Construction and Comparison of Recombinant Plasmids Encoding Type 1 Fimbriae of Members of the Family Enterobacteriaceae

Transcription

1 NFECTON AND MMUNTY, May 1985, p /85/ $02.00/0 Copyright 1985, American Society for Microbiology Vol. 48, No. 2 Construction and Comparison of Recombinant Plasmids Encoding Type 1 Fimbriae of Members of the Family Enterobacteriaceae STEVEN CLEGG,'* SHELA HULL,2 RCHARD HULL,2 AND JANET PRUCKLER1 Department of Microbiology, University of owa, owa City, owa 52242,1 and Department of Microbiology and mmunology, Baylor College of Medicine, Houston, Texas Received 25 October 1984/Accepted 21 January 1985 The genes encoding type 1 fimbriae of Salmonella typhimurium, Enterobacter cloacae, and Serratia marcescens were cloned in Escherichia coli. All transformants possessing recombinant plasmids were shown to be fimbriate and demonstrated mannose-sensitive hemagglutinating activity. A comparison of the physical maps of these plasmids revealed little similarity among them, although plasmids encoding type 1 fimbriae of Escherichia coli and Klebsiella pneumoniae appeared similar with respect to restriction enzyme sites. The fimbrial gene cluster ranged in size from 5.5 to 9.0 kilobase pairs as determined by transposon mutagenesis. Plasmid-containing E. coli strains were shown to produce species-specific fimbrial antigens with little or no cross-reactivity between genera. Therefore, it was presumed that each plasmid contained the gene encoding the fimbrial subunit. Complementation was not detected between nonfimbriate insertion mutants of different species but was seen with mutants of the same species. Many different genera within the family Enterobacteriaceae have been found to produce type 1, or mannose-sensitive (MS), fimbriae (4, 6). Although type 1 fimbriae appear morphologically similar regardless of the genus in which they are observed, they have been shown to be serologically unrelated (4, 9, 16). Differences in the polypeptide subunits of the fimbrial protein may explain this antigenic diversity. ndeed, the amino acid compositions of fimbriae purified from Escherichia coli and Salmonella typhimurium have been shown to be different (13). Recently, the E. coli type 1 fimbrial gene cluster was cloned, and with the techniques of recombinant DNA technology, it was demonstrated that a minimum of four proteins, organized into at least three transcriptional units, is necessary for fimbrial expression in E. coli (18). Studies of the molecular biology of type 1 fimbrial production have been restricted to the fimbriae of E. coli and Klebsiella pneumoniae (19). t is presently unknown whether the genetic components which encode type 1 fimbriae in other enteric bacteria are similar to the E. coli and K. pneumoniae systems or whether they differ significantly at the molecular level. This paper describes the cloning of the DNA sequences encoding fimbriae in S. typhimurium, Enterobacter cloacae, and Serratia marcescens. MATERALS AND METHODS Bacteria, growth conditions, and HA tests. The species of enteric bacteria used to prepare genomic libraries are shown in Table 1. A nonfimbriate strain of E. coli HB101 (hsdm hsdr reca) was used in all transformation experiments unless otherwise stated. Bacterial cultures were grown in Luria broth or on Luria agar (15) for 18 to 24 h at 37 C. Culture media were supplemented with the appropriate antibiotics at the following concentrations: ampicillin, 100,u.g/ml; tetracycline, 20 jig/ml; kanamycin, 20,ug/ml; chloramphenicol, 25,.g/ml (200,ug/ml for plasmid amplification). Spectinomycin (200 ug/ml) was used for the amplification of plasmids conferring chloramphenicol resistance. * Corresponding author. 275 Hemagglutination (HA) tests were performed on glass microscope slides with 1 drop of a freshly prepared suspension (3% [wt/vol] in phosphate-buffered saline) of guinea pig erythrocytes mixed with a small amount of bacteria. After the slide was gently rocked for 10 to 20 s, MSHA+ cultures agglutinated the erythrocytes but showed no activity when the erythrocyte suspension contained 2% (wt/vol) D-mannose. The HA activity of a bacterial suspension was quantitated in microtiter trays with serial twofold dilutions (25 [L1) mixed with an equal volume of erythrocytes. After they were mixed, the trays were placed on ice for 60 min, and the highest dilution causing visible agglutination was designated as 1 HA unit. HA inhibition tests were performed by mixing a diluted bacterial suspension (3 HA units in 25,ul) with dilutions of antifimbrial serum (25,ul), followed by incubation at 37 C for 60 min. Subsequently, 25,ul of erythrocytes was added to all reactions, and the titer of the serum was recorded as the highest dilution responsible for inhibition of an HA reaction. Preparation of antifimbrial serum. Serum raised against fimbriate strains was prepared as described by Duguid et al. (5). Nonfimbrial agglutinins were removed from the sera by absorption with nonfimbriate phase cultures. Preparation of DNA. Chromosomal DNA was prepared from overnight broth cultures (250 ml) after incubation at 37 C. After centrifugation, the bacterial pellet was suspended in 1.5 ml of 0.15 M NaCl-0.1 M EDTA (ph 8.0) containing 4.0 mg of lysozyme, and after incubation for 15 min at 37 C, the cells were frozen at -60 C. A solution (12.5 ml) of 0.1 M NaC-0.1 M Tris (ph 8.0)-1.0% sodium dodecyl sulfate was added to the frozen mixture which was thawed at 65 C, and after a second freeze-thaw cycle, a 0.5 volume of phenol saturated with TE buffer (10 mm Tris [ph 8.0], 1 mm EDTA) was added. After extraction, the DNA was precipitated by the addition of 2 volumes of 95% ethanol, centrifuged for 30 min (17,400 x g) and redissolved in 5 ml of TE buffer. After treatment with RNase (50 p.g) for 60 min at 37 C, the DNA preparation was deproteinized by extraction with phenol. Finally, the precipitated DNA was redissolved in 1.0 ml of TE buffer.

2 276 CLEGG ET AL. NFECT. MMUN. TABLE 1. Origin of enteric bacterial species used and name of recombinant plasmid derived from bacterial chromosomes Strain Source Reference Plasmid S. marcescens A506 Sputum This paper pmh2 S. typhimurium 6704 Feces This paper psf101 E. cloacae A524 Wound This paper pech3 K. pneumoniae A561 Urine 13 pbp7 E. coli J96 Urine 12 psh2 Plasmid DNA was purified from amplified cells (3) by sodium dodecyl sulfate lysis and ethidium bromide-cesium chloride equilibrium density gradient centrifugation (10, 20). Rapid detection of plasmids from transformed cells was performed by the method of Holmes and Quigley (11). The construction and expression of plasmids pbp7 (19) and psh2 (12) have been described in detail elsewhere. Restriction enzyme digestions. All restriction endonucleases used in these studies were purchased from commercial sources and used according to the instructions of the manufacturers. Enzyme digestions were performed in 25-,l volumes, and digestion patterns were observed after electrophoresis through 0.7% agarose gels as previously described (19). Preparation of genomic libraries. The cosmid cloning technique was used to prepare chromosomal libraries of the enteric bacteria (14, 19). Cosmid pmf7 (8) was kindly supplied by M. G. Feiss. Partially digested chromosomal DNA fragments were prepared after electrophoresis through and elution from a low-gelling-temperature agarose (FMC nserted DNA from S. typhimurium plfl01 E. cloacae S. marcescens pmmh2 No A a Ez : AAA a a w. Corp., Marine Colloids Div., Rockland, Maine). After ligation of these fragments to pmf7 which had been digested by the appropriate enzyme, recombinant molecules were packaged into lambda phage particles (14). E. coli transductants were selected on Luria agar containing ampicillin and were subsequently screened for HA activity. solation of TnS insertions. The phage A b221 cl857::tn5, carrying the TnS element encoding kanamycin resistance, was used to infect recombinant strains as previously described (19). Subsequently, plasmid DNA was prepared from these strains, and after transformation of the nonfimbriate E. coli HB101, bacteria were plated on agar containing kanamycin and the appropriate antibiotic. The location of TnS insertions was determined by analysis of restriction endonuclease fragments. RESULTS Cloning of the DNA sequences encoding type 1 fimbriae of S. typhimurium, S. marcescens, and E. cloacae. Each of the three species was shown to be strongly fimbriate as evidenced by its ability to agglutinate guinea pig erythrocytes and by electron microscopy. After construction of genomic libraries from each species, Ampr transductants were screened for MSHA activity. The frequency of transduction to the MSHA+ phenotype ranged from 1 in 200 for S. typhimurium to 1 in 400 for E. cloacae. Plasmids prepared from hemagglutinating clones transformed E. coli HB101 to Ampr and HA activity with 100% efficiency. a -i a c c O S -- A A AA A A A A 8 - E, x o S i S LB E a U S a u 0uw. z E a 0. a au 5 (A.6 K. pneumoniae pp7 0 w 0o 3uU E. coli a CL0_ E > - 40X w psm2 5 S FG. 1. Physical maps of plasmids encoding type 1 fimbriae. The sites of TnS insertions and their HA phenotypes are shown by triangles (open, MSHA+; solid, MSHA-). The thick black lines represent cloning vehicle DNA sequences. 1 kb

3 VOL. 48, 1985 PLASMDS ENCODNG FMBRAE OF ENTERC BACTERA 277 psf101 3 L S {3 A) i S A A (23 (23 A A. A E a A A AA A A (] (13 (13 tl( (23 (23 a Recipient strain possssing plasmid fromn Recipient strain transformed by plasmid of pbp7 group psf101 group pbp7 group (13 (2) psf101 group (l3 (2) to (23 + _ () - + ( FG. 2. Complementation of insertion mutants of pbp7 and psf101. Tn5-insertion sites are indicated by triangles. The numbers in parentheses indicate the complementation group. Plasmids from a single group do not complement each other to express MSHA activity (-), whereas plasmids of different complementation groups of the same species retain the MSHA phenotype (+). Type 1 fimbrial genes from the three species were further subcloned by conventional techniques by using the cloning vehicles pbr322 and pacyc184. The smallest plasmids constructed which retained the coding information for S. typhimurium, E. cloacae, and S. marcescens fimbrial production were designated psf101, pech3, and pmh2, respectively (Fig. 1). Thus, psf101 contained a kilobase pair (kb) EcoR fragment derived from the chromosome of S. typhimurium Removal of either the 9.5- or the 2.5-kb Hindll fragment from this plasmid resulted in the production of MSHA- transformants. Plasmid pech3 was constructed from a 15-kb E. cloacae chromosomal DNA fragment, and further subcloning of the 8.3-kb Cla fragment failed to transform E. coli HB101 to HA activity. Similarly, Sall, Hind, and BamH fragments of pech3 did not possess complete genetic information to encode functional type 1 fimbriae. A 15-kb chromosomal DNA segment isolated from S. marcescens A506 constitutes the inserted DNA of pmh2, and once again, further subcloning did not result in the construction of plasmids which retained the hemagglutinin-coding capacity. Physical maps of these plasmids are shown in Fig. 1, which also includes the recombinant plasmids encoding type 1 fimbriae of K. pneumoniae (pbp7) and E. coli (psh2). A comparison of the number and locations of restriction endonuclease sites within these plasmids revealed little similarity except for pbp7 and psh2. Estimation of the size of the DNA fragment necessary for fimbrial expression. The transposon TnS was used to determine the approximate length of DNA essential for expression of the MSHA phenotype in the three species (Fig. 1). Transformants were selected on agar containing kanamycin and the appropriate second antibiotic, the site of the Tn5 insertion being mapped by restriction enzyme analysis. For each plasmid, a minimum of 10 independent insertions was mapped, including at least five which resulted in the loss of HA activity. As reported for E. coli and K. pneumoniae (12, 19), a DNA length of 5.5 to 8.0 kb, depending upon the species, was necessary to contain the genes encoding phenotypic expression of fimbriae. n all cases, the insertion of the TnS element into these regions eliminated MSHA activity, and representative isolates, when observed by electron microscopy, were found to be nonfimbriate. n the case of plasmid pech3, no Tn5 insertion, which retained HA activity, could be mapped in the 2.0-kb BamH-EcoR region (Fig. 1). Complementation of TnS insertion mutants of pbp7 and psf101. MSHA- insertion mutants of pbp7 and psf101 were used to detect complementation groups in these plasmids (Fig. 2). At least two complementation groups were found with mutants of pbp7 alone, and strains possessing two different plasmids from each group, with the compatible cloning vectors pacyc184 and pbr322, were converted to HA+ phenotypes. Similar results were observed for psf101. Transformants containing one derivative of each plasmid (pbp7 and psf101) were never converted to the production of fimbriae as determined by HA. Such transformants possessed both plasmids, however, as shown by the number and sizes of restriction fragments after digestion. Serological specificity of fimbrial antigens produced by recombinant strains. Fimbrial antisera raised against the wild-type enteric bacteria were used to determine the type of fimbriae encoded by the recombinant plasmids. The HA inhibition titers of the antisera, titrated against plasmid-containing MSHA+ strains of E. coli, are shown in Table 2. n all cases, the HA inhibition titer of a fimbrial antiserum was at least fourfold greater with the recombinant strain encoding the homologous antigen than with the other strains. Thus, Serratia and Enterobacter antifimbrial sera only inhibited MSHA by E. coli containing plasmids pmh2 and pech3, respectively. Some cross-reactivity was observed by using the K. pneumoniae fimbriae with the E. coli antiserum, and

4 278 CLEGG ET AL. TABLE 2. HA inhibition titers of fimbrial antisera E. (oli Fimbrial antisera raised against": HB101 S. (vp/li- S. marcescontaining: E. coli K. pnieinotii(ie E. cloacae /t71li'li-il# (1cens psh <20 <20 20 pbp <20 psf <20 pech3 <20 <20 <20 80 <20 pmh2 <20 <20 <20 < ` Underlined titers are those of the homologous antigen-antibody system. Lowest dilution of a serum used in test was 1 in 20. similarly, the S. typhimurium antigen reacted with the E. coli antiserum; however, the titer in each case was significantly lower than with the homologous system. DSCUSSON The genes encoding type 1 fimbriae of S. marcescens, S. typhimurium, and E. cloacae were cloned in E. coli. A comparison of the physical maps of recombinant plasmids, along with those encoding fimbriae of E. coli and K. pneumoniae, demonstrated little similarity. For example, pbp7 possesses unique enzyme sites (e.g., BamH) within the inserted DNA when compared with psf101. Similarly, both plasmids possess two EcoR and Cla sites, but the map positions of these sites with respect to each other are different. The differences in the restriction maps of these plasmids may indicate differences at the molecular level of the gene clusters involved in fimbrial expression. However, a more detailed comparison between these genes and their gene products must be made before conclusions can be drawn concerning the homology (or lack thereof) of the type 1 fimbrial genetic apparatus in enteric bacteria. Experiments are currently in progress, using the minicell system of E. coli and deletion derivatives of these plasmids, to determine the number and location of polypeptides encoded by the recombinant molecules. As in other type 1 fimbrial gene clusters (12, 19), those of S. marcescens, E. cloacae, and S. typhimurium appeared to possess more coding information than is necessary to determine the fimbrial subunit (Fig. 1). For example, the molecular weight of the S. typhimurium fimbrial subunit has been reported to be 21,000 (13), whereas the amount of DNA necessary for its phenotypic expression was found to be approximately 5.0 kb. Thus, the length of DNA is significantly greater than that required to contain only the gene encoding the fimbrial subunit. This suggests that accessory gene products are contained within the chimeric plasmids and that these polypeptides are involved in phenotypic expression of fimbriae. Previous reports have shown that at least four polypeptides are essential for fimbrial expression in E. coli (18). Therefore, the question arises whether these accessory proteins are genus or species specific for fimbrial expression. nitial evidence, using the compatible cloning vehicles pbr322 (1) and pacyc184 (2), suggests that genes of S. typhimurium and K. pneumoniae cannot act in trans to encode functional fimbriae. However, complementation was found with mutants derived from the same species (Fig. 2). A more extensive analysis of complementation reactions among all the fimbrial plasmids is presently being undertaken. With serological tube agglutination assays, it was found that the type 1 fimbriae produced by different genera of the Enterobacteriaceae are serologically distinct (4, 16, 19). NFECT. MMUN. Therefore, if the plasmids described in this report possess the genes encoding the fimbrial subunit of each genus, E. coli strains carrying chimeric plasmids would be expected to react with the serum raised against the fimbriae of the wild-type strains. ndeed, such specificity was observed with the HA inhibition test. Cross-reactivity was observed only in two instances. A sharing of fimbrial antigenic determinants among isolates of E. coli and K. pneumoniae was reported previously (4); the reciprocal cross-reaction was weak in our experiment. Nevertheless, the similarity between fimbrial antigens of these two genera was reflected by a close degree of homology between the physical maps of psh2 and pbp7. The results described herein demonstrate that foreign fimbrial appendages can be synthesized in E. coli and are serologically identical to those of the wild-type strains. The precise role of type 1 fimbriae of enteric bacteria remains unknown, although the adhesive properties of these organelles undoubtedly convey an important property on the bacterial cell, as evidenced by the widespread occurrence of fimbriation among this group of bacteria. Speculation has centered around the possible role of these adhesins in facilitating infection of animals and humans (6, 7, 17). Alternatively, fimbriae may be responsible for attachment to surfaces by commensal or free-living members of the Enterobacteriaceae. The cloning of the fimbrial genes from different genera within this family will facilitate a comparative analysis of this genetic system. The recombinant plasmids described above will be used to provide information concerning the conservation of DNA elements coding for structurally and functionally related bacterial appendages. ACKNOWLEDGMENTS This work was supported in part by Public Health Service grants A and A-AM21009 from the National nstitute of Allergy and nfectious Diseases. We thank Marcia Reeve for typing the manuscript and. P. Crawford for comments on the manuscript. LTERATURE CTED 1. Bolivar, F. R., L. Rodriguez, P. J. Greene, M. C. Betlach, H. L. Heyneker, J. H. Crosa, and S. Falkow Construction and characterization of new cloning vehicles.. A multipurpose cloning system. Gene 2: Chang, A. C. Y., and S. N. Cohen Construction and characterization of amplifiable multicopy DNA cloning vehicles derived from the P1SA cryptic miniplasmid. J. Bacteriol. 134: Clewell, D. B Nature of ColE, plasmid replication in Escherichia coli in the presence of chloramphenicol. J. Bacteriol. 110: Duguid, J. P., and. Campbell Antigens of the type 1 fimbriae of Salmonella and other enterobacteria. J. Med. Microbiol. 2: Duguid, J. P., S. Clegg, and M.. Wilson The fimbrial and nonfimbrial hemagglutinins of Escherichia coli. J. Med. Microbiol. 12: Duguid, J. P., and D. C. Old Adhesive properties of Enterobacteriaceae, p n E. H. Beachey (ed.), Bacterial adherence (receptors and recognition series B: vol. 6). Chapman and Hall, Ltd., London. 7. Fader, R. C., and C. P. Davis Effect of piliation on Klebsiella pneumoniae infection in rat bladders. nfect. mmun. 30: Feiss, M. G., D. A. Siegele, C. F. Rudolph, and S. Frackman Cosmid DNA packaging in vivo. Gene 17: Gillies, R. R., and J. P. Duguid The fimbrial antigens of Shigella flexneri. J. Hyg. 56: Guerry, P., D. J. LeBlanc, and S. Falkow General method

5 VOL. 48, 1985 PLASMDS ENCODNG FMBRAE OF ENTERC BACTERA 279 for the isolation of plasmid deoxyribonucleic acid. J. Bacteriol. 116: Holmes, D. S., and M. Quigley A rapid boiling method for the preparation of bacterial plasmids. Anal. Biochem. 114: Hull, R. A., R. E. Gill, P. Hsu, B. H. Minshew, and S. Falkow Construction and expression of recombinant plasmids encoding type 1 of D-mannose-resistant pili from a urinary tract infection Escherichia coli isolate. nfect. mmun. 33: Korhonen, T. K., K. Lounatmaa, H. Ranta, and N. Kuusi Characterization of type 1 pili of Salmonella typhimurium LT2. J. Bacteriol. 144: Meyerowitz, E. M., G. M. Guild, L. S. Prestidye, and D. S. Hogness A new high capacity cosmid vector and its use. Gene 11: Miller, J. H Experiments in molecular genetics. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y. 16. Nowotarska, M., and M. Mulczyk Serologic relationship of fimbriae among the Enterobacteriaceae. Arch. mmunol. Ther. Exp. 25: Ofek,., and E. H. Beachey Mannose binding and epithelial cell adherence of Escherichia coli. nfect. mmun. 22: Orndorff, P. E., and S. Falkow Organization and expression of genes responsible for type 1 piliation in Escherichia coli. J. Bacteriol. 159: Purcell, B. K., and S. Clegg Construction and expression of recombinant plasmids encoding type 1 fimbriae of a urinary Klebsiella pneumoniae isolate. nfect. mmun. 39: Selker, E., K. Brown, and C. Yanofsky Mitomycin C-induced expression of trpa of Salmonella typhimurium inserted into the plasmid ColEl. J. Bacteriol. 129: Downloaded from on October 10, 2018 by guest