been added to the culture. However, 264 because of a number of accompanying technical difficulties, each of these methods is somewhat limited and can be misleading at times. This suggests that a combination of both types of studies may be necessary to accurately describe the survival, implantation and dissemination of Q. elhieehe from the GI tract. Moreover, because of the complexity of the interactions between Candi ee and the indigenous microflora, it is important that models that simulate, as closely as possible, the events as they occur in yiye be used if meaningful results are to be obtained (Freter, 1983). The anaerobic continuous-flow (CF) culture model of the ecology of the large intestinal flora described here meets these stringent requirements in that it reproduces a number of interactions as they occur in 11313 (Freter a ei. 1983b) . For instance, mixed populations of mouse cecal bacteria in CF cultures were maintained in the same proportions as in the mouse large intestine. Furthermore, the bacterial populations in CF cultures were able to resist implantation by ”invader" organisms (e.g., Qiostridieh M or Eggheniehie eeli) and reduced these populations to levels similar to those found after being fed to conventional mice (Freter et ei., 1983 a8b: Wilson 8 Freter, 1986). Moreover, it was found that Q. eibicens was eliminated from CF cultures of mouse cecal flora at a rate indistinguishable from that of the mouse cecum of conventional adult mice. A number of other 265 similarities between interactions involving indigenous and invader organisms, and of the ecosystems themselves, have been described by Freter (1983). Nevertheless, because of the similarities of interactions between Q. elhieans and bacteria in CF cultures and in the gut, this model was used to elucidate the bacterial mechanisms that control Caneida in the GI tract. These studies strongly support the hypothesis that the suppression of meiee in the gut is the result of a complex series of interactions that is regulated by an equally complex "group" of bacteria (Kennedy 8 Volz, 1985a) . Although several distinct mechanisms appear to collectively regulate the survival, implantation and dissemination of C_. eihieehe and other fungi from the GI tract (Kennedy 8 Volz, 1985a), these tend to fall into two general mechanisms: (i) antagonizing the growth of Q. aihicens and (ii) inhibiting the association of theiee with the intestinal mucosa. These have been summarized in Table 9. It is interesting to note from this table that a number of mechanisms are inhibitory to thdide. Discussions in the early literature of the possible inhibitory mechanisms which suppress thgiiee have suggested that anaerobiosis or competition for nutrients were the most important factors (Isenberg e; ei. , 1969: Paine, 1958). However, in studies using CF cultures it was found that anaerobic conditions did not prolong the 266 Table 9. Mechanisms by which the intestinal microflora inhibits theiee elhieene colonization and dissemination from the gastrointestinal tract Type Mechanism I Inhibition of mucosal association: formation of thick layers of bacteria that inhibit penetration into the mucosa competition for adhesion sites production of inhibitory substances (e.g., SCFA & DCBA) increased rate of disassociation modification of metabolic activity of Q. elhicans so that adhesins or digestive enzymes could not be produced II Regulation of the theiee population size: anaerobic condition of the gut production of inhibitory substances (e.g., SCFA, DCBA & H28) competition for limiting nutrients restricting substrate availability prolonging the lag phase and doubling time 267 lag phase of Q. eihieehe as did CF culture filtrates, indicating that other inhibitory mechanisms were also operating to suppress the growth of Caheiea. Moreover, the inhibitory effect on growth in CF culture filtrates could not be reversed by adding carbon or nitrogen sources, vitamins, or trace nutrients (individually or collectively) to CF culture filtrates and incubating the resulting cultures (under varying pHs) aerobically. Therefore, the suppression of Q; Me by intestinal bacteria was not due solely to depletion of nutrients and anaerobiosis, but appeared to be due to a combination of these factors in addition to the production of inhibitory substances as a result of the metabolic activity of intestinal anaerobes. Several breakdown products from the large number of primary dietary or host-derived substrates in the gut may antagonize Candide species, as might a number of substances synthesized by the microflora itself. The most likely inhibitory substances, however, would probably be those that are produced by several bacteria in the gut and, therefore, those that are produced in the highest concentrations. One such set of substances may be the short-chain fatty acids (SCFAs) . These are produced in relatively high concentrations by intestinal anaerobes (Argenzio, 1981), and it was found that both the total and individual levels of SCFAs present in CF cultures coincided with the quantity of acids found in the large 268 intestine of conventional mice or human feces used to establish the CF cultures. Because of this, because much has been speculated about the role of SCFAs in the control of the intestinal microflora (Argenzio, 1981; Bergeim e_t ei., 1941; Levinson, 1973) and since SCFAs have been shown to be inhibitory to £1eieeh_m_eh'_s_ yeast (i.e., Saccharomyees W) (Bergeim, 1941: Bergeim 95 e1” 1941) , these acids were tested for inhibitory activity against Q. albicans. These studies indicated that at normal concentrations SCFAs were inhibitory to M1313 growth, prolonging both the lag phase and doubling times. Further comparison between minimal inhibitory concentrations for individual acids and "gut levels", or levels found in CF cultures, showed that both acetic and butyric acid could antagonize theiee growth at the levels found in the GI tract. A similar argument could be made to support the role of deconjugated bile acids (DCBAs) in the suppression of emjee in the gut (Kennedy 8 Volz, 1985a) . For instance, deconjugated bile acids have also been shown to antagonize Caheide growth and to induce morphologic changes in the fungus when grown ih yiJr_o (Marshall e; eh, 1987; Kennedy, unpublished data). As will be discussed below, SCFAs and DCBAs have also been found to inhibit the adhesion of Q. eihieehe to intestinal mucosa (Kennedy 8 Volz, 1985a). It is likely that there are several substances which may antagonize the growth of theiee in the gut, however, 269 Freter (1983) has suggested that it is theoretically impossible to account for the total suppression of a sensitive microorganism in the gut solely on the basis of the production of growth inhibitors. If this were the case, Q. aihieans would be eliminated if too much inhibitor were present. Conversely, when too little inhibitor were present the theige population would increase until it became limited by some other mechanism (Freter, 1983). Although it has been reported that theiee population levels do show some increases or decreases from day to day in experimental animals (Ekenna 8 Sherertz, 1987), it still may be argued that other inhibitory mechanisms are also involved with the regulation of meme populations in the gut. For one thing, depending upon an individual's eating habits and the type of foodstuffs ingested at different meals, there must be fluctuations in the production of growth inhibitors. Moreover, studies with CF cultures have shown that Q. elbicehs can adapt itself and become resistant to certain "toxic" substances when it is implanted simultaneously with the intestinal microflora. The finding that Qendide, although replicating (albeit very slowly), could not persist in CF cultures of mouse cecal flora was probably due to a lack of adhesion sites or the inability to attach to the glass culture vessel. Similar studies performed in antibiotic-treated mice showed that (Q. aihieahs populations predominated 270 initially in the lumen but that mucosal-associated populations predominated after the intestinal microflora reestablished. Thus the adhesion and association of theiee with the GI mucosa may be an important determinant to colonization, since mucosal association would probably allow Q. eihieehe to remain in the gut even at a severely depressed growth rate (Kennedy, 1988) . Furthermore, it may also be that colonization of the mucosa would provide a more sheltered site from growth inhibitors than the lumen. The ability of the indigenous microflora to inhibit the association of Qemige with the intestinal mucosa, then, also represents an important defense mechanism that inhibits theiee colonization and dissemination from the GI tract. This seems especially true when considering that the first step in dissemination from the GI tract must be association with the mucosa. Although the dynamics of lumenal and mucosal- associated Q. alhicahs populations have not been completely defined, the importance of three distinct but interdependent steps in the association and passage of Candida through the mucosa have been implicated (Kennedy eh ei, 1987). These include: (i) contact with the surface of the mucus gel, (ii) penetration or trapping in the mucus gel, and (iii) adhesion and penetration of the epithelial surface proper. The first step in this process is probably an entirely random event, whereas the latter steps are not. The latter steps apparently involve the 271 interaction of specific adhesin-receptor binding and enzymatic modification of the mucus gel and epithelial surface for attachment to and penetration of GI mucosa (Cole e_t ei., 1988: Kennedy e_t el., 1987). Complicating this is the finding that Q. elhieehe could associate with the mucosa by at least five distinct mechanisms. These included: adhesion to epithelium, adhesion to mucus, co- adhesion to adherent fungi, co-adhesion to adherent bacteria, and entrapment in the mucus gel overlying the epithelium (Kennedy et el., 1987) . The physicochemical bases and ecological role each of these mechanisms may play in GI colonization by Qeneiee have been described in detail elsewhere (Kennedy, 1988). Despite all of this, the dense layers of bacteria lining the mucus gel do appear to play an important role in the resistance to GI colonization and dissemination by Candide (Kennedy 8 volz, 1985a). It has been noted that the first step in mucosal association, whether it is for Qendida or any microorganism, must be the penetration of the mucus gel (Freter, 1982). In infant mice, which lack a complete bacterial flora including the dense bacterial populations in the mucus gel (Davis e; ei., 1973: Schaedler eh ei., 1965), Q. eihieehe can readily associate with and pass through the mucosa to initiate systemic infection (Field eh el., 1981: Pope e; ei., 1979). Similarly, in adult mice treated with certain antibiotics to eliminate wall-associated populations of indigenous 272 bacteria, Q. eihieerie was able to associate with the GI mucosa and spread to visceral organs (Kennedy 8 Volz, 1983 8 1985b). In contrast, it was shown that intestinal tissues that possessed an indigenous wall-associated microflora strongly inhibited mucosal association (in yihge and in 2139) and dissemination of Q. elhicens from the GI tract (Kennedy 8 Volz, 1985a). Mucosal association by Candida was also found to be inhibited by a number of mechanisms (Kennedy 8 Volz, 1985a). For example, mucosal association by Q. albicens was blocked by competing for adhesion sites and physically blocking the larger yeast cells from penetrating into the mucus gel. In addition, data obtained from "adhesion assays" and examination of GI mucosa obtained from experimental animals challenged with Q. lbi ans, suggested that the mucosa-associated flora inhibited C_aeiide cells from penetrating deep into the mucus gel (Kennedy 8 Volz, 1985a). Moreover, the rate of disassociation from intestinal tissues by Q. eihieehe from antibiotic-treated animals was nearly three times that of untreated animals (Kennedy 8 Volz, 1985a). Nearly 75% of the .theiee_ cells that. had. associated. with. the cecal mucosa of untreated adult animals were removed within 1 h, whereas only about 30% of the thgiee cells were removed in the same time from animals treated with antibiotics (Kennedy 8 Volz, 1985a). Finally, certain chemical factors produced in the gut by the metabolic activity of the normal flora can inhibit 273 the association of Q. elhieene with intestinal mucosa (Kennedy 8 Vblz, 1985a). SCFAs and DCBAs, for instance, significantly reduced mucosal association of theiee cells in an adhesion assay that was performed under anaerobic conditions similar to that found in the gut (Kennedy 8 Volz, 1985a). Since environmental parameters are known to influence Qahdiee adhesion to vaginal and buccal epithelial cells (Lee 8 ‘King, 1983), it follows that modification of the GI environment by intestinal bacteria also modified attachment of Q. eihieehe to intestinal mucosa (Kennedy, 1988). Although the exact nature of this inhibition by SCFAs and DCBAs has not been elucidated, it has been suggested that these substances may have reduced the mucosal association of 93111111 by modifying Qandige adhesins or mucosal receptors, or both (Kennedy 8 Volz, 1985a). Alternatively, it may also be that these chemicals alter one or more physicochemical factors (e.g., the size of the double diffusion layer or the surface charge of yeast cells) that may cause a reduction in Qandida adhesion (Kennedy, 1988). NOnetheless, a correlation between the drop in the concentration of SCFAs and. DCBAs in 'the intestinal tract. and. an increase in mucosal association by Q. eihieehe following antibiotic treatment has clearly been shown (Kennedy 8 Volz, 1985a). Several lines of evidence suggest that the anaerobic populations, which dominate the intestinal flora (Moore, 1977), normally suppress thdiee in the gut (Kennedy 8 274 Volz, 1985a) . Although one recent publication suggests that the aerobic portion of the flora may also be involved with suppression of Q. elhieehe in the gut (Ekenna 8 Sherertz, 1987), the overwhelming evidence from both patient and animal studies indicate that resistance to colonization and dissemination by theiee is due to intestinal anaerobes. For instance, treatment of experimental animals with antibiotics that decreased the total populations of strictly anaerobic bacteria in the cecum, allowed an increase in facultative bacteria and promoted Q. bi s adhesion, colonization and dissemination from the GI tract (Kennedy and Volz, 1983 8 1985 a8b). In contrast, colonization and dissemination from the GI tract by M could not be induced by antibiotics that were active against aerobes or facultative bacteria but were sparing of the anaerobes (Andermont e1; e1., 1983; Wingard e_t; el., 1980). Studies in infant mice, furthermore, have shown that Q. a_lhieehe to maintain a stable population of about 103 to 104 per gram in the cecum and large intestine of the mice for the first two weeks of life, after which theiee counts decline to undetectable levels by four weeks (Field e3 e1” 1981). The timing of elimination of Q. eihigehe in infant mice, then, coincides with the acquisition and increase of populations of anaerobic bacteria, previously shown to occur during the second and third weeks after birth (Davis e_t; e_2l_.., 1973: Lee 8 Gemmelle, 1972; Lee Lt 275 Iei., 1971: Schaedler e; e1., 1965). It is also interesting to note that the increase in SCFAs and DCBAs in the large intestine also coincides with the acquisition of intestinal anaerobes (Lee 8 Gemmell, 1972) . Other evidence for the role of intestinal anaerobes in the regulation of theiee in the GI tract has been reviewed by Odds (1988). The studies described herein of the role of intestinal microflora in the control of theiee populations in appropriate models, therefore, may well clarify inconsistencies in the literature and indicate productive avenues for regulating GI colonization by Candiea in such a way that the beneficial effects are retained and the detrimental effects are eliminated. LITERATURE CITED LITERATURE CITED Abdelghaffar AA, Russell C: Role of strain difference and tetracycline on production of subcutaneous lesions by Ceneida eihieehe in experimental rats. fiehgeheegie 17:210-217. Anderson JD, Gillespie WA, Richmond MH: Chemotherapy and antibiotic resistance transfer between enterobacteria in the human gastrointestinal tract. WW. 6:461- 473, 1973. Anderson ML, Odds FC: Adherence of M eihigehe to vaginal epithelial: significance of morphological form and effect of ketoconazole. hyheeeh 28:531-540, 1985. Andermont A, Raiband P, Tancrede C: Effect of erythromycin on microbial antigonism: a study in gnotobiotic mice associated with a human fecal flora. _.'_.[ Ihfth Dis. 148:579-587, 1983. Argenzio, RA. Short-chain fatty acids and the colon. Qigesh Dis Sci 26:97-99, 1981. Baine WB, Koenig MG, Goodman JS: Clearance of theide eihicans from the bloodstream of rabbits. We 10:1420-1425, 1974. Bak AL, Stenderup A: DNA homology in yeasts: genetic relatedness within the genus QeQQiL. ,1 th hicrehiel 59:21-30, 1969. Balish E, Balish MJ, Salkowski CA, Lee KW, Bartizal KF: Colonization of congenitally athymic gnotobiotic mice by theiee albicens. App; Envirh hierob 47:647-652, 1984. Balish E, Phillips AW: Growth, morphogenesis and virulence of Qahdide elhieehe after oral inoculation in the germ-free and conventional chick. J heeteriol 91:1736-1743, 1966. Balish E, Phillips AW: Growth and virulence of Candiee elbicane after oral inoculation in the chick with a monoflora of either Escherichia coli or figreptococees feeceiis. g gectegiol 91:1744-1749, 1966. Balish E, Shih CN, Yale CE, Mandel AD: Effect of a prolonged stay in a locked environment on the microbial flora in dogs. Aerospace hed 45:1248-1254, 1974. 276 277 Ballou C: Structure and biosynthesis of the mannan component of yeast cell envelope. heg_huerehiel_£hyeiel 14:93-158, 1976. Barnes JL, Osgood W, Lee JC, King RD, Stein JH. Host- parasite interactions in the pathogenesis of experimental renal candidiasis. Leh_1hgee; 49:460-467, 1983. Barrett-Bee K, Hayes Y, Wilson RG, Ryley JF: A comparison of phospholipase activity, cellular adherence and pathogenicity of yeast. W 131:1217-1221, 1985. Bartnicki-Garcia S: Fungal cell wall composition, in Laskin AI. Lechevalier HA (e68-) 820W Vol. Ii; Microbiei Qompeeieieh. CRC Press, Cleveland, pp 201-214, 1973. Beneke ES. Rogers AL: WW. Burgess Publishing Company. Minneapolis, 1980. Bergeim O: Toxicity of intestinal volatile fatty acids for yeast and Esch. eeli. J.IDIQQL_Qi§ 66:222-234, 1940. Bergeim O, Hanszen AH, Pincussen L, Weiss E: Relation of volatile fatty acids and hydrogen sulphide to the intestinal flora. i_lhgeeh_hie 66:155-166, 1941. Bianchi DE: The lipid content of cell walls obtained from juvenile, yeast like and filamentous cells of Caneiee cthLc ns. WWW 33:324-332. 1967. Bishop CT, Blank F, Gardner P: The cell wall polysaccharides of theige eihieene, glucan, mannan and Chitin. th J Chem 38:869-881, 1960. Blyth W: Host-parasite relationships in experimental moniliasis. I. Candiee em. mam 10:262- 282, 1959. Bodey GP, Rodriguez V: Infection in cancer patients on a protected environment-prophylatic antibiotic program. Am J Med 59:497-504, 1975. Braun P, Calderone R: Chitin synthesis in theiee QM: Comparison of yeasts and hyphal forms . ,1 Beehegiei 133:1472-1477, 1978. Brawner DL, Cutler JE. Changes in surface topography of Qendide elhieehe during morphogenesis. . 2e; flyggl 23:389-393, 1985. EAL- 278 Brawner DL, Cutler JE. Variability in expression of cell surface antigens of theiee elhieehe during morphogenesis. Ihfect Ihhuh 51:337-343, 1986. Cabib E, Roberts R, Bowers B: Synthesis of yeast cell wall and its regulation. Ann_Bey__nicchcm 51:763-793, 1982. Calderone RA, Lehrer N, Segal E: Adherence of m eibicehs to buccal and vaginal epithelial cells: Ultrastructural observations. W 30:1001- 1007, 1984. Cantrell HR, Widra A: Experimental candidiasis in cortisone-treated mice. i_fieeheniel 87:1532, 1964. Carpenter AM: Studies on 910216; 1. Identification of 100 yeastlike fungi isolated from children. Am__J__Ql_ih Bath 25:98-101, 1955. Cassone A: Improved visualization of wall structure in §ecehezomycee cerevisiee. Expegieheie 29:1302-1309, 1973. Cassone A, Kerridge D, Gale EF: Ultrastructural changes in the wall of QeLdiee m following cessation of growth and their possible relationship to the development of polyene resistance. MW 110: 339- -,349 1979. Cassone A, Mattia E, Boldrini L: Agglutination of blastospores of theige eihieehe by Concanavalin A and its relationship with the distribution of mannan polymers and ultrastructure of the cell wall. i_th_hie;ehiei 105:263- 273 , 1978 . Cassone A, Simonette N, Strippoli V: Ultrastructural changes in the wall during germ-tube formation from blastospores of c albicans i_ccn_nicrcbicl 77: 417-426 1973. Caugant DA, Levin BA, Selander RK: Genetic diversity and temporal variation in the E coli population of a human host. Qenehics 98:467-490, 1981. Cawson RA, Rajasingham KC: Ultrastructural features of the invasive phase of Q. eihigehe. W 87:435-443, 1972. Centeno A, Davis CP, Cohen MS, Warren MM: Modulation of Qemiiee elbicahs attachment to human epithelial cells by bacteria and carbohydrates. W 39:1354-1360, 1983. 279 Chaffin LW, Stocco DM: Cell wall proteins of W albicans. can_l_Eicrcbicl 29:1438-1444. 1983- Chattaway FW, Bishop R, Holmes MR, Odds FC, Barlow AJE: Enzyme activities associated with carbohydrate synthesis and mycelial forms of C_ncica clbiccn_- i_ccc_nicrccicl 75: 97-109, 1973. Chattaway FW, Holmes MR, Barlow AJE: Cell wall composition of the mycelial and blastospore forms of Qaheida alhieahs. Meniiemhiei 51:367-376, 1968. Chattaway FW, Shenolikar S, O'Reilly J: Changes in the cell surface of the dimorphic forms of theiee eihicahs by treatment with hydrolytic enzymes. i_geLMiegehei 95:335- 347, 1976. Cheng KJ, Irvin RT, Costerton JW: Autochthonous and pathogenic colonization of animal tissues by bacteria. Qah e Microbiol 27:461-490, 1981. Chiew YY, Shepherd MG, Sullivan PA: Regulation of chitin synthesis during germ-tube formation in Cahdiea eibicahe. ALCh Microbioi 125:97-104, 1980. Clark JD: Influence of antibiotics or certain intestinal bacteria on orally administered theige elhieehe in germ- free and conventional mice. W 4:731-737, 1971. Cohen R, Roth FJ, Delgado E, Aheard DG, Kalser MH: Fungal flora of the normal human small and large intestine. New Egg {J Med 280:638-641, 1969. Cole GT, Seshan KR, Pope LM, Yancey RJ: Morphological aspects of gastrointestinal tract invasion by Candida elhieehe in the infant mouse. J Med Veg hyeoi 26: 173- 185. Collins-Lech C, Kalbfleisch JH, Franson TR, Sohnle PG: Inhibition by sugars of Caheide _a_ihicehe adherence to human buccal mucosal cells and corneocytes in X13119- Infeet imun 46:831-834, 1984. Cooper BH, Silva-Hutner M: Yeasts of medical importance, In Lennette EH (ed): Man 0 Cl' 'cal M o '01 , 4th Edition. American Society for Microbiology, Washington, pp526-541, 1985. Costerton JW, Irvin RT, Cheng K-J: The role of bacterial surface structures in pathogenesis. QgQ Q11; Eev in Mierehiol 8:303-338, 1981. 280 Costerton JW, Marrie TJ, Cheng K-J: Phenomena of bacterial adhesion, in Savage DC, Fletcher M (eds): ' ‘ 'a ‘22‘;'O!' ,- Qi!.3H: cl! 9! -='- 0- sighigieance. New York, Plenum Press, pp 3-43, 1985. Cox F: Adherence of theiee elhieehe to buccal epithelial cells in children and adults. i_Leh_QliD_Mc§ 102:960-972, 1983. Critchley IA, Douglas LJ. Isolation and partial characterization of an adhesin from We elbieans. ,1 Gen fliQEObigl 1987; 133:629-636. Dastidar SG: Studies on the antigens of the genus thgige by gel diffusion. Ineien_g_ueg_3e§ 56:247-249, 1968. Davenport JC: The oral distribution of theiee in denture stomatitis. fir Den; Q 129:151-156, 1970. Davidson S, Brish M, Rubinstein E: Adherence of We albieens to buccal epithelial cells of neonates. Mycopathoiegie 85:171-173, 1984. Davis CP, McAllister JS, Savage DC: Microbial colonization of the intestinal epithelium in suckling mice. infect Immhh 7:666-672, 1973. DeMaria A, Buckley H, von Lichtenberg F: Gastrointestinal candidiasis in rats treated with antibiotics, cortisone, and azathroprine. ID£§2£_IEEBD 13:1761-1770, 1976. Dewilde T, De Vroey C, Bosmans J, Hubens A, Vanbreuseghem 1R: Prevalence of Cendida aibieahs in patients receiving total parenteral nutrition. Sabeeraudia 20:169-171, 1982. Derjagiun BV, Landau L: Theory of the stability of strongly charged lyophobic sols and of the adhesion of strongly charged particles in solutions of electrolytes. Accc_2h¥cicchim_Q§§B 14:633-662. 1941- Diamond RD, Krzesicki R: Mechanisms of attachment of neutrophils to theiee eibicahs pseudohyphae in the absence of serum, and of subsequent damage to pseudohyphae by microbial processes of neutrophils i_n 1151‘..- J_Ql_i_n Ingeeg 61:360-369, 1978. Djaczenko W, Cassone A: Visualization of new ultrastructural components in the cell wall of QM elhi_ceh§ with fixatives containing tapo. g Qeii Bioim 52:186-190, 1971. 281 Domer JE: Intragastric colonization of infant mice with We e_lhieene induces systemic immunity demonstrable upon challenge as adults. LW 157: 950-958, 1988. Domer JE, Hector RF: Enhanced immune responses in mice treated with penicillin-tetracycline or trimethoprim- sulfamethoxazole when colonized intragastrically with Qandide elbicans. Antimicrcc_bc_chcmc. 31:691-697. 1987. Douglas LJ: Adhesion of pathogenic thgiee species to host surfaces. hicrohiel §ci 2:243-247, 1985. Douglas LJ: Surface composition and adhesion of theiee elbiQans. Bigcheh §OQ IIQDS 13:982-984, 1985. Douglas LJ, Houston JG, McCourtie J: Adherence of_theiee elbicehe to human buccal epithelial cells after growth on different carbon sources. W 12: 241- 243,1981. Douglas LJ, McCourtie J: Effect of tunicamycin treatment on the adherence of theigie e_lhieehe to human buccal epithelial cells. W 16: 199-202 , 1983. Doyle RJ, Nesbitt WE, Taylor KG: On the mechanism of adherence of We eehghie to hydroxylapatite. EEE§_Eicrccicl_Lccccr§ 15: 1-5 1982. Drasar BS, Hill MJ: MW. Academic Press, New York, 1974. Drasar BS, Shiner M, McLeod GM: Studies on the intestinal flora. I. The bacterial flora of the gastrointestinal tract in healthy andachlorhydric persons. gaeeroehterology 56:71-79, 1969. Dubos R: Staphylococci and infection immunity. Am s thig 105:643-645, 1963. Dubos R, Schaedler RW, Costello R, Hoef P: Indigenous, normal, and autochothonous flora of the gastrointestinal tract. I_E£E_M§§ 122:67-76, 1965. Dubos R, Schaedler RW, Stevens M: The effect of antibacterial drugs on the fecal flora of mice. LEM 117:231-243, 1963. Ekenna O, Sherertz RJ: Factors affecting colonization and dissemination of Qenfiifie elm; from the gastrointestinal tract of mice. ,‘Lnjee; 1mg 55:1558- 1563, 1987. 282 Elorza MV, Mugui A, Sentandraeu R: Dimorphism in Qeneme eihicahe: Contribution of mannoproteins to the architecture of yeast and mycelial cell walls. Ljen HiQIQQin 131:2209-2216, 1985. Elorza MV, Ricco H, Gozalbo D, Sentandreu R: Cell wall composition and protoplast regeneration in Qeneide elbicens. W1: 49:457-469. 1983. English MP: The saprophytic growth of keratinophilic fungi on keratin. fieheegeeeie 2:115-130, 1963. Epstein JB, Truelove EL, Izutzu KT: Oral candidiasis: pathogenesis and host defense. W 6:96-106, 1984. Eras P, Goldstein MJ, Sherlock P: theIide infection of the gastrointestinal tract. neglieine (Baltimore) 51:367- 379, 1972. Field LH, Pope LM, Cole GT, Guentzel MN, Berry LJ: Persistence and spread of Qeedifie eibicehs after intragastric inoculation of infant mice. InfeeeJmen 31:783-791, 1981. Finegold SM, Mathisen GE, George WL: Changes in human intestinal flora related to the administration of antimicrobial agents. In Hentge DJ (ed) Human inteseihei Wm- Aoademio Press. New York, 1983, pp 355-446. Finegold SM, Sutter VL, Mathisen GE: Normal indigenous intestinal flora. In Hentge DJ (ed) W ' ro o ' hea t d 's . Academic Press, New York, 1983, pp 3-31. Fletcher M: The question of passive versus active attachment mechanisms in non-specific bacterial adhesion, in Berkley RCW, Lynch JM, Rutter PR, Vincent B (eds): ° a s' t a . London, Ellis Horwood Limited, 1980, pp 197-210, Franklin MA, Skoryna SC: Studies on natural gastric flora. I. Bacterial flora of fasting human subjects. Qah nee ASSOQ Q 95:1349-1355, 1966. Freter R: Bacterial association with the mucus gel system of the gut. In: Schlessinger D (ed) hierehigiegy. American Society for Microbiology, Washington, pp 278-281, 1982 . 283 Freter R: Experimental enteric Shigelle and We infections in mice and guinea pigs. M 104:411- 418, 1956. Freter R: Factors controlling the composition of the intestinal microflora. In Stiles HW, Loesche WJ, O'Brien TC (eds): Egeceedinge: mierobie], aspecee ef dente], a__ies -. se - :_-.- ’u'! 0 ° 0.0- 005 L3: ra : . Washington, DC, vol 1, Information Retrieval, 1976, pp 109-120. Freter R: Interdependence of mechanisms that control bacterial colonization of the large intestine. Mieneeeql eeelhe; 14:89-96, 1984. Freter R: Mechanisms that control the microflora in the large intestine. In Hentge DJ (ed) We], m ra h is . Academic Press, New York, 1983, pp 33-54. Freter R, Brickner H, Botney M, Cleven D, Aranki A: Mechanisms that control bacterial populations in continuous-flow culture models of mouse large intestinal flora. Infect Immun 39:676-685, 1983. Freter R, Brickner H, Fekete J, O'Brien PCM, Vickerman MM: Testing of host-vector systems in mice. W Bell 2:68-76, 1979. Freter R, Brickner H, Fekete J, Vickerman M, Carey KE: Survival and implantation of £_s_eher_iehie ELL in the intestinal tract. Infiee§_1mmen 39:686-703, 1983. Freter R, Jones GW. Adhesive properties of yihzie choiege: nature of the ineteraction with intact mucosal surfaces. Infee;_;hmeh 6:918-927, 1976. Freter R, Jones GW: Models for studying the role of bacterial attachment in virulence and pathogenesis. Rey Inf§Q£_Di§ 5:5647-5658, 1983. Freter R, O'Brien PCM, Macsai MS. Role of chemotaxis in the association of motile bacteria with intestinal mucosa: fitness and virulence of nonchemotactic 1121.19 eheleze mutants in infant mice. Infieeh_1mmeh 34:222-233, 1981. Freter R, O'Brien PCM, Macsai MS. Role of chemotaxis in the association of motile bacteria with intestinal mucosa: ih vive studies. Infeeh Immun 34:234-240, 1981. Freter R, Stauffer E, Cleven D, Holdeman LV, Moore WEC: Continuous-flow cultures as in yiege models of the ecology of large intestinal flora. infee;_ihheh 39:666-675, 1983. 284 Gardiner R, Podgorski C, Day AW: Serological studies on the fimbriae of yeasts and yeast-like species. Behenieel QQZQLLQ 143:534-541, 1982. Garrison RG: Vegetative ultrastructure, in Arnold WN (ed): WW. Booa Raton, Florida, CRC Press, pp 139-160, 1981. Geck P, Novak EK: Immunofluorescent studies on yeasts. Actc_Hicrcbicl_Accc_§ci_chc 14:13-22. 1967- Gerhardt NB, Iglewski WJ: Alterations in the fecal flora of an individual frequently lacking coliforms. J__I_n£e9_t Difi 134:490-494, 1976. Ghannoum MA, Burns GR, Elteen K, Radwin SS: Experimental evidence for the role of lipids in adherence of QM spp. to human buccal epithelial cells. Mime 54:189-193, 1986. Ghannoum MA, Elteen KA: Correlative relationship between proteinase production, adherence and pathogenicity of various strains of Qandide M. 1 Meg Vet; Myeel 24:407-413, 1986. Ghannoum MA, Janini G, Khamis L, Radwan SS: Dimorphism- associated variations in the lipid composition of theiee elbicens. i_ccn_nicrccicl 32:2367-2375. 1986- Giannella RA, Broitman SA, Zamcheck N: Gastric acid barrier to ingested microorganisms in man: Studies in yiye and in yihxe. Q3; 13:251-256, 1972. Gibbons RJ, Etherden I, Peros W: Aspects of the attachment of oral streptococci to experimental pellicles, in Mergenhagen SE, Rosan B (eds): a a 's f WM. American Society for Microbiology, Washington, pp 77-84, 1985. Gopal P, Sullivan PA, Sheperd MG: Isolation and structure of glucan from regenerating spheroplasts of thgiee elbiccnc- i_ccn_Eicrchicl 130:1217-1225. 1984- Gordon MA, Elliot JC, Hawkins TW: Identification of Qandida eihicans, other Qeneiee species and Tereiepeie glehrege by means of immunofluorescence. eeeeegeeeie 5:323-328, 1983. Gotlieb-Jensen K, Anderson J: Occurrence of Qandide in gastric ulcers: significance for the healing process. ccccrccntcrclch 85:535-537. 1983- 285 Gracey M, Stone DE, Suharjono BS, Sunoto MD: Isolation of theiee, species from the gastrointestinal tract in malnourished children. Am_Q_Qlin_nh§;i 27:345-349, 1974. Guentzel MN, Cole GT, Pope LM: Animal models for candidiasis. In McGinnis MR (ed): Qe;zen§_Iepiee_ih ' V . New York, Springer-Verlag, pp 57-116, 1985. Guentzel MN, Herrera C: Effects of compromising agents on candidosis in mice with persistent infections initiated in infancy. ihjee§_iemeh 35:222-228, 1982. Hartley CL, Clements HM, Idnton KB: Effects of caphalexin, erythromycin and clindamycin on the aerobic gram-negative fecal flora in man. Mien; 11:125- 135,1978. Hasenclever HF: The antigens of thgiee eihigehg. Am_B§X B§_E_Di§_§2221 923150 158 1965- Hasenclever HF, Mitchell WO: Antigenic studies of theiee. I. Observation of two antigenic groups in theiee clbiccn_- i_fiecccricl 82:570-573. 1961a- Hasenclever HF, Mitchell WO: Antigenic studies of theiee. III. Comparative pathogenicity of Qenfiige eihieene group A group B and Ccn_ica ctcllcccicce- E.Bacccricl 82: 578- 581, 1961b. Hasenclever HF, Mitchell WO, Loewe J: Antigenic relationship of Qandide eithehe group A and group B to Qendida stellatoidea and Ccncicc cronicclic. J.Eecccricl 32: 547- 577, 1961. Hazen EL, Brown R, Mason A: Protective action of Fungicidin (Nystatin) in mice against virulence enhancing activity of Oxytetracycline on Qengi e elhieehe. Ahhihie; eh§_QhemQ 111:1125-1128, 1953. Hazen KC, Plotkin BJ, Klimas DM: Influence of growth conditions on cell surface hydrophobicity of theige alhicanc and ceccide glchrcca. InfchImmcn 54:269-271. 1986. Hector RF, Domer JE: Control of systemic spread of Qengiee eihieene with ketoconazole in the stomachs of mice treated with cytarabine. i_lhfieehenie 147:946-950, 1983. Hector RF, Domer JE: Mammary gland contamination as a means of establishing long-term gastrointestinal colonization of infant mice with QQDQiQ§.§lhi§§n§- infiee; Immgn 38:788-790, 1982. 286 Helstrom PB, Balish E: Effect of oral tetracycline, the microbial flora and the athymic state on gastrointestinal colonization and infection of BALB/c mice with Caneiee elbicen_. Infect_1mmcn 23:764-774. 1979- Holdeman LV, Cato EP, Moore WEC (eds) Aheerehe_Lehe;etery Meheei, 4th ed Anaerobe Lab. Virginia Polytechnic Institute and State University, Blacksburg, 1977. Howlett JA: The infection of rat tongue mucosa i_n m with five species of Qeneiee. e_neg_nierehiel 9:309-316, 1976. Howlett JA, Squier CA: Qenfiiee elhieehe ultrastructure: Colonization and invasion of oral epithelium. infieej', IEEEB 29:252-2600, 1980. Hummel RP, Maley MP, Miskell PW, Altemeier WA: Soppression of Genghis albicens by Escherichia celi- .1 Izeume 15:413-413. 1975. Hummel RP, Oestreicher EJ, Maley MP, Macmillan BG: Inhibition of canine: elbicans by Escherichie chi in 21:12 and the germfree mouse. e §ezg Ree 15:53-58, 1973. Huppert M, Cazin J, Smith H: Pathogenesis of Caneige elhieehe infection following antibiotic therapy. III The effect of antibiotics on the incidence of thQi e elhieehe in the intestinal tract of mice. i_fieehexiel 70:440-447, 1955. Isaacson RE: Pilus adhesins, in Savage DC, Fletcher M (eds): a 'a s : si lo ' signifiicance. New York, Plenum Press, pp 307-336, 1985. Isenberg H, Allerhand J, Berkman JJ, Goldberg D: Immunological and toxic differences between mouse virulent and mouse avirulent theiee elhieehe. Q flecteriel 86:1010-1016, 1963. Isenberg HD, Pisano MA, Carito SL, Berkman JI: Factors leading to overt monilial disease. I. Preliminary studies of the ecological relationship between We elhieehe and intestinal bacteria. Ahhihie§_Qheme 10:353- 363, 1960. Jenkins WM: Oral infections with theiee eihieehe. §eeh nee_e 18:192-200, 1973. Jennison RF: theiee infections in a maternity hospital, in Winner HI. Hurley R (eds): W Infeeeiehe. London, Livingston, pp 44-51, 1966. 287 Jones GW: The attachment of bacteria to the surfaces of animal cells. in Reissig JL (ed): Eicrcbiel_lnterectiene1 , series B vol 3. New York, Chapman and Hall, pp 139-176,1977. Jones GW: Adhesion to animal surfaces, in Marshall KC (ed): W. New York. Springer-Verlag, pp 71-84, 1984. Jones GW: Mechanisms of the attachment of bacteria to animal cells, in Klug MJ, Reddy CA (eds): Qeerenh W- Washington. DC. pp 136-143,1984. Jones GW, Abrams GD, Freter R. Adhesive properties of 112119 eheiere: adherence to isolated rabbit brush border membranes and hemagglutinating activity. mm 14:232-239, 1976. Jones GW, Freter R. Adhesive properties of 11211.9 ehqiere: nature of the interaction with isolated rabbit brush border membranes and human erythrocytes. 1mm 14:240-245, 1976. Jones GW, Isaacson RE: Proteinaceous bacterial adhesins and their receptors. QC Cei; gev hicgohio], 10:229-260, 1983. Jones GW, Richardson LA: The attachment to, and invasion of HeLa cells by Selineheile hyphiheeiim: The contribution of mannose-sensitive and mannose- -resistant haemagglutinating activities. MM 127: 361- 370, 1981. Jones GW, Richardson LA, Uhlman D: The invasion of HeLa cells by mug, hyphimuriuh: Reversible and irreversible bacterial attachment and the role of bacterial motility. W 127: 351-360,1981. Jonsen J: Serological studies in fungi: comparison of agglutinins and precipitins in anti- -Qenei_de sera. beige Eechcl_Eicrcbicl_§cen_ 37:79-83. 1955- Joshi SN, Garvin PJ, Sunwoo YC: Candidiasis of the duodenum and jejunum. Qeehgeeheezeiegy 80:829-833, 1981. Kapica L, Blank F: Growth of thgiee eihieene on keratin as a sole source of nitrogen. DQIEQLQL 115:81-105, 1957. Karaev ZO, Velichko EV, Bykov VL: Main characteristics of the process of adhesion of W to human epithelial cells. WW Ihheheiegii 7:59-61, 1986. 288 Kearns MJ, Davies P, Smith H: Variability of the adherence of Qeneide flhieene strains to human buccal epithelial cells: Inconsistency of differences between strains related to virulence. fiehehzehgie 21:93-98, 1983. Kemp G, Solotorvsky M: Localization of antigens in mechanically disrupted cells of certain species of theiee and Tcrclceei_. J.Immcncl 93: 305- 314. 1964- Kennedy MJ: Adhesion and association mechanisms of cancice elbicans- In HoGinnis HR (ed): Win W. New York. Springer-Verlag. pp 73-169, 1988. Kennedy MJ: Role of motility, chemotaxis and adhesion in microbial ecology- WW 506:260-273, 1937. Kennedy MJ, Bajwa PS, Volz PA: Gastrointestinal inoculation of mu schenckii in mice. Eyceccthclcgie 78:141-143. 1982- Kennedy MJ, Volz PA: Dissemination of yeasts after gastrointestinal inoculation in antibiotic-treated mice. Sabeugaudie 21:27-33, 1983. Kennedy MJ, Volz PA: Ecology of mm eihieehe gut colonization: Inhibition of Qeeeiee adhesion, colonization, and dissemination from the gastrointestinal tract by bacterial antagonism. Infiee§_lhmhn 49:654-663, 1985a. Kennedy MJ, Volz PA: Effect of various antibiotics on gastrointestinal colonization and dissemination by thgige elbicens- szw 23:265-274. 1985b. Kennedy MJ, Volz PA, Edwards CA, Yancey RJ: Mechanisms of association of theiee eihieehe with intestinal mucosa. Q neg_hie;ehiel 24:333-341, 1987. Kessler G, Nickerson WJ: Glucomannan-protein complexes from cell walls of yeast. uieLQhem 234:2281-2285, 1959. Kimura LH, Pearsall NN. Adherence of theiee eihicahs to human. buccal epithelial cells. legee;__1meeh 21:64-68, 1978. Kimura LH, Pearsall NN: Relationship between germination of Qeneige, eihieehe and increased adherence to human buccal epithelial cells. Ingee;_3mmhn 28:464-468, 1980. 289 King RD, Lee JC, Morris AL. Adherence of Qengiee eihieene and other theige species to mucosal epithelial cells. IDI§9L_IEEQQ 1980: 27:667-674. Klotz SA, Drutz DJ, Zajic JE: Factors governing adherence of M species to plastic surfaces. 1m 50:97-101, 1985. Kobayashi K, Suginaka H: Comparison of cell wall and membrane proteins from eight theiee species. Sebccreccie1_e_nec_2et_E2ccl 22:341. 334. 1984- Kolarova N, Masler L, Sikl D: Cell wall glycopeptides of Qendida albigne serotypes A and 3- nicchimJicchyLActe 323:221-227, 1973. Koznin PJ, Taschdjian CJ: Enteric candidiasis: diagnosis and clinical considerations. Eeeiehziee 30:71-85, 1962. Krause W, Matheis H, Wulf K: Fungaemia and funguria after oral administration of M elhieehe. meet i:598- 599, 1969. Kreger-van Rij NJW (ed): Ihe_yeeetc_e_texcnemic_ctcdy. 3rd Edition. Elsevier, 1984. Kulkarni RK, Hollingsworth PJ, Volz PA: Variation in cell surface features of .QQDQLQQ. elhieehe, With respect to carbon sources. Sabourehdie 18:255-260, 1980. Kuo-SC, Lampen JO: Tunicamycin - an inhibitor of yeast glycoprotein synthesis. giochem fiiephye Res thh 58:287- 295, 1974. Lee A, Gemmelle E: Changes in the mouse intestinal microflora during weaning: role of volatile fatty acids. Infect_1mmcn 5:1-7. 1972- Lee A, Gordon J, Lee CJ, Dubos R: The mouse intestinal microflora with emphasis on the strict anaerobes. e_fizp nee 133:339-351, 1971. Lee JC, King RD. Adherence mechanisms of We albicans. in Schlessinger D (ed): W Washington, DC, American Society for Microbioogy pp 269- 272 , 1983 . Lee JC, King RD: Characterization of Qeedige eibieehs adherence to human vaginal epithelial cells in vitro. MIME]! 41: 1024-1030, 1983 . 290 Lee KW, Balish E: Effect of T-cells and intestinal bacteria on resistance of mice to candidosis. RES; ,1 Eeticelceedcth_§cc 31:233-240. 1982. Lee KW, Balish MJ, Balish E: Resistance of germfree nude and thymus-bearing mice to systemic and gastrointestinal candidosis . v rmfr esea Proceedings of the VIIth International Symposium on Gnotobiology. Sasaki S, Ozawa A, Hashimoto K (eds). Tokai Univ Press, Tokyo, pp 429-432, 1981. Lee KL, Buckley HR, Campbell CC. An amino acid liquid synthetic medium for development of mycelial and yeast forms ofC e_ncice albicens Sabouraudia 13: 148-153. 1975. Lehrer N, Segal E, Barr-Nea L. In vitro and in vivo adherence of Cahdiee elhi cans to mucosal surfaces. Ann MiQEObin 134:293-306, 1983. Levison ME: Effect of colon flora and short-chain fatty acids on growth in vitro of Peeueemenee eeregihege and Enterobacteriaceae. 1h£e9§_1mmeh 8:30-35, 1973. Liljemark WF, Gibbons RJ: Suppression of We by human oral Steeptoeecci in gnotobiotic mice. £1129; Immen 8:846-849, 1973. Locci R, Peters G, Pulverer G: Microbial colonization of prosthetic devices. IV. Scanning electron microscopy of intravenous catheters invaded by yeasts. Zeheeelhl W 1733419424. 1981- Loeb GI: The properties of nonbiological surfaces and their characterization, in Savage DC, Fletcher M (eds): '1 e _°_ 111‘80122‘1111811: 1 '. ‘15.. 0'. °. eigniiieehee. New York, Plenum Press, pp 111-129, 1985. Ludlam GB, Henderson JL: Neonatal thrush in maternity hospital. Leheeh i:64-70, 1942. Luft JH: Ruthenium red and ruthenium violet. II. Fine structural purification, methods for use for electron microscopy and localization in animal tissues. hnat gee 171:369-416, 1971. MacDonald F, Odds FC: Virulence for mice of a proteinase- secreting strain of W eihieene and a proteinase- deficient mutant. J Gen hiegohiei 129:431-438, 1983. Maisch PA, Calderone RA. Adherence of Qeneige eihigeee to a fibrin-platelet matrix formed in vitre. w 27: 650- 656, 1980. 291 Maisch PA, Calderone RA: Role of surface mannan in the adherence of theige elhieene to fibrin-platelet clots formed in gigge. Infieeh_1mmeh 32:92-97, 1981. Marples MJ: Oral candidiasis. mm 80:58-59, 1968. Marrie TJ, Costerton JW: The ultrastructure of Caneiee eihieehe infections. §§n_l_H1£IthQl 27:1156-1164, 1981. Marshall KC: Interfecefluicrcbielmlm- Cambridge. MA, Harvard University Press, 1976. Marshall KC (ed): Weig- Berlin, Springer-Verlag, 1984. Marshall SE, Marples BA, Salt WG, Stretton RJ: Aspects of the effect of bile salts on Candida albicans. e Med ye; hyeel 25:307-318, 1987. McCourtie J, Douglas LJ. Relationship between cell surface composition of theiee QJQLQQDE and adherence to acrylic after growth on different carbon sources. 111W 32:1234-1241, 1981. McCourtie J, Douglas LJ. Relationship between cell surface composition, adherence and virulence of theiee eihieehe. Ihfiee§_lhmeh 45:6-12: 1984. McCourtie J, Douglas LJ. Extracellular polymer of Caneiee eihieeh_: isolation, analysis and role in adhesion. i_an mm 1985a: 131:495-503 . McCourtie J, Douglas LJ. Unequal distribution of adhesins within populations of m eihieehe.12§h§ hicrobie], Lee; 1985b: 27: 111-115. McCourtie J, MacFarlane TW, Samaranayake LP: Effect of chlorhexidine gluconate on the adherence of theiee species to denture acrylic. W 20:97-104, 1985. Meunier-Carpenter F, Kiehn TE, Armstrong D: Fungaemia in the immunocompromised host: changing patterns, antigenemia, high mortality. AE_J_H§Q 71:363-370, 1981. Miles MR, Olsen L, Rogers A: Recurrent vaginal candidiasis: Importance of an intestinal reservoir. 1.5m hee_heeee 238:1836-1837, 1977. Miller GR, Sarachek A: Absence of photoreactivating enzymes in Qendida elbicens. candida steueccidee and candida trccicalis- InfecLIcmun 10:398-399. 1974- 292 Minagi S, Miyaka Y, Fugioka Y, Tsuru H, Suginaka H: Cell- surface hydrophobicity of thgige species as determined by the contact-angle and hydrocarbon-adherence methods. ,1 th_nierehiel 132:1111-1115, 1986. Minagi S, Miyake Y, Inagaki K, Tsuru H, Suginaka H: Hydrophobic interaction in mete; eihieehe and Qeeeiee £192; ealis adherence to various denture base resin materials. _Iefeethn 47:11-14, 1985. Miyake Y, Fujita Y, Minagi S, Suginaka H: Surface hydrophobicity and adherence of Qengige to acrylic surfaces. _uieeghiee 46:7-14, 1986. Montes LF, Wilborn WH. Ultrastructure features of host- parasite relationship in oral candidiasis. ,1 Beetenie], 96:1349-1356, 1968. Montes LF, Wilborn WH. Fungus-host relationship in candidiasis. Arch Dermatol 1985: 121:119-124. Montplaisir S, Nabarra B, Drouhet E: Susceptibility and resistance of Qandida to 5-Fluorocytosine in relation to the cell wall ultrastructure. Wm 9: 1028- 1032, 1976. Moore WEC: Anaerobes as normal flora: gastrointestinal tract. WW. Montreal. 1977. pp 222-223. Moore WEC, Cato EP, Good IJ, Holdeman LV: The effect of diet on the human fecal flora. BQDDBIX Report 1; es ' a e ° c 5 pp 11-23, 1981. Moore WEC, Holdeman LV: The normal flora of 20 japanese Hawaiians. W 27: 961-979, 1974. Muller HL, Kirchhoff G: Serologische typen von theme elbicens. zentrelbl_Bektericl_2ereeicenk_1nfekcicnekr_flyg he; 210:114-121, 1969. Murray IG, Buckley HR: Serological study of theige species, in Winner HI, Hurley R (eds): gympesium en candica_1nfecticne. London. Livingston. pp 44-51. 1966- Myerowitz RL: Gastrointestinal and disseminated candidiasis: An experimental model in the immunosuppressed rat. WM 105:138'143. 1981- Myerowitz RL, Pazin GJ, Allen CM: Disseminated Candidiasis: Changes in incidence, underlying diseases, and pathology. W 68:29-38, 1977. 293 Nishikawa T, Hatano H, Ohnishi N, Sasake S, Nomura T: Establishment of theiee eihieehe in the alimentary tract of the germ-free mice and antagonism with Egghegiehie eeli after oral inoculation. gep_g_uiezehie1 13:263-276, 1969. Nystatin Multicenter Study Group: Therapy of candidal vaginitis: The effect of eliminating intestinal Qandiee. Amel_Qh§§§§_§¥n§QQl 1553551'655. 1936- Odds FC. Biotyping of medically important fungi, in McGinnis MR (ed): W 1. Springer-Verlag, New York, pp 157-171, 1985. Odds FC: theiee_ehe_Qeneieeeie. Baltimore, University Park Press, 1988. Odds FC, Gonzales A: Acid phosphatase levels in the genus Candida and their application to the taxonomy and identification of pathogenic theige species. fieheegeheie 12:287-294, 1974. Odds FC, Hall CA, Abbott AB. Peptones and mycological reproducibility. fieheereeeie 16:237-246, 1978. Okubo Y, Honma Y, Suzuki 8: Relationship between phosphate content and serological activities of the mannans of Qeneiee eihieehe strains NIH A-207, NIH B-792, and J-1012. J BQCEQrin 137:677-680, 1979. Ozawa K, Yabu-Uchi K, Yamanaka K, Yamachita Y, Ueba K, Miwatani T: Antagonistic effects of heeiilee natte and cum femlie on growth of candida elb__neica . Eicrcbicl_1mmcccl 23:1147-1156. 1979. Paerl HW: Influence of attachment on microbial metabolism and growth in aquatic ecosystems, in Savage DC, Fletcher M (eds): WWW eighifiieehee. New York, Plenum Press, pp 363-400,1985. Paine TF: The inhibitory actions of bacteria on Candiee growth. AhtibiOQiQs Chehg 8:273-281, 1958. Persi MA, Burnham JC, Duhring JL: Effects of carbon dioxide and pH on adhesion of theiee elhieehe to vaginal epithelial cells. ih;e9h_ihmhh 50:82-90, 1985. Phillips AW, Balish E: Growth and invasiveness of theiee eihieehe in the germfree and conventional mouse after oral challenge. hppi_hie;ehiel 14:737-741, 1966. Phillips MW, Lee A: Microbial colonization of rat colonic mucosa following intestinal pertubation. W 10: 79-88, 1984. 294 Plaut AG, Gorbach SL, Nahas L, Weinstein L, Spanknebel G, Levitan R: Studies of intestinal microflora. III. The microbial flora of human small intestinal mucosa and fluids geehereenherelegy 53:868-873, 1967. Ponton J, Jones JM: Analysis of cell wall extracts of Qeneiee elhjeene by sodium dodecyl sulfate-polyacrylamide gel elecrophoresis and western blot techniques. Infee; Immun 53:565-572, 1986a. Ponton J, Jones JM: Identification of two germ-tube- specific cell wall antigens of Qeneiee eihieeh_. ihiee; Immun 54:864-868, 1986b. Pope LM, Cole GT, Guentzel MN, Berry IJ: An experimental model of candidiasis in infant mice: Systemic infection following gastrointestinal colonization, in Kuttin ES, Baum GL (eds): WW- Amsterdam. Escerpta Medica, 1979, pp 71-74. Pope LM, Cole GT, Guentzel MN, Berry IJ: Systemic and gastrointestinal candidiasis of infant mice after intragastric challenge. We}; 25:702-707, 1979. Pope LM, Cole GT: SEM studies of adherence of Candiee eihieeee to the gastrointestinal tract of infant mice. WM 3:73-80. 1981- Pope LM, Cole GT: Comparative studies of gastrointestinal colonization and systemic spread by Qeneiee eihieehe and nonlethal yeast in the infant mouse. geeh Eiectreh hiezeee 4:1667-1676, 1982. Poulain D, Hopwood V, Vernes A. Antigen variability of candida albicens. Wicrchicl 12:223-271. 1985. Poulain D, Tronchin G, Jouvert S, Herbant J, Biguet J: Architecture pariétale des blastospores de Qandiee e_ihieehe localization de composants chimiques et antigéniques. hhhiuierehiei 132A:219-223, 1981. Pugh D, Cawson RA: The cytochemical localization of phospholipase and lysophoslipase in theiee eibicehe. eeheegeeeie 13:110-115, 1975. Pugh D, Cawson RA: The cytochemical localization of acid hydrolases in four common fungi. W 22:125- 132, 1977. Pugh D, Cawson RA: The cytochemical localization of phospholipase in Qefliee eihigehe infecting the chick chorio-allantoic membrane. geheereeeie 15:29-35, 1977. 295 Pugh D, Cawson RA: The surface layer of Qengiee eibicene. fiinghige 23:19-23, 1978. Ray TL: Fungal infections in the immunocompromised host. hefl_§lin_fl_fim 64:955-968, 1980. Rebora A, Leyden JJ: Napkin (diaper) dermatitis and gastrointestinal carriage of Qandide elhieene. grin ,1 Qezm 105:551-555, 1981. Reinhart H, Muller G, Sobel JD: Specificity and mechanism of in xicrc adherence of candida albicens W Sci 15: 406-413, 1985. Reiss E: Ecleccler__imnunclcgv of mycctic__end eccincmycctic_infecticne- New York. Elsevier. 1986- Richardson MD, Smith H: Resistance of virulent and attenuated strains of QenQiQe eihieene to intracellular killing by human and mouse phagocytes. q Infect pie 144:557-564, 1981. Rippon JW: Eedicel_Eycclcgy1__Ihe_pathogenic_fcngi_en__the IEELhQEEHLLQ..§UULLDQEUEE§LJE . WB Saunders Company, Philadelphia, 1988. Rogers TJ, Balish E: Immunity to Qeneiee eibicene. HIEIQD_B§¥ 44:660-682, 1980. Rossano F, Tufano MA: Effect of concanavalin A on the adhesiveness of We elhieene to vaginal cells. Beee hegLfiee; 22:313-317, 1975. Rotrosen D, Calderone RA, Edwards JE: Adherence of Qeneiee species to host tissues and plastic surfaces. Bey 1njee;_pie 8:73-85, 1986. Rotrosen D, Edwards JE, Gibson TR, Moore JC, Cohen AH, Green I: Adherence of Qeneige to cultured vascular endothelial cells: Mechanisms of attachment and endothelial cell penetration. ,1 Infect Dis 152:1264- 1274, 1985. Rotrosen D, Gibson TR, Edwards JE: Adherence of Qeneide species to intravenous catheters. ,1 Infech his 147:594, 1983. Rutter PR, Vincent B: Physiochemical interactions of the substratum, microorganisms, and the fluid phase, in Marshall KC (ed): ' e ' n e a ' . New York, Springer-Verlag, pp 21-38, 1984. Rutter PR, Vincent B: The adhesion of microorganisms to surfaces: Physiochemical aspects, in Berkley RCW, Lynch 296 JN. Rutter PR. Vincent B (eds): EicrcbieLechesicm enxfeeee. London, Ellis Horwood Limited, pp 79-92, 1980. Ryley' are lPathogenicity of Qeneiee eihieene with particular reference to the vagina. J Med Veg hyeol 24:5- 22, 1986. Samaranayake LP, MacFarlane TW: An in 213:9 study of the adherence of Qeneiee eihieene to acrylic surfaces. Arehe QI§1_BlQl 25:603-609, 1980. Samaranayake LP, MacFarlane TW: Factors affecting the in yihro adherence of the fungal oral pathogen Qandiee eihieene to epithelial cells of human origin. Anchs One; 5101 27:869-873, 1982. Samaranayake LP, MacFarlane TW: The effect of dietary carbohydrates on the in 212:2 adhesion of Qeneiee albicene to epithelial cells. J Med hiezehiol 15:511-517, 1982. Samaranayake LP, MacFarlane TW: On the role of dietary carbohydrates in the pathogenesis of oral candidosis. EEM§_H12IthQl_L§LE 2731-5. 1935- Samaranayake LP, McCourtie J, MacFarlane TW: Factors affecting the i_n vitno adherence of QeMiee eibicans to acrylic surfaces. h;eh_Q;e1_fiiel 25:611-615, 1980. San-Blas G: The cell wall of fungal human pathogens: Its possible role in host-parasite relationships. HYQQRQEDQLQQie 793159-134. 1982- Sandin RL. Studies on cell adhesion and concanavalin A induced agglutination of Qendiee elhieene after mannan extraction. J nee hicnobioi 24:145-150, 1987a. Sandin RL. The attachment to human buccal epithelial cells by' Qendida .eihieen_: an in 'vitro kinetic study 'using concanavalin A. Myeenehheiegie 98:179-184, 1987b. Sandin RL, Rogers AL. Inhibition of adherence of Cendiee eihieene to human buccal epithelial cells. hycepathelogie 77:23-26, 1982. Sandin RL, Rogers AL, Patterson RJ, Beneke ES. Evidence for' mannose-mediated adherence of Qeneige, eibieens to human buccal epithelial cells in vitro. infect Innnn 35:79-85. 1982. Saugar VK, Cortes JM, Self LW: Acid phosphatase production as an aid in rapid characterization of Qandiee species. Am_l_M§Q_I§QDEQl 41:327’332. 1975- 297 Savage DC: Adherence of normal flora to mucosal surfaces. Beachey EH (ed) Bacteriel_Acherence. London. Chapman and Hall, pp 33-59, 1980. Savage DC: Associations of indigenous microorganisms with gastrointestinal epithelial surfaces, in Hentges DJ (ed): IfluEuL_inEe§EineA1JALEIQIlQIA_JJL_h§§lEh_JuEL_Qi§§e§_n New York, Academic Press, pp 55-78, 1983. Savage DC: Associations of indigenous microorganisms with gastrointestinal mucosal epithelia. Am__i__Q1in__Nntr 23: 1495-1501, 1970. Savage DC: Localization of certain indigenous microorganisms on the ileal villi of rats. LBe_c§enie], 97:1505-1506, 1969. Savage DC: Microbial ecology of the gastrointestinal tract. Ann Rev hicnobioi 31:107-133, 1977. Sawai T, Hehre EJ: The distribution of transglucosyl- amylase among Cendida yeasts- MW Hicrcbicl_§ercl 29:377-385.1963- Schaedler RW, Dubos R, Costello R: Association of germfree mice with bacteria isolated from normal mice. J Ezn_heg 122:77-83, 1965. Schaedler RW, Dubos R, Costello R: The development of the bacterial flora in the gastrointestinal tract of mice. n E3p_fle§ 122:59-66, 1965. Schertwitz C, Martin R, Ueberberg H. Ultrastructural investigations of the formation of Qeneiee eihieene germtubes and septa. gehenneneie 16:115-124, 1978. Schwartz DS, Larsh HW: An effective medium for the selective growth of yeast or mycelial forms of Qandide Biochemical aspects of the two forms. Mycopathologia 70: 67-75. 1980. Schwartz DS, Larsh HW: comparative activities of glycolytic enzymes in yeast and mycelial forms of Qeneiee elbicens. Mycopathclccie 78:93-98. 1982- Sears HJ, Brownlee I: Further observations on the persistence of individual strains of Eeeheziehie eeii in the intestinal tract of man. J hectezioi 63:47-57, 1952. Sears HJ, Brownlee I, Uchigama JK: Persistence of individual strains of Eeeheniehie eeLi in the intestinal tract of man. e_fieeeeniel 59:293-301, 1949. 298 Segal E, Lehrer N, Ofek I: Adherence of Qeneiee eihieene to human vaginal epithelial cells: Inhibition by amino sugars. Enp_Qe11_§ie1 50:13-17, 1982. Segal E, Savage DC: Adhesion of Qeneiee elbicens to mouse intestinal mucosa in yinze: Development of the assay and test of inhibitors. I_H§Q_!§£_MYQ21 24:477-479, 1986. Segal E, Soroka A, Schechter A: Correlative relationship between adherence of Qen_dige elhieene to human vaginal epithelial cells in_ yi;;e_ and Qeneieel vaginitis. fiehenneneie 22:191-200, 1984. Segal E, Vardinon N, Schwartz J, Eylan E: Antigenic relationships between Qeneiee elhieene and various yeasts as reflected by immunoglobulin-class specificity. Benhel hienehiel 42:49-58, 1975. Shecter Y, Landau JW, Dabrowa N: Comparative electrophoresis and numerical taxonomy of same Qengjee species. hyeologie 64:841-853, 1972. Shedlofsky S, Freter R: Synergism between ecological and immunological control mechanisms of intestinal flora. Q; IngeQ§_Qi§ 129:296-303, 1974. Shepherd MG, Poulter RTM, Sullivan PA: Qeneiee eihieene: BiologY. genetics. and pathogenicity- AnnieLHicrchicl 39:579-614, 1935. Shepherd MG, Surarit R, Gopal PK, Sullivan PA: Cell wall metabolism of Q. elhieen_: B(13) and B(16) glucan synthesis, in Nombela C (ed). hierobiel eell yell eyntheeie_ene_en;eiyeie. Elsevier Science Publishers, New York, pp 73-83, 1984. Sikl D, Masler C, BAuer S: Mannan from the extracellular surface of Candida albicans Berkhout- Experiencie 20:456. 1964. Skerl KG, Calderone RA, Segal E, Sreevalsan T, Scheld WM: in yigne binding of Qeneiee eihieene yeast cells to human fibronectin. Qen_g_niezehiel 30:221-227, 1984. Slutsky B, Buffo J, Soll DR. High frequency switching of colony morphology in Candida. albicans. science 230:666- 669, 1935. Slutsky B, Staebell M, Anderson J, Risen L, Pfaller M, Soll DR. 'White-opaque transition': a second high- frequency switching system in Qeneiee elhieene. g fieQLezigl 169:189-197, 1987. 299 Smail EH, Jones JH. Demonstration and solubilization of antigens expressed primarily on the surfaces of Qengiee albicens germ tubes. Infect_1mmcn 45:74-81.1984. Smyth CJ, Jonsson P, Olsson E, Soderlind O, Rosengren J, Hjerten S, Wadstrom T: Differences in hydrophobic surface characteristics of porcine enteropathogenic Escherichie gen, with or without K88 antigen as revealed by hydrophobic interaction chromatography. LnfeeLInnnn 22:462-472, 1978. Sobel JD, Myers PG, Kaye D, Levison ME: Adherence of Qeneiee elhieene to human vaginal and buccal epithelial cells. _1_1n§ee;_nie 143:76-82, 1981. Spencer JF, Gorin PAJ: Systematics of the genus Qeneiee Berkhout. PMR spectra of the mannans and mannose- containing polysaccharides as an aid in classification. Antcnie_yen_Leeu_enhcek_l_nicrcbicl_§ercl 35: 33-44. 1969. Stewert-Tull DES, Timperley WK, Horne CHW: Some immunochemical observations of the cell walls of yeasts. §aboureueie 5:104-109, 1966. Stone HH: Studies in the pathogenesis, diagnosis and treatment of QenQiQe sepsis in children. ,1 Peeietr Snng 9:127-133, 1974. Stone HH, Geheber CE, Kolb LD, Kitchens WR: Alimentary tract colonization by Qeneiee elhieene. 1 §nrg Bee 14:273-276, 1973. Stone HH, Kolb LD, Currie CH, Gebeher CE, Cuzzell JF: Qeneiee sepsis: pathogenesis and principles of treatment. ADD.§QIQ 179:697-711, 1974. Sullivan PA, Yin CY, Molloy C, Templeton MD, Shepherd M: An analysis of the metabolism and cell wall composition of Qendiee elbieens during germ tube formation. Qan ,1 ninehiel 29:1514-1525, 1983. Sundstrom PM, Kenny GE. Characterization of antigens specific to the surface of germ tubes of Qeneiee eihieene by immunofluorescence. 1n:ee;_1nnnn 43:850-855, 1984. Sundstrom PM, Nichols EJ, Kenny GE. Antigenic differences between mannoproteins of germ tubes and blastospores of Candida albicans. Infect_lmmun 55:616-620. 1987- Suzuki M, Fukazawa Y: Immunochemical characterization of Qeneid_a elhigne cell wall antigens: Specific determinant of.§encice albicens serotype A mannan. Microbicl_immcncl 26:387-402, 1932. 300 Suzuki S, Sunayama H: Studies on the antigenic activities of yeasts. We], 13:95-101, 1969. Suzuki S, Usuyama M: Effect of W -mannosidase digestion on the immunochemical properties of two serotypes of candida yeastS- Tchcku_1ekka_niegeke_zenkyu Henge 18:55-62, 1971. Sweet CE, Kaufman L: Application of agglutinins for the rapid and accurate identification of medically important Candice species. M 19:830-835. 1970- Tadros TF: Particle-surface adhesion, in Berkley RCW, Lynch JM, Rutter PR, Vincent B (eds): WEE-3.8.1911 he enrfaces. London, Ellis Horwood Limited, pp 79-92, 1980. Thomas E, Reddy KR: Nonhealing duodenal ulceration due to candida. i_Cin_§eetrc 5:55-58. 1983- Thienpont D, vanCutsem J, Gough DA: Treatment of gastrointestinal candidosis in predisposed guinea pigs and in conventional mice with miconazole. hyheeen 21:417- 424, 1978. Tronchin G, Poulain D, Herbaut J, Biguet J. Localization of chitin in the cell wall of Qengiee elhieene by means of wheat germ agglutinin. Fluorescence and ultrastructural studies. W 26:121-128, 1981. Tronchin G, Poulain D, Vernes A. Cytochemical and ultrastructural studies of Qeneiee eihieen_. III. Evidence for modifications of the cell wall coat during adherence to human buccal epithelial cells. A;eh_hierehiel 139:221- 224, 1984. Tsuchiya T, Fukazawa Y, Kawakita S: Serological classificaiton of the genus Qengige, in Tokyo: Research Committe on Candidiasis, Education Ministry of Japan: 8ccciee_cn_cendidieeie_in_lepen. pp34-46. 1961- Turner JR, Butler TF, Johnson ME, Gordee RS: Colonization of the intestinal tract of conventional mice with Ceneige elhieene and treatment with antifungal agents. _ADLiEiQIQh Agente_cnemcther 9:787-792. 1976- Umenai T: Systemic candidiasis produced by oral Qeneiee administration in mice. W 126:173-175, 1978. Umenai T, Konno S, Ishida N: Systemic candidiasis by Qeneiee eihieene colonized in the gastrointestinal tract of mice. Ennexiennie 35:1331-1332, 1979. 301 Umenai T, Konno S, Yamauchi A, Iimura Y, Fuj imoto H: Growth of Qengiige in the upper intestinal tract as a possible source of systemic candidiasis in mice. Tohokn 1 Ezp_fleg 130:101-102, 1980. van der Waaij D, Berghuis de Vries JM, Lekkerkerk-van der Wees JEC: Colonization resistance of the digestive tract in conventional and antibiotic-treated mice. 1 Hyg 69: 405- 511, 1971. Verwey EJW, Overbeek JTG: W. Elsevier, Amsterdam, 1948. Vudhichamnog K, Walter DM, Ryley HC: The effect of secretory immunoglobulin A on the in 113,119 adherence of the yeast Qeneiee elhieene to human oral epithelial cells. Arch_crel_§icl 27:617-621. 1982- Waid JS: Physiological and biochemical adjustment of fungi to their environment, an advanced treatise, in Ainsworth GC, Sussman AS (eds): The Fungi. Vol 3. Academic Press, New York, pp 298-301, 1978. Walsh TJ, Merz WG: Pathologic features in the human alimentary tract associated with invasiveness of Candiea gnenieeii_. An 1 Ciin Beth 85:498-502, 1986. White LO, Gibbs E, Newman HC, Richardson MD, Warren RC: Comparison of the growth of attenuated strains of We eihieene in the kidneys of normal and cortisone-treated mice by chitin assay. Myeepeehelegie 67:173-177, 1979. Wilborn WH, Montes LF: Scanning electron microscopy of oral lesions in chronic mucocutaneous candidiasis. M 244:2294-2297, 1980. Wilson KH, Freter R: Interaction of Qleeimiem eifiLieiJe and Escherichia coli with microfloras in continuous-flow cultures and gnotobiotic mice. InfegLimnnn 54:354-358, 1986. Wingard JR, Dick JD, Merz WG, Sanford GR, Saral R, Burns WH: Differences in virulence of clinical isolates of Candiee we and Candida elbicans in mice. Infeee lngnn 37:833-836, 1982. Wingard JR, Dick JD, Merz WG, Sanford GR, Saral R, Burns WH: Pathogenicity of Candida trepicelic and Candida aihicene after gastrointestinal inoculation in mice. 1nfiee§_1nnnn 29:303-313, 1930. Winner HI, Hurley R: Qengi_d_a_ eihieene. London, Churchill, 1964. 302 Yamaguchi H, Kanda Y, Iwata K: Biochemical properties of mitchondria from Candida albicans. Sabouraudia 9:221-230. 1971. Yu R, Bishop C, Cooper F, Hasenclever H: Glucans from Candida albicans (serotype B) and from Candida narepsilesiso _Qan_l_§hem 45:2264-2267. 1967- Yu R, Bishop C, Cooper F, Hasenclever H, Blank F: Structural studies of mannan from Qandida albicans (serotype A at B). Qandida paransilgsis. Qandida ' . and Qandida_tronisalis- Qan.£.§hem 45:2205- 2211 , 1967. Zubrzycki L, Spaulding EH: Application of the continuous- flow culture method to studies on the normal fecal flora. Bacteriol_£rgs , p 101. 1957- Zubrzycki L, Spaulding EH: Studies on the stability of the normal human fecal flora. Wig; 83:968-974, 1962.