Willi!Illlllllllllll’llillllill 3 1293 00877 3370 III? This is to certify that the dissertation entitled GENETIC BASIS FOR THE HIGH-FREQUENCY MUTATION TO NONPIGMENTATION IN YERSINIA PESTIS AND ANALYSIS OF THE PIGMENTATION PHENOTYPE presented by Thomas Stephen Lucier has been accepted towards fulfillment of the requirements for Ph . D . degree in Microbiology WW Major professor Date 5/14/92 MS U is an Affirmative Action/Equal Opportunity Institution 0—12771 LIBRARY Michigan State University PLACE IN RETURN BOX to remove this checkout from your record. TO AVOD FINES return on or before due due. DATE DUE DATE DUE DATE DUE MSU is An Affirmative ActiorVEquei Opportunity Institution cMMmG-nt GENETIC BASIS FOR THE HIGH-FREQUENCY MUTATION TO NONPIGMENTATION IN YERSINIA PESTIS AND ANALYSIS OF THE PIGMENTATION PHENOTYPE BY Thomas Stephen Lucier A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Microbiology and Public Health 1992 -32/x d73 ABSTRACT GENETIC BASIS FOR THE HIGH-FREQUENCY MUTATION TO NONPIGMENTATION IN YERSINIA PESTIS AND ANALYSIS OF THE PIGMENTATION PHENOTYPE BY Thomas S. Lucier The pigmentation virulence determinant of Yersinia pestis encompasses the ability to bind hemin at 26°C, pesticin sensitivity, and a 'possible :mechanism for iron acquisition. Their role in virulence and the genetic basis for the high-frequency mutation to nonpigmentation with concomitant loss of these traits remain obscure. We used pulsed-field gel electrophoresis (PFGE) to demonstrate that this mutation is due to a specific chromosomal deletion of approximately 100 kb. The deletion had not occurred, but was able to do so, in a rare mutant capable of hemin storage, but not Pgm associated iron transport suggesting a genetic, but not functional linkage of these traits. Total genome size was found to be 4397.9 +/- 134.6 kb, and intraspecific variation in macrorestriction patterns confirmed the existance of three distinct biotypes. Studies of the accumulation of labelled inorganic and hemin-bound ii iron showed hemin binding to be mediated in! the» outer membrane of Pgm+ cells at 260C, and to be a storage function distinct from hemin uptake. Inorganic, but not hemin-bound iron was preferentially stored on a bacterioferritin-like cytoplasmic protein. Uptake of inorganic and hemin-bound iron were non-competitive and involved different mechanisms. Gradiant plates were developed in which cells were exposed to increasing concentrations of various iron chelators to better define the Pgm asociated iron uptake system. Results indicated this was an inducible, non-specific system which mediates sensitivity to the bacteriocin pesticin in all- pathogenic Yersinia species. Analysis of protein expression using SDS-PAGE and two-dimensional gel electrophoresis identified several potential components of this system. In if; pestis, but not enteropathogenic Yersinia species the mutation ix) Pstr was correlated with loss of expression of three of these peptides. Temperature profoundly influenced the ability of X; pestis to obtain iron in a Pgm independent manner. This correlated with expression of proteins in the outer membrane and periplasm. iii ACKNOWLEDGMENTS I wish to ‘thank the members of nu! committee; Dr. Kathyrn Brooks, Dr. Paul Coussens, and Dr. Loren Snyder for all of their efforts and advice throughout this project. I would also like to thank Dr. Susan Conrad, Dr. Jerry Dodson, Dr. Robert Hausinger and Dr. Coleman Wolk for making supplies and equipment available to me to perform this work.~ Special thanks to Dr. John Gerlach for his help with pulsed- field gel electrophoresis. I would also like to thank Dr. Robert Brubaker for teaching me so much about how to be a scientist, and Janet Fowler for her invaluable help. Finally I gratefully acknowledge the patience and understanding of my wife, Julie without whose love and support this would not have been possible. iv TABLE OF CONTENTS LIST OF TABLES O O O O O O O O O O O O O O O 0 O O O 0 LIST OF FIGURES O O O O O O O O O O O O O O O O O O O 0 LITERATURE REVIEW . . . . . . . . . . . . . . . . . . . Virulence Factors of Yersinia pestis . . . . . . . Iron Uptake in Yersiniae . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . CHAPTER 1 ARTICLE: Determination of genome size, macrorestriction pattern polymorphism, and nonpigmentation-specific deletion in Yersinia pestis by pulsed-field gel electrophoresis . . . . . . . . . . . . . . . . . . . . Abstract . . . . . . . . . . . . . . . . . . . . . Introduction . . . . . . . . . . . . . . . . . . . Materials and Methods . . . . . . . . . . . . . . Results . . . . . . . . . . . . . . . . . . . . . Discussion . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . CHAPTER 2 ARTICLE: Iron and hemin storage in Yersinia pestis . . Abstract . . . . . . . . . . . . . . . . . . . . . Introduction . . . . . . . . . . . . . . . . . . . Materials and Methods . . . . . . . . . . . . . . Results . . . . . . . . . . . . . . . . . . . . . Discussion . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . CHAPTER 3 Association of a system for iron acquisition with sensitivity to pesticin in pathogenic yersiniae, and the effect of iron deficiency on protein expression . . Abstract . . . . . . . . . . . . . . . . . . . . . Introduction . . . . . . . . . . . . . . . . . . . Materials and Methods . . . . . . . . . . . . . . Results . . . . . . . . . . . . . . . . . . . . . Discussion . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . vi vii 21 25 36 37 39 43 50 74 80 85 86 87 89 94 111 116 121 122 124 128 134 177 184 LIST OF TABLES LITERATURE REVIEW uhbJNl-J CHAPTER 1 2 3. CHAPTER 1 2 CHAPTER 1 Virulence factors of X; pestis . . . . . . . . Plasmids in Yersinia . . . . . . . . . . . . . Properties and distribution f Yop§ . . . . . . Phenotypic comparison of Pgm , Pgm and Pgm , Pst Z; pestis . . . . . . . . . . . . . 1 Characterization of X; pestis strains . . . . . Restriction fragment sizes and estimate of genome size for l; pestis KIM . . . . . . . . . Intraspecific variation in macrorestriction fragment patterns . . . . . . . . . . . . . . . 2 Specific activities of hemin and inorganic iron in subcellular fractions . . . . . . . . . . . Isolation of bacterioferritin . . . . . . . . . 3 Iron repressible peptides of X; pestis . . . . vi \lO‘ 20 44 54- 61 94 102 148 CHAPTER 1 2 3 4 CHAPTER bWNH CHAPTER 1 2 mums-w \l 10 11 12 l3 14 LIST OF FIGURES 2 Single dimensional PFGE of Pgm+ DNA . . . . . Two dimensional PFGE of Pgm DNA . . . . . . Intraspecific variation in SpeI macrorestriction patterns . . . . . ._. . . Single dimensional PFGE of Pgm and Pgm pestis KIM DNA . . . . . . . . . . . . . . Two dimensional PFGE of Pgm DNA . .+. . . . SfiI/SpeI two dimensional PFGE of Pgm and Pgm DNA 0 O O O 4. O O + O O - O O O O O I O PFGE of DNA from Pgm ,_Pgm ,Pst Y; pestis and their isogenic Pgm mutants . . . . . . . 3 A- 1. 5m column profiles for Pgm: Y; pestis . . A- 1. 5m column profiles for Pgm Y. pestis . . Isolation of inorganic iron binding protein . Influence of exogenous hemin on inorganic iron acquisition . . . . . . . . . . . . . . Influence of exogenous inorganic iron on hemin acquisition . . . . . . . . . . . . . . 4 Y. pestis gradiant plates . . . . . . . . . . zquseudotuberculosis and Y; enterocolitica gradiant plates . . . . . . . . . . . . . . . Complete gradiant late results . . . . . . . Iron uptake by Pgm and Pgm, Y; pestis . . . Temperature and iron uptake by 1; pestis . . Expression of outer membrane proteins at 26 C by X; pestis: two-dimensional gels . . . . . Stained gel of Y. pestis outer membrane proteins . . . . . . . . . . . . . . . . . Stained gel of Y. pestis periplasmic proteins Stained gel of £_ pestis cytoplasmic membrane proteins . . . . . . . . . . . . . . Stained gel of Y. pestis cytoplasmic proteins Autoradiograms of two- -dimensional gels of pestis outer membranes . . . . . . . . Autoradiograms of two- -dimensional gels of Y. pseudotuberculosis outer membranes . . . . Autoradiograms of two-dimensional gels of Y. enterocolitica outer membranes . . . . . . Autoradiograms of lane gels of Yersinia outer membranes . . . . . . . . . . . . . . . vii 53 56 6O 64 67 70 73- 99 101 105 108 110 136 138 140 145 147 152 155 158 161 163 166 169 172 175 LITERATURE REVIEW The genus Yersinia (genus IX of the family Enterobacteriaceae) contains seven recognized species, three of which are important human pathogens (19,23,32). Yersinia pseudotuberculosis and X; enterocolitica are enteropathogenic organisms which cause chronic gastrointestinal disease :hi human hosts. They are transmitted via contaminated food or water supplies. l: pestis is the causative agent. of bubonic and pneumonic plague, a disease transmitted among various species of rodents and occasionally from rodents to humans by the bites of certain species of fleas. This highly virulent organism can be fatal to primates and most rodents at infecting doses of 10 cells or less (30). While most commonly infecting rodents its considerable influence on human history results from periodic devastating pandemics responsible for the deaths of perhaps as many as 200 million people (23,84). Although exhibiting considerable differences in ecology and the diseases they cause, 1; pestis and Y;_pseudotuberculosis chromosomal DNA is nearly indistinguishable in hybridization studies (7,14,73). Detectable genetic discrepancies are primarily in plasmid content (5,47) making their study a potentially powerful tool for understanding the genetic basis for virulence and pathogenicity. Chromosomal DNA of l 2 L enterocolitica is much less closely related to that of the other two species suggesting that the divergence of Y; enterocolitica from the lineage occurred at an earlier time than did the separation of Y; pestis and Y. pseudotuberculosis (7,14,73). All three species are geographically widespread with numerous distinct strains (32). Intraspecific relationships within 1; pestis has not been extensively studied, however, a classification system which recognizes three biotypes (mediaevalis, antigua, orientalis) based on the ability to ferment glycerol and reduce nitrate has been proposed (45). ‘1; pestis is acquired by its vector when it obtains a blood meal from an infected mammalian host (usually a rodent). While multiplying in the gut the bacteria produce coagulase which results in blockage of the proventriculus and prevents utilization of the blood meal by the flea. The hungry vector is induced to feed. on another host where regurgitation of contaminated blood causes introduction of the bacteria into the subcutaneous tissues of the previously uninfected host. From here bacteria are transported in the dermal lymphatics to regional lymph nodes where they undergo rapid multiplication. Inflammation and necrosis at the infected nodes produces the "bubos" characteristic of bubonic plague. Spillover into the blood allows dissemination of the bacteria and establishment of foci of infection at other locations associated with the reticuloendothelial system. Establishment of foci of 3 infection within the lungs results in the pneumonic form of plague. Spillage of the organism into the vascular system permits ingestion by the vector and lethality (32). Differences in pathogenicity of Y_._ pestis versus the enteropathogenic yersiniae may relate in part to the unique ability (n? the former to cause death from peripheral sites of infection (23). Wild type Y; pestis exhibited LDSOS in mice of 10 for intravenous, intraperitoneal, subcutaneous, and intradermal routes of injection, while the enteropathogenic yersiniae demonstrated comparable pathogenicity only after intravenous injection‘ (23,24,30,37,38,108). Within 24 hours of intravenous injection with any of these species into mice organisms disappeared from the blood and large populations had become established 1J1 the liver, spleen and lungs. Once populations of 106 bacteria per gram of tissue were achieved bacteremia was again present presumably due to spillover from the heavily infected tissues (104,108,109). Invasion and multiplication of bacteria within the spleen and liver resulted in the appearance of focal necrotic lesions. A typical inflammatory response to infection failed to develop in mice injected with Y;_ pestis and. the lesions progressively enlarged and eventually involved entire organs (104,110). This pathology of infection suggests suppression of cell mediated immunity. Although enteropathogenic yersinial infections also proved fatal, the necrotic lesions were more typical with some degree of neutrophil recruitment (104,110). Examination of these lesions by electron microscopy showed '1; pseudotuberculosis remained extracellular (96). Given the number of in vitro studies which indicate yersiniae can survive or grow within professional phagocytes (39,41,60,83,105,106) this was a surprising observation. While there may be interactions in vivo between yersiniae and professional phagocytes which are as yet undetected, it seems probable that yersiniae exist primarily as extracellular parasites. VIRULENCE FACTORS OF Yersinia pestis Although effective treatments have been devised for diseases caused by the pathogenic yersiniae, their continued study 515 of significance since these organisms provide an important model for determining the mechanisms of bacterial virulence and pathogenicity. Methods of in vitro culture are well established, and since rodents are natural hosts a realistic 111 vivo system is readily available. Studies of yersiniae have provided initial observations on several mechanisms of virulence that. were subsequently found to apply to other pathogenic species including i) auxotrophs blocked in purine biosynthesis are avirulent (18), ii) plasmids can mediate bacterial virulence (5,47), iii) the ability to bind hemin is associated with virulence (57,58), iv) the ability to restrict bacterial growth in vivo by 5 limiting iron availability acts as a host defense mechanism (59), v) hemin can serve as the sole source of iron for a prokaryote (80), vi) iron can be obtained by a non-siderophore, membrane bound specific ‘transport. system (80). Characterization of mutants with significant increases in L050 is one method of identifying virulence factors. In .1; pestis this approach has led to the recognition of several determinants of virulence. These are listed in table 14 While the factors which were the object of this study are chromosomally encoded, pathogenic yersiniae all- require the 70 kb Lcr plasmid for virulence, and Y; pestis is unique in requiring three different plasmids for expression of full virulence (5,47,108). These plasmids and their products are described in table 2. In the remainder of this review I will describe these recognized virulence factors of X; pestis. PURINE BIOSYNTHESIS Mutational loss (M? the ability to synthesize purines was first demonstrated to result in avirulence in Salmonella typhi and was hypothesized to relate to the lack of readily available purines in the tissues and serum of the host (1). This has since been shown to be an important virulence factor in a number of pathogens including Y_. pestis (22). In 1; pestis it was demonstrated that if the block occurred prior 11) de novo synthesis of inosine monophosphate (IMP) Table 1. LD 50 Effect of loss of virulence factors in Yersinia pestis on the in guinea pigs and mice. Virulence Factora LD50b Lcr Pst Pgm Fra Pur mouse guinea pig mouse+Fe3+C + + + + + 10 10 10 o + + + + 107 108 107 + O + + + 105 106 10 + + o + + 107 108 10 + + + o + 10 104 10 + + + + Co 102 104 102 + + + + 0e 107 108 107 abbreviations for virulence factors Lcr: Pst: Pgm: Fra: Pur: Presence of Lcr plasmid Presence of Pst plasmid o absorb exogenous hemin at 26 C Fraction 1 Antigen ability to synthesize purines cells introduced by intraperitoneal injection injection of sufficient iron to saturate serum transferrin mutation blocks purine synthesis prior to inosine monophosphate mutation blocks purine synthesis after inosine monophosphate TABLE 2. Plasmids in Yersinia. Plasmid Proteins Encoded Species (kilobases) 9.5 Pesticin L Estis Pesticin Imunity Plasminogen Activator/ Coagulase 70 V Antigen Y_._ Estis \ Yersinia Outer Membrane .Y_._ pseudotuberculosis Proteins (Yops) _Y_. enterocolitica Lcr regulatory proteins (lcr or vir genes) 100 Fraction 1 L Estis Murine Toxin 8 there was only a relaatively small reduction in virulence, but if the block occurred at the conversion of IMP to guanosine monophosphate (GMP) the organisms were avirulent (18). The discovery that Y; pestis unable to convert IMP to GMP could not survive or reproduce in mouse peritoneal macrophages in vitro without the addition of hypoxanthine or guanosine to the culture medium supports the idea that this purine is normally unavailable to the parasite when it is within the mammalian host (105). FRACTION 1 ANTIGEN Fraction 1 antigen is a protein-carbohydrate complex associated with the capsule of Y; pestis (2). Complexes are composed of approximately 15 kda protein-carbohydrate (fraction 1A) and free protein (fraction 18) subunits which will assemble into structures of up to 300 kda (6,51). Maximum expression occurs at body temperatures of mammalian hosts (6). The significance of fraction 1 antigen as a virulence factor is unclear. While Fra- strains injected intraperitoneally into guinea pigs demonstrated reduced virulence, lethality is not reduced in mice (table 1) (30). Also encoded on the lOO-kb Tox plasmid is an exotoxin which is highly lethal to mice and rats, but non-toxic in other hosts (16). This is a: 120 kdal protein composed of subunits which may be 12 kDa (72), however, modified purification schemes suggest they may be larger (Lucier, unpublished data). While the role of the exotoxin in vivo 9 is uncertain, it run! account for the more rapid death of mice following intravenous injection with L pestis, than with enteropathogenic species of Yersinia which do not produce exotoxin (104,108). LCR PLASMID Early studies of virulence in Y;_pestis were hampered by a rapid shift to avirulence when cells were cultivated in vitro at 37°C but not at 26°C in a variety of commonly used media (50). Modification of media components showed this to be the result of an unexpected dependency on the presence of. at least 2.5 mM Ca2+ for unrestricted growth of virulent cells at 37°C (56). While zinc and strontium also facilitated growth, only calcium is present. in vivo in sufficient quantities to alleviate the restriction. Further studies showed restriction of growth to be Mg2+ dependent with lower concentrations of Ca2+ being permissive at magnesium concentrations below 20mM (27). 4A role for the Lcr plasmid in this response became apparent with the observation that calcium dependence for growth was lost in cells lacking the plasmid (5,47). Along with. the restriction in vegetative growth bacteria were discovered to begin synthesizing a set of Lcr plasmid encoded putative virulence factors including Yersinia outer membrane proteins (YOPS) and V antigen (10,11,12,20,27,100,102,118). This is termed the Lcr (low calcium response) and is mediated by similar plasmids in all pathogenic species of Yersinia 10 (5.47.86). The growth restriction exhibited by virulent yersiniae grown in conditions simulating the environment within host 2+) is the cells (37°C, 20mM Mg2+, approximately 0 mM Ca result of an ordered nutritional stepdown beginning with cessation of stable RNA synthesis and reduction of adenylate energy charge (40,119). One additional round of DNA replication and cell division occurs (109), and most protein synthesis is greatly reduced or eliminated, with the exception of most factors encoded on the Lcr plasmid and a few chromosomally encoded proteins which may be necessary. for full expression of virulence (69,119). In L pestis growth restriction is complete, although addition of Ca2+ or shifting the temperature down to 260C restores normal growth (119). Similar responses occur in the enteropathogenic yersiniae, however, the restriction of growth is not as extreme (21,38). Regulation of the Lcr is complex and not well understood. Structural genes for YOPS are located in several Operons scattered about the plasmid and function as a coordinately controlled regulon (10,13,100). Regulatory genes designated lcr (vir in Y; enterocolitica) are encoded within an 18 kb Ca2+ dependence region (10,52,85,1l7). Five thermally inducible lcr genes have been identified in this region and transposon insertions within these loci usually result in a loss in the ability to perform the Lcr just as in cells from which the plasmid has been cured (13,52). ll Spontaneous avirulent mutants which no longer expressed the Lcr were assumed to have lost the Lcr plasmid. Yet as many as 50% were found to still carry the plasmid, but with various mutations within the calcium dependence region(ll7). At least two other regulatory factors and the V antigen are encoded in ant adjacent lchVH operon (88). Insertions in this region often eliminate the regulatory effects of calcium so bacteria grow and express YOPS at 37°C regardless of calcium levels (117). On the opposite side of the calcium dependence region of Y; enterocolitica is the vir C operon which encodes a group of proteins necessary for the. secretion of YOPS (70). This 8.5 kb region contains 13 open reading frames and is regulated in the same manner as the YOP operons. Its existence in the plasmids of the other pathogenic yersiniae has not yet been reported. While the Lcr plasmids from these three species are very similar, homology within the two regulatory regions is especially high (nearly 100%) suggesting this mechanism is conserved within the pathogenic yersiniae (88). The arrangement and nucleotide sequence of the YOP genes is more variable (10,100), yet exchanging the Lcr plasmids of Y; pestis and .1; pseudotuberculosis still results in a normal Lcr in both species (116). Thus the Lcr and its regulatory mechanisms are highly conserved and essential for Yersinia virulence. Several comprehensive models describing regulation of the Ixn: have been proposed and while they disagree in many details, there are consistent similarities (8,42,87). All 12 recognize two separate regulatory loops one sensitive to temperature and the other sensitive to calcium concentration. .A shift to 370C initiates production of an activator encoded by the lcr F gene (42). Additional genes in the calcium dependence region may also be involved (8). This encourages transcription of plasmid encoded genes. Counteracting the thermally induced activator system in low calcium conditions is a: calcium sensitive repressor system involving at least lcr iiiof the lcr GVH operon (8,87,89). These systems determine the level of transcription at the YOP and lcr GVH operons. Non-polar‘ lcr \I mutants. have» recently been shown to lose their dependency on calcium for growth at 370C, yet they still produce YOPs (87). Thus accumulation of V antigen in the cytoplasm may be sufficient to promote restriction of growth. Association of V antigen with a regulatory function was unexpected. V and W antigens were initially shown to be produced by virulent yersiniae in vitro and in laboratory animals (31). V antigen is a 38 kDa monomer which may undergo aggregation, a property which may account for its initial size estimate of 90 kDa. It is found within the cytoplasm and in culture supernatants (93,102). Studies on the role of secreted V have been hampered by its autoproteolytic property (25), however, the passive protection provided by polyclonal V antiserum against infection by all pathogenic yersiniae argues for an essential role in virulence (lll). Histopathology studies suggest it may suppress cell mediated l3 immunity and granuloma formation (109,110). Thus V may be an unusual bifunctional protein serving both a regulatory role in the cytoplasm and as a virulence factor following secretion (87). W was originally identified as a 140 kDa lipoprotein which accumulated in the supernatant (64). Little is known of W, however, recent attempts at its isolation resulted in retrieval of a protein with characteristics of GroEL, a chaperone protein which facilitates secretion (65, Mehigh, unpublished data). W may be the GroEL - V complex. Yops are produced by all yersinial pathogens during the- Lcr, but there are species specific differences in expression. Table 3 provides a summary of identified Yops and their known characteristics. While these become major outer membrane proteins in enteropathogenic species early studies failed to discover them in the outer membrane of L pestis (102). This was a puzzling observation since serum from convalescent plague patients contained anti-Yop immunoglobulins (67), and when the Lcr plasmid of L pestis was transferred to Y; pseudotuberculosis normal outer membrane Yop expression occurred (116). With the discovery that mutants cured of the pesticin plasmid exhibited stable expression of Yops in their outer membrane (93) it became apparent that the plasminogen activator protein catalyzed the proteolysis of Yops (98). So although Yops are produced by L pestis they are immediately hydrolyzed upon reaching the outer membrane in pst+ strains. This may explain the .14 Ahaa.ecu.~c~.ooH.Nm.Hm.~m.~o.mm.mv.m~.mu.m.mv mooceueuom noncom 333898 + + + em . 2 no» 9:320 use cofluemoummo umaouea 39.22.... «an 538%»? uoaeuea o... .2..ng ouzuoaun ans—Mum + o o + me 2 no» moaned» 5 530nm unread new 83.69. + o o + 3 a now mosmmfiu cw 530nm ”5.5.3 new 60.5%... + o o + NN x no» 0 + + Hm n. now + 0 me H do» mamouaoommfi 3325 “omega—among 9.39;... cfimuoum + + + + we 2 now + 0 mm o no» + + 2. a no» .mwmogoomecm 3325 «so: 838?: E seepage... €32. mac» non—owns” . xzemcofiummmuucfi mo 8:33? 33mm 4». cw ”Baboon cogmpeumop e33...“ + + + + mm u now + + + «M o 90> + + + we 0 now + + + z. n do» 83932.. 2829.8 ”.32.: .3330 3:950 uo cannon—98 nuw3 ooumaoomwm cofimwznc o + + o t new oocodsufi> :3 too 835 0968 238 :85 830:6 venomous 33:38 ”E 80:88 33 B: 538.. .88 mo 8332.33 e5 83% .n mics 15 observation that a polyclonal preparation of anti-Y; pestis Yops provided complete passive protection against X; pseudotuberculosis but was ineffective against plague (23). The continual renewal of major surface antigens may mediate the removal of bactericidal molecules involved in host defense and allow for the expression of a more acute disease. Numerous studies have dealt with the possible roles of various Yops as virulence factors. Genetic studies involving insertional mutagenesis have helped identify Yops whose expression is necessary for full virulence (13,99,101). and a growing number of studies have addressed the specific functions of these Yops. These and other characteristics of the known Yops are summarized in table 3. Several of these have recently been the object of intensive study. Yad A is unusual in that its expresion is only temperature regulated; Ca2+ concentration has no effect (61). It is not expressed in Y; pestis due to a frameshift mutation and its mutational loss in I; pseudotuberculosis enhances lethality (92). Yad A may promote chronic rather than acute disease. Both Yop E and Ii may inhibit the activity of professional phagocytes using different mechanisms. Yop E may interfere with microtubular function (91) while Yop H is a tyrosine phosphatase which can remove phosphate from certain host proteins and perhaps interfere with signal transduction pathways necessary for phagocytosis (9,53). Recently there has been controversy surrounding the location at which Yops l6 exert their effects. Since they lack signal sequences and hydrophobic regions typical of membrane spanning proteins (71) they may be released proteins which exert their effects following their release from the bacterial cell. During the Lcr a few chromosomally encoded proteins continue to be expressed at high levels (69). The two expressed in greatest amounts are GroEL which would be necessary for ‘transport. of the large number cfif secreted proteins being produced and a large protein with catalase activity; Whether the latter is necessary for the expression of disease in Y. pestis remains to be resolved. PESTICIN PLASMID Ben-Gurion and Hertman (4) first recognized that virulent strains (ME 1; pestis produced a bacteriocin which they named pesticin. It is a 44 kDa protein which exhibits n-acetylglucosidase activity (46) and has a host range limited 11>;g; pseudotuberculosis serotype I, 08 strains of Y. enterocolitica, a few strains of Escherichia coli, and Pgm+, Pst_ strains of Y;_ pestis (28). Pesticinogenic strains are protected from the effects of the bacteriocin by an immunity protein absent in non-pesticinogenic mutants (55). It seems unlikely that a bacteriocin which targets envelOpe associated molecules in prokaryotes could have a role in virulence, so it was surprising that Pst- strains were no longer able to disseminate from peripheral sites of 1? infection in mice (24). The discovery of linked expression of a plasminogen activator protein with potent fibrinolytic activity helped explain this observation (28). It is located in the outer membrane as a transient 37 kDa monomer which autodegrades and forms a 35 kDa peptide (98). The protein was also shown to exhibit coagulase activity, however this may be an artifact restricted to rabbit blood (97), cu: an activity expressed only at 26°C which may be significant for survival in the gut of the flea vector (67). Recent studies also demonstrated the role of this protein in the hydrolysis of Yops as previously described in this. report. The 9.5 kb pst plasmid is unique to L pestis and encodes pesticin, the pesticin immunity protein, and the plasminogen activator (97). Loss of the plasmid results in significantly reduced virulence by peripheral routes of infection (Table 1), presumably due to reduced ability of the bacteria to reach the lymph or blood without expression of plasminogen activator. Injection of mice with 40 ug of iron can restore virulence to Pst- strains introduced intraperitoneally. Although this might imply existance of a lesion in an iron uptake mechanism. associated with the pesticin plasmid, no evidence for such a mechanism has been found. This effect could also result from inhibition of the activity of host professional phagocytes by (excess iron (111). 18 PIGMENTATION Virulent isolates of L pestis can absorb exogenous hemin (57) or congo red (107) to their surface when grown on solid media at 26°C thus forming pigmented (Pgm+) colonies. Strains which suffer mutational loss of this ability (Pgm- ) grow as white colonies and are avirulent via peripheral route (n3 injection (58). Interestingly they remain fully virulent showing normal progression of the disease if introduced intravenously (108) indicating that the Pgm associated virulence factors are necessary for survival in or dissemination from peripheral tissues. That virulence of. peripherally introduced Pgm- strains can be restored by 3+) to saturate serum injecting sufficient iron (Fe transferrin (58) raised the possibility that mutation to Pgm- resulted in time loss of an iron uptake mechanism required for full virulence. Another association exists between pigmentation and pesticin. Pgm+, Pst- strains are sensitive to pesticin due to the lack of the Pst plasmid encoded immunity protein. Yet Pgm-, Pst- strains are resistant (170. Even in Pgm+, Pst- cells the effects of pesticin are inhibited by sufficient Fe3+ or hemin in the medium (26). It. was suspected that pesticin could bind to iron repressible protein components of a high affinity iron uptake system in the cell envelope (28). Type B colicins of g; 22;; use this type of mechanism to attach to target cells (44). That ton g mutants of pesticin sensitive E. coli strains are also l9 resistant provides support for this idea (48). If this is the case, then a lesion in at least one iron uptake system would occur when cells become Pgm-. Initial studies found no differences jJ12.156.4 >2.475.0 ' dem SpeI. 0: Nod. 0:5)“. 4: “Au-1. l. 1 I“ l“...- L k ‘EnuJ-wmnhhuuuamunmmmmmwm 'Plobahlclmltiphfw. ‘mammwmawuam 'WdWmMquyWPFGE. 55 Figure 2. Two-dimensional pulsed—field. gel (of chromosomal DNA from Pgm+ cells of X; pestis KIM(Dl42) after digestion with §pgl (first dimension) and then 522i (second dimension) under the conditions defined in the text. The lanes across the top of both figures represent a single dimensional gel of the §Ee_l_ digest shown at the same scale as the two-dimensional gel illustrated directly beneath. (A) Photograph of the» two-dimensional. gel and (B) interpretive diagram showing the locations of fragments in the photograph; the three fragments drawn as Open circles are those absent in similar gels of chromosomal DNA from isogenic Pgm- organisms (see Fig. 5). Numbers in the margin indicate the location of lambda HindIII and concatemeric lambda DNA molecular size markers (in kilobases). 56 l Figure 2 57 with which the generation of small fragments prohibited performance of PFGE in a second dimension. It also exceeded the estimate of 3853.8 kb made by summation of individual fragments detected in the two-dimensional gel owing to Apal- mediated loss of small undetectable pieces of DNA, as discussed below. N231 gave sufficient fragment resolution for a second determination of genome size; however, the bands were closely spaced within the 20 11) 150 kb size range, making the analysis more difficult than that performed with §pg£. Single-dimensional pulsed-field and conventional gels indicated the presence of 37 bands with an apparent total length of 2751.6 1a» However, the width and intensity of several bands indicated the existence of comigrating fragments (Fig. 1). Two-dimensional gels performed with §2£g_ and then 523$ resulted in the appearance of 92 fragments with a total length of 3330.8 kb. Analysis as described for §pe_I-A&I two-dimensional gels permitted the identification of 10 bands composed of multiple fragments, ‘which are indicated in Table 2. Correcting for this comigration of the single-dimensional gels resulted in a chromosome size estimate of 3996.5 kb, which is similar to that described above for the analysis of §pgl digests. The Tox plasmid was isolated and subjected to hydrolysis by gag; ELL A32, and Lil. Attempts to detect tflua products of these digestions following PFGE of normal agarose electrophoresis were not successful , (nu Ila—Immun- a... 58 indicating the absence of corresponding recognition sites. Similarly, PFGE of purified but undigested Tox plasmid failed to form a visible band, indicating that the plasmid in circular form is unable to emerge from the well (33), and two-dimensional PFGE of an isolate lacking all three plasmids of the species (_Y_. pestis KIM(P15)) yielded results identical to those obtained with strains possessing the Tox plasmid (Not illustrated). Accordingly, all of the fragments detected by PFGE were of chromosomal origin. Addition of the known sizes of plasmid DNA (189.5 kb) to that determined with §pe_I for the chromosome (4208.4 kb) thus provides an estimate of 4397.9 kb for the genome of wild-type L pestis. Adding plasmids to the M estimate (3996.5 kb) results in a comparable total genome size of 4186.0 kb. Intraspecific variation. The macrorestriction patterns of a few additional strains of L pestis, selected on the basis of diverse geographical ofigin and biotype, were compared with that of strain KIM (Fig. 3). Although each strain exhibited a unique pattern of fragments, their number and range of sizes were similar, suggesting that there were no major variations in genome size. The patterns corresponded well with the ability to ferment glycerol, a primary determinant in assigning biotype (16). For example, the patterns of glycerol-negative strains (_Y_._ pestis subsp. orientalis) were significantly more similar to each other 59 Figure 3. Pulsed-field gels of genomic DNA from selected strains of X; pestis after digestion with £221 on (left) a 1% SeaKem GTG agarose gel run under the conditions defined in the legend to Fig. l (favoring resolution of fragments of greater than 100 kb) and (right) a 1.2% SeaKem GTG agarose gel run under conditions of 180 V at 150C with a 2- to 12-s pulse ramp for 27 11 (favoring .resolution «of fragments of less than 100 kb). Lanes: 1, lambda DNA concatemeric molecular size markers (shown in kilobases in both left margins); 2, strain KIM(Dl); 3, strain Kuma; 4, strain Yokohama; 5, strain Salazar; 6, strain EV76; 7, strain TS; 8, strain Dodson; 9; strain A12; 10, lambda HindIII size markers (shown in kilobases in the far right margin). 1lllHélI. 111131111! 3'11! 11‘ I" 11.1“ II I- 1-11111111ln .11 11111" Figure 3 61. TABLE 3. Feirwise comparisons of SpeI-generated total genomic DNA fragments from selected strains of rersinia pestis after single-dimensional pulsed-field gel electrophoresis strain Strain Rune Yokohona Salazar 3V76 rs Dodson Al? xxx 46170‘ 42166 28164 26164 24164 20164 28166 65.7b 63.6 43.8 40.6 37.5 31.3 42.4 Kwna 64170 22168 24168 20168 20168 24170 91.4 32.4 35.3 29.4 29.4 34.3 Yokohoma 22164 24164 18164 18164 22166 34.4 37.5 28.1 28.1 33.3 861626: . 54162 48162 42162 44164 87.1 77.4 67.7 68.8 -sv76 48162 42162 44164 77.4 67.7 168.8 rs 54162 52164 87.1 81.3 Dodson ‘ I 56164 87.5 ‘lo. of fragments scored as identical [sun of fragments detected in both strains. thtio x 100. 62 than they were to those of any of the glycerol-positive isolates (Table 3). The members of L pestis subsp. orientalis tested all exhibited fragments of 232.4, 213.0, 169.7, 60.5, 54.6, and 51.2 kb plus those between 40 and 30 Ida. Within this variety, the patterns of the two isolates from North America (strains A12 and Dodson) showed greater similarity to that of the TS strain from Java than to those of strains from South America (Salazar) and Madagascar (EV76). The patterns of the two glycerol-positive strains of I; pestis subsp. antiqua (Kuma and Yokohama) were highly similar to each other and more closely resembled that of strain KIM (L pestis subsp. mediaevalis) than those of L pestis subsp. orientalis. Digests of the three glycerol- positive isolates tested exhibited bands with sizes corresponding to those of fragments B, C, G, H, X, and Y listed in Table 2. The Pgm--specific deletion. Total DNA from Pgm+ (substrain D142) and isogenic Pgm- (substrain D35) cells of L pestis KIM was digested with SpeI, NotI, AscI, or _S_f_il and compared after single-dimensional PFGE. At least one fragment was always missing in the digests of Pgm- mutants (Fig. 4). These missing fragments were the 182.8 kb band B for Q, and 30.6 kb band Y for SE, and the 72.5 kb band M for _A_s_c_I_. In digests prepared with £19333, the 78.4 kb band Q became narrower and less intense owing to probable loss of a multiple comigrating fragment. This consistent 63 Figure 4. Pulsed-fiehd gels of chromosomal DNA from Pgm+ (strain KIM(Dl42), lanes l,3,5,and 7) and Pgm- (strain KIM(D35) lanes 2,4,6, and 8) derivatives of L pestis KIM digested with M (lanes 1 and 2), _I\1_o_t_I_ (lanes 3 and 4), gig; (lanes 5 and 6), and §£ll (lanes 7 and 8). Electrophoresis was performed in 1% SeaKem GTG agarose under the conditions described in the legend to Fig. 1. Positions of lambda concatemeric molecular size markers are indicated (in kilobases). Solid arrowheads indicate the positions of Pgm+-specific bands, and the open arrow designates the location of the 77-kb junctional fragment in DNA from Pgm- mutants digested with SfiI. Figure 4 64 1111111111, 1." I ' - 1111-11111 1 1.1111111 . 1.1 111111 Cr 65 disappearance of fragments in digests of DNA afrom Pgm- organisms is in accord with the concept that the mutation reflects the occurrence of a large deletion. Examination of DNA from six Pgm- mutant clones of L pestis KIM(Dl42) revealed identical macrorestriction patterns for SfiI, NotI, and §p_el (only one isolate was analyzed with _A_§_C_I_), suggesting the occurrence of a deletion with precise or nearly precise endpoints. Analysis of the deletion by single-dimensional PFGE after digestion with _Spfi Natl, or A_sc_I_ failed to reveal the existence of new junctional fragments in digests of Pgm- mutants. Accordingly, two-dimensional PFGE was used to provide an estimate of the size of the deletion. Analysis of Pgm- mutants by digestion with m and then M showed the disappearance of the 31.4 kb band Z (shown as two smaller fragments after cleavage with 5223 in Fig. 2) and of an additional fragment of about 40 kb that constitutes part of one of the three 95.9 kb fragments composing band P (Fig. 5). A Pgm--specific fragment is visible in a location indicating that it is slightly larger than band Y (Fig. 5), with an estimated size of approximately 34.8 kb. The loss of two fragments totaling 127.3 kb and the appearance of a 34.8 kb fragment suggest a deletion of 92.5 kb. Digests of Pgm- mutants prepared with $13 lacked the 182.8 kb band B. If most or all of the deleted Pgm+- specific sequence is contained within this large fragment, then one or possibly two junctional fragments should be 66 Figure 5. Two-dimensional pulsed-field gel of chromosomal DNA from Pgm“ cells of Y; pestis KIM(D35) after digestion with Spgl (first dimension) and then Apgl (second dimension) under the conditions defined in the text. The lane across the top of the figure represents a single-dimensional gel of the §pg£ digest shown at the same scale across as the two-dimensional gel illustrated directly beneath. An interpretive diagram showing the location of missing fragments in comparison to a similar gel of chromosomal DNA from the parental Pgm+ strain is given in Fig 28. The white arrow within the field indicates the location of the Pgm--specific junctional fragment. Solid arrowheads show the location of missing Pgm+-specific fragments, and the open arrowhead indicates the position of the junctional fragment on the single-dimensional gel. Numbers indicate the location of lambda HindIII and concatemeric lambda DNA. molecular size markers (in kilobases). 67 Ham.w . . v a 11.1.1111- .. . o .. Figure 5 68 detected in digests of Pgm- mutants following two-dimensional PFGE. Furthermore, use of £3.33}. for the second dimension of gig-generated digests of Pgm+ organisms should result in the appearance of the 95.9 kb and 31.4 kb fragments noted above in the column previously occupied in the first dimension by _S_f_i_I_ band B. Both of these assumptions were correct. Fragments of 95.9 and 31.4 kb were detected when _S_f_i_l band B was digested with M and then subjected to second-dimensional PFGE (Fig. 6A). DNA from Pgm- mutants, or course, lacked all of the §Le_1_- generated fragments originating at the location of S_f£ band B (Fig. 6B). However, two new fragments of 38 and 23 kb were observed at a location that corresponded to the position of the 77 kb junctional fragment first detected in single-dimensional gels of DNA from Pgm- mutants digested with 551—1 (Fig. 4). No further differences were evident in single-dimensional ($3) or two-dimensional (_S_g and M) gels, an observation consistent with occurrence of most or all of the Pgm--specific deletion within §_f_i_I band B. The 77 kb junctional fragment would thus arise following deletion of a 105.5 kb segment of DNA from _SLi_I_ band B. This value corresponds closely to the independent 92.5 kb estimate determined by two-dimensional PFGE with SpeI and ApaI. Mutation £2 pesticin resistance. DNA was prepared from the rare Pgm+ Pst- (pesticin-resistant) mutant L pestis 69 Figure 6. Two-dimensional pulsed-field gels of chromosomal DNA from (A) Pgm+ strain KIM(Dl42) and (B) Pgm- strain KIM(D35) cells of L pestis after digestion with Siil (first dimension) and S23; (second dimension). Lanes across the top represent single-dimensional gels of E digests at the same scale as the two-dimensional gels shown directly beneath. The white arrowheads in the field of panel B indicate the locations of Pgm--specific junctional fragments. The solid arrowhead shows the location of the missing 183-kb §£££ fragment, and the open arrow indicates the position of the 77-kb junctional fragment on the single- dimensional gel of DNA from Pgm- cells. Numbers indicate the locations of lambda HindIII and concatemeric lambda DNA molecular size markers (in kilobases). 70 dumb I cw APM Mu: 0.5 9m 1 Figure 6 71 KIM(KllS-B) and subjected to digestion by appropriate restriction endonucleases. DNA composing the Pgm+-specific 31.4 kb S23; band Z and 182.8 kb §£l£ band B was retained in this isolate (Fig. 7). However, DNA from an isogenic Pgm— (Y; pestis KIM(llS—C)) mutant isolated on Congo red agar, like that of typical Pgm- mutants (e.g., Y; pestis KIM(D35)), lacked both of these Pgm+-specific fractions. These results would be expected if the event accounting for the loss of IrpB, IrpC, IrpD, and IrpE in the rare mutant was a mutation undetectable by PFGE in the Irp operon contained within the 93 to 106 kb deletable sequence of typical Pgm+ yersiniae. This could be a small deletion or insertion or a polar mutation (introduction of a nonsense codon or frameshift). 72 Figure 7. Pulsed-fiehd gels of chromosomal DNA from Pgm+ strain KIM(Dl42) (lanes 1 and 6), Pgm- strain KIM(D35) (lanes 2 and 7), Pgm+ (pesticin-resistant) strain K115- 8 (lanes 3 and 8), and Pgm- (pesticin-resistant) strain K115- C (lanes 4 and 9) derivatives of I; pestis KIM digested with §EE£ (lanes 1 to 4) or SE3; (lanes 6 to 9). Conditions were as described in the legend to Fig. 1A. Positions of lambda concatemeric molecular size markers (lane 5) are given in kilobases. Arrows indicate the location of Pgm+-specific fragments. 73 1—' I—‘NNw 0° \1 m bmwxo U1 U‘I U") U7 TIMI“! (1811 1 1111111 11111 1 1123:! 21831 I 111111111111 . . (37‘111182 1 2:71 1 Milt?) ' Figure 7 74 DISCUSSION Y; pestis is related to E; ggli, as judged by significant DNA homology and the existence of numerous shared physiological determinants (7,8). Smith et a1. (47) used PFGE of NEE; restriction digests to estimate a genome size of 4550 kb for E; 32;; K-12 that was later refined to 4700 kb (15). The size of 4208.4 kb reported here for the chromosome of X; pestis, an obligate intracellular parasite restricted to fixed niches (7), is thus 89.5% of that determined for the more metabolically active E. coli. This smaller size may reflect the absence in ‘2; pestis of catabolic and regulatory functions (7) utilized by E; 22;; for existence as a commensal organism or for survival in soil and water. In this context, it is of interest that a value of 5900 kb was estimated for the genome of Pseudomonas aeruginosa (42), £3 robust saprophyte capable of growth in numerous natural environments. Genomic analysis of L pestis was difficult due to a lack of restriction endonucleases that cut at only a few sites on the chromosome. 1k) enzyme-generated fragments larger than the 312.6 kb band A were detected after digestion with SE. This observation contrasts with the case of E; coli, for which NotI digests contained fragments 75 of up to 1000 kb in size (47). Although N23; has been used for analysis of the genome of L pestis as well (23), the closely packed bands and larger number of comigrating fragments observed in Egg; digests suggested that Spgl might be a better choice for use in PFGE. Our difficulty in finding restriction endonucleases that produced few eneough fragments 133 permit detailed analysis was probably due in part to the 45.6% GC content (36) of the yersinial chromosome and to its relatively large size. It thus became necessary to find enzymes that recognized specific rare sequences rather than relying on those distinguishing 6- or 8-base sequences containing either all C/G or .A/T. Indeed, all of the enzymes tested that recognized 6-base sequences containing CTAG (SpeI, XbaI, NheI, and AvrII) produced digests with comparatively few fragments, although only those prepared with £221 could be accurately analyzed. This finding with X; pestis is in accord with the observation that CTAG is a rare sequence in bacteria with GC contents of greater than 45% (37). Our results also emphasized the need for careful evaluation of single-dimensional pulsed-field gels when using this method for determining the size of a bacterial genome. Digests prepared with Aggl initially seemed adequate, but the small sum of the visible fragments suggested considerable comigration, resulting in a low estimate of total size. Similar situations have been described for other species, including _P_:_ aeruginosa, for 76 which initial determinations of genome size by single- dimensional PFGE were 2700 kb (19). Later work involving two-dimensional PFGE (3) and additional restriction enzymes (42) resulted 111 higher estimates (n? 5,300 and 5,900 kb, respectively. Two-dimensional PFGE was useful in this study for confirming the genome size estimated from single-dimensional gels and for identifying comigrating fragments. Nevertheless, the sum of fragments located within a given column of their progenitor as determined by single-dimensional PFGE. This difference was caused by both generation in the second dimension of fragments of 10 kb or less that were too faint for reliable resolution and to a larger error in measuring fragments in two-dimensional gels as they approached sizes of 150 kb. These errors probably account for the discrepancy of 311 1d: observed between the values obtained by single-dimensional PFGE after correction for comigrating fragments (4,208.4 kb) and by summing the sizes of visible fragments after two-dimensional PFGE (3,853.8 kb). The smaller value obtained from the two- dimensional gel makes it unlikely that the length determined by single-dimensional PFGE is a significant underestimate of the correct genome size. The usefulness of macrorestriction pattern analysis in identification of unknown strains of a given species and in determining intraspecific relationships and genetic history has been established (1-3 , 6 , 31) . For example , 77 macrorestriction patterns determined by PFGE for both Lactococcus lactis and. Streptomyces ambofaciens exhibited 70 to 80% band identity, as opposed to 23 to 30% hand identity for distantly related strains (6, 31). These values are very similar to those determined in this study for selected isolates of L pestis. Our results are in accord with the accepted division of the species into glycerol-negative (X; pestis subsp. orientalis) and glycerol— positive strains and suggests that division of the latter into X; pestis subsp. antigua and Y; pestis subsp. mediaevalis (16) is equally legitimate. These determinations, however, were primarily undertaken to define the limits of diversity within the species. Further work will be required to fully exploit the potential of macrorestriction pattern analysis to help trace the historical spread of plague. In this context, it is possibly significant that strain T8 of L pestis subsp. orientalis, isolated :h1 Java, exhibited significant. band identity with the two North American isolates tested, Dodson (87.1%) and A12 (81.3%), and that the latter two ‘were equally similar to each other (87.5%). This relationship favors the hypothesis that plague has not always been present in the New World but rather was introduced from Indochina at the start of the present century (18, 22, 34). Although detection of a large deletion in Pgm- organisms was anticipated (41, 44-46), its extent was difficult to determine by PFGE because the size of missing 78 fragments depended Luxn1 which restriction endonuclease was used. New junctional fragments in digests of Pgm- mutants were not always evident after single-dimensional PFGE, and their detection usually required migration in a second dimension. Our determination of the size of the deletion is not precise because fragments of 10 kb or less could not be accurately resolved in two-dimensional gels. The estimate of 92.5 to 105.5 kb for the deletion may thus be a significant underestimate, although the concurrence of the two independent analyses suggests that this value is reasonably accurate. The deletion is certainly large enough to include all of those structural genes encoding the peptides required for pigmentation (ll, 39, 41, 44) and assimilation of iron (11, 44) not expressed in Pgm- mutants. Larger high-frequency deletions (250 to 2,100 kb) occur at specific regions of the chromosome in species of Streptomyces, although the exact positions of their endpoints and their extents appear to be variable. Amplification of DNA sequences associated with the deletion may also occur in these organisms (4, 5, 30). We found no evidence in this study for similar augmentation or for occurrence of variation in either the size or location of the deletion. These observations suggest that the Pgm- specific deletion involves a DNA sequence found as a single copy on the chromosome and that its endpoints are highly Specific. Chromosomally encoded virulence determinants similarly vulnerable to high-frequency deletion have been 79 described for other’ pathogenic bacteria. These factors include two two groups of hemolysin and fimbrial genes of uropathogenic ‘E; Iggll that are lost as part of exact deletions of 75 and 100 kb, respectively (20, 26, 27). The precision of these events reflects the occurrence of short direct repeats at each end of the DNA segment that evidently mediate site-specific recombination. Similarly, genes encoding the synthesis and retrieval of the siderophore aerobactin occur in a cluster found either in the chromosome or on the ColV plasmid of several enteric species (14). Flanking ISl elements give this segment the appearance of a composite transposon (29, 38, 40). Although they no longer function in transposition, these elements also mediate nearly precise deletions of the aerobactin genes through site-specific recombination. The high frequency and apparent precision of the Pgm- specific deletion in ‘1; pestis may also occur as a consequence (n3 site-specific recombination. Further study of the deletable segment may show 31: U: be an iron uptake and storage gene cluster that allows the bacterium to colonize its host. Its current instability in the chromosome may reflect an ability in the past to invade bacterial genomes. 80 REFERENCES Allardet-Servent, A., (L. Bourg, M1Ramuz, M. Pages, M. Bellis, and G. Roizes. 1988. DNA polymorphism in strains of the genus Brucel la . J . Bacteriol . 170:4603-4607. Arbeit, R. D., M. Arthur, R. Dunn, C. Kim, R. K. Selander, and R. Goldstein. 1990. Resolution (of recent evolutionary divergence among Escherichia coli from related lineages: the application of pulsed field electrophoresis to molecular epidemiology. J. Infect. Dis. 161:230-235. Bautsch, W., D. Grothues, and B. Tummler. 1988. Genome fingerprinting of Pseudomonas aeruginosa by two-deminsional field inversion gel electrophoresis. FEMS Micro. Letts. 52:255-258. Birch, A., A. Hausler, C. Ruttener, and R. Hutter. 1991. Chromosomal deletion and rearrangement in §Ereptomyces glaucescens. J. Bacteriol. 173:3531-3538. Birch” A., A. Hausler, M. ‘Vogtli, W; Krek, and. R. Hutter. 1989. Extremely large chromosomal deletions are intimately involved in genetic instability and genomic rearrangements in Streptomyces glaucescens. Mol. Gen. Genet. 217:447458. Bourgeois, P. L., M. Mata, and P. Ritzenthaler. 1987. Genome comparison of Lactococcus strains by pulsed-fiebd gel electrophoresis. FEMS Micro. Letts. 59:65-70. Brubaker, 1L. R. 1991. Factors promoting acute and chronic diseases caused by yersiniae. Clin. Microbiol. Rev. 4:309324. Brubaker, R. R. 1983. The Vwa+ virulence factor of Yersiniae: the molecular basis of the attendant nutritional requirement for Ca+2. Rev. Infect. Dis. 5, suppl 4:S748S758. Brubaker, R. IL. 1971. The genus Yersinia: biochemistry and genetics of virulence . Curr Topics Microbiol. and Immunol. 57:111-158. 10. ll. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 81 Brubaker, R. R. 1969. Mutation rate to nonpigmentation in Pasteurella pestis. .J. Bacteriol. 98:1404-1406. Butler, T. 1983. Plague and Other Yersinia Infections. Plenum Medical Book Co., N.Y. Carniel, E., J.-C. Antoine, A. Guiyoule, N. Guiso, and H. 1L. Mollaret. 1989. Purification, location, and immunological characterization of 13m; iron-regulated high molecular weight proteins of the highly pathogenic yersiniae. Infect Immun. 57:540-545. Carniel, E., O. Mercereau-Puijalon, and S. Bonnefoy. 1989. The gene coding for the 190,000-dalton iron-regulated protein of Yersinia species is present only in the highly pathogenic strains. Infect. Immun. 57:1211-1217. Carniel, E., D. Mazigh, and H. H. Mollaret. 1987. Expression of iron-regulated proteins in Yersinia species and their relation to virulence. Infect. Immun. 55:277280. Casse, F., C. Boucher, J. S. Julliot, M. Michel, and J. Denarie. 1979. Identification and characterization of large plasmids in Rhizobium meliloti using agarose gel electrophoresis. J. Gen. Microbiol. 113:229-242. Crosa, J. H. 1989. Genetics and molecular biology of siderophore-mediated iron transport :hi bacteria. Microbiol. Rev. 53:517-530. Daniels, D. I“ 1990. The complete AvrII restriction map (Hi the Escherichia coli genome and comparisons of several laboratory strains. Nuc. Acids Res. 18:2649-2651. Devignat, R. 1951. Varietes de l'espece Pasteurella pestis. Nouvelle hypothese. Bull. Wld. Hlth. Org. 4:247263. Ferber, D. M. and R. R. Brubaker. 1981. Plasmids in Yersinia pestis. Infect. Immun. 31:839-841. Girard, G. 1955. Plague. Ann. Rev. Microbiol. 9:253- 276. Grothues, D. and B. Tummler. 1987. Genome analysis of Pseudomonas aeruginosa by field inversion gel electrophoresis. FEMS Micro. Letts. 48:419-422. Hacker, J3, IL. Bender, M. Ott, J. Wingender, B. Lund, 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 82 R. Marre, and W. Goebel. 1990. Deletions of chromosomal regions coding for fimbriae and hemolysins occur in vitro and in vivo in various extraintestinal Escherichia coli isolates. Microb. Pathog. 8:213-225. Hirst, L. F. 1953. The Conquest of Plague. Clarendon Press. Oxford. Jackson, S. and T. W. Burrows. 1956. The pigmentation of Pasteurella pestis on a defined medium containing haemin. Brit. J. Exp. Path. 37:570-576. Jackson, S. and T. W. Burrows. 1956. The virulence- enhancing effect of iron on non-pigmented mutants of virulent strains of Pasteurella pestis. Brit. J. Exp. Path. 37:577-583. Knapp, S., J. Hacker, T. Jarchau, and W. Goebol. 1986. Large, unstable inserts in the chromosome affect virulence properties of uropathogenic Escherichia coli 06 strain 536. J. Bacteriol. 168:22-30. Knapp, S., J. Hacker, I. Then, D. Miller, and W. Goebel. 1984. Multiple copies of hemolysin genes and associated sequences in the chromosome of uropathogenic Escherichia coli strains. J. Bacteriol. 159:1027-1033. Krawiec, S. and M. Riley. 1990. Organization of the bacterial chromosome. Microbiol. Rev. 54:502-539. Lawlor, K. M., and S. M. Payne. 1984. Aerobactin genes in Shigella spp. J. Bacteriol. 160:266-272. Leblond, P., P. Demuyter, J. Simonet, and B. Decaris. 1991. Genetic instability and associated genome plasticity in Streptomyces ambofaciens: pulsed-field gel electrophoresis evidence of large DNA alterations in a limited genomic region. J. Bacteriol. 173:4229-4233. Leblond, P., F. X. Francou, J-M. Simonet, and B. Decaris. 1990. Pulsed-field gel electrophoresis analysis of the genome of Streptomyces ambofaciens strains. FEMS Micro. Letts. 72:79-88. Lee, J. J. and H. O. Smith. 1988. Sizing of the Haemophilus influenzae Rd genome by pulsed-field agarose gel electrophoresis. J. Bacteriol. 170:4402-4405. Levene, S. B. and B. H. Zimm. 1987. Separations of open circular DNA using pulsed-field gel electrophoresis. Proc. Natl. Acad. Sci. USA 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 83 84:4054-4057. Lipson, L. G. 1972. Plague in San Francisco in 1900. Annals Int. Med. 77:303-310. Maniatis, T., E. F. Fritsch, and J. Sambrook. 1982. MolecuLar cloning: a laboratory manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y. Marmur, J., S. Falkow, and 14. Mandel. 1963. New approaches to bacterial taxonomy. Ann. Rev. Microbiol. 17:329372. McClelland, M., R. Jones, Y. Patel, and M. Nelson. 1987. Restriction endonucleases for pulsed field mapping of bacterial genomes. Nuc. Acids Res. 15:5985-6005. McDougall, S. and J. B. Neilands. 1984. Plasmid and chromosome-coded aerobactin synthesis in enteric bacteria: insertion sequences flank; operon in plasmid-mediated systems. J. Bacteriol. 159:300-305. Pendrak, M. L., and R. D. Perry. 1991. Characterization of a hemin-storage locus of Yersinia pestis. Biol. Metals 4:41-47. Perez-Casal, J. F. and J. H. Crosa. 1984. Aerobactin iron uptake sequences in plasmid ColV-K30 are flanked by inverted ISl-like insertion elements and replication regions. J. Bacteriol. 160:256-265. Perry, R. D., M. L. Pendrak, and P. Schuetze. 1990. Identification and cloning of a hemin storage locus involved in the pigmentation phenotype of Yersinia pestis. J. Bacteriol. 172:5929-5937. Protsenko, O. A., P. I. Anisimov, O. T. Mosarovc, N. P. Donnov, Y. A. Popov, and A. M. Kokushkin. .1983. Detection and characterization of Yersinia pestis plasmids determining pesticin I, fraction I antigen and mouse toxin synthesis. Genetika 19:1081-1090. Ramling, [LW 1). Grothues, W. Bautsch, and B. Tummler. 1989. A physical map of the genome of Pseudomonas aeruginosa PAO. EMBO. 8:4081-4089. Schaffer, H. E. and R. R. Sederoff. 1981. Improved estimation of DNA fragment lengths from agarose gels. Anal. Biochem. 115:113-122. Sikkema, D, J. and R. R. Brubaker. 1989. Outer membrane peptides of Yersinia pestis mediating siderophore-independent. assimilation cfif iron. Biol. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 84 Metals 2:174-184. Sikkema, D. J. and R. R. Brubaker. 1987. Resistance to pesticin, storage of iron, and invasion (n3 Hela cells by yersiniae. Infect. Immun. 55:572-578. Smith, 1:. L. and G. Condemine. 1990. New approaches for the physical mapping of small genomes. J. Bacteriol. 172:1167-1172. Smith, C. L., J. G. Econome, A. Schutt, S. cho, and C. R. Cantor. 1987. A physical map of the Escherichia coli K12 genome. Science 236:1448-1453. Smith, C. L., S. cho, and C. R.Cantor. 1988. Pulsed field gel electrophoresis and the technology of large DNA. molecules, p. 41-72. 131 K. Davis (ed) Genome analysis: a practical approach. IRL Press, Oxford, England. Sodeinde, O. A. and J. D. Goguen. 1988. Genetic analysis of the 9.5-kilobase virulence plasmid of Yersinia pestis. Infect. Immun. 56:2743-2748. Staggs, T. M. and R. D. Perry. 1991. Identification and cloning of a fur regulatory gene in Yersinia pestis. J. Bacteriol. 173:417-425. Straley, S. C. and R. R. Brubaker. 1981. Cytoplasmic and. membrane proteins of yersiniae cultivated under conditions simulating mammalian intracellular environment. Proc. Natl. Acad. Sci. USA 78:1224-1228. Surgalla, M. J. and E. D. Beesley. 1969. congo red plating medium for detecting pigmentation in Pasteurella pestis. Appl. Microbiol. 18:834-837. Une, T. and R. R. Brubaker. 1984. IQ vivo comparison of avirulent Vwa- and Pgm- or Pst phenotypes of yersiniae. Infect. Immun. 43:895-900. Wee, S., J. B. Neilands., M. L. Bittner, B. C. Hemming, B. L. Haymore, and R. Seethram. 1988. Expression, isolation, and properties of fur (ferric uptake regulation) protein of Escherichia coli. Biol. Metals 1:62-68. Zahorchak, R. J., and R. R. ubaker. 1982. Effect of exogenous nucleotides on Ca dependence and V antigen synthesis in Yersinia pestis. Infect. Immun. 38:953- 959. CHAPTER I I (ARTICLE) Iron and Hemin Storage in Yersinia pestis by Robert D. Perry, Thomas S. Lucier, and Robert R. Brubaker (Manuscript to be submitted for publication) 86 ABSTRACT The pigmentation (Pgm+) phenotype,a required virulence determinant of Yersinia, pestis, encompasses 21 number «of characteristics that now appear to be genetically but not necessarily physiologically linked. We now use hemin storage (Hms+) phenotype to define a Pgm+ trait involved in the temperature-dependent storage of exogenous hemin. In this study, we have identified the outer membrane as the site of iron storage which occurs only at 26°C in Pgm+ cells of L pestis grown with hemin as the iron source. Outer membrane storage of hemin does not occur in spontaneous Pgm- cells nor in Pgm+ cells cultured at 37°C. We also identified a soluble inorganic iron storage pool in both Pgm+ and Pgm- cells. Expression of inorganic iron storage is temperature-independent and is associated with a bacterioferritin-like protein. At 37°C, these iron and hemin storage pools are relatively independent of each other. While the function(s) of these storage pools is undetermined, they may play important roles in the pathogenesis of the plague bacillus. 87 INTRODUCTION Although procaryotes must generally survive and grow in iron-deficient environments, pathogens living vdiflflx1 host tissues may encounter potentially iron-rich conditions. However, this iron will be bound to host molecules including high affinity ligands such as transferrin and lactoferrin or hemin (from hemoglobin, hemopexin, and haptoglobin) and ferritin (7,24,49). One possible iron rich environment is the flea gut (7) which is the site of multiplication for part of the life cycle of Yersinia pestis, the etiologic agent of bubonic and pneumonic plague (8). Hemolyzed blood would provide an abundance of hemin compounds and inorganic iron. ‘When grown in laboratory conditions simulating this environment (26°C) virulent cells of L pestis accumulate sufficient levels of exogenous hemin (20,21) or Congo red (44) from solidified media to form colored or "pigmented" (Pgm+) colonies. Spontaneous Pgm- mutants fail to accumulate these compounds, however, they still utilize all tested hemin compounds as nutritional sources of iron (33,36,37). Thus one defect in spontaneous Pgm- organisms lies in hemin storage but not in hemin utilization. While hemin storage appears to be rare in prokaryotes, inorganic iron storage on the cytoplasmic protein bacterioferritin may be widespread (45). This iron-storage 88 protein has been detected in six diverse genera: Azotobacter (6,11,39), Escherichia (1,50), Pseudomonas (29), Rhodopseudomonas (28), Rhodospirillum (2), and Streptomyces (15). The bacterioferritins from the first three genera have been more extensively characterized and show striking similarities to eucaryotic ferritins in subunit structure, consisting of multiple copies of a 15-18 kDa polypeptide, and in their nonhemin iron storage prOperties (1,27,39). Unlike eucaryotic ferritins, bacterioferritins contain one hemin moiety for every 2-5 subunit polypeptides. Although no function for bacterioferritins has been proven, iron storage could prevent formation of toxic radicals and provide iron for growth under subsequent iron starvation conditions (1). In this study' we examined hemin. and. inorganic .iron storage properties in isogenic Pgm+ and Pgm- cells of L pestis KIM. Hemin storage occurred primarily in the outer membranes cfif Pgm+ cells while storage of excess inorganic iron occurred in a soluble fraction of both Pgm)r and Pgm- cells. Peptides associated with the inorganic iron storage pool may represent a Y; pestis bacterioferritin. Under iron excess conditions incorporation of inorganic iron and iron from hemin appeared to be independent processes and incorporation from one source did not greatly repress incorporation from the other. 89 MATERIALS AND METHODS . . . + - . . Bacteria. An isogenic Pgm and Pgm pair of X; pestis KIM derivatives were used in this study. The Pgm- mutation results from the spontaneous deletion of approximately 102' kb of chromosomal DNA (16,26). Both derivatives possess endogenous toxin (me1) and pesticin plasmids, but are avirulent due to the absence of the low-calcium response (Lcr) plasmid (pCDl) (5,36,46). The Pgm+ determinant and the Lcr virulence regulon are genetically and biochemically unrelated (5,34). Cultivation and labelling of bacterial cells. Bacterial strains were stored at ~200C in buffered glycerol (3). The Congo red (CR) agar of Surgalla and Beesley (44) was used to test for the Pgm+ phenotype. Cells of Y; pestis were grown with aeration (200 rpm setting on a New Brunswick Model G76 gyratory shaker water-bath) in the synthetic medium of Higuchi et al (18) as modified by Zahorchak and Brubaker (51). In this study FeSO4 was omitted from the medium which was further deferrated by 8-hydroxyquinoline extraction (48) (hereafter called modified Higuchi ' 5 medium). Hemin and FeCl3 supplements were added after deferration. Glycerol stocked cells were inoculated onto tryptose blood agar (Difco, Detroit, MI) slants and 90 incubated for 48 h at 260C. Slant cultures were suspended in 33 mM phosphate buffer (pH 7.0) and used to inoculate modified Higuchi's medium (supplemented with 100 uM FeCl3 or 50-87 uM hemin) to an optical density (OD) of 0.1 at 620 nm. Cells were acclimated to the medium by growth at 26°C or 37°C by serial transfer for approximately eight generations (33). Acclimated cells were transferred to fresh modified Higuchi's medium and mid-log phase cells harvested for analysis. Growth of cultures was determined by OD620 measurements. Cells were labelled to a constant specific activity 55 with either FeCl (New England Nuclear Research Products 3 (NEN), Boston, MA) or (SSFe)hemin by addition of the isotope to modified Higuchi's medium. The radioactive medium was sterilized by filtration and used for cell cultivation as described above. The final concentration and specific activities in radioactive media were 100 uM with 13,000 cpm 55Fe per nmol inorganic Fe or 87 uM hemin with 1,700 cpm (55 Fe)hemin per nmol. Cell-associated radioactivity was quantitiated by filtration (0.22 um pore size, Millipore Corp., Bedford, MA) of approximately 109 CFU's followed by membrane washing’ with ice-cold nonradioactive ‘medium .and scintillation counting. Cell fractionation. Mid-log phase cells were fractionated into periplasmic, cytoplasmic, outer membrane, and inner membrane components as previously described 91 (41,42). Briefly, lysozyme-EDTA treatment and centrifugation separated spheroplasts from periplasm. Spheroplasts were lysed by sonication followed by successive isopycnic sucrose gradient centrifugations to separate cytoplasm, outer membranes and inner membranes. Although the sucrose density outer membrane banding pattern was significantly lower in all cells grown at 26°C, SDS-PAGE polypeptide profiles showed these fractions were authentic, isolated outer membranes (data not shown). Radioactivity associated with each cell fraction was determined by scintillation spectroscopy of appropriate aliquots. To analyze soluble cell components, mid-log phase cells were harvested by centrifugation, washed with 33 mM phosphate buffer (pH 7.0), and resuspended in 50 M N- 2-hydroxyethylpiperazine- N'-2'ethanesulfonic acid (HEPES) - 1.0 mM sodium citrate (pH 7.8, hereafter called column buffer). Cell suspensions were sonicated. on ice in .15 second bursts for 1 minute. Cellular debris and particulate membrane fragments were removed by centrifugation (10,000 x g, 15 minutes, 4°C). Protein concentrations were determined by the method of Lowry et al (25) and all samples diluted to 10 mg protein/ml. One ml samples were applied to Bio-Gel A-l.5m (100-200 mesh, Bio-Rad Laboratories, Richmond, CA) for molecular sieving. Samples were held at 4°C and eluted with column buffer at a flow rate of 0.15 ml/min. Protein Concentrations of column fractions (3 ml volumes) were estimated by OD280 measurements. Radioactivity of column 92 fractions was determined by drying 1.0 ml aliquots in scintillation vials overnight at 800C, adding 10 m1 Complete Counting Cocktail (Research Products International. Corp., Mount Prospect, IL), and quantitating cpm on a LS7500 Scintillation System (Beckman Instruments, Inc. Fullerton, CA). Identification of the soluble iron storage (peptide. Fractions eluting from the Bio-Gel A-l.5m column at an elution volume of approximately 270 ml displayed the highest radioactive peak under' most conditions (Fig. 1). These fractions from Pgm- cells grown at 37°C with FeCl3 were pooled for further analysis. Pooled fractions were dialyzed overnight with three changes of buffer against 50 mM Tris-HCl pH 7.8 and then loaded onto a DEAE cellulose column (Whatman Biosystems, Ltd., Maidstone, England) at room temperature. Proteins were eluted from the column using a 0 to 0.5 M NaCl gradiant in 50 mM Tris-HCl (pH 7.8) over a period of two hours with a flow rate of 1.0 ml per minute. Five m1 fractions were collected and their optical density and radioactivity were measured as described above for the A-l.5m column. Proteins in each fraction were identified by combining 25 1L1 from each fraction with an equal volume of SDS-PAGE sample buffer. These were used for SDS-PAGE and subsequent silver staining (23,30). 55 55 ( Fe)Hemin synthesis. ( Fe)Hemin was synthesized 93 from 55FeSO4 (NEN) and protoporphyrin IX (Sigma Chemical Co., St. Louis, MO) essentially by the methoed of Chang et a1. (10). Briefly, iron was incorporated into the prophyrin ring by incubation in a liitrogen atmosphere at 80°C in 55Fe)Hemin was pyridine-glacial acetic acid solvent. ( separated from reactants by chloroform extractions and acid precipitation (10). Analysis of authentic hemin (Sigma 55Fe)hemin product revealed virtually Chemical Co.) and our ( identical visible absorption spectra (data not shown). Hemin concentrations were determined spectrophotometrically at 580 nm against a standard' curve constructed with authentic hemin. The term hemin is used throughout this report as a generic term for iron-containing protoporphyrin IX without regard to oxidation of salt states. 94 RESULTS Subcellular localization of hemin and iron pools. The amount of label removed from the culture medium during growth was influenced by temperature, bacterial strain, and the form in which the label was present. All cells grown in the presence of inorganic iron removed less than 5% of the label from the medium. Cells grown with labelled hemin removed 19-25% with the exception of Pgm+ cells at 260C which removed approximately 85% of the hemin. Table 1 .reports the storage of inorganic (n: hemin- associated iron within subcellular fractions in specific activities (nmoles of iron/mg protein). Although several differences are notable, Pgm+ cells of l; pestis KIM grown at 26°C with hemin clearly stored excess iron in their outer membranes. Under these conditions outer membranes contained 51-fold more iron than Pgm- cells and 317-fold more iron than Pgm+ cells grown at 370C. Inner membranes and periplasms of Pgm+ cells grown at 26°C contained considerably more iron than these fractions from Pgm- cells grown under the same conditions. The large amount of hemin accumulated on the outer membrane could result in :more passing into time periplasm and inner membrane, or it could represent low level contamination. of ‘these samples with outer membranes during the fractionation procedure. 95 H>urm w. mvnnwmnn >0nnonwcnnnnac mccnnwwcwnn wma+ nowwm man- onHHm mnmnnnosu ~m.0 nnocn: u~.o nnozn: nm.o nnotn: uu.n Quota: no“: :63»: maul. :63»: no“: :83»: wow... :69»: ocnmn :navneomm oo.m mow.» .H.m ~.~ m.o pm.» Hm.a u.o Hanan :navnnsnm m.m aw.» m.w c.u u.w u.» u.» a.u 863366.. 5.» 8.4. :2. 38 up . m... 6.4. 5.4 awnovwmua m.~ u.~ H~.u u.c ~.a H.m a.u ~.u m nnwwm some ocannno on ~m.n on uu.n no a nosmnmsn momnwmpn mnnp0nw<. w wmoo1H.ma my o.ou H.mw w.~ Haw.~ pom. m» om>m Nw o.om H.~m m.o Hpq.w ppm.o om w. monomsnmom 0m coasnm M: nsm chemo mopcopm ome mxnnmon mnwpw pummman w: «so mmapwm. 104 Figure 3. Silver-stained 12.5% SDS-PAGE gel of pooled column chromatography fractions containing the soluble inorganic iron pool. Analysis of Pgm- cells grown with 55FeCl3 at 370C showing crude soluble cell extract (lane A), radioactive peak from Bio-Gel A-l.5m column at about 270 ml elution volume (lane B), and radioactive peak from DEAE cellulose column eluting at 0.35 M NaCl (lane C). Numbers on the left are the positions of molecular weight standards in kDa. The arrows indicates the locations of the presumptive bacterioferritin subunits. 1.: Nagy-’11;- ' " 1. 105 Figure 3 106 excess conditions, Pgm+ and Pgm- cells were labelled to a constant specific activity by growth for approximately eight generations at 370C in modified Higuchi medium with 55FeCl 55 3 or ( Fe)hemin. These cells were used to inoculate two separate cultures containing the same radiolabelled iron source with or without the other iron source, unlabelled and in excess. If there is an interaction between the two forms of iron pools, accumulation of radioactive label in cultures containing both iron sources should diverge from those with a single iron source. iEor both Pgm+ and Pgm- cells grown 55 with FeCl the addition of 5-fold excess unlabelled hemin 3! did not inhibit the continued accumulation of inorganic iron 55Fe)hemin, (Fig. 4). However, for cells grown with ( addition of 5-fold excess unlabelled inorganic iron resulted in an initial release of some accumulated hemin without significantly affecting subsequent hemin acquisition. This caused a lower specific activity in both Pgm+ and Pgm- cells exposed to both iron sources (Fig. 5). Thus inorganic iron may have an initial slight inhibitory affect on hemin accumulation without significantly affecting the continued rate of acquisition. 107 Figure 4. 370C growth and 55FeCl3 accumulation by Pgm+ (panels A and B) and Pgm- (panels C and D) cells of L pestis. Cells were grown to constant. specifit: activity prior to transfer to medium containing 50 uM 55FeCl3 alone (panels A and C) or with 250 uM unlabelled hemin (panels B and D). Open circles are cell growth measured by optical density at 620 nm. Closed circles are nmoles of accumulated iron per ml of culture. 108 Optical Density I'O .‘ O IU'I 930391 OI 9 0 I . O h d o w.- - . .1 heal a a nllnanl 1 LLLAIAII ~12. - O 3‘.‘ 0' d a':. 0 d OI gels. 8 I '° 13'1": O 0| -1 A 1 121411 1 1 LAAAIAI . 11:6. 3 o - 0| .- d O " ‘ 0| ' 3 -‘ M m a a ILAILAI 1 4 1116-11 . 980891019 0 55 3+ Fe (nmoles/ml) Figure 4 109 Figure 5. 370C growth and (55)hemin accumulation by + .— Pgm (panels A and B) and Pgm (panels C and D) cells of X; pestis. Cells were grown to constant. specific: activity prior to transfer to medium containing 20 uM (55)hemin alone (panels A and C) or with 1 mM unlabelled FeCl (panels B and 3 D). Open circles are cell growth measured by optical density at 620 rmu Closed circles are nmoles of accumulated iron per ml of culture. Figure 5 smoH 9303 9I OI 9 0 110 Optical Density 93039IOI 9 0 11111 9ZOZ9I OI 9 O Vi‘ii 1 All 93039I OI 9 O Uli'll' U. lllll . n lllllll I 4 Ill . . d .. o 9 O [55F9]hemh (nmoles/ml) ~, lll DISCUSSION Jackson and Burrows (20,21) first described the pigmentation phenotype as £1 required virulence determinant identified by the accumulation of sufficient exogenous hemin to form pigmented colonies at 26°C but not at 370C. The characteristics linked to the Pgm+ phenotype now include sensitivity to the bacteriocin pesticin, (4) growth at 37°C in an iron-chelated medium, (35) expression of unique iron-regulated outer membrane polypeptides (30), and expression of unique Hms-specific peptides (32,36,42). Spontaneous Pgm- cells lose all of the above characteristics due to a 102 kb chromosomal deletion (16,26). It is now apparent that the Pgm+ phenotype of L pestis consists of numerous, separable, and independent traits and we use the term hemin storage (hms) to specifically refer to the physiological trait of pigmented colony formation at 26°C in the presence of exogenous hemin or Congo red. The original observation by Jackson and Burrows (20) demonstrated that hemin molecules were retained for storage without apparent removal of inorganic IJIWh We have localized the site of hemin storage in Pgm+ cells to the outer membrane and shown that it is restricted to cells grown at 260C presented with hemin as the iron source. A number of Hms-specific outer membrane polypeptides have now been identified that may be 112 involved in hemin binding (32,36,42). The function of the Hms+ phenotype in L pestis is unknown. The extent to which hemin is removed from the medium by Pgm+ cells at 260C shows that hemin storage is enormous and raises several possible roles for the expression of such a system. The thermal regulation of this system suggests the flea gut as the environment where this phenotype would be highly expressed. Although several Hms- specific proteins are expressed at low levels at 37°C (32), the time period that stored hemin remains absorbed following shift up to 37°C upon entry into the mammalian host is unknown. However, any putative function in the mammalian environment would likely occur soon after entry; 'This is also suggested by the finding that Pgm- cells are only avirulent via pmmipheral (subcutaneous or intraperitoneal) routes of injection and retain full virulence if introduced intravenously (46). Iron stored during growth in the hemin-rich environment of the flea gut may be used to allow rapid initial multiplication prior to induction of systems capable of obtaining inorganic iron or hemin from mammalian sources. Storage may simply prevent free hemin from participating in generation of damaging oxygen radicals in either the flea or the mammalian host. A recent study suggests that the hemin storage aspect (32,34) of the Pgm+. phenotype may be necessary for long-term survival in the flea and for blockage of the flea proventriculus (22). Alternatively, hemin molecules on the surface of X; pestis 113 may facilitate its uptake by eucaryotic cells. Such a function has been proposed for Shigella flexneri where the CR+ phenotype correlates with increased ability to invade and infect HeLa cells (13,43) but not with hemin transport and nutritional utilization (31). However, the physiological characteristics and regulation of Cr+ phenotype of S_. flexneri is expressed at 370C, leaving the degree of functional and genetic similarities between these systems unresolved. A final putative function for hemin storage in L pestis is the inhibition of a variety of nonspecific host defenses (47) by an array of surface hemin molecules. Although none of the above putative functions are mutually exclusive, evidence for the validity of any of these functions awaits future experimentation. We speculate that the inorganic iron storage pool may represent :3 1; pestis bacterioferritin whose expression is independent of growth temperature and pigmentation phenotype. Native size, subunit molecular weight, behavior on DEAE and the relative amount of cytoplasmic iron associated with the molecule in cells grown with FeCl3 all closely match results for bacterioferritins from other species (l,2,ll,27,28,29,39,45). Bacterioferritins are often difficult to purify by traditional means and some minor contaminants remain in our preparations. Future work should result in highly purified preparations which will be used to analyze hemin content, iron binding capacity and absorption spectra data to help confirm our tentative 114 identification of this molecule as bacterioferritin. Although the purpose of bacterioferritins is unproven, stored iron in L pestis may allow rapid multiplication after injection into the mammalian host prior to expression of iron acquisition systems which can extract iron or hemin from host ligands. However, the relatively low amount of storage on this component when cells were grown with hemin at 26°C seems inconsistent with this idea. Alternatively, this pool may simply prevent excess free iron from damaging the bacterial cell through generation of oxygen radicals (45). While cells of L pestis can utilize either inorganic iron or hemin as sole sources of nutritional iron (33,36,37), our results indicate that accumulation of one form does not have a significant inhibitory effect on the accumulation of the other. Utilization of either inorganic iron or hemin compounds suggests that L pestis cells should possess enzymatic activities for inorganic iron insertion into and removal from porphyrin rings. Such activities would lead to a functional linkage between the hemin and inorganic iron pools. If such a linkage exists, it was not evident from the hemin and inorganic iron competition experiments performed here under iron and hemin surplus conditions. These results also suggest separate uptake mechanisms for hemin bound iron and inorganic iron. Results of the studies of the interaction of the iron pools, the preferential accumulation of hemin iron in fractions from 115 the Al.5m columns containing an enormous hemin binding protein at 37°C, and the lower specific activities of the presumptive bacterioferritin fractions from the Al.5m columns of cells grown with hemin all indicate a mechanism for uptake of exogenous hemin without prior removal of iron, and possibly delayed removal of iron from the hemin even after internalization. Other bacteria have been shown to employ hemin specific uptake systems which allow them to use hemin as EM) iron source (12,14,16,17,33,40). This would account for the inability of inorganic and hemin-iron sources to interfere with each other's uptake into X; pestis cells, however, the apparent continued separation of these pools following uptake may occur only under iron-surplus conditions. The significance of these observations and their relevance to growth in vivo is uncertain. Since cells were grown to constant specific activity intracellular iron availability would be high and it is unlikely that high affinity transport mechanisms which are probably expressed in the mammalian host would be operational (10,36,37,38). Under conditions of iron deprivation encountered in the host there may be greater unification of the iron pools following uptake. lo 11 116 REFERENCES Andrews, S. C., P. M. Harrison, and J. R. Guest. 1989. Cloning, sequencing and mapping of the bacterioferritin gene (bfr) of Escherichia coli K12. J. Bacteriol. 171: 3940-3947. Bartsch, R. G., T. Katuno, T. Horio, and M. D. Kamen. 19714 Preparation and properties of Rhodospirillum rubrum cytochromes c , cc', b5 7 5, and flavin Chem. mononucleotide protein. .J. Biol. 246: 4489- 4496. Beesley, EL 1)., R. R. Brubaker, W. A. Janssen, and M. J. Surgalla. 1967. Pesticins III. Expression cd' coagulase and mechanism of fibrinolysis. J. Bacteriol. 94: 19-26. Brubaker, Robert R" 1969. Mutation rate to nonpigmentation in Pasteurella pestis. .3. Bacteriol. 98: 1404-1406. Brubaker, R. IL. 1983. The Vwa+ virulence factor of Yersiniae: The molecular bafigs of the attendant nutritional requirement for (21 . Rev. Infect. Dis. 5: suppl. 4: $748-$758. Bulen, W. A., J. R. LeComte, and S. Lough. 1973. A hemoprotein from Azotobacter containing non-heme iron: Isolation and crystallization. Biochem. Biophys. Res. Commun. 54: 1274-1281. Bullen, J. J. 1981. The significance of iron in infection. Rev. Infect. Dis. 3: 1127-1138. Butler, T. 1983. Plague and other Yersinia infections. N. Y.: Plenum Press. Carniel, E., D. Mazigh, and H. H. Mollaret. 1987. Expression of iron-regulated proteins in Yersinia species and their relation to virulence. Infect. Immun. 55: 277-280. Chang, C. K., R. K. DiNello, and D. Dolphin. 1980. Iron porphines. Inorg. Synthesis. 20: 147-155. Chen, M. and R. R. Crichton. 1982. Purification and characterization of a bacterioferritin from Azotobacter chroococcum. Biochim. Biophys. Acta. 707: 1-6. 12. 13. 14. lSe 1.6. 22L 117 Coulton, J. W. and J. C. S. Pang. 1983. Transport of hemin by Haemophilus influenzae type b. Curr. Microbiol. 9: 93-98. Daskaleros, P. A. and S. M. Payne. 1987. Congo red binding phenotype is associated with hemin binding and increased infectivity of Shigella flexneri in the HeLa cell model. Infect. Immun. 55: 1393-1398. Daskaleros, P. A., J. A. Stoebner, and S. M. Payne. 1991. Iron uptake 1J1 Plesiomonas shigelloides: Cloning of time genes for the heme-iron uptake system. Infect. Immun. 59: 2706-2711. Dyer, D. W., E. P. West, and P. F. Sparling. 1987. Effects of serum carrier proteins on the growth of pathogenic neisseriae with heme-bound iron. Infect. Immun. 55: 2171-2175. Fetherston, J. D., P. Schuetze, and R. D. Perry. 1992. Loss of the pdgmentation phenotype in Yersinia pestis. is due to the spontaneous deletion of 102 kb of chromosomal DNA which is flanked by a repetitive element. Mol. Microbiol. submitted. Helms, S. EL, .1. D. Oliver, and J. C. Travis. 1984. Role of heme compounds and haptoglobin in Vibrio vulnificus pathogenicity. Infect. Immun. 45: 345- 349. Higuchi, K., L. L. Kupferberg, and J. L. Smith. 1959. Studies on the nutrition and physiology of Pasteurella pestis: III. Effects of calcium ions on the growth of virulent and avirulent strains of Pasteurella pestis. J. Bacteriol. 77: 317-321. Inoue, Y. and H. Kubo. 1965. The metabolism of Streptomyces griseus. cytochrome b. Biochim. Biophys. Acta. 110: 57-65. Jackson, S. and T. W. Burrows. 1956. The pigmentation of Pasteurella pestis on a defined medium containing haemin. Brit. J. Exp. Pathol. 37: 570-576. Jackson, S. and T. W; Burrows. 1956. The virulence- enhancing effect of iron on non-pigmented mutants of virulent strains of Pasteurella pestis. Brit. J. Exp. Pathol. 37: 577-583. Kutyrev, V. V., A. A. Filippov, O. S. Oparina, and O. A., Protsenko. 1992. An lysis of Yersinia pestis chromosomal determinants Pgm and Pst associated with virulence. Microbial Path. 12: in press. 23.. 24 - 25.. 26- 27- 28- 29. 30. 31. 32. 33. 34.. 118 Laemmli, U. K. 1970. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature (London). 227: 680-685. Lankford, C. E. 1973. Bacterial assimilation of iron. CRC Crit. Rev. Microbiol. 2: 273-331. Lowry, O. H., N. J. Rosebrough, A. L. Farr, and R. J. Randall. 1951. Protein measurement with the folin phenol reagent. J. Biol. Chem. 193: 265-275. Lucier, T. S. and R. R. Brubaker. 1992. Determination of genome size, macrorestriction pattern polymorphism, and non-pigmentation-specific deletion in Yersinia pestis by Pulsed-Field Gel electrophoresis. J. Bacteriol. 174: Mann, 8., J. M. Williams, A. Treffry, and P. M. Harrison. 1987. Reconstituted and native iron cores of bacterioferritin and ferritin. J. Mol. Biol. 198: 405-416. Meyer, T. E. and M. A. Cusanovich. 1985. Soluble cytochrome composition of the purple phototrophic bacterium Rhodopseudomonas sphaeroides ATCC 17023. Biochim. Biophys. Acta. 807: 308-319. Moore, G. K. S., S. Mann, and J. V. Bannister. 1986. Isolation and properties of the complex nonheme- .iron-containing cytochrome b5 7 (bacterioferritin) from liseudomonas aeruginosa. J. fborg. Biochem. 28: 329- .336. Ddorrisey, J. IL. 1981. Silver stain for proteins in E>olyacry1amide gels: .A modified procedure with eenhanced uniform. sensitivity. AnalJ Biochem. 117: 307-310. IPayne, S. PL 1989. Iron and virulence in Shigella. D401. Microbiol. 3: 1301-1306. IPendrak, M. L. and R. D. Perry. 1991. (Zharacterization of a: hemin-storage locus of Yersinia Egestis. Biol. Metals. 4: 41-47. I?erry, R. D. and R. R. Brubaker. 1979. Accumulation <2iated with the Pgm+ phenotype and not hemin binding may be SnKb-PnRUF.UIv Figure 6 gel nest _— 154 Figure 7. Comparison of peptides present in a stained gel (12% SDS-PAGE) of purified outer membrane proteins of X; pestis KIM grown at 37°C (lanes A-D) and 26°C (lanes E- H). Pgm+ grown in iron replete medium (lanes A and E), Pgm+ cells grown in iron deficient medium (lanes B 1and .F), Pgm+,Pstr cells grown in iron deficient medium (lanes C and G), and Pgm- cells grown in iron deficient medium (lanes D and H). Molecular weight markers are in the unlabelled middle lane and their sizes in kilodaltons are indicated along the left edge of the gel. Positions of detectable Irps (Table 1) are indicated along the right edge. The open arrow indicates expression only at 26°C, and asterisks show peptides not expressed by Pgm- cells. 155 116 97.4 .p 66.2 45.0 21.5 14.4 Figure 7 156 gels (Fig. 7,14). We were unable to confirm the presence of Irp E (65 kDa, isoelectric point = 5.98). At least eight more irps unrelated to pigmentation were identified (Fig. 6,7,11). The 36, 34 and 28.5 kDa peptides (Irps G, H, and I) were best seen in two-dimensional gels (Fig. 6,11), and the latter two (Irps H,I) were expressed only at 260C. The Irp Dd group included two or three peptides only visible in two-dimensional gels (isoelectric point = 6.2 - 6.6, Fig. 6,11). We confirmed previous observations that the HMWP's are visible in lane gels of outer membrane samples, but are not seen in two-dimensional gels (7,8,44). This would occur if the isoelectric points of these peptides were outside the 4.5 - 7.0 range of our analysis. Our results indicated that HMWP's are not restricted to the outer membrane as was previously suggested (6). HMWPs were seen in all fractions from Pgm+ and Pgm+, Pstr cells at both temperatures. Since most peptides could be clearly assigned to a: particular location it is unlikely to reflect poor separation of fractions during sample preparation. Twelve iron repressible peptides including the two HMWP's were found in periplasm samples (Table 1, Fig. 8). Several were among the nmst strongly expressed proteins in this compartment. Pirps G and J were only expressed at 26°C. A new Pgm+ specific peptide was identified at 19 kDa (Pirp L). It was clearly visible at 26°C, however, at 370C there was a slightly larger peptide expressed in enormous mmmq 157 Figure 8. Comparison of peptides present in a stained gel (12% SDS-PAGE) of purified periplasmic proteins of _Y_. pestis KIM grown at 370C (lanes A-D) and 260C (lanes E- H). Pgm+ cells grown in iron replete medium (lanes A and E), Pgm+ cells grown in iron deficient medium (lanes B and F), Pgm+,Pstr cells grown in iron deficient medium (lanes C and G), and Pgm- cells grown in iron deficient medium (lanes D and H). Positions of molecular' weight standards in kilodaltons are between the gels. Positions of detectable Pirps (Table 1) are indicated along the right edge. Open arrows indicate expression only at 260C, and asterisks show peptides not expressed by Pgm- cells. 158 M '~ - : ‘ W a} ‘ .4... V.- -—* cal-q} N «we 4» in " ‘ s Figure 8 200 116 97.4 66.2 45.0 31.0 14.4 < * ‘ be homologous with Irps B-D, G, and the M group. Irp A was present and expressed. at :much higher levels than in Y. pestis (Fig. 12,14), but Irp F was absent (Fig. 14). X; enterocolitica was more unique. Although it expressed peptides apparently homologous to Irps B-D, the peptide which could be Irp A had a slightly lower isoelectric point (Fig. 13). No peptides corresponding to Irp G or the M group were detectable, and unique Irps of 85 kDa and 67 kDa were expressed (Fig. 13,14). Carniel et a1 (8) described Irps of 240, 190, 89, 81, 79, 70, 68, and 27.5 kDa in outer membranes cfif;£; enterocolitica serovar 0:8. We identified peptides which probably corresponded to all of these except those at 89 and 27.5 kDa. Differences in cultivation (they harvested cells well into the stationary phase), sample preparation, or strain of cells being analyzed could be responsible. No definite differences were observed between Psts and Pstr strains of enteropathogenic species (Fig. 12,13, 14). 1A.- ‘I‘nll-nl -5. I la 165 Figure 11. Autoradiograms of two dimensional gels of purified outer membranes from X; pestis KIM grown at 37°C. (A) Pgm+ cells grown in iron replete medium, (E) Pgm+ cells grown in iron deficient medium, (C) Pgm- cells grown in iron deficient medium. Horizontal dimension is IEF (pH) and vertical dimension is SDS-PAGE (molecular weight in kilodaltons). Letters refer to Irps (Table 1). in at 3"; ' . f‘ (pd. 321 'v 1"}; 3‘ “Egan. “ 4‘ Figure 11 150 100 80 40 150 100 00 40 30 166 . 2.4....- 5'.- hf!“- Figure 11 150 100 80 40- so— A 167 1mm“ 168 Figure 12. Autoradiograms of two dimensional gels of purified outer membranes from X; psuedotuberculosis grown at 370C. (A) PstS cells grown in iron replete medium, (B) Psts cells grown in iron deficient medium, (C) Pstr cells grown in iron deficient medium. Horizontal dimension is IEF (pH) and vertical dimension is SDS-PAGE (molecular weight in kilodaltons). Letters refer to Irps (Table l). 169 I . .i ”4. . 0 “an Hu'aiki «II? h "an .zn r... an» 4,9 F‘. r . Ru. 1 «flu 5— \|| .6 ‘ n,£fi§ 1' c "" 38 bah gt- 7.- fi-‘ 1 p Q on! iflF’ u “‘9 ..'\4.Av a o 5 1 100 40 54 53 £8 59 Figure 12 170 h. li. a. a. 150 100 80 60 50 40 Figure 12 171 Figure 13. Autoradiograms of two dimensional gels of purified outer membranes from X; enterocolitica WA grown at 370C. (A) PstS cells grown in iron replete medium, (E) PstS cells grown le iron deficient medium, (C) Pstr cells grown in iron deficient medium. Horizontal dimension is IEF (pH) and vertical dimension is SDS-PAGE (molecular weight in kilodaltons). Letters refer to Irps (Table l) which may be homologous to those in other pathogenic Yersinia and arrowheads indicate Irps unique to this species. 172 w a . e . o N, a. _ _ _ 0 0 o. 0 0 0 .3 0 8. 6 5 4 1 «I 100 00 60 50 40 30 20 Figure 13 g-v al- I-b b H., u d g ”or It Up a». .o J. ll . .5: ~:- w IV. "Ind On.” «HYW a F. nu .-. l , . , . ”Ha r Fly 5 ”Pa NH.) 1 . «Ia. ”U4 1" ”I“ I? P MP» n V .11‘ 1“. I41 I 3. s .h .l s 173 o 5 1 Figure 13 174 Figure 14. Autoradiogram of SDS—PAGE gel of purified outer membranes from }(_. pestis KIM (lanes A-C), L pseudotuberculosis PBl (lanes D-F), and X; enterocolitica WA (lanes G-I). Pgm+ (lane A) or Psts (lanes D and G) cells grown in iron replete medium. Pgm+ (lane B) or PstS (lanes E and H) cells grown in iron deficient medium. Pgm- (lane C) or Pstr (lanes F and 1) cells grown in iron deficient medium. All cells were grown at 370C. Position of molucular weight markers (kilodaltons) are indicated on the left. Labelled arrowheads indicate the positions of Irps identified if) X; pestis (Table 1). Unlabelled arrowheads show time position of detectable Irps unique to Y) enterocolitica. of purifii A-C). g nd 61:63 Pstsllua Pgm- (h! )n defintz Position 3 :ated on 1‘5 ons of It: i arfOi'hafis cue t0 3 a Figure 14 200.0— 1 16.3-— 97.4- 66.2— 45.0— 31.0— 1'75 I\ I! (2 I) IS F: (i ii I 176 Since loss of Irps B-D has been linked to Pstr and the associated lesion le iron uptake in X; pestis the presence of these peptides in Pstr enteropathogenic strains was unexpected. (M1 two-dimensional, but not lane gels, enteropathogenic species exhibited a pair of 100 kDa peptides with isoelectric points of 6.4 - 6.6 in iron deficient medium. They are not visible in gels of Pstr X; enterocolitica. Since this change is not observed in the other species and it involves minor peptides its significance is unknown. Apparently the most likely mutational event causing loss of pesticin sensitivity in enteropathogenic species is not analogous to that in _Y_.. pestis although the simultaneous lesion in iron uptake does occur. 177 DISCUSSION Sikkema and Brubaker (43) demonstrated that Pgm+, but not Pgm- or Pgm+, Pstr Z; pestis were able to grow in iron deficient medium containing the ferric chelator citrate at 37°C. This could result from impaired ability of mutants to obtain iron through a lesion in a transport system, or a metabolic change could cause a greater iron requirement in Pgm- cells which cannot be :met under these conditions. Although we cannot discount the latter possibility, labelled iron uptake results are consistant with a lesion in iron transport. Observable changes :hi protein expression correlated with the mutation to nonpigmentation or pesticin resistance involve only envelope associated proteins which would be expected in a mutation affecting transport. Although these studies failed to clarify the nature of the Pgm+-specific transport system, similar results for citrate, pyrophosphate, EDDA, and NTA suggest it does not target a specific chelator. Why the difference in growth was greatest with EDDA is unclear, but may reflect some degree of specificity by this system for iron chelated in a particular manner, or it may relate to differences in the abilities of chelators to bind Fe3+ under conditions present in the agar. Metabolic activity of the cells could decrease the pH and oxygen levels in the agar and allow reduction of 178 some Fe3+ to Fe2+ as would the inclusion of low levels of dithiothreitol. Ferrous iron would be less tightly bound by most ferric chelators. However, very high affinity ferric chelators (desferrioxamine mesylate and conalbumin) made iron unavailable t°.X; pestis at 370C. We clearly demonstrated the‘ correlation. Jbetween pesticin sensitivity anui a system for iron acquisition in pathogenic Yersinia” Pgm+, Pstr showed growth essentially identical to that, of Pgmf 'Z;_ pestis. on gradient plates indicating the functional separation of hemin binding and iron uptake/pesticin sensitivity described in other studies (25,39,44). In previous experiments enteropathogenic' species of Yersinia failed to exhibit evidence for the association of pesticin sensitivity and ability to acquire iron which was demonstrated by Yersinia pestis (43,44). Use of citrate to chelate iron in these studies obscured the lesion since citrate actually enhances growth of enteropathogenic Yersinia in iron deficient. medium. In! using ferrous chelators we demonstrated a similar system associated with pesticin sensitivity in all three species. The effectiveness with which ferrous chelators retard growth in iron deficient media suggests a possible role for ferrous iron uptake. Reduction of ferric to ferrous iron is required in several siderophore and non-siderophore based acquisition systems for release of iron from the chelator and transport across the cytoplasmic membrane (14,17). Non- specific reductase systems are believed to play a role in . P-..— ' I .._' .i. ' v A-, - a 179 iron acquisition by several pathogenic bacteria (12,14,23). Growth and iron uptake by cells utilizing these mechanisms are effectively blocked by ferrous chelators such as ferrozine and 2,2 dipyridyl (14,27). Psts cells may express a ferrous uptake system which can better compete with these chelators for iron than can Pgm_ (Pstr) cells. It is also possible that the inclusion of these chelators reduces ferric iron concentrations to a point where PstS cells would have a significant advantage in obtaining iron due to the presence of an additional non-specific, high affinity ferric transport system. Additional studies of growth and uptake, and tests of Psts and Pstr strains for ferric iron reductase activity should help clarify this. Sikkema and Brubaker (44) speculated on a role for one or more of Irps B-E of Z; pestis in the PstS associated iron uptake system since their loss correlated with Pstr and inability to grow in iron deficient media containing citrate. This is supported by our finding that iron regulated peptides corresponding to Irps B-D were expressed by enteropathogenic Yersinia. However, it was surprising that no differences in outer membrane protein expression were evident following mutation to pesticin resistance. Since these species do not experience the high frequency deletion event seen in l; pestis loss of pesticin sensitivity may most commonly result from mutations only affecting as yet unidentified periplasmic or cytoplasmic membrane proteins needed for internalization of iron and/or 180 pesticin. Since these cells still readily acquire iron from most available sources , it is unlikely to be analagous to £222 mutations which result in tolerance of PstS strains of _E_._ g1; (16). There might also be mutational "hotspots" causing loss of function of critical peptides without concomitant loss of expression. Isolation and study of additional Pstr mutants is ongoing to resolve this. Unrestricted growth of enteropathogenic species in the presence of Fe3+ chelators at 370C (exception: conalbumin) suggest the operation of uptake systems not found in L pestis. Expression cu? several unique iron regulated outer membrane proteins by Y_. enterocolitica may reflect this, however, little difference was apparent between ‘1; pseudotuberculosis and X; pestis. Greater Irp A expression in enteropathogenic species may be involved, although there is run evidence as 1x) its function. Both enteropathogenic species can obtain iron from ferrienterobactin and ferrioxamine (3,36), and specific receptors of 81.0 kDa (Fep A) (41) and 13.7 kDa (1) respectively have been identified in the outer membrane of X; enterocolitica. The latter is to small to detect in our gels, but the former is similar in size to Irp A. However, since L pestis is unable to utilize this siderophore and it expresses Irp A in reduced, but significant amounts, it is unlikely to be the Fep A homolog. The dramatic effect of temperature on growth of i pestis in the presence of various chelators is luau-_L '— 2‘ 181 unprecedented. Earlier studies (demonstrated. reduced expression of systems for iron acquisition at higher temperature in E; coli and Salmonella typhimurium, however, these effects occurred primarily at febrile rather than normal host temperatures (19,24,54). Whether this reflects temperature regulation of expression or inability of the involved proteins to function at 370C is unresolved, however, the observation that the temperature at which cells are grown 1J1 iron deficient medium rather than the temperature at which the test is performed has a greater influence on kinetics of labelled iron uptake and the presence of outer membrane (Irps H and I) and periplasmic (Pirp G and J) iron regulated proteins expressed only at 26°C favors (flue former hypothesis. If this is the case, then it may be possible to obtain mutants which constitutively express 260 systems at 370C. Attempts to isolate mutants capable of growth in the presence of ferrioxamine at 37°C have so far been unsuccessful and leave the issue unresolved. Proteins comprising a particular system for iron acquisition are often coordinately regulated and encoded on a single operon (13,35). Thus identification of other iron repressible proteins associated with the Pgm phenotype might help define the components of the PstS linked uptake system. We confirmed the suspected relationship between HMWP's and pigmentation, but could not characterize them as essentially outer membrane in location as previously described (6). EL_ a: Hutu—um. __l __ -‘ 182 Their large size and significant expression in all compartments suggests they span the envelope, extending from the cytoplasm to the outer' membrane. Since iodination studies indicate they are not exposed on the surface of the cell they may only be associated with the inner leaflet of the outer membrane (6, Perry, pers. comm.). The reason for our discrepancy with Carnial et a1 (6) probably relates to differences in fractionation procedure. Their use of Triton X-100 to dissolve cytoplasmic membranes would destroy any association with this structure leaving only the attachment to and apparent preferential expression in the outer membrane. The newly identified. 113 kDa outer’ membrane A protein (Irp PW anui 19.0 kDa periplasmic protein (Pirp L) were the only additional Pgm-specific iron repressible peptides identified. While the absence of Irp F in the Pgm+, Pstr mutant suggests possible close linkage with genes for Irps BE: in l; pestis, its absence in enteropathogenic species argues against it having a critical role in pesticin sensitivity' or its associated iron uptake system. This system could utilize Irps B-D and possibly E as the outer membrane receptor for initial binding of iron or pesticin. If so this would be the first identified non-specific receptor for iron acquisition. Pirp L and HMWP's could facilitate transport across the periplasm and cytoplasmic membrane. This hypothesis is highly speculative since there is abundant room in the 100 kb Pgm+-specific DNA segment for these genes to be separately encoded, and serve different 183 functions while being coordinately regulated. Genetic studies are underway to address this. Iron is known to serve as a regulatory signal influencing the expression of a variety of proteins. In pathogens the low iron levels encountered in the host would cause expression of genes necessary for survival and pathogenicity (35). In ‘3;. pestis some) of the numerous iron-regulated envelope proteins are probably involved in Pgm-independent systems for iron acquisition such as utilization of iron from ferritin or hemin. Others may serve additional functions. Transcription of virulence factors including certain toxins (2,10) and hemolysins (49) are iron regulated in certain bacteria. However, none of these Pgm-independent iron regulated proteins correspond to known virulnce factors in Yersinia. Since expression of many of these peptides is subject to Fur regulation (53) it should be possible to identify their genes and study them in more detail. 184 REFERENCES Baumler, A. J. and K. Hantke. 1992. A lipoprotein of Yersinia enterocolitica facilitates ferrioxamine uptake in Escherichia coli. J. Bacteriol. 174: 1029- 1035. Boyd, J., M. N. Oza, and J. R. Murphy. 1990. Molecular cloning and DNA sequence analysis of a diphtheria to_x iron—dependent regulatory element (93 5) from Corynebacterium diphtheriae. Proc. Natl. Acad. Sci. USA. 87: 5968-5972. Brock, J. H. and J. N9. 1983. The effect of desferrioxamine on the growth of Staphylococcus aureus, Yersinia enterocolitica, and Streptococcus faecalis in human serum: uptake of desferrioxamine-bound iron. FEMS Microbiol. Lett. 20: 439-442. Brubaker, R. R. 1969. Mutation rate to nonpigmentation in Pasteurella pestis. .J. Bacteriol. 98: 1404-1406. Burrows,T. W. and G. A. Bacon. 1960. V and W antigens in strains of Pasteurella pseudotuberculosis. Brit. J. Exp. Pathol. 41: 38-44. Carniel, E., J-C. Antoine, A. Guiyoule, N. Guiso, and H. PL, Mollaret. 1989. Purification, location and immunological characterization of the iron—regulated high-molecular-weight proteins of the highly pathogenic yersiniae. Infect. Immun. 57: 540—545. Carniel, E., A. Guiyoule, I. Guilvout, and O. Mercereau- Puijalon. 1992. Molecular' cloning, iron-regulation and mutagenesis of the ire 2 gene encoding HMWP2, a protein specific for the highly pathogenic Yersinia. Mol. Microbiol. 6: 379-388. Carniel, E., D. Mazigh, and H. H. Mollaret. 1987. Expression of iron-regulated proteins in Yersinia species and their relation to virulence. Infect. Immun. 55: 277-280. Carniel, E., O. Mercereau-Puijalon, and S. Bonnefoy. 1989. The gene coding for the 190,000 dalton iron- regulated protein of Yersinia species is present only in the highly pathogenic strains. Infect. Immun. 57: 1211-1217. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 185 Calderwood, S. B. and J. J. Mekalanos. 1987. Iron regulation of shiga-like toxin expression in Escherichia coli is mediated by the £93 locus. J. Bacteriol. 169: 4759-4764. - Cox, C. D. 1980. Iron uptake with ferripyochelin and ferric citrate by Pseudomonas aeruginosa J. Bacteriol. 142: 581-587. Cowart, R. E. and B. G. Foster. 1985. Differential effects of iron on the growth of Listeria monocytogenes: minimum requirements and mechanism of acquisition. J. Infect. Dis. 151: 721-730. Crosa, J. H. 1989. Genetics and molecular biology of siderophore-mediated iron transport in bacteria. Microbiol Rev. 53: 517-530. Emery, T. 1987. Reductive mechanisms of iron assimilation. In: Iron Transport in Microbes, Plants and Animals. J. B. Neilands (Ed.) VCH Press. N.Y.,. N.Y. 235-250. Ferber, D. M. and R. R. Brubaker. 1981. Plasmids in Yersinia pestis. Infect. Immun. 31: 839-841. Ferber, D. M., J. M. Fowler, and R. R. Brubaker. 1981. Mutations to tmderance and resistance to pesticin and colicins :hl Escherichia coli CL J. Bacteriol. 146: 506-511. Fescher, E., B. Strehlow, D. Hartz, and. V. Braun. 1990. Soluble and membrane-bound ferrisiderophore reductases of Escherichia coli K-12. Arch. Microbiol. 153: 329-336. Frost, G. E. and H. Rosenberg. 1973. The inducible citrate-dependent iron transport system in Escherichia coli K-12. Biochim. Biophys. Acta. 330: 90-101. Garibaldi, J. A. 1972. Influence of temperature on the biosynthesis of iron transport compounds by Salmonella typhimurium. J. Bacteriol. 110: 262- 265. Griffiths, E. 1987. The iron-uptake systems of pathogenic bacteria. In: Iron and Infection. J. J. Bullen and E. Griffiths (Ed.) John Wiley and Sons. N.Y., N.Y. 69-137. Higuchi, K., L. L. Kupferberg and J. L. Smith. 1959. Studies on the nutrition and physiology of Pasteurella pestis: III. Effects of calcium ions on the growth of 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 186 virulent and avirulent strains of Pasteurella pestis. J. Bacteriol. 77: 317-321. Jackson, S., and T. W. Burrows. 1956. The pigmentation of Pasteurella pestis on a defined medium containing hemin. Brit. J. Exp. Pathol. 37: 570- 576. Johnson, W. , L. Varner, and M. Poch. 1991. Acquisition of iron by Legionella pneumophila: role of iron reductase. Infect. Immun. 59: 2376-2381. Kluger, DL .1. and B. A. Rothenburg. 1979. Fever and reduced iron: their interaction as a host defense response to bacterial infection. Science. 203: 374- 376. Kutyrev, V. V., A. A. Filippov, O. S. Oparina, and O. A. Protsenko. 1992. An lysis of Yersinia pestis chromosomal determinants Pgm and Pst associated with virulence. Microbial. Pathogen.: in press. Laemmli, U. K. 1970. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature (London). 227: 680-685. Lesuisse, E. and P. Labbe. 1989. Reductive and non- reductive mechanisms of iron—assimilation by the yeast Saccharomyces cerevisiae. J. Gen. Microbiol. 135: 257-263. Lowry, O. H., N. J. Rosebrough, A. L. Farr, and R. J. Randall. 1951. Protein measurement with the folin phenol reagent. J. Biol. Chem. 193: 265-275. Lucier, T. S. and R. R. Brubaker. 1992. Determination of genome size, macrorestriction pattern polymorphism, and non—pigmentation-specific deletion in Yersinia pestis by pulsed-field gel electrophoresis. J. Bacteriol. 174: 2078-2086. Massad, G., J. E. L. Arceneaux, and B. R. Byers. 1991. Acquisition of iron from host sources by mesophilic Aeromonas species. J. Gen. Microbiol. 137: 237- 241. McIntosh, M. A., S. S. Chenault, and C. R. Earhart. 1979. Genetics and physiological studies on the relationship between colicin B resistance and ferrienterochelin uptake in Escherichia coli K-12. J. Bacteriol. 137: 653-657. Morrisey, J. ft. 1981. Silver stain for proteins in polyacrylamide gels: a modified procedure with 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 187 enhanced uniform sensitivity. Anal. Biochem. 117: 307-310. Neilands, J. B. 1982. Microbial envelope proteins related. to iron. Ann, Rev; Microbiol. 36: 285- 309. O'Farrell, P. H. 1975. High resolution two-dimensional electrophoresis of proteins. J. Biol. Chem. 250: 4007-4021. Payne, S. M. 1988. Iron and virulence in the family enterobacteriaceae. CRC Crit. Rev. Microbiol. 16: 81- 111. Perry, R. D. and R. R. Brubaker. 1979. Accumulation of iron by yersiniae. J. Bacteriol 137: .1290- 1298. Perry, R. D. and R. R. Brubaker. 1983. Vwa phenotype of Yersinia enterocolitica. Infect. Immun. 40: 166- 171. Pendrak, M. L. and R. D. Perry. 1991. Characterization of a hemin-storage locus of Yersinia pestis. Biol. Metals. 4: 41-47. Perry, R. D., M. L. Pendrak, and P. Schuetze. 1990. Identification and cloning of a hemin storage locus involved in the pigmentation phenotype of Yersinia pestis. J. Bacteriol. 172: 5929-5937. Rogers, H. J. 1973. Iron-binding catechols and virulence in Escherichia coli. Infect. Immun. 7: 445- 456. Rutz, J. M., T. Abdullah, S. P. Singh, V. I. Kalve, and P. E. Klebba. 1991. Evolution of the ferric enterobactin receptor in gram-negative bacteria. J. Bacteriol. 173: 5964-5974. Schryvers, A. B. 1988. Characterization of the human transferrin. and lactoferrin receptors ix: Haemophilus influenzae. Mol. Microbiol. 2: 467-472. Sikkema, D. J. and R. R. Brubaker. 1987. Resistance to pesticin, storage of iron, and invasion of HeLa cells by Yersiniae. Infect. Immun. 55: 572-578. Sikkema, D. J. and R. R. Brubaker. 1989. Outer membrane peptides of Yersinia pestis mediating siderophore-independent assimilation of iron. Biol. Metals. 2: 174-184. .11] I 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 188 Staggs, T. M. and R. D. Perry. 1991. Identification and cloning of a fur regulatory gene in Yersinia pestis. J. Bacteriol. 173: 417-425. Straley, S. C. and R. R. Brubaker. 1981. Cytoplasmic and. membrane proteins of yersiniae cultivated under conditions simulating mammalian intracellular environment. Proc. Natl. Acad. Sci. USA. 78: 1224- 1228. Straley, S. C. and R. R. Brubaker. 1982. Localization in Yersinia pestis of peptides associated with virulence. Infect. Immun. 36: 129—135. Une, T. and R. R._Brubaker. _1984. Iq_vivo comparison of avirulent Vwa and Pgm or Pst phenotypes of Yersiniae. Infect. Immun. 43: 895-900. Waalwijk, C., D. M. MacLaren, and J. deGraff. 1983. In vivo function of hemolysin in the nephropathogenicity of Escherichia coli. Infect. Immun. 42: 245-249. Wake, A. , M. Misawa, and A. Matsui. 1975. Siderochrome production by Yersinia pestis and its relation. to ‘virulence. Infect. Immun. 12: 1211- 1213. Waring, W. S. and C. H. Werkman. 1942. Iron requirements of heterotrophic bacteria. Arch. Biochem. 1: 425—433. Wee, S., J. B. Neilands, M. L. Bittner, B. C. Hemming, B. L. Haymore, and R. Seethram. 1988. Expression, isolation, and properties of £25 (ferric uptake regulation) protein of Escherichia coli. Biol. Metals. 1: 62-68. Weinberg, E. D. 1984. Iron withholding: a defense against infection and neoplasia. Physiol. Rev. 64: 65-102. Worsham, P. L. and J. Konisky. 1984. Effect of growth ,temperature on the acquisition of iron by Salmonella typhimurium and Escherichia coli. J. Bacteriol. 158: 163-168. Zahorchak, R. J. and R. R. ubaker. 1982. Effect of exogenous nucleotides on Ca dependence and V antigen synthesis in Yersinia pestis. Infect. Immun. 38: 953- 959. "I(llllllllllllllll