|
Mem Inst Oswaldo Cruz, Rio de Janeiro, Vol. 99, No. 7, November, 2004, pp. 727-732 Differences in the Stability of the Plasmids of Yersinia pestis Cultures in Vitro: Impact on Virulence TC Leal-Balbino+, NC Leal, CV Lopes, AMP de Almeida Departamento de
Microbiologia, Centro de Pesquisas Aggeu Magalhães-Fiocruz, Campus
da UFPE, Cidade Universitária, 50670-420 Recife, PE, Brasil Received 2 March
2004 Code Number: oc04147 Plasmid and chromosomal genes encode determinants of virulence for Yersinia pestis, the causative agent of plague. However, in vitro, Y. pestis genome is very plastic and several changes have been described. To evaluate the alterations in the plasmid content of the cultures in vitro and the impact of the alterations to their pathogenicity, three Y. pestis isolates were submitted to serial subculture, analysis of the plasmid content, and testing for the presence of characteristic genes in each plasmid of colonies selected after subculture. Different results were obtained with each strain. The plasmid content of one of them was shown to be stable; no apparent alteration was produced through 32 subcultures. In the other two strains, several alterations were observed. LD50 in mice of the parental strains and the derived cultures with different plasmid content were compared. No changes in the virulence plasmid content could be specifically correlated with changes in the LD50. Key words: plague - Yersinia pestis plasmids - virulence Plasmid and chromosomal genes encode determinants of virulence for Yersinia pestis, the causative agent of plague (Perry & Fetherston 1997, Parkhill et al. 2001, Deng et al. 2002). A 102-kb, unstable chromosomal area (locus pgm) is essential for Y. pestis virulence (Fetherston et al. 1992, Hinnebusch et al. 1996, Buchrieser et al. 1998). Typical strains of Y. pestis harbour three plasmids: pPst (9.5kb), encoding a plasminogen activator protease (Pla) (Sodeinde & Goguen, 1988, 1989); pFra (90kb), encoding a capsular protein Fraction 1 (F1) with antiphagocytic activities (Du et al. 2002) and murine toxin (Ymt), required for survival in the flea (Hinnebusch et al. 2002); pYV (70kb), encoding the Yop virulon which comprises both the Yop effectors proteins and the proteins necessary for injecting them into host cells. The Yop virulon enables the bacteria to survive and multiply in the lymphoid tissues of the host (Cornelis 2002). Several insertion sequences (IS100, IS200, IS285) present in the three plasmids favour recombination events and genetic plasticity (Filippov et al. 1995, Parkhill et al. 2001, Deng et al. 2002). However, atypical strains lacking some plasmids have been found in several foci around the world. On the other hand, strains containing extra DNA bands or additional cryptic plasmids have also been found (Filippov et al. 1990, Chu et al. 1998). In vitro, Y. pestis genome is very plastic and several changes have been described: emergence of additional DNA-bands, increasing of plasmid molecular mass, and integration of plasmids into the bacterial chromosome with or without loss of functions (Zsigray et al. 1985, Protsenko et al. 1991). Furthermore, the typical plasmids may be eliminated spontaneously at a high frequency during storage in the laboratory or through successive subcultures (Protsenko et al. 1991). The study of the plasmid content of Y. pestis isolates from the plague foci of Northeast Brazil, stored in the laboratory for several decades, showed that some of them displayed an atypical plasmid profile characterized by the absence of some plasmids or by the presence of extra DNA bands (Leal et al. 1997a, Leal & Almeida 1999, Cavalcanti et al. 2002). The absence of plasmids and the emergence of extra-DNA bands in the Brazilian isolates could also have been produced during storage or handling in the laboratory. To observe possible alterations in plasmids of the cultures in vitro and the impact of the alterations to their pathogenicity, three low subcultured and highly pathogenic Y. pestis isolates were submitted to serial subcultures. Colonies selected after subcultures were analyzed for their plasmid content and the presence of some characteristic genes of each plasmid. LD50 in mice of the parental strains and derived cultures displaying different plasmid contents were compared. MATERIALS AND METHODS Bacteria and culture conditions - The study involved two Brazilian isolates: one old (P. Exu 340, originating from a finger bone marrow in a fatal human case in 1969), and the most recent Brazilian isolate (P. CE 882, originating from a hemoculture from a plague case in 1997), and one isolate from another South America focus for comparison (P. Peru 375 originating from a plague patient from Peru in 1994). For this work a loopful of each parental culture was grown in BHI (brain heart infusion broth, Difco Laboratories, Detroit, MI, US) at 28ºC for 24 h followed by streaking on BAB (blood agar base, Difco) for 48 h. Isolated colonies were picked and grown in BHI and the plasmid content of each culture was determined. The derivative cultures were named YP (for Y. pestis), the number of the parental culture followed by the number of the subculture (Table). Subcultures on Congo red agar plates (CRA) - Each derivative culture was submitted to serial CRA. This medium is used as a phenotypic marker for chromosomal pathogenicity enabling us to differentiate red-pigmented virulent (Pgm+) colonies and white avirulent (Pgm-) colonies (Jackson & Burrows 1956, Surgalla et al. 1970, Fetherston et al. 1992). Colonies selected from the CRA plates were grown in BHI and immediately analyzed for their plasmid content, the presence of specific plasmid genes and virulence in mice. DNA procedures - Plasmid DNA was extracted by the alkaline lysis method of Birnboim and Doly (1979) and electrophoresed in 0.6% agarose gels in Tris-borate buf-fer at a constant voltage of 100 V, followed by staining with ethidium bromide (10 mg/ml) and visualisation under UV. The plasmid sizes were determined by comparison with plasmids of known sizes (147, 63, 35.8, 6.9 kb) present in Escherichia coli 39R861 and the prototypical plasmids of the vaccine strain Y. pestis EV76 (~ 100, 70, 9.5 kb). Total DNA was extracted according to Maniatis et al. (1989) and quantified after electrophoresis in 1% agarose gels by comparison with a known amount of l HindIII DNA. PCR analysis - Amplification of the plasmidial genes caf1 (506 pb) located on pFra, pla (920 pb) located on pPst, lcrV (800 pb) located on pYV and the chromosomal gene irp2 (300 pb) was performed as described by Leal and Almeida (1999) using primers designed from published sequences (Sodeinde & Goguen 1989, Galyov et al. 1990, Motin et al. 1992, Guilvout et al. 1993). The gene irp2 was included in this study as a chromosomal pathogenicity marker (Buchrieser et al. 1998). Amplification of the gene ymt located on pFra was performed according to Leal et al. (1997b) with primers drawn by the program PrimerSelect (DNAstar, Inc.). Western-blot - Total proteins were extracted from cells collected by centrifugation from 1 ml of the broth culture that was incubated at 28ºC overnight. The pellet was suspended in 100 µl of Laemmli buffer, boiled for 10 min, and electrophoresed in homogeneous, 12.5% sodium dodecyl sulphate-polyacrylamide gels (SDS-PAGE) followed by staining with Coomassie-blue R or transferred to poly-vinylidene fluoride (PDVF) membranes (Immobilon-Millipore, MA, US). Molecular weight markers were (LMW): phosphorylase b (94 kDa) albumin (67 kDa), ovalbumin (43 kDa), carbonic anhydrase (30 kDa), trypsin inhibitor (20.1 kDa), and lactalbumin (14.4 kDa) (Amersham Biosciences of Brazil Ltd, SP, Brazil). For immunoblot, protein bands were transferred to PDVF membranes at 600 V. Membrane proteins were successively incubated with rabbit anti-F1 antigen polyclonal serum and peroxidase-conjugated goat anti-rabbit globulins (Amersham). Bound antibodies were detected by diaminobenzidine (DAB) (Sigma, MO, US). Rabbit anti-F1 antigen polyclonal serum was prepared following protocols described by Chu (2000). Virulence in mice - Animal experiments were carried out with the approval of the local ethics committee (Ceua/Fiocruz P0049-00). Tenfold serial dilution (10-1 to 10-9) in sterile normal saline was prepared from BAB slants of each culture grown at 28oC for 48 h. Four dilutions of each culture were tested in Swiss-Webster mice, five to six weeks of age, obtained from the facilities of the Centro de Pesquisas Aggeu Magalhães (CPqAM), Recife, Brazil. Four mice were inoculated with each dilution; each animal received 0.2 ml of suspension containing from 6 x 100 to 6 x 108 bacteria by the subcutaneous route. The animals were observed for two weeks. The number of bacteria per dose was determined by plate count of CFUs (colony forming units) on BAB plates. LD50 were calculated by the Reed and Muench method (1938). RESULTS AND DISCUSSION Analysis of the strain P. CE 882 - The determination of the plasmid content, after electrophoresis on agarose gels, of P. CE 882 and its derived cultures analyzed respectively after the 4th, 18th, 22nd, and 32nd subcultures, revealed the presence of the three typical plasmids: pFra, pYV and pPst in all of them. Neither curing of plasmids nor emergence of extra DNA bands was observed through 32 subcultures of that strain (Table, Fig. 1a, d). As expected, all DNA segments of sizes corresponding to the genes: caf1 (506 bp) and ymt (850 bp) located in pFra, lcrV (800 bp) in pYV, and pla (920 bp) in pPst were amplified in the parental strain P. CE 882 and its derived cultures. The gene irp2 (300 bp) used as a chromosomal pathogenicity marker was also amplified (Table, Fig. 2a). As shown in the Table there was no difference between the LD50 of P. CE 882 and the culture YP 882/32 obtained after 32 subcultures (LD50: > 4x103). However, it is worth pointing out that these LD50 values are higher than the one obtained at isolation (LD50: < 10) (Aragão et al. 2002). In spite of its apparent stability (no alteration had been detected though 32 subcultures) and for reasons still not explained, its virulence in mice was decreased. Analysis of the strain P. Exu 340 - After electrophoresis on agarose gels, only the plasmids pYV and pPst were visualised in the culture recovered from the bacterial collection (Table, Fig. 1b); also, the genes lcrV and pla located respectively in these two plasmids were amplified by PCR. The plasmid pFra was not visualised and the gene caf1 was not amplified. However, in spite of the absence of pFra and lack of amplification of caf1, the gene ymt was amplified in this culture and in its derived cultures until the 9th subculture. From the 10th subculture onward, there was no further amplification of this gene (Table, Fig. 2b). Three types of alterations of pFra have been described: integration into the chromosome with or without loss of function, mutation in the operon f1, and curing of this plasmid without integration into the chromosome (Protsenko et al. 1991). The integration of pFra into the chromosome seems to be quite frequent, occurring at multiple sites, and can be due to the presence of insertion elements (Filippov et al. 1995, Parkhill et al. 2001). The loss of pFra in P. Exu 340 might have occurred in the laboratory, because the presence of the gene ymt was still detected by PCR from the culture recovered from the collection and from its derived cultures until the 9th subculture, disappearing only after the 10th subculture. Possibly, there was a transiently integration of ymt into the chromosome and its elimination afterwards. An equally likely possibility is that a significant portion, but not all, of the bacteria in the colony have lost the plasmid. Despite the absence of pFra and caf1, P. Exu 340 was shown to be highly virulent in mice (LD50: 6x100) (Table). Chromosomal alteration in this culture, revealed by the growth of non-pigmented colonies on CRA plates and the loss of the gene irp2 (Pgm- irp2-) also occurred (Table). Brubaker (1969), related that spontaneous non-pigmented mutants arise at a frequency of ~105. However, this frequency may vary among the strains (Iteman et al. 1993, Cavalcanti et al. 2002). Non-pigmented (Pgm-) cultures are avirulent in mice by subcutaneous route of infection (Jackson & Burrows 1956, Une & Brubaker 1984) and Pgm- irp2- cultures are avirulent by both subcutaneous and intravenous routes (Almeida et al. 1993, Iteman et al. 1993). As expected, the culture YP 340/15 (Pgm- irp2-), lacking pFra, caf1 and ymt was shown to be avirulent (LD50: > 6x108). Analysis of the strain P. Peru 375 - P. Peru 375 harboured the three plasmids at isolation (Fig. 1e, line 3); however, in this work only pPst and pYV were visualized after electrophoresis, in the culture recovered from the stock. This suggests spontaneous curing of pFra during storage (Table, Fig. 1c, d). The genes caf1 and ymt were still amplified by PCR (Table, Fig. 2c); they were probably integrated into the chromosome. Plasmid and chromosomal alterations were not produced at the same time in the whole population of P. Peru 375. Four phenotypes were observed on CRA plates after its 4th subculture. One colony of each phenotype was analyzed and differences in their plasmid content was also found (Table, Fig. 2c). The culture YP 375/4a displayed the same characteristics as those of the parental strain and their LD50 were quite similar (8x100 and 1x101, respectively). In YP 375/4b, ymt was not amplified. In YP 375/4c, ymt was also absent and, additionally, this colony was Pgm- on CRA plates (chromosomal pathogenicity marker) and had a higher LD50 (> 1x104). In YP 375/4d and YP 375/10, caf1 was amplified, but not ymt; additionally, pPst was not visualized and the gene pla was not amplified. The LD50 of YP 375/4d lacking pFra and pPst was very low (3x101). The gene caf1 was amplified from the cultures lacking pFra obtained up to the 16th subculture of P. Peru 375; however, afterwards, it was no longer amplified at the same time that only Pgm- irp2- colonies were found on CRA plates (Table). As expected, the culture YP 375/22 (Pgm- irp2-) lacking pFra, caf1 and ymt was shown to be avirulent (LD50: 8x105). The presence of the gene caf1 was determined by PCR in cultures lacking pFra indicating that it could be integrated into the chromosome. To investigate if the f1 operon was functional, the protein content of P. Peru 375 and four derived cultures (pFra- caf1+) was analyzed. A protein band of approximately 17 kDa, corresponding to the F1 antigen encoded by the gene caf1, was visualized in all the cultures examined. Furthermore, anti-F1 specific antibodies recognized these protein bands by immunoblot (Fig. 3). Therefore, the loss of virulence in some cultures derived from P. Peru 375 was due rather to concomitant alterations at the locus pgm (to be published elsewhere) (Fetherston et al. 1992, Buchrieser et al. 1998) than to the curing of pFra or pPst plasmids (Table). Non-pigmented (Pgm-) cultures are avirulent in mice by subcutaneous route of infection (Jackson & Burrows 1956, Une & Brubaker 1984). Accordingly, Kutyrev et al. (1989), Drozdov et al. (1995), and Almeida et al. (2003) have already shown that the virulence of Y. pestis in laboratory or wild animals is not hampered by the loss of pPst or pFra. In addition, the occurrence of wild Y. pestis variants missing pFra in some plague foci and their participation in the epizootic processes was noted (Kutyrev et al. 1989). Furthermore, Y. pestis strains unable to produce some virulence factors were isolated from plague patients or from rodents and fleas and they were supposed to cause mild or even fatal plague (Williams et al. 1978). In conclusion, the stability of Y. pestis plasmids varied among the strains studied. As observed previously (Leal et al. 1997a), pYV was found to be the most stable and pFra the most unstable, followed by pPst. Plasmids of P. CE 882 were shown to be quite stable; no apparent alteration occurred through 32 subcultures. However, for reasons still unknown, its virulence in mice decreased. In the other two strains, serial passage resulted in loss of the three plasmids normally found in Y. pestis. No changes in the virulence plasmid content could be specifically correlated with changes in the LD50. Studies are in progress for better understanding the genetic mechanism of the stability of the plasmids of Y. pestis. ACKNOWLEDGEMENTS To Silvana Santos and Yara Nakazawa (CPqAM-Fiocruz) for their technical assistance. REFERENCES
Copyright 2004 Instituto Oswaldo Cruz - Fiocruz The following images related to this document are available:Photo images[oc04147f2.jpg] [oc04147f3.jpg] [oc04147t1.jpg] [oc04147f1.jpg] |
|