2; 1...}. E... er r :lv.¢:..lo 5:}..32 r. t.lr$.‘x)’ai iii-5‘. .r iiitslkfgi I. ‘55) 5.5.3.2159. I... Elsglt’E1fi Enfltflréti it . r (ILKCrfrfIzEIR 51? $.Yflrfi:tffurti (4.3%? (.x; . .37.<(1.7(.?I 11:41:2323. I. . ‘l{.{(¢r.§.. (. 13. rxf ,. Ian (.11: n. (I ‘ t . 5... (tiff t({§.\.€3{?tf it c!" if: £515.. ; 412.7 v1! :51! 3). . . . (4!. ‘ICCKISI. 2.3.1453...) 135(1)} 1 L! 515.. I (I xii .rll IL!!!)PLSCz 1:753 :15 5;; Ir. < f. (it . 1.9? Iii!!!)éflfr (1L1 €y (1.7 r. . é’r‘r’r ‘ mosaic)? . rill}: SEPIA}... :1 {7.1; (I, a‘il‘f.‘ . . (1:31)) I ‘ .. .L....rvl,rr(.rltrzu a; 26.1! (I r :3 5.2.3}? I 13(1)... Ii: 51151;.1411 lilftltllftl...) «it!!! (- 43.5.1. {ii In)? 7).? iii (it‘ll! (4'45!).(li..\a!: 11)). (t!!! 5.1; 1.4.. i grin-git}!!! iviulrc b. .‘4. 1 l ' 1 .31} «i Ii ‘ . girufiinrm'k 1'. 2 325710”? e 7 llllllllllllllllllllllllllllllllllllHllllllllllllllllNll w 31293 00788 5886 -_.__. F LIBRARY Michigan State 1 University p M“ This is to certify that the thesis entitled THE IDENTIFICATION AND SPECIATION OF SELECTED CORYNEBACTERIUM SPECIES presented by Hsiuo-Yi Su has been accepted towards fulfillment of the requirements for MS degree in CLS MM“ Major pr e Date February 15, 1990 0-7639 MS U is an Affirmative Action/Equal Opportunity Institution PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before date due. DATE DUE DATE DUE DATE DUE MSU Is An Affirmative Action/Equal Opportunity Institution CZHICWDM3-D.‘ .; LA ..-— 'T n , .‘r: 2.5—:e1‘25—‘da-T-‘J .YHM . . _ , . THE IDENTIFICATION AND SPECIATION OF SELECTED CORYNEBACTERIUM SPECIES BY Hsiuo-Yi Sn L THESIS Submitted to Michigan State University in partial fulfillment of the requirements of MASTER OF SCIENCE Department of Clinical Laboratory Science 1990 a W, a’” QU54IXB ABSTRACT THE IDENTIFICATION AND SPECIATION OF SELECTED CORYNEBACTERIUM SPECIES BY Hsiuo-Yi Su The nondiphtheria corynebacteria are a poorly classified group of bacteria commonly found on human skin and mucous membranes. Because of the expanding problem of infections in immunocompromised hosts, the isolation and identification of these bacteria are of increasing significance. Forty nine isolates of Corynebacterium spp. and Arcanobacterium haemolyticum were identified with the API 208 rapid identification system. These results were compared to those obtained with conventional biochemical methods as well as those obtained from gas-liquid chromatographic analysis of cellular fatty acids (GLCFA). The API 208 system cannot produce a definitive identification of Corvnebacterium spp. However, with the addition of GLCFA, it can identify and speciate most medically important corynebacteria. 6-aminolevulinic acid (6-ALA), hydrolysis of Tween 80 and susceptibility to antibiotics were also investigated. All Corvnebacterium spp. which could utilize 6-ALA, were susceptible to vancomycin but were resistant to fosfomycin. These unique characteristics will be helpful in the identification. ACKNOWLEDGEMENTS I gratefully acknowledge and appreciate my major advisor, Dr. Barbara E. Robinson, Diagnostic Microbiology Section, Bureau of Laboratory and Epidemiological Services, Michigan Department of Public Health, for her continuous guidance, support and encouragement in the course of my research. I would like to thank the members of my Master's degree committee, Dr. Douglas Estry, Dr. Martha Mulks and Dr. Alvin Rogers for their involvement in this research. Special thanks are given to the Diagnostic Microbiology Sections for their assistance in laboratory work and for their friendship. Sincere acknowledgement and appreciation are extended to Mrs. Beatrice E. Treleaven for her technical advice and encouragement. I thank Dannie G. Hollis, Centers for Disease Control, Atlanta, GA., for kindly providing cultures and gas-liquid chromatographic information. Last but not least, I would like to express my deepest gratitude to my family and friends for their support and love throughout my educational endeavor. ii TABLE OF CONTENTS PAGE LIST OF TABLES ....................................... iii LIST OF FIGURES ...................................... v I. INTRODUCTION ................................ 1 II. LITERATURE REVIEW ........................... 3 History of Corynebacterium Species .......... 3 History of Gas-liquid Chromatography ........ 6 Introduction ............................... 6 Identification and taxonomy ................ 7 Direct diagnosis of infection .............. 10 Identification of Corynebacterium Species by Gas-liquid Chromatography .................. 12 Identification of Corynebacterium Species by Commercial Test Kits ....................... 15 Antimicrobial Susceptibility of Corynebacterium Species .................................... 17 Unique Biochemical Characteristic of Corynebacterium Species .................... 18 III. MATERIALS AND METHODS ....................... 20 Bacterial Isolates .......................... 20 Carbohydrate Utilization Reactions .......... 24 Biochemical Reactivity ...................... 25 Growth Studies .............................. 25 Supplemental Tests Used for Speciation ....... 26 6-aminolevulinic acid ...................... 26 Hydrolysis of Tween 80 ..................... 27 Susceptibility to selected antimicrobial agents .................................... 27 API 208 Rapid Identification System ......... 28 Analysis of Cellular Fatty Acid by Gas- liquid Chromatography ...................... 31 Preparation of cellular fatty acid methyl esters extracts ........................... 31 Gas-liquid chromatography of cellular fatty aCid O0....O00......OOOOOOOOOOOOOOOOOOOOOOO 32 Iv. RESULTS OOOOOOOOOOOO0.00000000000000000000000 34 Colonial and Cellular Characteristics ....... 34 Carbohydrate Utilization Reactions .......... 35 Biochemical Reactivity ...................... 39 Growth Studies .............................. 41 Supplemental Tests Used for Speciation ...... 41 API 208 Rapid Identification System ......... 45 Gas-liquid Chromatography ................... 47 iii v. DISCUSSION 0 O O O O O O O O O O O O O O I O O O O O O O O O O O O O O O O O O 51 VI 0 CONCLUSION O O O O O O O O O O O O O O O O O O I O O I O O O O O O O O O O O O 6 4 VII O LITERATURE CITED O O O O O O O O O O O O O O O O O O O O O O O O O O O O 6 6 iv 10. 11. 12. LIST OF TABLES Bacterial Isolates Tested ....................... Type Strains for A. haemolyticum and Cogynebacterium Species ........................ Abbreviations of the API 208 Biochemical Tests .0.0...0.0.0.0....OOOOOOOOOOOOOOOOOOOOOOOOO Operating Parameters of Gas-liquid Chromatography ................................. Determination of Carbohydrate Utilizations for A. haemolyticum and Corynebacterium Species .... Abbreviations Used for Carbohydrate Utilizations and Biochemical Reactivity ..................... Determination of Biochemical Reactivity of A. haemolyticum and Corynebacterium Species ....... Growth studies of Salt Tolerance and Temperature Dependency for A. haemolyticum and Corynebacterium Species ....................... Supplemental Biochemical and Susceptibility Tests for A. haemolyticum and Corynebacterium SpeCies O...0......OOOOOOOOOOOOOOOOO0.0.0000... API 208 Profiles of A. haemolyticum and Corynebacterium Species ........................ Cellular Fatty Acid Composition of A. haemolyticum and Corynebacterium Species .................... Biochemical Tests for Separation of A. haemolyticum and Corynebacterium Species Within a GLC Group 0....OOOOOOOOOCOOOOOOOOOOO0.00.00.00.00. Page 21 23 30 33 36 37 4O 42 44 46 48 SO LIST OF FIGURES Figure page 1. Identification Scheme for Selected Corynebacterium 58 SPeCies O0.0...0.0.0....OOOOOOOOOOOOOOOOOOOOOOO vi INTRODUCTION Corynebacterium species may have a commensal relationship with other microorganisms in the normal pharyngeal, cutaneous and gastrointestinal flora of humans. However, they may also be implicated in a variety of infectious diseases such as endocarditis, meningitis, brain abscesses, pneumonia, empyema, lung abscesses, septicemia, osteomyelitis and skin and soft tissue infections. Infections tend to occur in patients with defective host defenses, particularly in those with malignancies, breaks in skin and mucous membranes or patients receiving immunosuppressive therapy. As the incidence of organ transplantation and acquired immunodeficiency syndrome (AIDS) increases, it is likely that serious disease due to Corvnebacterium species will also increase. Identification of the Corvnebacterium species may pose difficulties, because of (i) the requirements of some species for prolonged incubation, special media and exacting growth conditions [56] (ii) the difficulties in characterizing corynebacteria biochemically [62], serologically [3] and toxicologically [65] and (iii) the reigning confusion regarding the exact taxonomic status of many bacterial species that morphologically resemble members 2 of the genus Corvnebacterium [64]. Therefore, attempts to definitively identify these bacteria often end in frustration. Many species are relatively inactive in standard biochemical tests, thus producing too few "positive" results to provide adequate characterization. Even among the biochemically active organisms, biochemical tests usually provide varied results both within and among laboratories. Nevertheless, little attention has been focused on the development of new procedures or use of standardized test procedures that would be helpful in differentiating the Corynebacterium species. It has been proposed that Corynebacterium haemolyticum should be reclassified in a new genus, Arcanobacterium, which is comprised of a single species, A. haemolyticum. This organism appears distinct from all corynebacteria on the basis of numerical taxonomy and data from the analysis of fatty acids, peptidoglycans, menaquinones and DNA [10]. In this project, A. haemolvticum and selected Corvnebacterium species were examined for (1) conventional biochemical reactions, (ii) porphyrin production, (iii) hydrolysis of Tween 80, (iv) susceptibility to selected antibiotics, (v) identification by the use of a rapid commercially available system and (vi) the composition of cellular fatty acids present in the cell wall. We would like to establish a definitive identification scheme by modifying the present methods [28, 39, 62, 64]. LITERATURE REVIEW Corynebacterium species are often isolated from various clinical specimens. These bacteria are generally disregarded as contaminants. In recent years, there has been evidence that Corynebacterium species can cause serious and life- threatening infections in immunocompromised hosts. Therefore, the identification of the Corvnebacterium species has become increasingly important. History of Corvnebgcterium species Pleomorphic, nonsporulating, gram-positive bacilli are generally referred to as coryneforms and are common inhabitants of the human skin, particularly in the intertriginous regions [53]. They have been considered to have morphological similarities to Q. diphtheriae, hence they have also been termed "diphtheroids". However Pitcher in 1977 [52] using cell-wall analysis showed that only about 60% of skin coryneforms are true Corvnebacterium species and that there are several other coryneform genera among the bacterial flora of human skin. This finding has rendered previous classifications of the skin coryneforms invalid because they included isolates of several genera [20, 39, 4 62, 64]. In 1971, a report published by Cummins [14] defined Corvnebacterium species by the presence of mesodiaminopimelic acid, arabinose and corynomycolic acids in the cell wall was commonly accepted. However, a numerical taxonomy of Corynebacterium species based on conventional bacteriological tests still demonstrated a high degree of diversity. Infections caused by Corynebacteria other than Q. diphtheriae have been reported [37]. For example, Q. Ainutissimum can cause erythrasma of the skin [59]. In 1984, Berger e; di- [4] encountered a case of severe and recurrent breast abscesses caused by Q. minutissimum. Recently the medical significance of Corvnebacterium species has been stressed in reports of nosocomial infection in compromised patients caused by Q. xerosis [70], Q. pseudodiphtheriticum (Q. hofmannii) [19], and Corynebacterium group JK (Q. jeikeium) [16, 54, 55, 56, 68]. In 1986, Samies g; g;. [58] reported a patient with AIDS complicated by a lung abscess and persistent bacteremia due to Corynebacterium eggi. It has often not been possible to identify these organisms to the species level using morphological criteria or conventional biochemical tests. They vary in size from short coccobacilli to longer bacillary forms, some straight and some club—shaped. This variation in cellular morphology is affected by the conditions of growth. In addition, many species are relatively inactive in standard biochemical 5 tests and even among the biochemically active organisms, test results are unusually varied. It seems that there are no consistently positive biochemical reactions other than production of catalase [39]. In 1981, Suzuki A; AA. [67] divided 33 isolates of coryneform bacteria into seven groups by the use of DNA-DNA hybridization and found good agreement between the DNA homologies and the chemotaxonomic profiles reported to date for the Corvnebacterium species. However, the DNA base composition of the Corynebacterium species varied widely from 52 mol% to 70 mol% G+C. This range is too wide to use DNA-DNA hybridization as an identification method [67]. In addition, technical difficulties existed in harvesting sufficient DNA from the Corynebacteria to perform homology testing. Moreover, standard representatives had not been deposited in the various national culture collections which could have been used for comparison. It is possible to perform numerical taxonomy based on protein electrophoresis. Numerical taxonomy based on whole- cell proteins separated on polyacrylamide gel is becoming a method of classifying bacteria and shows a good correlation with DNA-DNA hybridization. However, in 1982, Jackman [30] failed to classify Corvnebacterium, Mvcobacterium, Gordona, Rhodococcus and Staphylococcus by analysis of electrophoretic protein patterns, because Staphylococcus aureus, Staphylococcus epidermidis and several 6 Corvnebacterium species shared the same protein patterns. Therefore, the protein patterns for corynebacterial isolates may not be genus specific. History of Gas-liquid Chromatography (GLC) Introduction Since the introduction of chromatography in 1905, microbiologists have used chromatographic procedures for the isolation, purification and separation of microbial components and products from metabolism. The concept of using GLC as an analytical tool for the identification of microorganisms was advanced in the middle 19605. Results from early studies on the analysis of organic acids and other microbial metabolites, cellular fatty acids and products from the pyrolysis of whole bacterial cells demonstrated the feasibility of this approach and provided a basis for subsequent detailed investigations of a large number of microorganisms. The reliability, precision and accuracy of recent studies have been improved significantly by continuing advancements in GLC instrumentation. In the early 19805, desktop computers became widely available in the laboratory to aid in the determination of GLC results. The use of modern chemical database management software has led to the development of a computer-controlled 7 GLC system that automates the analysis of acid identification, the determination of fatty acid composition and the comparison of prepared bacterial isolates to known profiles for identification [42, 43]. In recent years, GLC has been used to study the chemical composition and metabolic activity of microorganisms as a basis for their identification. Chemical compounds such as short—chain acids, amine metabolites and cellular fatty acids have provided valuable information for the recognition of both genus and species. Identification and taxonomy A major use of GLC in microbiology laboratories has been in the analysis of bacterial metabolites. Systematic studies have demonstrated the value of short-chain acids in the identification of both anaerobic bacteria [27, 66] and aerobic bacteria [17, 18]. For example, in 1978, Dees e; BI. [17] indicated that Achromobacter-like organisms produced large amounts of 2-ketoisocaproic acid and Flavobacterium meningosepticum produced large amounts of 2-ketoisovaleric acid. The presence and relative amounts of these acids were found to be useful for distinguishing these organisms. Bacterial metabolites other than short-chain acids can also be analyzed by GLC. Brooks [5] in 1969 and Coloe [12] in 1978 both reported the detection of aliphatic and 8 aromatic amines which were shown to be of taxonomic value in the identification of Clostridium species and Proteus species respectively. Another value of GLC is in the analysis of cellular fatty acids (GLCFA). This analysis will give a chromatogram having a distinct pattern of fatty acids for each group or species of bacteria. Closely related species may be differentiated by the presence, absence or large quantitative differences of one or more acids. The fatty acids which may be present in the bacterial cell vary in the chain length from C1° to cm, the mono-unsaturated forms from C1,,1 to Cm”, the branched chain fatty acids (both iso and antiiso), cyclopropanes (Cu-cyc and Cm-cyc), as well as a(2-hydroxy) and B(3-hydroxy) acids of Cm, Cm, Cu and C16 fatty acids. The identity of the separated peaks was established by comparison of the relative retention times with those of a standard mixture of straight chain saturated and mono-unsaturated fatty acid methyl esters. In 1981, Moss [47] reported that information that aids distinction between closely related bacterial species is obtained from the analysis of GLCFA composition and from short-chain acid products from various bacterial genera. Genera and species are therefore distinguishable by the presence or absence of particular fatty acids or by quantitative differences in the amounts of acid present. In an earlier paper, Moss [46] reported the GLCFA profile of 9 two species of Eseudomonas cepacia and Pseudomonas aeruqinosa. It was clear from comparison of the chromatograms that there were major differences between these organisms. 2. aeruginosa contained four acids (3-OH me, Cum, 2-OH cm0 and 3-OH Cum) which were not present in 2. cepacia. In addition, 2. cepacia contained three acids (3-OH Cum, 2-OH me and 3-OH me) which were absent in A. aeruginosa. A major quantitative difference was the relatively large amounts of 17 and 19 carbon cyclopropane acids in P. cepacia compared to that found in E. aeruginosa. During the past several years, techniques for the identification of Campylobacter to the species level by conventional methods has tended to be very difficult because these bacteria are asaccharolytic, inactive in most biochemical tests and may have atypical or inconclusive biochemical reactions. In 1987, Lambert A; AA. [36] analyzed GLCFA compositions of 368 isolates of Campylobacter species and Campylobacter-like organisms by GLC. The Campylobacter species or Campylobacter-like organisms could be placed into four groups based on differences in their fatty acid profiles and they could be speciated with the addition of few biochemical reactions. Therefore, GLCFA could be a presumptive test in differentiating Campylobacter species and Campylobacter-like organisms. In addition, automated GLC instrumentation was used to allow the analysis of large numbers of samples with minimal operator interaction. 10 Fatty acids profiles can also be used in combination with other methods to provide a strong taxonomic classification even with isolates which had atypical phenotypic characteristics. In 1988, Wallace A; AA. [71] used GLC to determine the fatty acid composition of three Kingella species, Cardiobacterium hominis and Eikenella corrodens. The ability to resolve 18-carbon isomers provided a clear means of differentiating A. denitrificans and A. kingae from Kingella indolo enes, Eikenella corrodens and Cardiobacterium hominis. Originally, Kingella indoloqeneg was added to Kingella species on the basis of biochemical similarities and percent DNA base composition [63]. The study of DNA-rRNA hybridization [57] showed that Kingella indologenes is not related on the generic level to other Kingella species and is not a Kingella species. The present of large amounts of Cm,1 in A. indoloqeneg but not in A. denitrificans or A. kingae further confirmed that this species was not closely related to the other two. Direct diagnosis of infection After GLC was applied to the identification of microorganisms, studies were initiated to use the technique for the direct examination of body fluids in the hope of discovering unique or characteristic compounds as markers of specific diseases. In 1983, Brooks [6] used frequency-pulsed 11 electron capture GLC (FPEC-GLC) to examine amine profiles of cerebrospinal fluid (CSF) from patients with acute tubercular meningitis. Brooks found that indole might be uniquely present in tubercular meningitis. Later studies [21] indicated that it was present in only 50% of those cases and also was present in some CSF from patients with acute bacterial infections, but not in patients with cryptococcal, viral or parasitive infections. Brooks [6] also applied FPEC-GLC techniques to the study of fecal specimens from patients with various types of diarrheal diseases. The preliminary data have shown marked differences in FPEC-GLC profiles among stools from confirmed cases of antimicrobial-associated pseudomembranous colitis due to Clostridium difficile, enterotoxigenic A. lei, rotavirus and adenovirus. Additional specimens are being studied to expand the data base and other efforts are being made to identify some compounds which appear to be useful chemical markers of those diseases. For example, in 1987, French and Chan [24] detected tuberculostearic acid [(R)-10- methyloctadecanoic acid (TBSA)] in appropriate clinical specimens (sputum) as a means of screening for the presence of Mycobacterium tuberculosis and other mycobacteria by using gas chromatography-mass spectrometry (GC-MS) combined with selected ion monitoring. Although the GLCFA patterns may or may not be sufficient for species identification, in some cases they 12 can provide additional data for the identification of microorganisms. The GLCFA patterns will group selected isolates which can usually be separated by a few biochemical tests [36, 44, 45, 48]. In addition, chromatographic information can also be used to identify microorganisms with atypical phenotypes [36] or which were previously misclassified [71]. Furthermore, the use of FPEC-GLC and combined GC-MS techniques in studies of body fluids might exist some limitations, but, in the future, these techniques could become potential analytical tools for the rapid diagnosis of infections diseases without the need to resort to cultures in the future. Identification of Corynebacterium Species by GLC Lipid markers have become increasingly useful in the classification and identification of bacteria. The potential usefulness of GLC for the classification and identification of Corynebacterium isolates and related taxa has been recognized. However, little research has been published on the use of GLC for routine identification and speciation of Corynebacterium species. In 1982, Collins and Goodfellow [11] used GLC to analyze the non-hydroxylated long-chain fatty acid methyl esters of cutaneous corynebacteria and found that they contained major amounts of straight-chain and mono-unsaturated fatty acids. For example, Q. 13 pseudodiphtheriticum and Q. xerosis had major amounts of hexadecanoic (Cum) and mono-unsaturated octadecenoic (Cmfl) acids. Furthermore, in 1985, Athalye and Noble [2] analyzed the fatty acid methyl esters of 19 unidentified pathogenic coryneform bacteria by GLC and compared the resulting profiles to the reference strains, namely Caseobacter 01 or bus, Corynebacterium ppyig, Q. diphtheriae, Q. xerosis and Rhodococcus egp;. All of the isolates had distinct fatty acid profiles but most conformed to a general pattern of high levels of C18,l with only trace amounts of 10-methyl octadecenoic acid (tuberculostearic acids). This observation is in general agreement with Collins and Goodfellow [11]. The profiles were very similar to those obtained from Q. diphtheriae and Q. xerosis but could be differentiated from Q. ppyis, Qgg. 01 or bus, 3. egg; and two unidentified pathogenic isolates which had significantly higher levels of tuberculostearic acid. Athalye and Noble also found that the Corynebacterium JK group was distinguishable from the other corynebacterial isolates examined [2] by the lack of tuberculostearic acid and the size of their mycolic acids (Cu-C“). These results strongly supported the proposal of Riley ep Ql- [56] to reclassify these organisms into a new species. In 1985, McGinley A; El- [40] working with the cutaneous corynebacteria, used a fused silica capillary column to quantify mycolic acids from C224,9 and from Cwas. 14 Analysis of extractable mycolic acids revealed that all species except Q. ppyig, which had 87% CAM”, had 73-92% C3,,»36 mycolic acid. Butler pp _A. [7] in 1986 used high- performance liquid chromatography (HPLC) of bromophenacyl esters of mycolic acid as an aid to assign a particular organism to one of four mycolic acid-containing genera: Corynebacterium, Rhodococcus, Nocardia and Mycobacterium. They found that the number of carbon atoms in mycolic acids of Corynebacterium species varied from C2° to C36 while those in Nocardia species and Rhodococcus species had chain lengths ranging from C36 to C“. They showed that HPLC would be a useful tool to separate the various classes of mycolic acids on the basis of chain length and the degree of saturation. In 1988, Butler A; AA. [8] differentially identified Mycobacterium species by reverse-phase HPLC to detect p- bromophenacyl mycolic acid esters of four major pathogenic species. They showed that it was possible to shorten the saponification time and still provide an effective, direct analysis of the mycolic acids of the slowly growing pathogenic mycobacteria. However, in another investigation done by Chevalier and Pommier [9], changes in the mycolic acid composition of three cutaneous isolates of Corynebacterium species were shown to be associated with the addition of Tween 80 (a polyethylene derivative of sorbitan monooleate) to the culture medium. Analysis of data from GC- 15 MS showed that the carbon chain length and the degree of unsaturation had been affected: the levels of corynomycolic acid with 36 carbon atoms and two double bonds increased significantly and this high level of C36,2 corynomycolic acid may have been due to the condensation of two oleic acid (Cmd) molecules. It became obvious that mycolic acid composition could be affected by environmental conditions. Indeed, previous results by Athalye [2] showed that the C13,1 fatty acid content was increased in bacteria grown with Tween 8 0 . Identification by Commercial Test Kits Generally, biochemical characterization of Corynebacterium species requires 24 to 72 h. Some species may require up to 2 weeks for a definitive identification by conventional carbohydrate utilization. In order to reduce the time required for biochemical identification, several rapid test methodologies have been developed. In 1983, Thompson [69] reported the use of a rapid fermentation test with Andrade's indicator and 10% rabbit serum to identify Q. diphtheriae and other medically important Corynebacterium species within 4 h. Later in 1984, Kelly A; Q1. [33] reported on the use of the API 208 system (Analytab Products, Plainview, NY) for the rapid identification of multiply antibiotic-resistant, aerobic Corynebacterium group 16 JK. The API system is based on rapid utilization of carbohydrates and on hydrolysis of substrates due to the presence of preformed enzymes in the test bacterium. The identification provided by the API 208 system was confirmed by conventional biochemical methods; however, the data generated from that investigation did not produce sufficient information on Corynebacterium other than the JK and DZ groups. In 1986, another rapid identification system, Minitek (BBL Microbiology Systems, Cockeysville, MD) [61] was used to correctly identify a wide range of Corynebacterium species, including group JK within 12 to 18 h of incubation. An exception to this was the acidification of maltose for group JK required up to 72 h. The observed results generally agreed with the Coyle's report [13]. Additionally, the Minitek system included both fructose and o-nitrophenyl-B-D- galactosidase (ONPG); therefore, the Q. pseudogenitalium biotype could be differentiated from Q. genitalium and the JK group from C. bovis respectively [13]. Furthermore, a positive ONPG response served as a useful hallmark for the identification of Q. aquaticum and Q. minutissimum [54]. However, the cost and time requirement for set-up were not economic compared to the other rapid identification systems. The Rapid Identification Method (RIM series; Austin Biological Laboratories, Inc., Austin, TX) was used to identify and differentiate Corynebacterium group JK [25]. 17 The RIM kit includes carbohydrate utilization, nitrate reduction and urease activity that requires less than 1 h of incubation and correlated well (292%) with conventional methods with the exception of the maltose test. However, when negative glucose reactions were obtained by RIM, all other RIM tests were negative. These negative reactions should alert one to retest. Because of the existing difficulties among rapid fermentation tests, the Minitek system and RIM, the good correlationship with conventional methods and easy handling make the API 208 a promising candidate for the rapid identification of the Corynebacterium species. Antimicrobial Susceptibility of Corynebacterium Species Most of the Corynebacterium species other than group JK are susceptible to the antimicrobial agents commonly used to treat infections caused by gram-positive microorganisms. Unfortunately, group JK can produce serious and even fatal infections, particularly in patients suffering from terminal debilitating diseases. These bacteria are resistant to many antibiotics; however, they remain susceptible to vancomycin [26]. Corynebacterium group JK can also be difficult to isolate in blood cultures. A semiselective medium for the detection of multiresistant corynebacteria was described in 18 1984 by Wichmann e; _l. [72]. The medium consisted of lecithin, histidine, glycerol, sodium thiosulfate, fosfomycin (phosphomycin, phosphonomycin, fosfomycin, Sigma Chemical Co., St. Louis, MO), ticarcillin and 5- fluorocytosine. This was especially useful for Corynebacterium group JK because of the low frequency of isolation in primary cultures. In 1988, Wirsing von Koenig e; _l. [73] used fosfomycin to selectively isolate corynebacteria from clinical specimens. It provided a simple, highly sensitive, specific aid for the isolation of Q. diphtheriae and the other Corynebacterium species from clinical specimens. Unique Biochemical Characteristic of Corynebacterium Species Other special media might be needed to enhance growth for Corynebacterium species. For example, Corynebacterium group JK usually requires blood or serum in the medium while Q. ppyig requires nutrient agar with Tween 80 in concentrations as low as 0.01%. Therefore, lipid may not be needed as a carbon source but, rather, might serve as a vitamin or cofactor. In 1969, Smith [62] found that Tween compounds were inhibitory to the corynebacteria at 1.0% concentrations. Inhibition was considered to have occurred when growth developed at higher but not lower substrate concentrations. In Smith's study, laureate (Tween 20), 19 palmitate (Tween 40), stearate (Tween 60) and oleate (Tween 80) were tested to find the concentrations permitting growth of all corynebacterial isolates. Smith also found that all Copynebacterium species produced porphyrin and showed coral- red fluorescence under a Wood's ultraviolet light (wavelength, approximately 360 nm), after incubation at 37°C for 18 to 36 h. The production of porphyrin can also be used as a clinical marker for mild skin infection erythrasma in which the scaly skin lesions fluoresce coral pink if exposed to UV light. Despite improvements in the identification of Corynebacterium species due to numerical phenetic analysis, tests based upon form and function have failed to yield a stable identification scheme for Corynebacterium species due to the immense number of biotypes that can be grouped only in an arbitrary way. Recently, attention has turned to lipid analysis for the provision of characteristics. This study describes the use of the modern analytical techniques (gas— liquid chromatography) in combination with selected conventional microbiological tests to establish a more efficient and accurate means of identification and speciation of the Corynebacterium species. MATERIALS AND ‘TEODS Bacterial Isolates Stock cultures of Corynebacterium species were obtained from clinical specimens and were stored in 1% skim milk at -70°C. Isolates were passaged twice on 5% sheep blood agar plates (BBL Division, Becton Dickinson Company, Cockeysville, MD) prior to use. Additional isolates were provided by Dannie G. Hollis, Centers for Disease Control (CDC), Atlanta, GA. Tested isolates are listed in Table 1. All isolates were subcultured onto 5% sheep blood agar, incubated at 35°C in 4% CIA for 24-48 h, then observed for colonial morphology, reaction on blood and then stained by the Gram's method [28]. Colonial morphology was observed for size, shape, colony surface, consistency and pigmentation of growth. Gram stain smears were examined for cellular morphology including cell shape, size, arrangement and spores. In order to provide confirmatory information, the biochemical features and cellular fatty acid profiles of the representative strains of each species were also examined. The type strains are listed in Table 2. 20 21 Table 1. Bacterial Isolates Tested Bacteria Culture Source Obtained No. from Arcanobacterium haemolyt icum ATCC 9 3 4 5 ATCC‘I stra in ATCC A 204-87 Foot Ulcer MDPHb A 854—87 Elbow Drainage MDPH A 229-89 Blood MDPH Corynebacterium group JK A 399-85 Blood MDPH A 655-85 Blood MDPH A 56-86 Blood MDPH A 57-86 Blood MDPH Q. minutissimum-I A 368-86 Blood MDPH A 606-86 Blood MDPH A 999-87 Blood MDPH F 6508-88 Unknown CDCc F 8508-88 Unknown CDC F 7602-88 Unknown CDC KC 1381-90 Unknown CDC Q. minutissimum—II F 6150-88 Unknown CDC F 8896-88 Unknown CDC F 9839-88 Unknown CDC G 286-88 Unknown CDC A 387-89 Blood MDPH A 388-89 Blood MDPH A 563-89 CSF MDPH KC 1291-90 Unknown CDC Q. pseudodiph- theriticum A 791-86 Nasal MDPH A 820-86 Nasopharyngeal MDPH A 51-88 Nasal MDPH A 67-88 Nasopharyngeal MDPH A 508—89 Sputum MDPH KC 1364-90 Unknown CDC 22 Table 1. (continued) Bacteria Culture Source Obtained No. from Q. renale-I B 2229-88 Unknown CDC C 1500-88 Unknown CDC KC 1366-90 Unknown CDC Q. renale-II B 3267-88 Unknown CDC B 3473-88 Unknown CDC Q. striatum A 499-85 Blood MDPH A 25A-88 Placenta MDPH A 58-88 Sputum MDPH A 191-88 Wound drainage MDPH F 5333-88 Unknown CDC G 614-88 Unknown CDC G 606-88 Unknown CDC KC 1871-88 Unknown CDC KC 1367-90 Unknown CDC Q. xerosis F 1511-88 Unknown CDC F 3785-88 Unknown CDC G 677-88 Unknown CDC KC 1368-88 Unknown CDC A 758-89 Unknown MDPH Atypical Q. xerosis A 22-87 Blood MDPH °ATCC: American Type Culture Collection, Rockville, MD. bMDPH: Michigan Department of Public Health, Lansing, MI. CCDC: Centers for Disease Control, Atlanta, GA. 23 Table 2. Type Strains for Arcanobacterium haemolyticum and Corynebacterium Species Species Isolate No. Resembles Type Strain A. haemolyticum Q. minutissimum-I KC Q. minutissimum-II KC Q. pseudodiph- theriticum KC Q. renale-I KC Q. striatum KC Q. xerosis KC 1381 1291 1365 1366 1367 1368 ATCC 9345 NCTC‘10288 ATCC 23346 ATCC 19410 ATCC 19412 ATCC 6940 ATCC 373 ‘NCTC: National Collection of Type Strain Cultures, Central Public Health Laboratory, London, England. 24 Carbohydrate Utilization Reactions The ability of the bacterial isolates to utilize carbohydrates was determined on triple sugar iron agar (T81) and carbohydrate test media. The TSI was used to detect bacterial growth, biochemical responses and the production of hydrogen sulfide. The production of acids from the carbohydrate test media was assessed according to accepted protocols [28] except as described below. The initial concentrations and substrates tested were 1% for galactose, glucose, lactose, maltose, mannitol, sucrose, xylose and 10% for lactose. Following studies with the API 208 system, fermentation of 1% trehalose was added. In addition, fermentation of 1% salicin and 2% starch were added to the carbohydrate test scheme [13]. Enhancement of growth with 10% rabbit serum supplementation was used only for Corynebacterium group JK. All carbohydrates except for 10% lactose which consisted of 10% lactose in phenol red broth base [28], were incorporated into bromocresol purple base at pH 7.0, autoclaved at 10 lb/in‘ at 121°C for 10 min and stored at 4°C for 6 months [28]. After inoculation, the carbohydrate test media were incubated at 35°C and examined daily for acid production for 3 days. 25 Biochemical Reactivity The determination of biochemical features of bacterial isolates was performed as described by Hollis and Weaver [28]. The biochemical tests included the hemolytic reaction on 5% sheep blood agar, production of catalase, reduction of nitrate, hydrolysis of gelatin, motility, hydrolysis of esculin, production of indole from tryptophan, production of stable acid from glucose (methyl red), production of acetoin from glucose (Voges-Proskauer), utilization of citrate and multiple metabolic reactions in milk medium (litmus milk). All tests were incubated in ambient conditions at 35°C. Cultures were observed and results recorded for a total of eight days of incubation except litmus milk, which was incubated for 15 days. Growth Studies The ability of the bacterial isolates to grow in the presence of various concentrations of salt was determined. Salt tolerance testing was performed by inoculating two drops (approximately 80 pl) of a 24-48 h old heart infusion broth culture into nutrient broth to which different concentrations (0%, 3%, 6% and 9%) of NaCl had been added [28]. The inoculated cultures were incubated at 35°C in ambient conditions and were observed for 7 days. Growth was 26 judged as inhibited when poor growth or no growth occurred in broth with salt concentrations from 0% to 9%. The ability of the bacterial isolates to grow at different temperatures was also assessed. Temperature studies were performed by inoculating two drops (approximately 80 pl) of a 24-48 h old heart infusion broth culture onto duplicate tubes of tryptone glucose yeast extract (TGYE) agar [28]. An exception to this was Corynebacterium group JK which was inoculated onto TGYE agar containing 5% sheep blood. All cultures were incubated at 25°C and 42°C respectively in an ambient atmosphere and were observed for 7 days. Results were recorded as growth or no growth occurring on the TGYE agar. Supplemental Tests Used for Speciation 6-aminolevulinic acid (6-ALA) The production of porphyrin was demonstrated by the appearance of coral—red fluorescence under a Wood's ultraviolet light [13]. The determination of porphyrin production was performed by incorporation of 0.5, 1.0, 2.0 and 4.0 mM 6-ALA (Sigma Chemical Co., St. Louis, MO) into M/15, pH 7.0 phosphate buffer [34]. The mixture was sterilized with a 0.45 pm filter (Nalge Co., Rochester, NY), distributed in 0.5 ml amounts in 4 m1 sterile capsule vials 27 and stored at -20°C up to 6 months. A 3-mm loopful of a 24- 48 h old bacterial culture was emulsified and incubated at 35°C in ambient conditions. The inoculum was examined for fluorescence under a Wood's lamp (wavelength, approximately 360 nm) after 18-24 h incubation. Hydrolysis of Tween 80 The ability of Corynebacterium species to produce esterase was determined by the formation of cloudy zones around bacterial colonies on trypticase soy agar containing 0.01% CaCl2 [62]. The determination of esterase activity was modified by incorporation of 0.05, 0.1, 0.2 and 0.4% Tween 80 (Difco Laboratories, Detroit, MI) into M/15, pH 7.0 phosphate buffer containing 0.1% neutral red dye [22]. The mixture was dispensed in 2 ml amounts in screw-capped tubes, autoclaved at 121°C for 15 min and stored at 4°C for two weeks. Tested isolates were heavily inoculated as described above into the phosphate buffer, incubated at 35°C in ambient conditions and observed up to 7 days. Tests were repeated once if they resulted in negative reactions. Susceptibility to selected antimicrobial agents Two antibiotics discs were tested on Mueller-Hinton agar plates containing 5% sheep blood: 200 pg fosfomycin and 28 30 pg vancomycin (Difco Laboratories, Detroit, MI) disks. The test methods were processed according to standard methods described in NCCLS M2-T4 [50]. Sensitivity and resistance of vancomycin were defined by the zone diameters listed in the NCCLS M2-T4 [50]. Because fosfomycin disks containing 200 pg of fosfomycin (Sigma Chemical Co.,St. Louis, MO) and 20pg of glucose-6-phosphate (Aldrich Chemical Co., Milwaukee, WI) were not commercially available, they were prepared and interpreted according to Forsgren's report [23]. API 208 Rapid Identification System The API 208 system (Analytab Products, Plainview, NY) was evaluated for the rapid identification of A. haemolyticum and Corynebacterium species. The test strips were processed according to the method of Kelly [33] which differs from the API 208 protocol by the means of inoculum preparation. A cell suspension was prepared by removing 4 to 5 colonies of a 24-48 h old isolate from a 5% sheep blood agar plate with a sterile cotton swab and emulsifying it in sterile saline to approximate the turbidity of a 5.0 McFarland standard. The API 20S strips were inoculated with a sterile Pasteur pipette by filling each cupule with two drops (approximately 80 pl) of the suspension. The strips were incubated at 35°C for 4 h in an ambient atmosphere. 29 Cultures of Streptococcus pyogenes and Streptococcus aqalactiae were used as controls as recommended by the manufacturer. A biotype profile number was constructed from the results of each isolate. The principle of coding is to condense the 20 binary pieces of information (+ or -) into a profile number. To do so the tests are divided into groups of three and each positive reaction is given a value equal to 1, 2 or 4 according to the position of the test in its group: first, second or third, respectively. The sum of these 3 values (0 for negative reactions) gives the corresponding digit with a value between 0 and 7 (208 Analytical Profile Index, Analytab Products, Plainview, NY). The identifications obtained with the 208 strip were confirmed by conventional methods including utilization of carbohydrates [28], hydrolysis of esculin [28], hydrolysis of hippurate [29], production of B-galactosidase [38] and production of B-glucosidase [51]. A listing of the 208 tests and their abbreviations can be found in Table 3. 30 Table 3. Abbreviations of the API 20S Biochemical Tests“ Abbreviation Reactions Bem bile-esculin hydrolysis Man mannitol utilization Gls p-nitrophenyl-B-D-g1ucopyranosidase Sor sorbitol utilization Ngs p-nitrophenyl-N-acetyl-B-D-glucosaminidase Gly glycerol utilization Npg o-nitrophenyl-B-D-galactosidase Sbs sorbose utilization Ina indoxyl-acetate Raf raffinose utilization Lac lactose utilization Phs disodium p-nitrophenyl-phosphatase Sac sucrose utilization Tre trehalose utilization Arg arginine utilization Hip hippurate utilization Leu L-leucyl-4-methoxy-B-naphthylamidase Ser L-seryl-B-naphthylamidase Pyr L-pyrrolindonyl-B-naphthylamidase Arl L-arginyl-8-naphthylamidase Hem presence or absence of hemolysis is recorded as the final test. “Reprint from API 208 package insert. 31 Analysis of Cellular Fatty Acids by Gas-liquid Chromatography (GLC) The cellular fatty acid composition of the bacterial isolates was examined and the results were compared to the GLC patterns of the reference strains. A limited assessment was made of the possibility of distinguishing between organisms at both species and subspecies levels on the basis of cellular fatty acid composition as revealed by GLC fingerprinting. Preparation of cellular fatty acid methyl esters (FAME) extracts Bacterial isolates were grown on 5% sheep blood agar plates at 35°C for 24-48 h. After incubation, approximately 1.0 ml of sterile distilled water was added to the surface of the plate and the bacterial growth was removed by gentle scraping with a sterile bent glass rod. The cell suspension was placed in a borosilicate screw-cap tube fitted with a teflon-lined cap and saponified (alkaline hydrolyzed) with 4 ml of 5% NaOH in 50% aqueous methanol at 100°C for 30 min. After the saponificate was cooled to room temperature, the pH was lowered to 2.0 with 4 ml of 15% hydrochloric acid- methanol reagent. The FAMEs were heated at 100°C for 15 minutes to liberate and methylate any bound or amide-linked 32 fatty acids which were not released by saponification. After cooling to room temperature, 8 ml of diethylether-hexane (1:1, vol/vol) was added and the contents were mixed by shaking. The phases were allowed to separate by standing 1 to 2 min and the organic phase (top layer) was carefully transferred to a second teflon-lined capped tube. A second extraction with 8 ml of diethylether-hexane solvent removed all but trace amounts of the methyl esters. The combined organic layers containing the FAME were concentrated to a volume of 1.0 ml under nitrogen evaporation. A small amount of disodium sulfate was added to the remaining liquid to absorb traces of moisture from evaporation. The FAME extract was then further evaporated to 0.3 ml and analyzed by GLC. If the analysis was not done within 30 min, samples were stored at -20°C in 1.5 ml vials with teflon-lined capped/silicone septa (Shamrock Glass Co., Linwood, PA). Gas—liquid chromatography of cellular fatty acids The FAME samples were analyzed by gas-liquid chromatography (3500 system, Varian Instrument Group, Walnut Creek, CA). This system includes a gas chromatograph with a flame ionization detector, an electronic integrator and a minicomputer (US-650 Series Data System). The gas chromatograph was equipped with a 08-1 fused silica capillary column (0.32mm diameter, 0.25pm thick film, J & W 33 Scientific Inc.) with cross-linked methylphenyl silicone as the stationary phase. The operating parameters of the instrument that were automatically controlled by the computer software are described in Table 4. Table 4. Operating Parameters of Gas-liquid Chromatograph Column Temperature 150°C for 2 min 150°C to 230°C at 4°C/min for 20 min 230°C for 5 min Flow Rate 1.6 ml/min nitrogen at 12 psi column back pressure and 33.2 cm/sec velocity Sample Size 1 pl of extract Detector Flame Ionization Sensitivity 11f” carbon g/sec The FAMEs were identified by comparing retention times with those of reference standards: FAME standard (Supelco Inc., Bellefonte, PA) and Legionella pneumophila extract characterized by mass spectrometry (Michigan Department of Public Health, Lansing MI). The chromatograms with retention times and peak areas were recorded with the electronic integrator and transferred to a computer for calculation, storage and the final reports. RESULTS The identification of A. haemolyticum and Copynebacterium species involved the determination of the Gram reaction, cellular and colonial characteristics, hemolytic reaction on 5% sheep blood and whether the isolate was capable of utilizing carbohydrates fermentatively, oxidatively or not at all. Once these characteristics had been determined, the organism was tested for additional biochemical reactivities, susceptibility to antimicrobial agents and the analysis of cellular fatty acids. Colonial and Cellular Characteristics Colonies of the Corynebacterium species ranged from pinpoint and grey to large, smooth and butyrous on 5% sheep blood agar. Some isolates of Q. renale and Q. xerosis produced pale yellow to tan pigments; some isolates of Q. xerosis formed rough dry colonies. Four isolates of A. haemolyticum had similar colonial morphologies, however, they grew slower on 5% sheep blood agar than the Corynebacterium species. Corynebacterium species are irregularly shaped Gram positive rods, which were coccoid to slightly curved, with a 34 35 tendency to form clubs and pointed forms. Endospores were not present. Generally, there was a lack of metachromatic granules in the organisms. Microscopically, isolates of A. haemolyticum were slender, irregular bacillary forms with many of the cells arranged at an angle to give V-formations during the first 18 h of growth. As growth proceeded, the bacteria became granular and segmented so that they resembled small and irregular Copynebacterium species. Both the rods and the coccoid cells were gram-positive and asporogenous. Because of the variations present in the cellular morphologies these properties were not used in the diagnostic scheme. Carbohydrate Utilization Reactions Biochemical responses on triple sugar iron (TSI) agar and carbohydrate fermentation patterns were determined. The results are shown in Table 5 and their abbreviations are in Table 6. Isolates were arranged into four groups based upon their reactions on TSI agar. 36 .Hme pom ummoxo mmumuphnonuco Han ou cuppa Enumm Hannnu wed. .m>fiufinom maxcoS .H um>fiuflmom nonpaonfl mo wmmlaa .> uo>fluflmom nonmaomfl mnoa Ho woa .I um>AUflmom umpoaomfl whoa Ho wow .+ undonfi>m. .0 dance Ca poumwa mum mcoflunfl>ounnd. I i + H + + + i I i + 4 s H Ramowmaumv mwmoumx .w n u > + u > + u r u + 4 a m mfimopox .m I i + I > I + i i I + m d m Enumflpum .w u u + u + u i u a n + 4 m m HHumHmcmp .w n a + u n u u u u n + 4 m m Hroamcou .w u r + n u u u u u a u z m e sneeuflpmnu Igmflpoosmma .w u r + i I > n u I n + a m m HHusssHmmAuoces .m u a + n a + + u a u + < m s Husssflmmflpscfla .w u r > u + n n n u I + z z 6 sh ozone .Eswnmuomnoc>poo i i + I + + > + i I 4+ < 4 w 590wu>HOEomn .m Una Hem mum one How an: 05m onq an: H>x 0H0 “Dan Hanan pounce woa .Hwe .oz mmfloomm mmfloomm Sawuouomnoc>poo one EDOHDNHOEoon EsflnopomnocnOMm How mCOHumNHHHuD mucuo>conuoo uo COHHMCAEHmme .m manna 37 Table 6. Abbreviations Used for Carbohydrate Utilizations and Biochemical Reactivity Abbreviation Reaction Gal acidification of galactose Lac acidification of lactose Mal acidification of maltose Man acidification of mannitol Sal acidification of salicin Sta acidification of starch Suc acidification of sucrose Tre acidification of trehalose Xyl acidification of xylose NO3 reduction of nitrate Mot motility MR methyl red VP Voges-Proskauer TSI triple sugar iron agar A acid production on TSI K alkaline reaction on TSI N no response on TSI Mlk litmus milk Ure production of urease Cit utilization of citrate Gel hydrolysis of gelatin Esc hydrolysis of esculin Ind production of indole Cat production of catalase 38 Q. pseudodiphtheriticum was not able to utilize glucose (K/N) while A. haemolyticum, Q. minutissimum-I, Q. renale, Q. striatum and Q. xerosis utilized glucose as well as sucrose by a fermentative mechanism (A/A). Q. pinutissippp- II fermented glucose only (K/A). Corynebacterium group JK was not able to acidify the carbohydrates present in T81 (N/N) . None of the isolates were able to ferment 10% lactose, mannitol, salicin or xylose. A. haemolyticum, Corynebacterium group JK, atypical Q. xerosis and 88% of the Q. striatum were able to ferment galactose. Q. renale-II was able to weakly acidify this sugar. The ability to utilize galactose fermentatively was useful in differentiating Q. xerosis from an atypical isolate of Q. xerosis. The isolate listed as atypical Q. xerosis was biochemically similar to Q. xerosis with the exception of galactose fermentation. The atypical Q. xerosis was able to ferment galactose while Q. xerosis was unable to do so. Q. xerosis was the only species which produced acid from trehalose. Q. minutissimpp showed variable responses to sucrose. Because of this, it was possible to place Q. minutissimum into two subgroups. Q. minutissimum-I was able to ferment sucrose whereas Q. minutissippm-II was unable to do so. Corynebacterium group JK was biochemically inert. It required serum supplementation with 10% rabbit serum for 39 growth. It also did not produce any acids from carbohydrate test media unless 10% rabbit serum was added. Biochemical Reactivity The abbreviations and results obtained from performing biochemical tests for differentiation of the species are presented in Tables 6 and 7 respectively. Isolates of all Corynebacterium species were non—hemolytic and produced catalase whereas isolates of A. haemolyticum were 8- hemolytic and did not elaborate the enzyme. Q. pseudodiphtheriticum reduced nitrate and hydrolyzed urea. Q. striatum and Q. xerosis also reduced nitrate but could not hydrolyze urea while Q. renale could only hydrolyze urea and was unable to reduce nitrate to nitrites or free nitrogen gas. A. haemolyticum, Corynebacterium group JK and Q. minutissimum could neither reduce nitrate nor split urea. Results obtained from the methyl red test allowed the separation of Q. renale into two subspecies. Q. renale-I produced stable acid while Q. renale-II could not produce stable acid. The bacterial isolates were uniformly negative in the following characteristics: motility, hydrolysis of gelatin and hydrolysis of esculin. All isolates were nonreactive in litmus milk after 15 days incubation. Hydrogen sulfide and indole were not produced by any bacteria tested. 4O .m>fluflmom nonmaomfl Mo «meHH .> am>flufimom moumHomH mmoH no on .I uo>w9flmom mwumHouH whoa Ho woo .+ undonfi>mn .o OHDMB Ga cm>wm who m:0fluofl>mhnn<. + u u u I n u u u n + r admondmomo mflwommw .M Di + I i I I i i i i i + i mwmmmwm . UI + I I I i I > I I i + i finwmwmwm . + I I i i I I I + I i I HHIMAmmMM . + I u u u r + r + a n u Huuquum . 0| 0| + u r u n > u r + r + u ssoflmflnumw ummflummmumm . OI + I l I > > > .I I. I I l HHIEDEflmmHHDCHE . + l l l l l + I I l l I Hlaggflmmflpsfififi . OH» + I I I i i i I i I I i Mb msoum Edflnmuowno: uoo i I I I I I I I i I I + ammwwNHOEonn .d n umo gas own How ufio m> ms 62H who cosmoz mflmsaosm: no mowoomm mofloomm Esfluouomnoc moo can EsowMNfloEumm EdwuouomQOCSUAQ mo >uw>wuommm Hmoflaonoowm mo coflumcwfiumpmo .h OHQMB 41 Growth Studies The ability of the Corynebacterium species and A. haemolyticum to grow in the presence of high concentrations of salt and at 25°C and 42°C was determined. The results obtained from salt tolerance testing and temperature studies are presented in Table 8. Except for A. haemolyticum and the atypical Q. xerosis, all bacterial isolates were able to grow in 6% NaCl broth. Both A. haemolyticum and the atypical Q. xerosis were inhibited by concentrations of NaCl greater than 3%. Some Corynebacterium species such as Q. minutissimum, Q. renale and Q. xerosis could survive in the salt broth up to a concentration of 9%. All of the isolates were able to grow at 25°C but only Corynebacterium group JK, Q. renale-II, Q. xerosis and the atypical Q. xerosis were able to survive at 42°C. Supplemental Tests Used for Speciation Additional tests were performed to determine if they would aid in the identification of the A. haemolyticum and Corynebacterium species. 42 Table 8. Growth Studies of Salt Tolerance and Temperature Dependency for Arcanobacterium haemolyticum and Corynebacterium Species Species NaCl Growth at 0% 3% 6% 9% 25°C 42°C A. haemolyticum +‘ -+ — - + - Corynebacterium group JK + + + v + + Q. minutissimum-I + + + + + v Q. minutissimum-II + + + + + v Q. pseudodiph— theriticum + + + v + v Q. renale-I + + + + + v Q. renale-II + + + + + + Q. striatum + + + + + v Q. xerosis + + + + + + Q. xerosis (atypical) + + - - + + aSymbols: +, 90% or more isolates positive; -, 10% or less isolates positive; v, 11-89% of isolates positive. 43 The supplemental tests included production of intermediates of the porphyrin pathway by the use of 6- aminolevulinic acid (6-ALA) and hydrolysis of Tween 80. Susceptibility to two antimicrobial agents, 30 pg vancomycin and 200 pg fosfomycin was also determined. The results of these assays are shown in Table 9. Several different concentrations of 6-ALA could be utilized by Corynebacterium species. These ranged from 0.5 mM to 4.0 mM. However, only 6-ALA at a concentration of 2.0 mM resulted in a clearly differentiated color reaction. A. haemolyticum was the only species which was unable to produce porphyrin intermediates from the 6-ALA medium. The test bacteria had variable abilities to grow and hydrolyze Tween 80 at concentrations from 0.05% to 0.4%. A distinct color change was detected when Tween 80 at a concentration of 0.1% was utilized. Concentrations greater than or equal to 0.2% inhibited bacterial growth and resulted in false negative reactions. A. haemolyticum and Corynebacterium group JK grew well in 0.1% Tween 80 medium and were capable of hydrolyzing Tween 80 after two days of incubation. Q. renale, Q. xerosis and atypical Q. xerosis were not able to hydrolyze Tween 80 after 7 days of incubation. The other isolates of Corynebacterium species gave variable results when tested for the ability to hydrolyze Tween 80. 44 Table 9. Supplemental Biochemical and Susceptibility Tests for Arcanobacterium haemolyticum and Corynebacterium Species Species 6-ALA Tween 80 Fosfo- Vanco- 2 mM 0.1% mycin mycin A. haemolyticum -‘ + Susb Sus Corynebacterium group JK + + Res Sus Q. minutissimum-I + v Res Sus Q. minutissimum-II + v Res Sus Q. pseudodiph- theriticum + v Res Sus Q. renale-I + - Res Sus Q. renale-II + - Res Sus Q. striatum + v Res Sus Q. xerosis + - Res Sus Q. xerosis (atypical) + - Sus Sus aSymbols: +, 90% or more isolates positive; -, 10% or less isolates positive; v, 11-89% of isolates positive. bAbbreviations: Sus, susceptible to antimicrobial agents; Res, resistant to antimicrobial agents. 45 All of the Corynebacterium species were uniformly susceptible to vancomycin and resistant to fosfomycin with the exception of A. haemolyticum and the atypical Q. xerosis. These isolates were susceptible to fosfomycin and vancomycin. API 208 Rapid Identification System In an attempt to find a more rapid method of identifying Corynebacterium species, the API 208 system was evaluated. Reactions were interpretable after 4 h of incubation. The results obtained by the use of API 208 are illustrated in Table 10. Only Q. pseudodiphtheriticum (0000072), Q. striatum (0040172) and Q. xerosis (0040372) had specific profiles. Two profiles were obtained for isolates identified as A. haemolyticum (0216007 and 0612107) and Corynebacterium group JK (0040012 and 4040012) respectively. The two profiles differed only in the ability of some isolates of Corynebacterium group JK to produce B-glucosidase and in the ability of A. haemolyticum to utilize glycerol and sucrose and to produce phosphatase. 46 N s m H v o o H AHmonaumv mHmoumx .w m s m o v o o m wHwoumx .m N s H o e o o m asumHuum .M N m o o q o v m HHImHmcmu .w N H o o o o o m HImHocmH .m N s o o o o o m EsoHuHumnu IcmHooosme .w N b o o m o v H N b o o v o o H N m o o v o o H N H o o o o o H N m o o o o o H m H o o o o o m HHIssaHmmHuocHs .w N h H o v o o H N h H o o o o m N H H o v o o m HIEDEHmmHuDCHE .M m H o o v o a m Mb ozouo m H o o v o o m ssHumuomowcsuoo h o H N H w o N h o o w H N o N EDOHU>HOEGMQ .fl EmmHuduhmummomqummudouBoommnmomquomochnmmmszomozuomwchozamm omumma mHHuoum mom Ham .oz meomom mmHoQO EsHumuomnmc>uou can EooHp>H0Emmn sownouomoocmou< no mmHHmoum mom Hod .oH oHnma 47 Moreover, A. haemolyticum was the only species which produced o-nitrophenyl-B-D-galactosidase. Q. renale could be placed into two subspecies based on production of p- nitrophenyl-B-D-galactosidase, hydrolysis of indoxyl acetate and utilization of leucine. The 15 isolates of Q. minutissimum were a more diverse group. The profiles obtained from these isolates were similar to those obtained from testing Q. renale-I (0000012) and Q. striatum (0040172). Atypical Q. xerosis was the only isolate utilizing raffinose while the other Corynebacterium species were unable to do so. Gas-liquid Chromatography The elution patterns of the fatty acid methyl esters were determined by gas-liquid chromatography for 49 isolates of Corynebacterium species and A. haemolyticum. The results are shown in Table 11. All of the isolates contained major (24-46%) amounts of octadecenoic acid (Gun) and significant (20-46%) amounts of hexadecanoic acid (Cum). They could be differentiated into four GLC groups based on the differences in cellular fatty acids. 48 Table 11. Cellular Fatty Acid Composition of Arcanobacterium haemolyticum and Corynebacterium Species GLC groups/ Cellular Fatty Species Acids 14:0‘ 16:0 17:0 18:2 c18:l 18:0 GLC group A A. haemolyticum 2.5b 31.3 T 16.2 26.5 17.3 Corynebacterium group JK T 31.2 T 29.2 24.0 17.7 GLC group B Q.minutissimum-I T 36.7 T T 33.1 8.1 Q.minutissimum-II T 31.4 T T 40.0 13.9 Q. pseudodiph- theriticum T 37.2 T T 46.2 9.9 Q. renale-I T 37.5 T ND 45.2 7.6 Q. renale-II T 30.6 T T 37.0 21.5 Q. striatum 2.8 46.0 T ND 39.0 4.0 GLC group C Q. xerosis T 20.2 12.7 T 29.4 15.3 GLC group D Q. xerosis (atypical) 5.2 50.5 T 3.6 16.0 8.1 6Number to left of colon refers to number of carbon atoms; number to right refers to number of double bonds. bNumber refers to percentage of total acids; T, less than 2%; ND, not detected. 49 The first GLC group, A, was characterized by the presence of C18:2 (16% and 29% respectively). A. haemolyticum and Corynebacterium group JK were assigned to this group. A. haemolyticum had smaller amounts of cis-Cm:1 than those of Corynebacterium group JK. The next group of isolates, GLC group B, had a fatty acid profile similar to that of group A, except that C15,:2 was absent and the relative concentration of C15,:1 was greater than that found in group A isolates. GLC group B included Q. minutissimum, Q. seudodi htheriticum, Q. renale and Q. striatum. Q. minutissimum and Q. renale could be further divided into two subspecies by the relative percentages of me. Subspecies I had higher levels of me than those of the subspecies II. GLC group C consisted of only one species Q. xerosis. This group was characterized by the presence of C1,:o and relatively small amounts of Cum. Isolate A22 was assigned to the additional GLC group D, because of the relatively large amounts (51%) of cum and low levels (16%) of cis- Cm”. By integrating the results of selected biochemical tests (Table 12) with the fatty acid results, identification of these organisms could be made to the species level. There were good agreement between reference strains and tested isolates in both biochemical characteristics and fatty acid components (Tables 5, 7 and 11). 50 Table 12. Biochemical Tests for Separation of A. haemolyticum and Corynebacterium Species Within a GLC Group GLC groups/ Species Biochemical Reactions GLC group A A. haemolyticum Corynebacterium group JK GLC group B Q. minutissimum-I Q. minutissimum-II Q. pseudodiph- theriticum Q. renale-I Q. renale-II Q. striatum GLC group C Q. xerosis GLC group D Q. xerosis (atypical) Nit‘ Ure Cat 6 -ALA B-hemolvsis -b _. _ .. + + + + + — - + + + — - + + + - + - + + — + - + + - + - + + - aAbbreviations are listed in Table 6. bSymbols: +, 90% or more isolates positive; -, 10% or less isolates positive. DIBCUSSION Most workers agree that the standard methods for the identification of the genus Corynebacterium are not practical for some of the species within that genus. The identification of Corynebacterium species is hindered because: (i) pleomorphism offers little as an aid to species recognition or grouping (ii) great biochemical similarities exit between most of the Corynebacterium species and (iii) some of the Corynebacterium such as group JK are relatively biochemical inert. Several identification schemes have been proposed in which properties such as sugar fermentation, nitrate reduction, lipase activity and urease production are used [39, 62, 64]. Marples [39] suggested that the last three characteristics were closely linked to the activity of the organisms in vivo and are therefore more applicable than other characteristics to an identification scheme. However, none of the proposed schemes proved satisfactory when used for the study of large numbers of isolates because in each case only a small proportion of the isolates could be identified. Another diagnostic scheme [62], using catalase, motility, nitrate reduction, urease, glucose, maltose and sucrose reactions combined with the reactions obtained from 51 52 triple sugar iron agar, morphological and growth characteristics could speciate Corynebacterium species only in an arbitrary way. For example, in the differentiation between Q. xerosis and Q. striatum, Smith [62] indicated that Q. xerosis was able to ferment glucose, maltose and sucrose and was unable to reduce nitrate or split urea. These characteristics were used to separate Q. xerosis from other closely related human species such as Q. striatum, which had similar characteristics but usually was not capable of fermenting maltose. However, exceptions were found in some isolates of Q. xerosis which could not produce acid from maltose and therefore were unable to be distinguished from Q. striatum. Because of this, these bacteria could not be identified with confidence. The differentiation of the Corynebacterium species is often based upon a laboratory's experience with the observation of gram-stain morphologies and growth responses on conventional media [13]. Q. striatum is a slender rod and grew slower than Q. xerosis. However, these criteria are disputed and are not well-accepted. Therefore, a simpler but more comprehensive range of tests was developed to produce a useful and accurate diagnostic scheme for the identification and speciation Corynebacterium. These included several supplemental reactions (production of porphyrin intermediates from 6-ALA medium, hydrolysis of Tween 80 and susceptibility to selected antimicrobial agents), the use of 53 a commercial test kit and the analysis of cellular fatty acids as additional tests. The results obtained from carbohydrate fermentation testing often did not agree with published standards [13]. Results obtained from fermentation of salicin and starch were particularly in opposition to published findings [13] which found that Q. xerosis was able to ferment salicin and Q. striatum was able to ferment starch. The findings from this study showed that all isolates were salicin-negative and had variable abilities to ferment starch. This could be explained by the fact that various indicators such as phenol red, iodine and bromocresol purple, or different concentrations of carbohydrates were used. For example, Q. striatum grows slowly and may not produce sufficient amounts of acid to change a less sensitive indicator. We tested fermentation of both 1% and 2% starch and found that 2% gave more definitive results. In addition, different indicators (phenol red, iodine and bromocresol purple) were used in the starch fermentation test. We found that bromocresol purple gave a distinct color change after 24-48 h of incubation while the other indicators required more than 48 h. Fermentation of galactose was useful only for differentiating A. haemolyticum and Corynebacterium group JK [28]. No further information has been documented on the use of galactose for the identification of the other Corynebacterium species. In this study, 88% of the Q. 54 striatum and 100% of the Q. renale-II fermented galactose. The results obtained from fermentation of carbohydrates suggested that the carbohydrate acidification patterns could not be solely relied upon to separate closely related species, such as Q. striatum and Q. xerosis. The results from salt tolerance and temperature dependency studies are too variable to be used reliably for the differentiation of the Corynebacterium species. Coyle gt AA. reported that if skin scrapings from Q. minutissimum infected lesions are cultured in a medium containing porphyrin precursor, the colonies will fluoresce coral-red when examined with a Wood's lamp [13]. However, little attention has been paid to the other Corynebacterium species. This test is based on the biosynthetic pathway of heme: 6-aminolevulinic acid porphobilinogen porphyrins(fluoresces) heme Many more isolates produced porphyrins than were expected. In addition to Q. minutissimum, Q. pseudodiphtheriticum, Q. renale, Q. striatum and Q. xerosis, showed the typical coral-red fluorescence with the various concentrations (0.5, 1.0, 2.0 and 4.0 mM) of 6-ALA tested. This finding confirmed that more than one species is capable 55 of producing porphyrins and that additional variations in biochemical activity may be found among the fluorescent isolates. A. haemolyticum was the only organism which was not able to utilize 6—ALA. In addition, only 2.0 mM 6-ALA showed the best color differentiation after 4 h of incubation. Esterase activity was determined by observing the ability of the bacterial isolates to hydrolyze various concentrations (0.05, 0.1, 0.2 and 0.4%) of Tween 80 (polyoxyethylene derivative of sorbitan monooleate). The results indicated that 0.1% Tween 80 provided the clearest color differentiation. Tween 80 at 0.05% was too dilute to form the neutral red-Tween compound, therefore the test was difficult to interpret. Tween 80 compounds inhibited growth at concentrations greater than or equal to 0.2%. Inhibition of growth resulted in false negative results. The time required for the appearance of a pink color is 24-48 h. A. haemolyticum (100%), Corynebacterium group JK (100%) and 14% of the Q. minutissimum-I, 25% of the Q. minutissimum-II, 17% of the Q. pseudodiphtheriticum and 11% of the Q. striatum were able to hydrolyze 0.1% Tween 80 while the other Corynebacterium species were unable to do so. Because of this, utilization of 2.0 mM 6—ALA and hydrolysis of 0.1% Tween 80 can be used as additional tests for the confirmatory identification of slowly growing isolates such as A. haemolyticum and biochemically inert isolates such as 56 Corynebacterium group JK. Fosfomycin is a broad-spectrum antibiotic that is particularly effective against gram-positive rods (such as Listeria, Rhodococcus and Erysipelothrix) and gram-negative rods. Glucose-6-phosphate must be added to fosfomycin disk to potentiate its activity [23]. The results from this study showed that all isolates except A. haemolyticum and the atypical Q. xerosis were resistant to fosfomycin. Vancomycin is effective against most isolates of Staphylococcus and Streptococcus. The microorganisms used in this study were susceptible to vancomycin. The use of these two unique antimicrobial patterns will aid in the differentiation and speciation of A. haemolyticum and the Corynebacterium species. In attempts to shorten the incubation period and to search for more useful biochemical characteristics for the identification of Corynebacterium species, the API 208 system was also tested in this investigation. The API 208 system was less helpful than expected because Q. minutissimum yielded many codes. However, the other test organisms (Q. pseudodiphtheriticum, Q. renale-I, Q. striatum and Q. xerosis) had unique profiles which were also associated with the type strains. The production of o- nitrophenyl-B-D-galactosidase and the presence of B- hemolysis in A. haemolyticum readily distinguished it from the other Corynebacterium species. Corynebacterium group JK 57 had two specific codes (0040012 and 4040012) which were different from the other Corynebacterium species. Moreover, the API 208 strip was easy to use and interpret because it did not required serum supplement and gave a clear color reaction after 4 h of incubation. In addition, test results were corroborated by agreement with conventional biochemical methods (90% consistency) [28, 29, 51]. The investigation concluded that carbohydrate fermentation profiles of API 20S can complement the conventional identification methods as a means to facilitate the identification and speciation of some isolates of the Corynebacterium species. Although difficulties were encountered in the identification of Q. minutissimum with the API 208 system, when the API results were used in conjunction with other tests (such as catalase, urease and nitrate) the problem could be resolved. In addition, the API results of Q. minutissimum might suggest that Q. minutissimum should be further divided into several subspecies according to the variations in biochemical metabolisms. Furthermore, the difficulties which exist in differentiation of Q. striatum and Q. xerosis were resolved by the addition of the trehalose fermentation test. Therefore, the identification and speciation scheme can be simplified by using a few selective biochemical reactions without any special instrumentation (Figure 1). 58 Figure 1. Identification Scheme for Selected Corynebacterium Species. Catalase-positive Corynebacterium species - glucose or TSI + 7 I I Q. pseudodiphtheriticum Corynebacterium group JK Q. minutissimum Q. renale Q. striatum C.xerosis. I - Nitrate + kfikHO group JK Q. striatum minutissimum Q. xerosis renale | I - Trehalose + Urea + , I group JK Q. renale I I 1 Q.striatum C.xerosis Q. Q. minutissimum I I Galactose - Serum required - Methyl red + I I I Q. minutissimum I - Sucrose + 1 Q. renale—II Q. renale-I I Q. group JK + I I Q. minutissimum—II I I Q. minutissimum-I 59 Generally, if the biochemical test results obtained with unknown microorganisms match those of the type strains, then the gas-liquid chromatographic fatty acid (GLCFA) profiles should also be the same. Basically, the overall GLCFA profiles of Corynebacterium species were consistent with those of the type strains (Table 11) which contained large amounts of Cm,o and cis-Cmn. The GLCFA results indicated that the relative percentages of the fatty acid are useful in dividing the 44 isolates of Corynebacterium species, 4 isolates of A. haemolyticum and one atypical isolate of Q. xerosis into four GLC groups. GLC group A was characterized by the presence of high levels of Cum, cm2 and cis-Cmu. A. haemolyticum and Corynebacterium group JK were included in GLC group A. These two species could be easily differentiated because A. haemolyticum was B-hemolytic and did not produce catalase or utilize 6-ALA (Table 12). GLC group B was a diverse group because this group represented at least four different species (Q. minutissimum, Q. seudodi htheriticum, Q. renale and Q. striatum). By integrating the results of selected biochemical tests (Table 12) with the fatty acid results, identification of these organisms could be made to the species level. For example, Q. pseudodiphtheriticum could be easily speciated from the other isolates because it did not utilize glucose (Table 5). The other three species could be 60 differentiated by their ability to reduce nitrate and hydrolyze urea (Table 12). Q. renale had two GLCFA profiles. The GLCFA profiles of Q. renale-I matched those of the type strain KC 1366, but not those of Q. renale-II. Q. renale-I contained smaller quantities (7.6%) of Cm:o and was methyl red-positive while Q. renale-II contained higher levels (21.5%) of C13:o and was methyl red-negative. However, no published report has suggested the separation of Q. renale into two subspecies and no reference strain of Q. renale-II is available to confirm our finding. A similar finding in Q. minutissimum was that the GLCFA profiles and biochemical reactions of Q. minutissimum-I (sucrose-positive) were consistent with those of the type strain KC 1381 while the GLCFA profiles and biochemical reactions of Q. minutissimum-II (sucrose- negative) were consistent with those of the type strain KC 1291. Therefore, both Q. renale and Q. minutissimum could be divided into two subspecies based upon their biochemical reactions and GLCFA profiles. Four isolates included in GLC group C were identified by conventional testing as Q. xerosis. A previous study [2] indicated that the fatty acid composition of this species was similar to that of the other Corynebacterium species. However, the slightly higher level of C1,:o was observed and could be an important characteristic for the distinction between Q. xerosis and the other Corynebacterium species. 61 The fatty acid profiles of the CDC cultures were essentially identical to those of the data provided by Centers for Disease Control (D. G. Hollis, personal communication). This enabled the use of GLCFA as a simple, accurate, reproducible and reliable method to identify and speciate the Corynebacterium. A. haemolyticum has many biochemical similarities to the Corynebacterium species except the lack of catalase production and 6-ALA utilization. In this project, the results obtained from supplemental tests and the API 208 confirm that A. haemolyticum differs greatly from the other Corynebacterium species and should be removed from this genus. Although the GLCFA profile of A. haemolyticum cannot be distinguished from that of Corynebacterium group JK, biochemically, these bacteria differ in the presence of B- hemolysis, production of catalase and utilization of 6—ALA. Some isolates were originally identified as "Corynebacterium species unable to speciate" or "possible Corynebacterium species" because most of the biochemical tests match those of the reference strains. However, the GLCFA profiles were not consistent with those of the type strains. For examples, isolate A22 was originally identified as Q. xerosis by conventional biochemical methods. It was reidentified as a variant of Q. xerosis after the analyses of cellular fatty acids and the API 208 system were performed. Because of the presence of relatively small 62 amounts of C1,,o and the utilization of raffinose, isolate A22 was assigned to an additional GLC group D as a variant isolate. However, due to the limitations of the present diagnostic technique, no conclusion could be made as to which species isolate A22 truly belongs. Therefore, further investigations of the biological features of Corynebacterium species are necessary. Isolate 6286 could not be confidently identified as Q. minutissimum because it was the only isolate of Q. minutissimum which was unable to ferment maltose. After the GLCFA results had been interpreted, G286 was confirmed to be Q. minutissimum. These GLCFA results demonstrate the usefulness of cellular fatty acid analysis for the differentiation of suspected Corynebacterium species. Isolate A499 was identified as Q. xerosis because of biochemical and morphological similarities to the type strain. However, the API 208 code was the same as that of Q. striatum and there was an absence of C1,:o by GLCFA. These two additional tests confirmed that this isolate should be identified as Q. striatum. Some isolates with weak metabolic activity (A. haemolyticum) or which required special growth conditions (serum requirement for Corynebacterium group JK) could be differentiated with the addition of supplemental tests, such as utilization of 2.0 mM 6-ALA, hydrolysis of 0.1% Tween 80, fermentation of galactose and selected tests of 63 antimicrobial susceptibility. However, GLCFA analysis readily distinguished A. haemolyticum and Corynebacterium group JK from the other Corynebacterium species, thus eliminating the need for these additional tests. CONCLUSION During the past several years, techniques for the isolation of Corynebacterium species have improved and more of these bacteria are being detected in clinical sources. Corynebacteria are difficult to identify to the species level by conventional methods since there are inconclusive biochemical reactions. Because the Corynebacterium cannot always be confidently identified to the species level, many types of biochemical tests have been used for differentiation. This study describes the use of cellular fatty acid analysis, in addition to biochemical tests, and their ability to differentiate Corynebacterium species. The proposed identification scheme utilizes the following parameters: presence of hemolysis on 5% sheep blood agar; Gram stain morphology; growth and metabolic response on triple sugar iron agar; production of catalase; motility; utilization of carbohydrate (galactose, glucose, maltose, sucrose and trehalose); reduction of nitrate; production of urease; utilization of 2.0 mM 6-ALA; hydrolysis of 0.1 % Tween 80; susceptibility to selected antimicrobial agents; rapid identification by the use of API 208 system and analysis of GLCFA. Moreover, the identification and speciation scheme can 64 65 be simplified either by (i) using a few selected biochemical reactions such as reduction of nitrate, production of catalase, fermentations of sucrose and trehalose (Figure 1) or (ii) analyzing GLCFA profiles which can divide Corynebacterium species into four GLC groups and use this information in combination with reduction of nitrate, production of catalase and urease (Table 12) or (iii) the use of a rapid identification kit. Although the API 208 system alone cannot provide a definitive identification of Corynebacterium species, it can reduce the incubation period of some biochemical tests (such as fermentation of carbohydrates) and may replace some of the conventional biochemical methods. The results of API 208 also suggest that Q. renale and Q. minutissimum should be divided into subspecies. These findings are supported by GLCFA results. Further exploration will be helpful to characterize these heterogeneous groups. LITERATURE CITED Athalye, M., w. C. Noble, A. I. Mallet, and D. E. Minnikin. 1984. Gas chromatography-mass spectrometry of mycolic acids as a tool in the identification of medically important coryneform bacteria. J. Gen. Microbiol. 130:513-519. Athalye, M., w. C. Noble, and D. E. Minnikin. 1985. Analysis of cellular fatty acids by gas- chromatography as a tool in the identification of medically important coryneform bacteria. J. Appl. Bacteriol. 58:507-512. Barksdale, L. 1970. Corynebacterium diphtheriae and its relatives. Bacteriol. Rev. 34:378-422. Berger, 8. A., A. Gorea, J. Stadler, M. Dan, and M. Zlberman. 1984. Recurrent breast abscesses caused by Corynebacterium minutissimum. J. Clin. Microbiol. 20:1219-1220. Brooks, J. B., and W. E. C. Moore. 1969. Gas chromatographic analysis of amines and other compounds produced by several species of Clostridium. Can. J. Microbiol. 15:1433-1437. Brooks, J. B. 1983. Gas-liquid chromatography as an aid in rapid diagnosis by selective detection of chemical changes in body fluids, p. 313-334. In I. D. Coonrod, L. J. Kunz, and M. J. Ferrano (ed.), The direct detection of microorganisms in clinical samples. Academic Press, Inc., NY. Butler, w. R., D. G. Ahearn, and J. O. Kilburn. 1986. High-performance liquid chromatography of mycolic acids as a tool in the identification of Corynebacterium, Nocardia, Rhodococcus and Mycobacterium species. J. Clin. Microbiol. 23:182- 185. Butler, w. R., and J. O. Kilburn. 1988. Identification of major slowly growing pathogenic Mycobacteria and Mycobacterium qordonae by high-performance liquid-chromatography of their mycolic acids. J. Clin. Microbiol. 26:50-53. Chevalier, J., M. T. Pommier, A. Cremieux, and J. Michel. 1988. Influence of Tween 80 on the mycolic 66 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 67 acid composition of three cutaneous corynebacteria. J. Gen. Microbiol. 134:2457-2461. Collins, M. D., D. Jones ,and G. M. Schofield. 1982. Reclassification of Corynebacterium in the genus Arcanobacterium as Arcanobacterium haemolyticum. J. Gen. Microbiol. 128:1279-1281. Collins, M. D., M. Goodfellow, and D. E. Minnikin. 1982. Fatty acid composition of some mycolic acid- containing coryneform bacteria. J. Gen. Microbiol. 128:2503-2509. Coloe, P. J. 1978. Headspace gas liquid chromatography for rapid detection of Escherichia coli and Proteus mirabilis in urine. J. Clin. Pathol. 31:365-369. Coyle, M. B., D. G. Hollis, and N. B. Groman. 1985. Corynebacterium species and other coryneform organisms, p.193-204. In E. H. Lennette, A. Balows, W. J. Hausler,Jr., and H. J. Shadomy (ed.), Manual of clinical microbiology, 4th ed. American Society for Microbiology, Washington, DC. Cummins, C. S. 1971. Cell wall composition in Corynebacterium bovis and some other corynebacteria. J. Bacteriol. 105:1227-1228. Damato, R. F., and L. Hochstein. 1982. Evaluation of a rapid inoculum preparation method for agar disk diffusion susceptibility test. J. Clin. Microbiol. 15:282-285. Dan, M., I. Somer, and B. Knobel. 1988. Cutaneous manifestations of infection with Corynebacterium group JK. Rev. Infect. Dis. 10:1024-1027. Dees, S. B., and o. W. Moss. 1978. Identification of Achromobacter species by cellular fatty acids and by production of keto acids. J. Clin. Microbiol. 8:61-66. Dees, S. B., C. W. Moss, D. G. Hollis, and R. E. Weaver. 1986. Chemical characterization of Flavobacterium odoratum, Flavobacterium breve, and Flavobacterium-like groups IIe, IIh, and IIf. J. Clin. Microbiol. 23:267-273. Donaghy, M., and J. Cohen. 1984. Pulmonary infection with Corynebacterium hofmannii complicating systemic lupus erythematosus. J. Infect. Dis. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 68 147:962. Evans, N. M. 1968. The classification of aerobic diphtheroids from human skin. Br. J. Dermatol. 80:81-83. Edman, D. C., and J. B. Brooks. 1983. Gas-liquid chromatography-frequency pulsed-modulated electron-capture detection in the diagnosis of infectious diseases. J. Chromatogr. 274:1-25. Finegold, S. M., and E. J. Baron. 1986. p. 621-622. Bailey and Scott's diagnostic microbiology, 7th ed. C. V. Mosby Co., St. Louis, MO. Forsgren, A., and M. Walder. 1983. Antimicrobial activity of fosfomycin in vitro. J. Antimicrob. Chemother. 11:467-473. French, G. L., C. Y. Chan, and S. W. Cheung. 1987. Diagnosis of pulmonary tuberculosis by detection of tuberculostearic acid in sputum by using gas chromatography-mass spectrometry with selected ion monitoring. J. Infect. Dis. 156:356-362. Grasmick, A. B., and D. A. Bruckner. 1987. Comparison of rapid identification method and conventional substrates for identification of Corynebacterium group JK isolates. J. Clin. Microbiol. 25:1111- 1112. Gronemeyer, P. 8., A. S. Weissfeld, and A. c. Sonnenwirth. 1980. Corynebacterium group JK bacterial infection in a patient with ax: epicardial pacemaker. Am. J. Clin. Pathcfld 74:838- 842. Guerrant, G. 0., M. A. Lambert, and C. wu Moss. 1982. Analysis of short-chain acids from anaerobic bacteria by high-performance liquid chromatography. J. Clin. Microbiol. 16:355-360. Hollis, D. G., and R. E. Weaver. 1981. Gram-positive organisms: a guide to identification. Special Bacteriology Laboratory, Centers for Disease Control, Atlanta, GA. Hwang, M. N., and G. M. Ederer. 1975. Rapid hippurate hydrolysis method for presumptive identification of group B streptococci. J. Clin. Microbiol. 1:114-115. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 69 Jackman, P. J. H. 1982. Classification of Corynebacterium species from axillary skin by numerical analysis of electrophoretic protein patterns. J. Med. Microbiol. 15:485-492. Jones, D., and M. D. Collins. Irregular, nonsporing gram-positive rods. 1986, p.1261-1276. In P. H. Sneath, N. S. Mair, M. E. Sharpe and J. G. Holt. (ed.), Bergey's Manual of Systematic Bacteriology, vol. 2, Waverly Press, Inc. Baltimore, MD. Kahan, F. M., J. S. Kahan, D. J. Cassidy, and H. Kropp. 1974. The mechanism of action of phosphomycin. Ann. N. Y. Acad. Sci. 235:364-367. Kelly, M. C., I. D. Smith, R. J. Anstey, and J. H. Thornley. 1984. Rapid identification of antibiotic-resistant corynebacteria with API 20S system. J. Clin. Microbiol. 19:245-247. Kilian, M. 1984. Haemophilus species, p.390-391. In E. H. Lennette, A. Balows, W. J. Hausler, Jr., and H. J. Shadomy (ed.), Manual of clinical microbiology, 4th ed. American Society for Microbiology, Washington, DC. Lambert, M. A., and C. W. Moss. 1980. Production of p- hydroxyhydrocinnamic acid from tyrosine by Peptostreptococcus anaerobius. J. Clin. Microbiol. 12:291-293. Lambert, M. A., C. M. Patton, T. J. Barrett, and C. W. Moss. 1987. Differentiation of Campylobacter and Campylobacter-like organisms by cellular fatty acid composition. J. Clin. Microbiol. 25:706-713. Lipsky, B. A., A. C. Goldberger, L. S. Tompkins, and J. J. Plorde. 1982. Infections caused by nondiphtheria Corynebacteria. Rev. Infect. Dis. 4:1220-1235. MacFaddin J. F. 1980. Biochemical tests for identification of medical bacteria, 2nd ed. Waverly Press, Inc., Baltimore, MD. Marples, R. R. 1969. Diphtheroids of normal human skin. Br. J. Dermatol. 81(suppl. 1): 47-53. McGinley, K. J., J. N. Labows, J. M. Zeckman, R. M. Nordstrom, G. P. Webster, and J. J. Leyden. 1985. Analysis of cellular components, biochemical 41. 42. 43. 44. 45. 46. 47. 48. 49. 70 reactions, and habitat of human cutaneous lipophilic diphtheroids. J. Invest. Dermatol. 85:374-377. McGinley, K. J., J. N. Labows, J. M. Zeokman, K. M. Nordstrom, G. F. Webster, and J. J. Leyden. 1985. Pathogenic JK group Corynebacteria and their similarity to human cutaneous lipophilic diphtheroids. J. Infect. Dis. 152:801-806. Merrick-Gass, M. T. 1986. Gas chromatography in bacterial identification. Am. Chem. Prod. Rev. 12:8-15. Meyer, D., and P. Cohan. 1986. Designing new compounds with PC databases. Am. Chem. Prod. Rev. 12:16. Moss, C. W. 1978. New methodology for the identification of nonfermenters: gas-liquid chromatographic chemotaxonomy, p.171-203. In Gilardi G. L. (ed.), Glucose nonfermenting gram- negative bacteria in clinical microbiology. CRC Press, Inc., West Palm Beach, FL. Moss, C. W., C. L. Hatheway, M. A. Lambert, and L. M. McCroskey. 1980. Production of phenylacetic and hydroxyphenylacetic acids by Clostridium botulinum type. J. Clin. Microbiol. 11:743-745. Moss, C. W. 1980. Gas-liquid chromatography of bacterial fatty acids with a fused-silica capillary column. J. Clin. Microbiol. 12:127-130. Moss, C. W. 1981. Gas-liquid chromatography as an analytical tool in microbiology. J. Chromatogr. 203:337-347. Moss, C. W., P. L. Wallace, D. G. Hollis, and R. E. Weaver. 1988. Cultural and chemical characterization of CDC groups EO-2, M-5, and M-6, Moraxella species, Oligella urethralis, Acinetobacter species, and Psychrobacter immobilis. J. Clin. Microbiol. 26:484-492. Musial, C. E. 1987. Tuberculosis and other mycobacterioses. p.564-565 and p.828-829. In B. B. Wentworth, V. S. Baselski, C. P. McArthur, C. A. Needham, G. V. Doern, J. M. Miller, and D. L. Sewell (ed.), Diagnostic procedures for bacterial infections, 7th ed. American Public Health Association, Inc., Washington, DC. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 71 National Committee for Clinical Laboratory Standards. 1988. Performance standards for antimicrobial disk susceptibility tests. M2-T4. National Committee for Clinical Laboratory Standards Susceptibility Testing Supplement Program, Villanova, PA. Piot, P. 1980. Gardnerella va inalis, p.876. In E. H. Lennette, A. Balows, W. J. Hausler, Jr., and H. J. Shadomy (ed.), Manual of clinical microbiology, 4th ed. American Society for Microbiology, Washington, DC. Pitcher, D. G. 1977. Rapid identification of cell-wall components as a guide to the classification of aerobic coryneform bacteria from human skin. J. Med. Microbiol. 10:439-445. Pitcher, D. G. 1978. Aerobic cutaneous coryneforms: recent taxonomic findings. Br. J. Dermatol. 98:363-370. Quinn, J. P., P. M. Arnow, D. Weil, and J. Rosenbluth. 1984. Outbreak of JK diphtheroid infections associated with environmental contamination. J. Clin. Microbiol. 19:668-671. Riebel, W., N. Frantz , D. Adelstein, and P. J. Spagnuolo. 1986, Corynebacterium group JK: A cause of nosocomial device-related infection. Rev. Infect. Dis. 8:42-49. Riley, P. 8., D. G. Hollis, G. B. Utter, R. E. Weaver, and c. N. Baker. 1979. Characterization and identification of 95 diphtheroid (group JK) cultures isolated from clinical specimens. J. Clin. Microbiol. 9:418-424. Rossau, R., A. Van Landschoot, W. Mannheim, and J. De Ley. 1986. Inter- and intrageneric similarities of ribosomal ribonucleic acid cistrons of the Neisseriaceae. Int. J. Syst. Bacteriol. 36:323- 332. Samies, J. H., B. N. Hathaway, R. M. Echols, J. M. Veazey, and V. A. Pilon. 1986. Lung abscess due to Corynebacterium egg'. Am. J. Med. 80:685-688. Sarkany, 1., D. Taplin, and H. Blank. 1961. The etiology and treatment of erythrasma. J. Invest. Dermatol. 37:283-290. Sherris, J. C., and T. Clyde. 1984. Laboratory tests in 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 72 chemotherapy, p.959-987. In E. H. Lennette, A. Balows, W. J. Hausler,Jr., and H. J. Shadomy (ed.), Manual of clinical microbiology, 4th ed. American Society for Microbiology, Washington, DC. Slifkin, M., G. u. Gil, and c. Engwall. 1986. Rapid identification of group JK and other corynebacteria with the minitek system. J. Clin. Microbiol. 24:177-180. Smith, R. F. 1969. Characterization of human cutaneous lipophilic diphtheroids. J. Gen. Microbiol. 55:433-443. Snell, J. J. S., and S. P. Lapage. 1976. Transfer of some saccharolytic Moraxella species to Kingella Henriksen and Bovre 1976, with descriptions of Kingella indologenes sp. nov. and Kingella denitrificans sp. nov. Int. Syst. Bacteriol. 26:451-458. Somerville, D. A. 1973. A taxonomic scheme for aerobic diphtheroids from human skin. J. Med. Microbiol. 6:215-224. Soucek, A., A. Souckova, M. Mara, and F. Patocka. 1972. Inhibition of the hemolytic action of a and B lysins of S. pyogenes by Q. haemol ticum, Q. ovis and Q. ulcerans. Toxicon. 10:501-509. Sutter, V. L., D. M. Citron, and S. M. Finegold. 1980. Wadsworth anaerobic bacteriology manual, 3rd ed. C. V. Mosby Co., St. Louis. MO. Suzuki, K., T. Kaneko, and K. Komagata. 1981. Deoxyribonucleic acid homologies among coryneform bacteria. Int. J. Sys. Bacteriol. 31:131-138. Telander, B., R. Lerner, Palmblad J., and 0. Ringertz. 1988. Corynebacterium group JK in a hematological ward: Infections, colonization and environmental contamination. Scand. J. Infect. Dis. 20:55-61. Thompson, J. 8., D. R. Gates-Davis, and C. T. D. Yong. 1983. Rapid microbiochemical identification of Corynebacterium diphtheriae and other medically important corynebacteria. J. Clin. Microbiol. 18:926-929. Valenstein, P., A. Klein, C. Ballow, and W. Greene. 1988. Corynebacterium xerosis septic arthritis. 71. 72. 73. 73 Wallace, P. L., D. G. Hollis, R. E. Weaver, and C. W. Moss. 1988. Cellular fatty acid composition of Kingella species, Cardiobacterium hominis, and Eikenella corrodens. J. Clin. Microbiol. 26:1592- 1594. Wiohmann, 8., C. H. Wirsing von Koenig , E. Becker- Boost, and H. Finger. 1984. Isolation of Cogynebacterium group JK from clinical specimens with a semiselective medium. J. Clin. Microbiol. 19:204-206. Wirsing, von Koenig C. H., T. Krech, H. Finger, and M. Bergmann. 1988. Use of fosfomycin disks for isolation of diphtheroids. Eur. J. Clin. Microbiol. Infect. Dis. 7:190-193. : "TIEIfiI/EIII'IIIIIII