,. ‘ 7. .- “11‘s. _ _.,, ‘ .- , A. ._. $4211.?” an x «5, ~11”: » “van-“W‘- - 7.“.3 y ‘2‘... .52“. ...r in. rand; 1-. .v .44 0.! - . 9, A ‘a £43 ..-,‘... ...,_-.¢‘. ._ - . ', -. 941fl‘ . ' '1‘?“ - ‘ V ' , . rc‘ . ‘ ,. ‘ ‘ . #1,»...47, ‘ k _ , . V. ~ .U.» n nAlI‘I . , u u A . \ - - -¢.~..¢.."“ . ~'l'.\’:y€'"7‘ W‘- 3 ‘ - " _ "’ ..A“ ‘f ‘ k’l4<-r.’. g “2" , . ‘ - . V ‘ :1. _;‘ . . . ~......‘,J..‘\‘_,_ “ . , p ’ 3.3 “,2," ‘ ._. 3"..- ‘. ' JI_ ‘ ,' ‘ - . a. , . ‘. u .‘ ...r.. .gfizv ‘ 4-. r ... N .. ..--.Lc v19. "‘1 a n . . . - 7... - .. 1 ' I‘d" “ ‘ ‘ "‘ “In ‘ .u, ‘ , ~3.nrn}..\” . . , - ‘ ~ . ‘r » «rum. .- ‘ ~ . < _~. 4 rm. ‘ 1c... _ 4 4'. _ .‘.< ‘,- 7,. _ v' 1% x r -- V A < h . - . s; a!“ " ”'3 -~' ' ‘ ~. . ‘ v y , 15:3?! -LLI v w, ‘ ~> . . -.. .1. “#1.. _w~=€ has“! a - , , 't ' - _‘I , ‘. " It» . ‘ ¢ ‘ ‘ \hIL .591. , "m1 -- ‘r ' J - * - , . ,, ”Vu‘ ~ _ ~-‘ 31:57 .yfimunw. . ' ‘ . . ' 41* ‘ V. .N _. ‘ N~ , .- ‘ “an: A .423“. v v .AJE’ILII f.“ , ‘ .5. . _- r“ J. V {,5 "v.0.“ ' L‘- “I .331 0.. . - ,4 _ ! r... ‘. ~ 5 «w “ .. ’ 1.. ’ ‘ t: ’ A ... 3- ~71 - » ‘ v—I; ‘ I ., :vsz F‘ I ‘r v l “7. .3 .. v.7 ‘ l\ . ' ~"~¥~.~n ' - tau..u . ‘ tan _ _ % ..-‘_,,,., 3‘1”- , . ‘4" , » A. J 74‘. L , 9 v A .‘v u . ,N. . - .1- mm, «:A ‘ .‘ 1‘. ‘.‘ '. "r '2 r: vii/>35, J nit”). J I . . I I “"3"? ‘4 5 “NJ: r.‘.‘" I. V l-l '1 f ,_.u ., , “3”,... ' l . '. 1 ' ”pl-22' 33:3 “I, mi“. .p-udu‘ Allah .f n," 'FP‘0~>; u - v . l -1‘:.(' .-.w..’r' .. ., II!“ , lam”: "ff,“ v'-_"l’ I" .. V" . IJIIILUW I’lllllilllllllllhlllllllllllllllllllllll 1293 00882 5709 This is to certify that the dissertation entitled GENETIC ANALYSIS OF HOST—INDEPENDENT MUTANTS OF BDELLOVIBRIO BACTERIOVORUS 109J presented by TODD WILLIAM COTTER has been accepted towards fulfillment of the requirements for Ph . D . Microbiology and degree in Public Health M - jor professor Date 3/ 23/ 92 MSULI an Affirmative Action/Equal Opportunity Institution 0-12771 fl- ——_.————_—-._—-5 - ~ .._.-.———-~ _ __ .. LIBRARY Mlchlgan State 3 University WW PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before date due. DATE DUE DATE DUE DATE DUE lfl _JL_ ir— l MSU Is An Affirmative Action/Equal Opportunity Institution cmmMut s JUL GENETIC ANALYSIS OF HOST-INDEPENDENT MUTANTS OF BDELLOVIBRIO BACTERIOVORUS 109.] BY TODD WILLIAM COTI‘ER A DISSERTATION Submitted to Michigan State Universi in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Microbiology and Public Health 1992 ABSTRACT GENETIC ANALYSIS OF HOST-INDEPENDENT MUT ANT S OF BDELLOVIBRIO BACTERIOVORUS 109.] By Todd William Cotter Members of the genus Bdellovibrio are obligate intraperiplasmic parasites of other gram-negative bacteria. Certain spontaneous Bdellovibrio mutations eliminate the host cell requirement, giving rise to host-independent (H-I) mutants that can grow axenically in media containing peptone and yeast extract. This dissertation describes the genetic analysis of three such mutants derived from B. bacteriovorus 1091 . In order to allow for such an analysis, a system for the genetic manipulation of Bdellovibrio was developed. In Chapter 1 I describe the conjugal transfer of broad host range cloning vectors, namely IncP and Ian derivatives, from E. coli to B. bactefiovoms. Ian derivatives were maintained in B. bacteriovorus via autonomous replication. IncP derivatives could only be maintained after integration, through recombination between cloned bdellovibrio sequences and the recipient genome. This system was used to deliver an IncP based wild-type B. bactefiovoms 109] genomic library into an H-I mutant, BBS, resulting in the identification of wild-type sequences that significantly enhanced plaque formation by the H-1 mutant. These ”enhanced-plaques" were dramatically larger and clearer than those formed by BBS, and were very similar to those formed by the wild-type except for a moderate reduction in diameter. Further genetic analysis, described in Chapter 2, narrowed the plaque-enhancing sequences down to a 959 bp EcoRI-Xbal fragment, and showed that the equivalent region in BBS contained a mutation. Two additional H-I mutants were isolated and shown to carry mutations in the same 959 bp fragment. Because of its affect on plaquing ability, and the occurrence of mutation in three independent H-I mutants, this locus was termed hit, for host interaction. Merodiploid recombinants that contained both wild-type and mutant derived hit sequences displayed ”intermediate“ phenotypes. BBS recombinants that contained wild~type hit sequences displayed the enhanced-plaque phenotype, but still formed colonies that were indistinguishable from BBS. The reciprocal experiment, where mutant hit sequences were recombined into the wild-type, gave similar but slightly different results. These recombinants formed plaques that were very similar to wild-type, and also displayed a capacity for ”density-dependent" axenic growth. These results are considered with respect to the genetic basis for host-independent growth. ACKNOWLEDGMENTS I am grateful to Barry Chelm for giving a ”green-hom" the opportunity to work in his lab, for instilling in me elements of toughness and perseverance, and for creating an excellent working atmosphere. With the addition of friendship, the same feelings apply to several people who worked with me in Barry's lab: Prudy Hall, Tom Adams, John Sommerville, Bill Holben, Greg Martin, John Scott-Craig and Elizabeth Verkamp. For their contributions to my development as a scientist and as a person, I would like to thank my co-workers in Mike Thomashow's lab: Julia Bell, Deane Lehman, Rom Bada, Ravindra Hajela, Tim Lynch, Nancy Artus, Wei-Wen Guo, Brett McLamey, Dave Horvath, Kathy Wilhelm, Stokes Baker, Chen-Tao Lin, Sue Hammar, Steve Krebs, Sarah Gilmour and Jim Marks. I have never before known a east of characters like this one. I would like to thank my committee members, John Breznak, Wendy Champness, Bob Hausinger and Peter Wolk for sharing in my interest in Bdellovibrio. A debt of gratitude a mile deep and a mile wide is owed to Mike Thomashow, my major professor for the last 4 years. Words can not do justice to what Mike has taught me about writing, communication, diplomacy, and critical thinking; he has been the best of mentors. And besides all of that, he knows how to have a good time. Special thanks also go to Bob Meeley and Jim Marks for helping me to vent graduate school frustrations, and for being the finest of friends. Finally, to Sue, Celia and Natalie: I have the best family in the world. iv TABLE OF CONTENTS LIST OF TABLES LIST OF FIGURES INTRODUCTION LIST OF REFERENCES CHAPTER 1 Identification of wild-type Bdellovibrio bacteriovorus sequences that enhance the plaque forming ability of a spontaneous host-independent mutant SUMMARY INTRODUCTION MATERIALS AND METHODS Bacterial strains and plasmids Media and growth conditions Isolation and characterization of H-I mutant BBS Chemicals and reagents Matings DNA manipulations, Southern analysis and library construction RESULTS Conj ugal transfer of RSFlOlO and RK2 derivatives into B. bacteriovorus Identification of wild-type sequences that enhance plaquing activity in H-I mutant BBS DISCUSSION LIST OF REFERENCES 10 ll 12 12 12 14 15 15 16 17 17 21 30 32 CHAPTER 2 Identification of a Bdellovibrio locus, hit (host-interaction), that contains a mutation in three spontaneous host- independent mutants SUMMARY INTRODUCTION MATERIALS AND METHODS Bacterial strains, plasmids, media and culture conditions Chemieals and reagents Matings DNA manipulations and construction of pVKa-l DNA sequencing RESULTS Identification of a mutation in H—I mutant BBS The hit locus is alterred in other spontaneous mutants of B. bacteriovorus 109J The hit mutations affect multiple open reading frames Recombination of the wild-type hit locus into the H-1 mutants does not restore a wild-type phenotype Recombination of mutant hit loci into wild-type B. bacteriovoms 109J yields recombinants displaying ”density-dependent" axenic growth DISCUSSION LIST OF REFERENCES SUMMARY APPENDIX A Changes in gene expression during intraperiplasmic growth of B. bacteriovorus 109J QUESTION METHODS RESULTS AND DISCUSSION APPENDIX B Other methods of introducing DNA into Bdellovibrio APPENDIX C Delivery of TnS into Bdellovibrio bacteriovonts vi Page 34 35 36 36 39 39 39 4O 40 43 48 51 53 S6 62 69 70 72 Page APPENDIX D Trans complementation experiments 74 APPENDIX E Characterization of the putative ORF2 gene product 76 vii LIST OF TABLES CBMflflHflKl Table 1. Bacterial strains and plasmids Table 2. Conj ugal transfer of plasmids into B. bacteriovoms CBMJHHDRZ Table 1. Bacterial strains and plasmids Table 2. Plating characteristics of hit recombinants viii Page 13 18 37 52 LIST OF FIGURES CHAPTERI Figure 1. Figure 2. Figure 3. Figure 4. Southern analysis of plasmid containing B. bacteriovorus Enhanced-plaque phenotype of BBS(pTC l2) RFLP analysis of cosmid containing exconjugants Plasmid constructs containing B. bacteriovorus 109J DNA and their ability to enhance plaque formation in H-I mutant BBS CHAPTERZ Figure 1. Figure 2. Figure 3. Figure 4. Figure 5. Figure 6. pTC32 does not confer an enhanced-plaque phenotype upon BBS Plasmid constructs containing B. bacteriovoms 109J DNA and their ability to enhance plaque formation in PM mutant BB5 IIDN A sequence of the 959 bp EcoRI-Jatal fragment from the hit ocus Potential coding regions within the 959 bp Beam-m1 fragment Mutated hit locus does not confer the mutant plaque phenotype upon 109J .2 Two classes of BBS (pTCSO) recombinants APPENDIX A Figure 1. SDS-PAGE analysis of in-vivoe3SS-1abelled Bdellovibrio proteins. APPENDIXE Figure 1. Computer analysis of the putative ORF2 gene product ix Page 20 23 26 29 42 45 47 50 55 59 68 INTRODUCTION Bdellovibn'o Biology The genus Bdellovibrio, originally described by Stolp in 1962 (24), comprises a group of gram-negative bacteria that are obligate intraperiplasmic parasites of other gram-negative bacteria. Bdellovibrios are small (approximately O.3x1.S um) curved rods that are highly motile by means of a single polar flagellum. Members of this genus are widespread, having been isolated from soil, fresh water, marine water and sewage. The normal Bdellovibrio life cycle is bi—phasic, consisting of a free living search phase and an intraperiplasmic growth phase. While in search phase, the free swimming bdellovibrio actively seeks a suitable host, but does not multiply. A host cell encounter is an apparently random event, where the rapidly swimming parasite collides with, and attaches to, the host outer surface. Initial attachment is reversible, but is followed by irreversible attachment if the attack is to be succesful (34). The basis for host specificity is unknown; a particular bdellovibrio strain may parasitize a broad or narrow range of gram-negative species (25, 26). Approximately 10 min after attachment, a bdellovibrio rapidly breaches the host outer envelope and becomes lodged within the host periplasm (34). Concomitant with entry, the host cell usually rounds up into a sphere. The rounded cell containing the bdellovibrio has been termed the bdelloplast. The mechanism for penetrating the host outer membrane is unknown. However, penetration of the host peptidoglycan is thought to be eaused by bdellovibrio directed enzymatic activities (27, 31). The 2 primary activity is that of a glycanase, which solubilizes glucosarnine moieties present in the host peptidoglycan and LPS. The appearance of this activity correlates with the time of entry, and is thought to create a localized hole in the host peptidoglycan. While most bdellovibrios express the glycanase activity, an exception is B. bacteriovorus strain W, which gains entry without any detectable glycanase (31). In this case a peptidase activity that solubilizes diaminopimelic acid residues is thought to allow entry by reducing the cross-linking of host peptidoglycan. A more detailed review of these and other activities directed against the host peptidoglycan can be found elsewhere (28, 31). Also during the period of attachment and entry, the attacking bdellovibrio causes the cessation of metabolic activity within the host cell. Eleven min after attachment, the host cell has lost the ability to synthesize protein, RNA and DNA (32). Within 30 min after attachment the host cell can no longer respire exogenous substrates and its cytoplasmic membrane becomes permeable to small molecules (16). Neither the order of these events nor the mechanisms by which they occur are known. At some point during the early stages of attachment and penetration, the parasite undergoes a transition from search phase to growth phase. After this transition, a bdellovibrio spends the initial 60 min of a 3-4 hr growth cycle degrading host macromolecules into their monomeric units. For example, host DNA (12, 17) and RNA (6) are degraded into intermediate sized fragments (approx. 1x105 daltons) and soluble nucleotides. These processes are regulated such that over 90% of the host nucleic acids are retained within the bdelloplast. This insures that these materials will remain available for bdellovibrio biosynthetic needs. The fact that DNA degrading activities can be blocked by addition of chloramphenicol, and that bdellovibrios can grow on heat killed host cells, has led to the conclusion that these degradative activities are catalyzed by enzymes of bdellovibrio origin. 3 The next stage of the cycle, which begins after the first hour, is the period of bdellovibrio growth. Growth is detectable through an increase in the respiration rate (5) and the incorporation of host derived nucleotides into bdellovibrio DNA (12) and RNA (6). In order to generate the energy for growth, the bdellovibrio respires acetate (8, 11), ribose moieties (7) and certain amino acids (8); all of which are derived from the host. The energy obtained from these compounds drives the biosynthesis of bdellovibrio macromolecules. Besides DNA and RNA, host derived monomers are also used for the synthesis of proteins (28) and lipids (11). This pattern, however, is not absolute. The bdellovibrios do have the ability to alter, as well as synthesize de novo, a number of cellular components. It is known that host peptidoglycan components are not used for the synthesis of the corresponding bdellovibrio polymer (27). It has also been shown that some bdellovibrio DNA is synthesized from host RNA components (6, 17), and that some bdellovibrio fatty acids are produced by alteration of host fatty acids or synthesized de novo from acetate (1 1). As the bdellovibrio synthesizes new cell material, it elongates into a coiled filament. At the end of the growth phase this filament fragments into 3—5 individual cells. A terminal lytic activity is produced at the very end of growth (27) that causes dissolution of the bdelloplast wall and release of the progeny bdellovibrio. In addition to the normal bi-phasic Bdellovibrio life cycle just described , some strains can enter a resting stage under certain environmental conditions. For example, B. bactefiovoms strain W form resting structures called bdellocysts, in response to changes in growth conditions (29). After penetration and bdelloplast formation, bdellocyst formation can be initiated by transferring strain W cultures from growth medium to a phosphate buffer. Bdellocysts remain in the bdelloplast until germination conditions (addition of NH4") are created (30). Several marine bdellovibrio isolates have also been shown to enter a stable resting stage while remaining in the bdelloplast (19). Return to vegetative growth for these strains occurs after addition of yeast extract 4 to the growth medium. Structurally, the two types of resting stages differ in that the bdellocysts formed by strain W contain an additional outer wall layer. It has been suggested (19) that the formation of a resting stage by marine bdellovibrios is a survival strategy that occurs in response to nutrient-poor, and therefore host-poor, conditions. The capacity to enter a resting stage may be common among bdellovibrios, even though the proper in vitro conditions for its occurence in other strains have not been established. Intraperiplasmic growth is a highly efficient process. Approximately 50% of the host cell carbon content is converted into bdellovibrio cell material. In addition, Y ATP values ealculated for intraperiplasmic growth (14) indicate that relatively little energy is required for the synthesis of bdellovibrio cell material. The Y ATP value (grams dry weight of cell material formed per mole ATP produced) of 19-26 obtained in these experiments compares very favorably to the generally accepted value of about 10 for heterotrophic growth in rich media, and approaches the theoretical maximum of about 30 ealculated for an organism expending energy only for the polymerization of monomeric units into cell polymers (14). This impressive efficiency of ATP utilization has been attributed to a combination of: the conservation of phosphate bonds in host nucleic acids (6, 15,) and phospholipids (l l); a sequestered, complete nutrient supply (14); and the close coupling of energy generation and utilization (14). Regulation of BdeHovibrio growth and development The ”switch” between the search and growth phases of the Bdellovibrio life cycle is stringently controlled. Search phase bdellovibrios cannot grow in complex commercial media (13, 21, 23), nor do they synthesize DNA in the absence of host cells (35). Further, growth phase bdellovbrios immediately differentiate into search phase cells 5 after premature release from the bdelloplast (18), even if they are suspended in rich growth medium. Thus, bdellovibrios appear to require specific, continuous cues provided by the host cell to remain in the growth phase. Understanding the mechanism that regulates the alternation between search and growth phase is fundamental to our knowledge of Bdellovibrio growth and development. This issue has been addressed in a number of physiological studies. Several groups have reported that high concentrations of host derived, cell-free extracts (1-6 mg protein per ml) can support the axenic growth of bdellovibrios (3, 9, 13), while media containing various combinations of rich commercial media cannot (13, 21 , 23). The active component(s) in the extract is heat stable (100°C) and pronase sensitive (3, 9). Further identification of the active component(s) has been unsuccesful. Similarly, a factor involved in Bdellovibrio cell division, derived either from the host (4) or the parasite (2) has been suggested, but not identified. Although these approaches retain the potential to further our understanding of bdellovibrio host- dependence, to date they have revealed few specifics. The control of the Bdellovibrio life-cycle has also been studied through the isolation of spontaneous mutants that are capable of growth in the absence of host cells or host extracts. These mutants, termed host-independent (H-I), have been isolated from the three recognized Bdellovibrio species, B. bacteriovoms (10, 33), B. stolpii (1) and B. starrii (20, 22). H—I mutants are normally grown on complex bacteriological media containing various mixtures of peptone, yeast extract and nutrient broth. Under these conditions their growth and development is very similar to that of the wild-type grown intraperiplasmically (1, 20). Most H-I isolates retain limited intraperiplasmic (IP) growth eapabilities, and form plaques that are much smaller and more turbid than wild- type plaques. After repeated subculture, variants often arise that have lost all IP growth eapabilities, presumably through the accumulation of additional mutation(s) (20, 33). 6 H-I mutants occur at a frequency of 10"6 - 10'7, which suggests that they arise from a single mutational event (10, 20, 33, this work). The lack of systems for genetic analysis in Bdellovibrio has prevented the identification of such mutations, and thus it is unknown whether the H-1 phenotype results from mutation at one or multiple loci. The identification and characterization of H-I mutations could potentially reveal how a single mutation enables a normally host-dependent Bdellovibrio to switch back and forth between search and growth phase in the complete absence of normal host cell cues. Two possibilities are: 1) activation of new metabolic capabilities or 2) inactivation of normal regulatory activities associated with host-dependence. Ultimately, defining the basis for H-I growth may, in turn, enhance our understanding of Bdellovibrio host-dependence. Amongst procaryotic obligate intracellular parasites, the Bdellovibrio—host system is the only case where anything is known about the genetic basis for host-dependence. Even in the most extensively studied genera, Chlamydia, Coxiella and Rickettsia, no mutants capable of axenic growth have never been isolated and there is a complete lack of systems for genetic manipulation (T. Hackstedt, Rocky Mountain Lab, NIH, personal communication). In light of this, understanding the phenomenon of H—I growth in Bdellovibrio is significant, for it could potentially result in the first description of the genetic basis for obligate intracellular parasitism. The focus of this Dissertation has been to deepen our understanding of bdellovibrio host dependence through the genetic analysis of H-I mutants. Chapter 1 describes the development of a conjugation system for B. bacteriovorus 1091, and the use of this system to identify a 5.6 kb DNA fragment from wild-type B. bacteriovoms 109] that significantly improves the IP growth of an H-I mutant, BBS. Chapter 2 demonstrates that BBS and two other H-I mutants contain a mutation within the 5.6 kb fragment. DNA sequence analysis precisely located the mutation in each H-I mutant and defined a locus designated hit (host-interaction). 10. 11. LIST OF REFERENCES Diedrich, D. L., C. F. Denny, T. Hashimata, and S. F. Conti. 1970. ffiuétgggg parasitic strain of Bdellovibrio bacteriovorus. J. Bacteriol. Eksztejn, M., and M. Varan. 1977 . Elongation and cell division in Bdellovibrio bacteriavorus. Arch. Microbiol. 114: 175-181. ‘ Friedberg, D. 1978. Growth of host-dependent Bdellovibrio in host cell free system. Arch. Microbiol. 116:185-190. Gray, K. M., and E. G. Ruby. 1990. Prey-derived signals regulating duration of the developmental growth phase of Bdellovibrio bacteriovorus. J. Bacteriol. 172:4002-4007. . Hespell, R. B. 1976. Glycolytic and tricarboxylic acid cycle enzyme activities during intraperiplasmic growth of Bdellovibrio bacteriovoms on Escherichia coli. J. Bacterial. 128:677-680. Hespell, R. B., G. F. Miazzari, S. C. Rittenberg. 1975. Ribonucleic acid destruction and synthesis during intraperiplasmic growth of Bdellovibrio bacteriovorus. J. Bacteriol. 123:481-491. Hespell, R. B., and D. A. Odelson. 1978. Metabolism of RNA-ribose by Bdellovibrio bacteriovoms during intraperiplasmic growth on Escherichia coli. J. Bacteriol. 136:936-946. Hespell, R. B., R. A. Reason, M. F. Thomashow, and S. C. Rittenberg. 1973. Respiration of Bdellovibrio bacteriovoms strain 1091 and its energy substrates for intraperiplasmic growth. J. Bacteriol. 113: 1280-1288. Horowitz, A. T., M. Keesel, and M. Shilo. 1974. Growth cycle of predacious bdellovibrios in a hast-free extract system and some properties of the host extract. J. Bacterial. 117:270-282. Ishigura, E. E. 1973. A growth initiation factor for hast-independent derivatives of Bdellovibrio bacteriovorus. J. Bacteriol. 115:243-252. Kuenen, J. G., and S. C. Rittenberg. 1975. Incorporation of long-chain fatty acids of the substrate organism by Bdellovibrio bacteriovorus during intraperiplasmic growth. J. Bacteriol. 121:1145-1157. Martin, A., and S. C. Rittenberg. 1972. Kinetics of deoxyribanucleic acid destruction and synthesis during growth of Bdellovibrio bacteriovoms strain 109D on Pseudomonas putida and Esherichia coli. J. Bacteriol. 111:664-673. 13. 14. 15. 16. 17. 18. 19. 20. 21. 25. 27. 8 Rainer, A. M., and M. Shilo. 1969. Host-independent growth of Bdellovibrio bacteriovorus in microbial extracts. J. Gen. Microbiol. 59:401-410. Rittenberg. S. C., and R. B. Hespell. 1975. Energy efficiency of ilnltgasperiplasmic growth of Bdellovibria bacteriovorus. J. Bacterial. 121: 1158- Rittenberg, S. C., and D. Langley. 1975. Utilization of nucleoside manophosphates per se for intraperiplasmic growth of Bdellovibrio bacteriovorus. J. Bacteriol. 121:1137-1144. Rittenberg, S. C., and M. Shilo. 1970. Early host damage in the infection cycle of Bdellovibrio bacteriovorus. J. Bacteriol. 102: 149-160. Raesan, R.A., and S. C. Rittenberg. 1979. Regulated breakdown of Escherichia coli deoxyribanucleic acid during intraperiplasmic growth of Bdellovibrio bacteriovorus 109J. J. Bacterial. 140:620-633. Ruby, E. G., and S. C. Rittenberg. 1983. Differentiation afier premature 11% 3f) intraperiplasmically growing Bdellovibrio bacteriovoms. J. Bacterial. Sinchez—Amat, A., and F. Tarrella. 1990. Formation of stable bdelloplasts as gpgrvation-survival strategy of marine bdellovibrios. J. Bacteriol. 56:2717- Seidler, R., and M. P. Starr. 1969. Isolation and characterization of host independent bdellovibrios. J. Bacteriol. 100:769-785. Shilo, M. 1969. Morphologieal and physiological aspects of the interaction of Bdellovibrio with host bacteria. Curr. Top. Microbiol. Immunol. 50:174-204. Shilo, M., and B. Bruff. 1965. Lysis of gram-negative bacteria by host i‘nodgqqngggt ectoparasitic Bdellovibrio bacteriovorus isolates. J. Gen. Microbiol. Starr, M. P., and J. C. C. Huang. 1972. Physiology of bdellovibrios. Advan. Microbial. Physiol. 8:215-261. Stolp, B., and H. Petzhold. 1962. Untersuchungen iiber einen obligat parasitischen Mikrarganismus mitl 'scher Activitat ffir Pseudomortas- Bakterien. Phytopath. Z. 45:364-3 . Stolp, H., and M. P. Starr. 1963. Bdellovibrio bacteriovorus gen. et sp. n. , a predatory cataparasitic and bacteriolytic microorganism. Antonie van Leeuwenhoek. 29:217-248. Taylor, V. 1., P. Baumann, J. L. Reichelt, and R. D. Allen. 1974. Isolation, enumeration and host range of marine bdellovibrios. Arch. Microbial. 98: 101- l 14. Thomashow, M. F., and S. C. Rittenberg. 1978. Intraperiplasmic growth of Bdellovibrio bacteriavorus 109J : solubilizatian of Escherichia coli peptidoglycan. J. Bacteriol. 135:998-1007. 31. 32. 33. 35. 9 Thomashow, M. F., and S. C. Rittenberg. 1979. The intraperiplasmic growth cycle-the life style of the bdellovibrios, p. 115-138. In J. H. Parish (ed.), Devkeelppmental biology of procaryotes. University of California Press, Ber ey. Tudor, J. J ., and S. F. Conti. 1977. Characterization of bdellocysts of Bdellovibrio sp. J. Bacterial. 131:314-322. Tudor, J. J ., and S. F. Conti. 1978. Characterization of germination and activation of Bdellovibrio bdellocysts. J. Bacteriol. 133: 130-138. Tudor, J. J ., M. P. McCann and I. A. Acrich. 1990. A new model for the penetration of prey cells by bdellovibrios. J. Bacteriol. 172:2421-2426. Varan, M., I. Drucker and M. Shilo. 1969. Early effects of bdellovibrio infection on the synthesis of protein and RNA of host bacteria. Biochem. Biophys. Res. Commun. 37:518-524. Varon, M., and J. Seiiffers. 1975. Symbiosis-independent and symbiosis- incgompetent mutants of Bdellovibrio bacteriovorus 109]. J. Bacterial. 124: 1 191- 11 7. Varon, M., and M. Shilo. 1968. Interaction of Bdellovibrio bacteriovorus and host bacteria. 1. Kinetic studies of attachment and invasion of E. coli B by Bdellovibrio bacteriovorus 109J. J. Bacteriol. 95:744-753. Varan, M., and M. Shilo. 1969. Interaction of Bdellovibria bacteriovorus and host bacteria. II. Intracellular growth and development of Bdellovibrio bacteriovorus strain 109 in liquid cultures. J. Bacterial. 99:136-141. Chapter 1 Identification of wild-ty Bdellovibrio bacteriovorus sequences t at enhance the plaque formin abilit of a s ntaneous host-in epen ent mu t SUMIWARY Bdellovibrio bacteriovoms is an obligate intraperiplasmic parasite of other gram- negative bacteria. Spontaneous mutants of Bdellovibrio that can be cultured in the absence of host cells, called host-independent (I-I-I), occur at a frequency of 10'6 - 10”. Most H-I strains display diminished intraperiplasmic growth capabilities and farm plaques that are smaller and more turbid than those formed by wild-type strains on lawns of host cells. In order to begin a genetic analysis of H-I growth, a system for the conjugal transfer of broad host range cloning vectors from E. coli into B. bacteriovoms 109] was developed. Ian type plasmids were capable of autonomous replieation in B. bacteriovorus 109J . IncP derivatives did not replicate in Bdellovibrio, but could be maintained via integrational recombination through cloned Bdellovibrio sequences. A wild-type B. bacteriovorus 109J genomic library was constructed and transferred into the hast-independent (H-I) mutant BBS . These experiments resulted in the identification of a 5.6 kb BamHI fragment of wild-type DNA that signifieantly enhanced the plaque forming ability of BB5. 10 11 INTRODUCTION Bdellovibrio bacteriovorus is an obligate intraperiplasmic parasite of other gram- negative bacteria. Its bi-phasic life cycle revolves around the availability of suitable host cells (reviewed in 22). After contact with a host cell, the bdellovibrio penetrates the host outer envelope and takes up residence within the host periplasm. At some point during the early stages of this interaction, the bdellovibrio undergoes a shift from search phase to growth phase, initiating a complex, temporally regulated series of activities required for growth in this unique environment. At the completion of growth, the progeny return to search phase and lyse the host, eausing their release back into nature. Certain spontaneous Bdellovibrio mutations obviate the host requirement, giving rise to mutants that can be cultured in complex bacteriological media (3, 8, 18, 19, 23). Such host-independent (H-I) mutants complete the transition from search phase to growth phase and back again without the normal signals associated with the intraperiplasmic niche. Upon initial isolation, the vast majority of these mutants retain limited intraperiplasmic (1P) growth capabilities, and are termed facultative. When plated on lawns of host cells, facultative H-I mutants form plaques that are smaller and more turbid than plaques formed by wild-type bdellovibrios (3, 19, 23). H-I mutants generally mimic wild-type development when cultured on various mixtures of yeast extract, peptone and nutrient broth. It has been suggested that the H-1 phenotype results from a single mutational event, since H-I mutants occur at a frequency of 10’6 - 10'7 (8, 18, 23, this work). Additional genetic characterization of H-I mutants has been hindered by the absence of systems for the genetic manipulation of bdellovibrios. The long term goal of this research is to better understand how Bdellovibrio regulates the switch between search phase and growth phase, and back again. H-I mutants of B. bacteriovorus 109J , such as BB5 , have lost the tight control over growth 12 initiation. Mutation to host-independence may, therefore, affect some fundamental aspect of how bdellovibrios sense the presence of host and control their life cycle. For this reason, I have chosen to investigate the genetic basis for H-I growth in BBS. Towards this end, I have developed a system for the conjugal transfer of broad host range plasmids into wild-type Bdellovibrio bacteriovoms 109] and its H-I derivatives. Using this system, I have identified wild-type B. bacteriovorus 109J sequences that enhance the plaquing ability of an H-I mutant. MATERIALS AND METHODS Bacterial strains and plasmids. The bacterial strains and plasmids used are listed in Table l. B. bacteriovoms 1091 and E. coli ML35 were obtained from S. C. Rittenberg. All bdellovibrio strains were single plaque or single colony purified and stored in 15% glycerol at -80°C. Media and growth conditions. All E. coli cultures were grown at 37°C in LB medium (11). When appropriate, E. coli cultures contained antibiotics at the following concentrations: ampicillin (Ap), 100 pg/ ml; chloramphenicol (Cm), 20 ug/ ml; kanamycin sulfate (Km), 25 ug/ml; rifampicin (Rt), 100 ug/ml; streptomycin sulfate (Sm), 50 ug/ml, and tetracycline (Tc), 12 pg/ml. Intraperiplasmic (IP) cultures of B. bacteriovorus 109J and its derivatives were grown in dilute nutrient broth (DNB) at 30°C, using E. coli ML35 as substrate. DNB consisted of 1 mM CaC12, 0.1 mM MgC12 and 0.8 g nutrient broth (Difco) per liter. Host cells were prepared by washing overnight cultures of ML35 once in an equal volume of DNB. Liquid IP cultures were set up by adding approximately 109 host cells and 107 bdellovibrios per ml, using single plaques or overnight cultures as the bdellovibrio inoculum. These cultures routinely lysed in 1 day. IP cultures were 13 Table 1. Bacterial strains and plasmids Strain or plasmid Description Source or reference B. bacteriovorus 1091 wild-type (14) 1091.1 Smr derivative of 1091 This study 1091.2 Rfr derivative of 1091.1 This study BB5 H-I derivative of 1091.2 This study E. coli ML35 B IacI lacY (15) SR—l Smr RfI derivative of ML35 This study DHS F' endAI real] (7) hstI 7(I‘K-mK+) deoR thi-I supE44 gyrA96 relAI SMIO supE44 hst thi-I thr-I euB6 (20) lach tonA21 recA Muc RP4-2Tc::Mu, Kmr Plasmids pKC7 ColEl Apr Kmr (13) pBR328 ColEl Apr Cmr Tcr (21) pRK2013 ColEl Kmr tra(RK2) (5) pSUP204 Ian Apr Cnnr Tar (12) pSUP304.1 1an Apr Kinr (12) pMMB33 Ian Kmr cos( ) (6) pRK290 IncP Ter (4) pVKIOO IncP 'l‘er Kmr (9) pVK102 IncP Tar Knit (9) pVKoz-l Derivative of pVKIOO containing 445 (Chapter 2) bp HaeII fragment from pUC19 pTC3 23.5 kb fragment of B. bacteriovorus This study 1091 DNA in EcoRI site of pVKIOO pTCS 4.9 kb fragment of B. bacteriovorus This study 1091 DNA in EcoRI site of pVKIOO pTC6 20.5 kb fragment of B. bacteriovorus This study 1091 DNA in EcoRI site of pVK100 pTC7 19.5 kb fragment of B. bacteriovorus This study 1091 DNA in EcoRI site of pVK100 pTC8 5.6 kb BamHI fragment from pTC7 in This study BarnHI site of pMMB33 pTC12 5.6 kb BamI-II fragment from pTC7 in This study BglII site of pVK102 . pTC50 0.96 kb EcoRI-Xbal fragment from (Chapter 2) pTC12 in BarnI-II site of pVKa-l 14 plated for plaque development by adding 0.1 ml of the appropriate bdellovibrio dilution and 101° washed host cells (in 0.3 ml) to 3 ml overlay (DNB plus 0.7% agar held at 50°C), and immediately spread on DNB plates that contained 1.5 % agar. Under these conditions plaques became visible after 3-4 days. Rf’ and Smr bdellovibrios were grown on E. coli SR-l, in the presence of 100 ug/ml Rf or 50 ug/ml Sm. Plasmid containing bdellovibrios were grown on E. coli ML35 carrying pKC7 or pBR328, in the presence of 35 ug/ml Km or 10 rig/ml Cm, respectively. H-I bdellovibrio cultures were grown at 30°C in PYE medium (18) that contained 10 g peptone and 3 g yeast extract per liter. PYE plates were solidified with 1.5% agar. When appropriate, antibiotics were used at the same concentrations as in I-P cultures. Isolation and characterization of H-I mutant BB5. Spontaneous H-I mutants of B. bacteriovorus 1091 were obtained at a frequency of 10'6 - 10'7 using the method of Seidler and Starr (18). Selection for H-I growth yielded yellow CFUs that varied from 0.2 mm - 3.0 mm in diameter. The ”class" of spontaneous H—I mutants that formed medium and large colonies could generally be sub-cultured on solid or liquid PYE medium, whereas the small and tiny colony "class" could occasionally be subcultured if several were pooled together to form a large inoculum. A well isolated large colony isolate, BB5, was selected for further characterization. BBS formed circular, smooth edged colonies that were 1-2 mm in diameter after 7 days incubation on PYE plates at 30°C. BBS also formed small plaques when plated in overlay lawns of host cells on DNB medium. These plaques were smaller and more turbid than plaques formed by wild-type B. bacteriovorus 1091 (see Results). Total PFUs formed by BBS were generally 10-100% the total number of CFUs. 15 Chemicals and reagents. Complex medium components were purchased from Difco. Restriction endonuclases and T4 DNA ligase were purchased from New England Biolabs. [a-32P]dCTP (800 Ci/mM) was purchased from DuPont/New England Nuclear. Matings. Individual matings were conducted on 3 cm sq pieces of nitrocellulose (Schleicher and Schuell) that were incubated on PYE plates. The nitrocellulose was autoclaved in water, placed on PYE plates and allowed to dry (30 min at RT). Wild- type Bdellovibrio recipients were prepared from freshly lysed IP cultures. Such cultures were concentrated 10-fold by centrifugation, and 0.1 ml of the suspension was spread on a nitrocellulose filter and allowed to dry (30 min at RT). H-I mutant recipients were prepared by placing 0.1 ml of an overnight culture on a nitrocellulose filter, letting it dry (30 min at RT) and then incubating the filter overnight at 30°C on a PYE plate. Donor E. coli cultures that had been washed once in DNB and concentrated 10 fold were spread (0.1 ml) on top of the recipients. After 16-24 hrs incubation at 30°C, individual matings (nitrocellulose filters) were transferred to 2 ml DNB and vortexed vigorously, followed by serial dilution and plating for PFUs and CFUs. When plating for axenic growth on PYE plates, Sm (50 jig/ml) was included in the medium to select against growth of the donor. All bdellovibrio recipient cultures were started from -80°C stocks. Donor strains were either E. coli SMIO or E. coli DHS. In the case of SM10, functions required for the conjugal transfer of Ian and IncP type plasmids are provided by an IncP plasmid that is integrated into the SM10 genome. When DHS was used as donor, the same transfer functions were provided by the helper plasmid pRK2013, which is a ColEl derivative and cannot replicate in Bdellovibrio (T. Cotter, unpublished observation). When DHS was the donor, overnight cultures of DHS containing the target plasmid and DHS (pRK2013) were mixed in equal volumes and treated as described above. 16 DNA manipulations, Southern analysis and library construction. Most DNA purifieation and recombinant DNA methods were standard (16). Bdellovibrio genomic DNAs were purified by a CTAB based extraction procedure ( 1). For Southern analysis, bdellovibrio genomic DNA was digested with the appropriate restriction enzymes, fractionated by electrophoresis in 0.7% agarose gels and transferred to Nytlan membranes (Schleicher and Schuell) using the capillary method. Prior to hybridization, membranes were prewashed in 0.1X SSPE (1X: 0.18 M NaCl; lmM EDTA; 10 mM NaPO4, pH=7.7) and 0.5% SDS at 65°C. Prewashed membranes were then prehybridized for 1 hr at 68°C in hybridization fluid that contained 6X SSPE, 0.5% SDS and 0.25% nonfat dry milk (Sanalac). Radiolabeled probe was then added and allowed to hybridize overnight at 68°C. All past- hybridization washes contained 0.5 % SDS and were done in the following order: twice in 2x SSPE at RT; twice in 0.1x SSPE at RT and three times in 0.1x SSPE at 68°C. Radiolabeled probes (approximately 10’7 dpm/ug DNA) were produced by nick translation (kit obtained from BRL) or random priming (16). The B. bacteriovorus 1091 genomic cosmid library TVL-l was constructed according to the method described by Ausubel et al. ( 1). Genomic DNA was partially digested with EcoRI, and size fractionated in 0.5% agarose gels; DNA fragments were electroeluted from the gel and purified with elutip—d columns (Schleicher and Schuell). The size fractionated DNA (2030 kb) was ligated into the EcoRI site of pVK100 (9), and the ligation products packaged with commercial extracts (Promega) according to supplier specifications. Packaged cosmids were transduced into DHS and stored at - 80°C. 17 RESULTS Corriugal transfer of RSF1010 and RK2 derivatives into B. bacteriovorus. Previous studies have shown that RK2 transfer filnctions supplied in trans can provide the means for conjugal transfer of Ian (RSF1010) and IncP (RK2) derived plasmids from one gram-negative species to another (4, 12). We attempted to transfer both types of plasmids into B. bacteriovorus 1091 by conjugation, using E. coli as the donor (see Materials and Methods). Conjugal transfer of the RSF1010 derivatives pSUP204, pSUP304.1 and pMMB33 produced antibiotic resistant recipients of host-dependent (H-D) and host-independent (H-I) Bdellovibrio strains (Table 2). Matings conducted in the absence of RK2 transfer functions did not yield antibiotic resistant Bdellovibrio recipients, indicating that plasmid transfer was conj ugal in nature (data not shown). Kmr (20—40 rag/ml) and Cmr (5-10 rig/ml) were effective in selecting for the transfer of RSF1010 derivatives, while Tc’ (2-25 rig/ml) and Apr (5-50 pg/ml) resistance were not. Two lines of evidence indicated that RSF1010 derivatives were maintained in B. bacteriovorus 1091 by autonomous replication. Southern analysis of total DNA isolated from BBS(pMMB33) revealed a single hybridizing band of 13.8 kb, representing linear pMMB33 (Figure 1A). In addition, BBS(pMMB33) total DNA was used to transform E. coli DHS to antibiotic resistance, and the transforrnants were shown to contain pMMB33 (data not shown). All attempts to conjugally transfer the RK2 derivatives pRK290 and pVK100 into H-D and H-1 strains of B. bacteriovorus failed to yield antibiotic resistant recipients (Table 2). Since RSF1010 plasmids were mobilized into bdellovibrio by RK2 transfer functions, it was possible that RK2 plasmids were also transferred but could not replicate. If true, insertion of B. bacteriovorus 1091 sequences into a RK2 derivative could potentially allow such constructs to be maintained in bdellovibrio after integration 18 Table 2. Conjugal transfer of plasmids into B. bacter'iovor'usa Plasmid Transfer Useful frequencyb antibiotic selections Ian pSUP204 1:110:3 Cmr, Kmr pSUP304.1 1x10’3 Kmr pMMB33 1x10‘3 Kmr IncP pRK290 0 -- pVKIOO 0 -- pTC3 1x104 Kmr a Data apply to matings involving H-D or H-I strains as recipient, and either SM10 or DHS(pRK2013) as donor. ° Transfer frequency expressed as number of antibiotic recipients per total recipients 19 Figure 1. Southern analysis of plasmid containing B. bacteriavorus. (A) BamHI digests of total DNA isolated from BB5 (lane 1) and BBS(pMMB33) (lane 2) probed with nick translated pMMB33. (B) BamHI digests of total DNA isolated from BB5 (lane 1) and BBS(pTCSO) (lane 2) probed with the random prime labelled 5.6 kb BamHI insert from pTC12. 20 m -—13.8 Figure 1. 2 kbp —5.6 - —4.l .... —2.4 21 by homologous recombination. Indeed, in contrast to the cloning vector alone, the conjugal transfer of pTC3, an RK2 derivative (pVK100) containing a random 23.5 kb fragment of B. bacteriavorus 1091 DNA, produced kanarnycin resistant recipients of both wild-type and H-I mutant derivatives of B. bacteriovorus 1091 (Table 2). The transfer frequency of pTC3 was approximately 10-fold less than that of RSF1010 derivatives, presumably due to the requirement for recombination. The integration of RK2 based constructs into the bdellovibrio genome was demonstrated by Southern analysis of total DNA obtained from BBS(pTCSO) (Figure 1B). pTCSO contains a 0.96 kb EcoRI-Xbal fragment of B. bacteriovorus 1091 DNA derived from the S .6 kb bdellovibrio DNA insert in pTC12 (described below). The natural XbaI and EcoRI termini of the 0.96 kb fragment were replaced with BamHI termini prior to cloning into the BamHI site of pVKa-l (Chapter 2). Based on the physical characterization of this region (Chapter 2), integration of pTCSO into the BBS genome by homologous recombination should transform the 5.6 kb BamHI fragment into two BamHI fragments of 4.1 and 2.4 kb. This prediction proved to be true (Figure 1B). Total DNA from BBS (lane 1) contained the expected single hybridizing band representing the 5.6 kb BamHI fragment, whereas a BamI-II digest of BBS(pTCSO) contained two smaller fragments of 4.1 and 2.4 kb. Identification of wild-type sequences that enhance plaquing activity in H-I mutant BB5. The H-I mutant BBS, like other described H-I mutants, retains a diminished capacity for intraperiplasmic (IP) growth, fornring small turbid plaques on lawns of host cells (Figure 2). If the mutations that result in the H-1 phenotype are due to gene inactivation, then it might be possible to identify the affected region by transferring a wild-type genomic DNA library into BB5 and screening the exconj ugants for plaquing ability that is significantly enhanced over that displayed by BBS. 22 Figure 2. Enhanced-plaque phenotype of BBS(pTC 12). Plaques formed by (A) 1091.2(pTC3) (wild-type), (B) BBS(pTC3) (H-1) and (C) BBS (pTC12) are compared. pTC3 does not affect the plaque phenotype of wild-type or mutant strains and is used here as a control to confer Km’. 23 Figure 2. 24 I first attempted to construct a wild-type genomic DNA library in pMMB33 (6), which could autonomously replicate in B. bacteriavoms 1091. Initially this construction appeared to be succesful, but continued propagation of several individual clones in E. coli indieated that pMMB33 could not stably maintain large Bdellovibrio DNA inserts (25-35 kbp). A stable cosmid library of B. bacteriovorus 1091 genomic DNA, however, was constructed in pVKIOO (see Materials and Methods). Given that the packaging process should restrict insert size to between 22 and 28 kbp and that the B. bacteriovoms 1091 genome is 2.0 x 106 bp (10), I calculated according to Clark and Carbon (2) that 366 clones would be required to assure 99% probability that any given sequence would be represented in the library. Based on that figure, several genomic libraries were made, one of which was TVL- l . TVL-l contained approximately 700 clones, and was constructed by combining six independent sub-libraries (VL-l , VL-2, VL-3, VL-4, VL-S and VL-6), each of which contained about 100 clones. VL-l through VL-6 were individually mated into BBS , and the recipients screened for improved plaquing ability. Sub-libraries VL-3 and VL-4, as well as pTC3 (which served as the negative control), did not confer an enhanced plaquing phenotype upon any BBS recipients. Sub-libraries VL-l , VL-2, VL-S and VL-6, however, gave rise to larger, clearer PFUs that comprised 1-2 % of the total Kmr recipients. If the enhancement of BBS plaquing activity resulted from homologous recombination of wild-type sequences into the recipient genome at the site of the H-1 mutation, then identical or related cosmids should have been present in the enhanced plaque recombinants. RFLP analysis was conducted to determine if this was the case. Total DNAs were digested with HindIII, an enzyme that cuts once within pVKlOO, and subjected to Southern analysis using pVKIOO as probe (Figure 3). Each digest would be expected to contain 2 bands that hybridized to the cloning vector, both of which which would have extended in opposite directions from the HindIII site within pVK100 to HindIII sites in the adjacent bdellovibrio DNA (Figure 3A). When 18 random 25 Figure 3. RFLP analysis of cosmid containing exconiugants. (A) Schematic diagram showing co-integration of a cosmid into the recipient Bdellovibrio genome, and the 2 hypothetical HindIII fragments that hybridize to the pVK100 probe. Restriction site: H, HindIII. Hypothetical cloned Bdellovibrio DNA represented by the solid box. pVRIOO represented by the hatched box. The thin line represents the recipient genome. (B) HindIII digests of total DNA isolated from 18 random BBS recipients probed with nick translated pVK100. Individual Kmr isolates were obtained after mating with sub- libraries VL-l (lanes 1-9) and VL-2 (lanes 10-18). (C) HindIII digests of total DNA isolated from 18 enhanced-plaque BB5 recipients probed with nick translated pVK100. Individual Kmr isolates were obtained after mating with sub-libraries VL-l (lanes 1-6), VL-2 (lanes 7-12) and VL-S (lanes 13-18). A H H B VL-l V132 kbp 23_................!9 8.5- c O . 5.6— . . ’ u - o - c v - . . " 3.7— C VL-l VL-2 VL-S kbp 23 "Hr-"'- ~r-'- - -o'o..- 9.4— ' u—-------~u-— 2.3 ‘- Figure 3. 27 recipients were analyzed in this way, 17 distinct patterns were observed (Figure 3B). These results contrasted with those obtained from RFLP analysis of enhanced-plaque exconjugants from VL-l , VL-2 and VL-S. All of the BBS recombinants that contained cosmids from VL-l (9 individuals) and VL-2 (8 individuals) displayed the same RFLP pattern. A second, distinct pattern was seen in 10 isolates that contained cosmids from VL-S. These two RFLP patterns are seen in Figure 3C, which shows 6 individuals each from VL-l, VL-2 and VL-S. These data indicated that specific regions of the B. bacteriovorus genome were involved in conferring the enhanced-plaque phenotype upon BBS . Further characterization of wild-type sequences responsible for plaque enhancement in BB5 required the isolation of an entire cosmid from an enhanced-plaque BBS exconj ugant. This was accomplished in several steps. From an enhanced-plaque BBS exconjugant that showed the predominant RFLP pattern (lane 5, Figure 3C), total DNA was isolated and digested with BamI-II. Since BamI-II does not out within pVK100, the integrated vector was released with flanking bdellovibrio sequences attached to each end. This linear, vector-containing BamI-II fragment was then circularized in a dilute ligation and transformed into E. coli DHS. A single plasmid, pTCS , was identified that contained pVKIOO plus 2 flanking EcoRI-BarnI-II fragments of 2.3 and 2.6 kb. The 4.9 kb EcoRI insert from pTCS was purified and used to probe colony lifts of VL-l . Two cosmids that hybridized to the pTCS insert, pTC6 and pTC7, were isolated and found to contain overlapping inserts of 20.5 kb and 19.5 kb, respectively (Figure 4). Cosmids pTC6 and pTC7 were tested for the ability to confer the enhanced-plaque phenotype upon BB5 . Recombinants containing pTC6 had the same plaque phenotype as the negative control, BBS(pTC3), whereas BBS (pTC7) exhibited the enhanced- plaque phenotype. From within pTC7, a 5.6 kb BarnHI fragment was identified, carried on pTC 12, that also enhanced plaque development in BBS (Figure 4). 28 Figure 4. Plasmid constructs containing B. bacteriovorus 1091 DNA and their ability to enhance plaque formation in H-I mutant BB5. Inserts from plasmid constructs that contain overlapping fragments of B. bacteriovorus 1091 DNA are shown. Restriction sites: B, BarnHI; E, EcoRI. 29 Plaque Enhancement Figure 4. 30 BBS(pTC12) plaques, pictured in Figure 2, appeared identical to plaques formed by BBS (pTC7). These plaques were very similar to those formed by the wild-type, except for a moderate reduction in diameter. DISCUSSION To date, studies on Bdellovibrio growth and development have involved biochemical, physiological and observational experimentation. These approaches can now be complemented by genetic analysis. In particular, we have shown that Ian and IncP plasmids can be conj ugally transferred from E. coli into B. bacteriovorus 1091. Ian plasmids are maintained in B. bacteriovorus 1091 by autonomous replication. IncP plasmids cannot replicate in B. bacteriovorus 1091, but can be maintained via homologous recombination and integration through cloned Bdellovibrio sequences. This system allows for the transfer of genomic libraries, which dramatically expands the potential for genetic analysis in Bdellovibrio. Conjugation should also facilitate the delivery of transposons into the bdellovibrio genome. Indeed, preliminary results (T. Cotter, unpublished data) indicate that Tn5 can be delivered via suicide plasmids, albeit at low frequency (10'8). Presumably, transposon mutagenesis could be further developed into a useful tool for the genetic analysis of B. bacteriovorus. Here I use the Bdellovibrio conj ugation system to initiate a genetic characterization of Bdellovibrio H-I mutants. H-I mutants have lost the stringent control over growth, being able to complete the transition from search phase to growth phase and back again, in the absence of host cells. Understanding the genetic basis for H-I growth may reveal fundamental information concerning the mechanism that regulates the switch between search and growth phases, and provide insight into the nature of host- dependency in Bdellovibrio. Towards this end, I have identified a 5.6 kb BamHI fragment of wild-type B. bacteriovoms 1091 DNA that confers an enhanced-plaque 31 phenotype upon H-I mutant BB5. The 5.6 kb fragment was isolated from one of the two cosnrids identified as conferring an enhanced-plaque phenotype upon BBS (Figure 3C), leaving open the possibility that the other cosmid contained unrelated sequences from another region of the Bdellovibrio genome. Additional Southern analysis ('1‘. Cotter, unpublished results) indicates that the 2 RFLP patterns seen in Figure 3C are derived from related cosmids, suggesting that the 5.6 kb BamI-II fragment is the only region of the B. bacteriovorus genome that can confer this phenotype. In the following chapter I further characterize the 5.6 kb fragment from wild-type, BBS and two additional H-I mutants. lo. 11. 32 LIST OF REFERENCES Ausubel, F. M., R. Brent, R. E. Kingston, D. D. Moore, J. G. Seidmon, J. A. Smith, and K. Struhl. 1987. Current protocols in molecular biology. Greene Publishing Associates, Wiley Interscience, New York. Clarke, L., and J. Carbon. 1976. A colony bank containing 8 nthetic ColEl hybrid plasmids representative of the entire E. coli genome. Ce 9:91-99. Diedrich, D. L., C. F. Denny, T. Hashimoto, and S. F. Conti. 1970. Facuétggséglg parasitic strain of Bdellovibrio bacteriovorus. 1. Bacterial. 1 1: - . Ditta, G., S. Stanfield, D. Corbin, and D. R. Helinski. 1980. Broad host range DNA cloning system for gram-negative bacteria: Construction of a gene bank of Rhizobium meliloti. Proc. Natl. Acad. Sci. USA 77:7347-7351. Figurski, D., and D. Helinski. 1979. Replication of an origin containing derivative of plasmid RK2 dependent on a plasmid function provided in trans. Proc. Natl. Acad. Sci. USA 76:1648-1652. Frey, J., M Bagdasarian, D. Feiss, F. C. H. Franklin, and J. Deshusses. 1983. Stable casnrid vectors that enable the introduction of cloned fragments into a wide range of gram-negative bacteria. Gene 24:299-308. Hanahan, D. 1985. p.109. In D. M. Glover (ed.), DNA cloning: a practical approach, vol. 1. IRL press, McLean Virginia. Ishiguro, E. E. 1973. A growth initiation factor for host-independent derivatives of Bdellovibrio bacteriovorus. 1. Bacterial. 115:243-252. Knauf, V. C., and E. W. Nester. 1982. Wide host range cloning vectors: a cosmid clone bank of an Agrobacteriurn Ti plasmid. Plasmid 8:45-54. Matin, A., and S. C. Rittenberg. 1972. Kinetics of deoxyribanucleic acid destruction and synthesis during growth of Bdellovibrio bacteriovorus strain 109D on Pseudomonas putida and Escherichia coli. J . Bacterial. 111:664-673 Miller, J. H. 1972. Experiments in molecular genetics. Cold Spring Harbor Laboratory, Cold Spring Harbor, New York. Priefer, U. B., R. Simon, and A. Puhler. 1985. Extension of the host range of Escherichia coli vectors by incorporation of RSF1010 replication and mobilization functions. 1 . Bacterial. 163:324-330. 17 18 19 20 13. 14. 15. 16. 17. 18. 19. 20. 21. 23. 33 Rao, R. N., and S. G. Rogers. 1979. Plasmid pKC7: A vector containing ten r‘rgstriction endonuclease sites suitable for cloning DNA segments. Gene 3:79- Rittenberg, S. C. 1972. Nonidentity of Bdellovibrio bacteriovorus 109D and 1091. 1. Bacterial. 109:432-433. Rittenberg, S. C., and M. Shilo. 1970. Early host damage in the infection cycle of Bdellovibrio bacteriovorus. J . Bacteriol. 102: 149-160. Sambrook, J ., E. F. Fritsch, and T. Maniatis. 1989. Molecular cloning: A laboratory manual, 2nd ed. Cold Spring Harbor Laboratory, Cold Spring Harbor, New York. Seidler, R. J ., M. Mandel, and J. N. Baptist. 1972. Molecular heterogeneity of the bdellovibrios: Evidence of two new species. 1. Bacterial. 109:209-217. Seidler, R., and M. P. Starr. 1969. Isolation and characterization of host- independent bdellovibrios. J . Bacterial. 100:769-785. Shilo, M., and B. Bruff. 1965. Lysis of gram-negative bacteria by host inodgpqnggrét ectoparasitic Bdellovibrio bacteriovorus isolates. J . Gen. Microbiol. : 1 - . Simon, R., U. Priefer, and A. Puhler. 1983. A broad host range mobilization system for in viva genetic engineering: transposon mutagenesis in gram- negative bacteria. Bio/Technology l :7 84-790. Soberon, X., L. Covarrubias, and F. Bolivar. 1980. Construction and characterization of new cloning vehicles. IV. Deletion derivatives of pBR322 and pBR32S. Gene 9:287-305. Thomashow, M. F., and S. C. Rittenberg. 1979. The intraperiplasmic growth cycle-the life style of the bdellovibrios, p. 115-138. In J. H. Parish (ed.), Developmental biology of procaryotes. University of California Press, Berkeley. Varon, M., and J. Seiiffers. 1975. Symbiosis-independent and symbiosis- incompetent mutants of Bdellovibrio bacteriovorus 1091. 1. Bacteriol. 124: 1191- 1 197 . Chapter 2 Identification of a Bdellovibrio bacteriovorus locus, _hit most-interaction) that contains a mutation in three spontaneous host-independent mutants. SUMMARY Previous work (Chapter 1) has identified a 5.6 kb BarnI-II fragment of wild-type Bdellovibrio bacteriovoms 1091 DNA that significantly enhances the plaque forming ability of the host-independent (H-I) mutant BBS . Further genetic analysis, described here, has narrowed down the sequences required to confer this phenotype to a 959 bp EcoRI-Xbal fragment. DNA sequence analysis of this fragment, and the equivalent fragment from BBS, indicates that BBS contains a single base pair deletion in this region. Two additional H-I mutants, BB3 and BB4, are also shown to contain deletion mutations within the 959 bp EcoRI-Xbal fragment. Because of its affect on plaque formation in H-I mutants, and the occurrence of mutation within this region in three H-I mutants, this locus is termed hit, for host-interaction. Merodiploid recombinants that contained both wild-type and mutant derived hit sequences displayed ”intermediate” phenotypes. BB5 recombinants that contained wild-type hit sequences formed plaques that were very similar to those of the wild-type (enhanced-plaque phenotype), but could still form colonies that were indistinguishable from those formed by BBS. Recombination of mutant hit sequences into the wild-type yielded recombinants that formed plaques which were also very similar to wild-type, but 34 35 displayed a capacity for "density-dependent” axenic growth. The genetic basis for the intermediate phenotypes is discussed. INTRODUCTION Members of the genus Bdellovibrio are obligate intraperiplasmic parasites of other gram-negative bacteria. Their unique life-cycle involves the alternation between a free swimming search phase and an intraperiplasmic growth phase (reviewed in 20). While in the search phase, bdellovibrios are highly motile and metabolically active, but do not replicate their DNA. Upon contact with a host cell, the parasite attaches to, and then rapidly penetrates, the host outer envelope and becomes lodged within the periplasmic space. During these early stages of the interaction with host cells, the bdellovibrio undergoes a transition from search phase to growth phase. In the initial stages of the growth cycle, the parasite conducts the partial degradation of many host cell components. These products are then used for biosynthesis and energy generation as the bdellovibrio elongates into a coiled, multicellular filament. At the cessation of growth, the filament divides into individual search phase progeny that are released back into the environment. In vitro the entire cycle requires about 3.5 hours to complete. The control of bdellovibrio growth is strict: all attempts to bypass the host cell requirement with commercial media have been unsuccessful (4, 10, 16). However, several studies have demonstrated that concentrated cellular extracts from hosts and other bacteria can induce wild-type bdellovibrios to enter the growth phase and support the completion of the entire growth cycle (4, 6, 10). No specific active factor has been identified in these extracts, but rudimentary analyses suggest that a heat-stable proteinaceous component may be involved (4, 6). To date, it is unknown how these extracts stimulate Bdellovibrio growth. 36 The requirement for a host cell can also be bypassed genetically. Spontaneous mutants of wild-type ”host-dependent” (H-D) Bdellovibrio have been isolated that are able to grow in the absence of hosts or host extracts. Such ”host-independent” (H-I) mutants have been obtained from each of the three recognized Bdellovibrio species, B. bacteriovorus (7, 21), B. stolpii (2) and B. .starrii (15, 17). In addition to being able to grow axenically on complex bacteriological media, most H-I strains retain limited intraperiplasnric (IP) growth capabilities. Such mutant strains are termed facultative, forming smaller, more turbid plaques than wild-type strains on lawns of host cells. Variants of facultative H-I strains that are incapable of IP growth have also been isolated (15, 21). The axenic growth and development of H-I mutants mimics that of wild-type strains growing intraperiplasmically. H-I mutants arise at a frequency of 10'6 - 10'7, suggesting that they can result from a single mutational event at one or multiple loci. A detailed characterization of the genetic locus or loci affected in H-I mutants should provide insight into the regulation of the bdellovibrio developmental cycle. Toward this end, we have identified vectors that can be used to transfer DNA into bdellovibrios by conjugation and have used the system to identify a 5.6 BarnHI fragment from wild-type B. bacteriovorus 1091 that greedy enhances the plaquing ability of an H-I mutant (Chapter 1). Here we describe further genetic and molecular characterization of this region from wild-type B. bacteriovorus 1091 and three independent H-I mutants. The data indicate that each mutant has suffered a mutation within this region of DNA and define a locus, hit (host- interaction), that affects the interaction of Bdellovibrio with host cells. MATERIALS AND METHODS Bacterial straim, plasmids, media and culture conditions. The bacterial strains and plasmids used are listed in Table 1. B. bacteriovorus 1091 and E. coli ML35 were 37 Table 1. Bacterial strains and plasmids BamHI site of pUC18 Strain or plasmid Description Source or reference B. bacteriavonls 1091 wild-type (1 1) 1091.1 Smr derivative of 1091 (other paper) 1091.2 Rf’ derivative of 1091.1 (Chapter 1) BB3 H-I derivative of 1091.1 This study BB4 H-I derivative of 1091.1 This study BBS H-I derivative of 1091.2 (Chapter 1) E. coli ML35 B lacI IacY (12) DHS F‘ emu] recAI (5) hSdRI7(11( mK+) deoR thi 1 supE44 gyrA96 relAI DHSa F' ¢80dlacZ M15 (lacZ YA Bethesda Research argF)U169 e 1 real] Laboratories hstI 7(rK rK'mK )deoR thi- I supE44 gyrA96 relAI DHSaF' F' ¢80dlacZ M15 (lacZYA Bethesda Research argF)U169 e 1 real] Laboratories hst17(rK mK )deoR thi- I supE44 gyrA96 relAI SM10 supE44 hst thi- I thr-I euB6 (18) lach tonA21 recA Muc RP4-2Tc: Mzu, Kmr Plasmids pKC7 ColEl Ap: Kmr (9) pUC18 ColEl Ap; (22) pUCl9 ColEl Aprr (22) pRK2013 ColEl Kmr tra(RK2) (3) pVKIOO IncP Tc: Km; (8) pVK102 IncP Tor Kmr (8) pVKa-l pVKIOO derivative containing the 445 bp This study HaeII fragment from pUCl9 in EcoRI site pTC3 23.5 kb fragment of B. bacteriovorus (Chapter 1) 1091 DNA in EcoRI site of pVKIOO pTC7 19.5 kb fragment of B. bacteriovorus (Chapter 1) 1091 DNA in EcoRI site of pVKIOO pTC12 5.6 kb BamHI fragment from pTC7 in (Chapter l) Bng site of pVK102 pTCl6 5.6 kb BamHI fragment of BB3 DNA in This study BamHI site of pUC18 pTC17 S .6 kb BamHI fragment of BB4 DNA in This study Table l (cont'd). 38 pTC20 pTC28 pTC30 pTC32 pTC3S pTC36 pTC37 pTCSO pTC56 5.6 kb BamHI fragment of BB5 DNA in This study BamHI site of pUC18 5.6 kb BamHI fragment from pTCl6 in This study BgIII site of pVK102 5.6 kb BamHI fragment from pTC17 in This study BgIII site of pVK102 5.6 kb BamHI fragment from pTC20 in This study BgIII site of pVK102 2.5 kb BamHI-Xbal fragment from pTC12 This study in BamHI-Xbal site of pVKa-l 3.1 kb BamHI-Xbal fragment from pTC12 This study in BamHI-Xbal site of pVKa-l 2.8 kb EcoRI fragment from pTC12 This study in EcoRI site of pVKa-l 0.96 kb EcoRI-Xbal ment from This study pTC12 in BamHI site a pVKa-l 0.96 kb EcoRI-Xbal fragment from This study pTC32 in BamHI site of pVKa-l 39 obtained from S. C. Rittenberg. Antibiotic resistant mutants and H-1 mutants of B. bacteriovorus 1091 were isolated as described by Seidler and Starr (15). All bdellovibrio strains were single plaque or single colony purified and stored in 15 % glycerol at -80°C. Media and culture conditions used for the propagation of all E. coli and Bdellovibrio strains was as previously described (Chapter 1). Chemicak and reagents. Complex medium components were purchased from Difco. Restriction endonucleases and T4 DNA ligase were purchased from New England Biolabs. [a-32P]dCTP (800 Ci/mM) and [a-3SS]dATP (500 Ci/mM) were purchased from DuPont/New England Nuclear. Matings. The conjugal transfer of plasmids from E. coli to B. bacteriovorus was done as previously described (Chapter 1). All of the genetic experiments described here involved the the conj ugal transfer of bdellovibrio DNA containing IncP plasmids, and their integration into the recipient bdellovibrio genome via homologous recombination. All of the recipients were therefore merodiploid for the cloned sequences. Southern blot analysis has previously shown that integration of Bdellovibrio DNA containing constructs occurs via homologous recombination between the cloned Bdellovibrio insert and the equivalent region of the recipient genome (Chapter 1). DNA manipulations and comtruction of pVKa-l. Most DNA purification and recombinant DNA methods were those of Sambrook et a1. (13). Bdellovibrio genomic DNAs were purified by a CTAB based extraction procedure (1). The 5.6 kb BamHI fragment that contains part or all of the hit locus was cloned from H-I mutants BB3, BB4 and BBS by isolating genomic BamHI fragments that ranged from 5.0 to 6.0 kb in size and ligating them into pUC18. BamHI fragments in that size range were isolated by electroelution (13) after agarose gel electrophoresis. 40 pTC 16 (BB3), pTC17 (BB4) and pTC20 (BBS) were identified as containing the correct 5.6 kb BamHI fragments by colony hybridization (13) using pTC12 as probe. To allow for their conjugal transfer into bdellovibrio, the three mutant derived BamHI fragments were subcloned into the BglII site of pVK102 yielding pTC28 (BB3), pTC30 (BB4) and pTC32 (BB5). pVKa-l was constructed by cloning the 445 bp HaeII fragment from pUC19 (that contains the polylinker and a-complementation sequences) into the EcoRI site of pVK100. Prior to ligation, the overhanging termini of each fragment were removed by treatment with the Klenow fragment. pVKa-l allows the use of ctr-complementation for cloning and contains unique BamHI, KpnI, PstI, SstI and 1001 restriction sites within the polylinker. DNA sequencing. DNA sequences were determined by the dideoxy chain- ternrination method (14) using the Sequenase procedure of United States Biochemical with single stranded M13 DNA as template. Sequence of the wild-type 959 bp EcoRI- XbaI fragment was obtained from both strands. The equivalent, mutant derived fragments were sequenced in their entirety on one strand, using primers that were designed according to the wild-type sequence. Discrepancies with the wild-type sequence were confirmed by sequencing on the complementary strand. Open reading frame and similarity analyses were performed on a VAX 8650 computer using University of Wisconsin GCG protein and DNA analysis software (version 6.0). RESULTS Identification of a mutation in H-I mutant BB5. We previously showed that recombination of pTC12 into the genome of B. bacteriovorus BBS greedy enhances the plaquing ability of this H-I mutant (Chapter 1; Figure 1). A likely explanation of this 41 Figure 1. pTC32 does not confer an enhanced-plaque phenotype upon BB5. Plaques formed by (A) 1091.2(pTC3) (wild-type), (B) BBS(pTC3) (I-I-I), (C) BBS(pTC12) and (D) BBS(pTC32) are compared. Although pTC12 and pTC32 contain equivalent 5.6 kb BamI-H inserts derived from wild-type and H-1 mutant BBS, respectively, pTC32 does not confer enhanced plaquing in BBS. pTC3 does not affect the plaque phenotype of H-D or H—I strains and is used here as a control to confer Kmr. 42 Figure 1. 43 result was that BBS contained a mutation within the region of the genome corresponding to the 5.6 kb BamHI fragment cloned in pTC12 and that recombination of the wild-type sequences at this locus resulted in at least a partial correction of the original genetic lesion. To test this hypothesis, we isolated the 5.6 kb BamHI region from BBS (carried on pTC32) and compared it to that of the wild-type fragment at the physical and genetic levels. Comparative restriction analysis of wild-type and mutant derived 5.6 kb BamHI fragments did not reveal any obvious deletions or rearrangements. Genetic experiments, however, indicated otherwise. Specifically, in contrast to pTC12, recombination of pTC32 into the BBS genome did not enhance the plaquing ability of BBS (Figure 1). To narrow down the site of the mutation, various portions of the 5.6 kb BamHI fragment from pTC12 were subcloned, yielding pTC3S, pTC36, pTC37 and pTCSO (Figure 2), and the smaller DNA fragments were tested for whether they could enhance the plaquing ability of BB5. When recombined into BBS, pTC36 had no effect on plaque morphology whereas pTC35, pTC37 and pTCSO mimicked the effects of pTC12 (Figure 2). These data suggested that BBS had suffered a mutation within the EcoRI-Xbal fragment cloned in pTCSO. This conclusion was confirmed by DNA sequence analysis (Figure 3). The data indicated that the corresponding EcoRI-Xbal fragments from wild-type B. bacteriovorus 1091 and its H-I derivative BBS differed by one base pair; the mutant had a single base pair deletion located 208 bp from the Xbal site (marked by the closed circle in Figure 3). As this region of the genome had an effect on Bdellovibrio plaquing, the locus was designated hit (host-interaction). The hit locus is altered in other spontaneous H-I mutants of B. bacteriovorus 1091. The occurence of a mutation within the hit locus of BBS raised the question of whether this locus was commonly affected in H-I mutants derived from B. bacteriovorus 1091. I addressed this issue by isolating two additional H-I mutants, 44 Figure 2. Plasmid constructs containing B. bacteriovorus 1091 DNA and their ability to enhance plaque formation in H-I mutant BBS. Inserts from plasmid constructs that contain overlapping fragments of B. bacteriovorus 1091 DNA are shown. Insert sizes are: pTC12, 5.6 kb; pTC35, 2.4 kb;pTC36, 3.2 kb; pTC37, 2.8 kb; pTCSO, 0.96 kb. Restriction sites: B, BamHI; E, EcoRI; X, 1601. pTCtz pTC$5 press pTC37 pTCSO 45 Plaque Enhancement H X H B X i.___1 X Figure 2 E + .l. + 46 .35 E coma—av Lo sees a 382 some see .5 2 .2865 ”escape: 53589.9 .95 c5 35 .95 E 23.52:. ease—ac Co 22.82 88%5 238.? £86 vac—o Ea 29.8.5 635m: name—U 6:8. as a... as. .5595 Exec—8m .3 e3 2: a 85:8... <2: .m can: 4fl7 .m «Sufi _¢ouu uhp¢9 (poo—uu—uo uc<~