 |
| About Bioline | All Journals | Testimonials | Membership | News |
|
||||||
|
||||||
Conjugation by mosquito pathogenic strains of Bacillus sphaericus Margarita Correa, Allan A Yousten+ Microbiology Section, Biology Department, Virginia Tech, Blacksburg, Virginia 24061, U.S.A. +Corresponding author. Fax: +1-540-231.9307. E-mail: Yousten@vt.edu Received 17 September 1996; Accepted 11 December 1996
Code Number:OC97076
Sizes of Files:
Text: 22.4K
Graphics: Tables (gif) - 34.8K
A mosquito pathogenic strain of Bacillus sphaericus carried out the conjugal transfer of plasmid pAM beta1 to other strains of its own and two other serotypes. However, it was unable to conjugate with mosquito pathogens from three other serotypes, with B. sphaericus of other DNA homology groups or with three other species of Bacillus. Conjugation frequency was highest with a strain having an altered surface layer (S layer). Conjugal transfer of pAM beta1 was not detected in mosquito larval cadavers. B. sphaericus 2362 was unable to mobilize pUB110 for transfer to strains that had served as recipients of pAM beta1. These observations suggest that it is unlikely that genetically engineered B. sphaericus carrying a recombinant plasmid could pass that plasmid to other bacteria. Key words: Bacillus sphaericus - mosquito pathogen - bacterial conjugation Mosquito pathogenic strains of Bacillus sphaericus have been assigned to DNA homology group IIA (Krych et al. 1980). Pathogenicity is caused by the production of one or more of four toxins (Charles et al. 1996, Liu et al. 1996, Thanabalu & Porter 1996). Despite the proven effectiveness of mosquito larvicides based on B. sphaericus 2362, it may be useful to introduce genes for additional toxins from other bacteria such as B. thuringiensis or to add genes for extracellular protease to enhance manufacturing efficiency on protein substrates. These genes would initially be introduced on plasmid cloning vectors carrying antibiotic resistance genes as markers. The transfer of the antibiotic resistance genes to other bacteria by conjugation in the aquatic environment would be undesirable. It is unknown to what extent B. sphaericus is able to transfer genes by conjugation. Grigorova et al. (1988) demonstrated transfer of plasmids from Streptococcus agalactiae and S. pyogenes to B. sphaericus but did not investigate further transfer among B. sphaericus strains. Orzech and Burke (1984) reported that B. sphaericus 1593 could accept conjugative plasmid pAM beta1 from Enterococcus faecalis JH2.2 and could transfer pAM beta1 to another strain of B. sphaericus. Whether B. sphaericus could transfer the plasmid to a wider range of B. sphaericus strains or to other bacteria was not reported. Also, it is unknown if B. sphaericus is capable of conjugation based upon plasmids it normally carries. We addressed these questions by using the broad host range plasmid pAM beta1 to investigate the range of bacteria to which a known, conjugative plasmid could be transferred. To determine if the large, cryptic plasmid carried by B. sphaericus 2362 is conjugative, the mobilizable plasmid pUB110 was used as a marker to indicate if conjugation took place. MATERIALS AND METHODS Bacteria and mosquito larvae - E. faecalis JH2.2 [pAM beta1] and B. sphaericus 2362 [pUB110] were obtained from W Burke. B. subtilis PSL1 UM4 [pLS20] and B. subtilis IG-20 UM1 were obtained from C Thorne. All other B. sphaericus isolates were from the Virginia Tech Culture Collection. Recipients in all matings were rifampicin resistant. B. sphaericus 2362 [pAM beta1] was constructed by filter mating E. faecalis JH2.2 [pAM beta1] with B. sphaericus 2362 and selecting erythromycin-resistant B. sphaericus colonies on BATS agar (Yousten et al. 1985) supplemented with 10 mg/ml erythromycin. The strain 2362 used as recipient was not restriction-deficient. B. subtilis PSL1 UM4 [pLS20, pUB110] was constructed by protoplast transformation (Chang & Cohen 1979) of PSL1 UM4 [pLS20] with plasmid DNA extracted from B. sphaericus 2362 [pUB110]. Conjugation - Matings in liquid were performed by a modification of the method of Battisti et al. (1985). Donor (B. sphaericus 2362 [pAM beta1]) and recipient (B. sphaericus 2362 rif^r,lys^-) were grown with shaking (160 rpm) for 5 hr at 30 C in tubes of NY broth (Difco nutrient broth supplemented with 0.05% yeast extract). The same volume (2 ml) of donor and recipient were added to 25 ml of NY broth and shaken (160 rpm) at 30 C for 20 hr. The number of donors and recipients were enumerated on NY agar containing 10 ug/ml erythromycin (donor) or 25 ug/ml rifampicin (recipient). Transcipients were enumerated on NYSM agar (Myers & Yousten 1978) containing both antibiotics. Conjugation on membranes was done by a modification of the method of Koehler and Thorne (1987). Donors and recipients were grown with shaking for 5 hr at 30 C. One ml each of donor and recipient were mixed and 0.1 ml of the mixture spread onto a 45 mm diameter, 0.45 um membrane filter (Millipore, Bedford, MA) placed on NYSM agar. Bacteria were incubated for 20 hr at 37 C. Growth was washed off the membrane into 2 ml NY broth, diluted and plated. Donors, recipients, and transcipients were enumerated as described above using 25 ug/ml rifampicin and either 10 ug/ml erythromycin (pAM beta1) or 5 ug/ml neomycin (pUB110). Conjugation in mosquito larval cadavers was tested by feeding third instar larvae of Culex quinquefasciatus, mixtures containing equal numbers of donor and recipient spores as previously described (Correa & Yousten 1995). The larvae were removed from the spore suspension, rinsed, and placed in clean water lacking spores. At 0 hr (immediately upon completion of feeding), at 48 hr and at 72 hr, 75 larvae were removed, rinsed in sterile water, and homogenized in a sterile glass tissue grinder. At 48 hr and at 72 hr, all the larvae removed were dead. The homogenate was divided, and part heated (80C, 12 min) and part held unheated prior to plating on NYSM agar containing erythromycin and rifampicin. Conjugation frequency in all experiments was the number of transcipients divided by the number of potential recipients. Verification of transcipient identity and plasmid transfer - To verify that the putative transcipients were derived from the recipients, recipients were either amino acid auxotrophs (strain 2362 lys^- and 1691 his^-) or possessed different sensitivity to bacteriophages (Yousten et al. 1980) than the donors. Representative transcipients were selected and tested for these traits. Plasmid extraction and electrophoresis was done as described by Seyler et al. (1993). RESULTS The effect of incubation time on membranes on conjugation frequency was tested by mating B. sphaericus 2362 [pAM beta1] with B. sphaericus 2362 lys^- and plating after 3, 6, 9 and 20 hr. There was an approximate 20-fold increase in frequency as the incubation was extended from 3 to 20 hr. The same pair of strains was used in 20 hr filter mating to demonstrate an approximate doubling in conjugation frequency as the temperature of membrane incubation was increased from 25 C to 30 C and another doubling with an increase from 30 C to 37 C (data not shown). Subsequent filter matings were carried out for 20 hr at 37 C. Conjugation frequency in broth and on membranes was compared by mating B. sphaericus 2362 [pAM beta1] with B. sphaericus 2362 lys^-. Conjugation frequencies in broth were much lower (7.2X10^-7+/-1.7X10^-7) than on membranes (1.9X10^-4+/-1.1X10^-4). B. sphaericus 2362 and other mosquito pathogens belong to DNA homology group IIA whereas the type strain of the species, ATCC 14577, is a nonpathogen and belongs to homology group I. Additional nonpathogens, all presently referred to as B. sphaericus, are found in four other homology groups. The pathogens of group IIA have been subdivided by serotyping. To determine the ability of mosquito pathogenic strain 2362 to transfer pAM beta1 to other bacteria, filter matings were performed and the results are reported in Table. In addition to being able to transfer the plasmid to 2362 at a mean frequency of 1.9X10^-4, 2362 [pAM beta1] successfully transferred pAM beta1 to strains 1593 and 1691 of the same serotype (5a5b) as the donor and to strains 2297 and 31-2 of serotypes 25 and 9a9c respectively. However, no transfer was detected to mosquito pathogens IAB460 (serotype 6), SSII-1 (serotype 2),or Kellen Q (serotype 1). Also, pAM beta1 was not transferred to nonpathogenic B. sphaericus strain ATCC 14577 (homology group I), to NRS 1199 (homology group V) or to a B. sphaericus (homology group V) isolated from mud in a local mosquito breeding site. To test for the possibility of interspecies conjugation, B. subtilis IG-20, a restrictionless variant, as well as B. mycoides and B. thuringiensis var. israelensis were used as recipients. No transfer of pAM beta1 was detected. The highest conjugation frequency detected in this series of experiments was found with B. sphaericus 1593-5-1 as recipient. The mean frequency was about 15-fold higher than with 2362 as recipient and about 7-fold higher than with the parental 1593 as recipient. Strain 1593-5-1 is the same homology group and serotype as the donor, 2362 [pAM beta1], but differs by possessing a surface protein layer, S layer, of lower molecular weight than the parent strain 1593 (Lewis & Yousten 1988). The mutant was isolated based on its resistance to phage 4, a lytic phage for both 2362 and the parental strain 1593. The phage resistance was used as positive identification of the transcipients. Evidence against transformation or transduction being involved in pAM beta1 transfer was provided by carrying out the mating on a membrane incubated on medium containing 100 ug/ml DNase and by substituting filter sterilized culture supernatant from the donor for the donor bacteria themselves. DNase did not affect the frequency of recovering erythromycin resistant colonies and no erythromycin resistant bacteria were found when culture supernatant was incubated in place of the donor bacteria. Several large, self-transmissible, usually cryptic plasmids detected in B. thuringiensis and B. subtilis can mobilize the transfer of smaller plasmids (Battisti et al. 1985, Koehler & Thorne 1987, Reddy et al. 1987). It was unknown if the large, cryptic plasmid residing in B. sphaericus 2362 was conjugative and if it might be capable of mobilizing plasmids. This was tested using strain 2362 carrying the large, cryptic plasmid as well as the 4.5 kb pUB110. The pUB110 present in the strain 2362 that was to be used as donor was proven to be mobilizable by transferring the plasmid to B. subtilis PSL1 UM4 [pLS20] by protoplast transformation. pLS20 had been shown to promote the transfer of pUB110 (Battisti et al. 1985). B. subtilis PSL1 UM4 [pLS20, pUB110] was filter mated with B. subtilis IG-20 UM1 and transcipients selected on a medium containing 10 ug/ml rifampicin and 5 ug/ml neomycin. Conjugation frequency obtained in this mating was equal to that reported by Koehler and Thorne (1987) who used the same B. subtilis strains. This demonstrated that the pUB110 used in experiments with B. sphaericus was capable of being mobilized. B. sphaericus 2362 [pUB110] was filter mated with B. sphaericus 2362, 1691, 2297 and 31-2, four strains that had been shown to be effective recipients for pAM beta1. It was also mated with strains Kellen Q and 1883 (serotype 2). In none of these experiments were transcipients recovered. If B. sphaericus was able to transfer recombinant plasmids by conjugation following its dispersal as a larvicide in the aquatic environment, the most likely site for this to take place would be in the larval cadaver. In this site the spores of B. sphaericus are known to germinate and grow vegetatively in the presence of large numbers of bacteria indigenous to the larvae (Correa & Yousten 1995). To test the suitability of the larval cadaver as a site for conjugation, spores of strain 2362 [pAM beta1] were fed with spores of either 2362 lys^- or 1593-5-1 to third instar mosquito larvae. Larvae accumulated between 10^5 and 10^6 spores of each strain per larva. In two trials with each recipient strain, no transcipients were recovered at either 48 hr or 72 hr after feeding, although these same combinations of bacteria had been shown to conjugate when filter mated. To test whether soluble substances in the larval cadaver might interfere with conjugation, larval homogenate was mixed with cells of 2362 [pAM beta1] and 1593-5-1 when they were placed on filter membranes for mating. There was no decrease in conjugation frequency in the presence of homogenate. DISCUSSION The broad host range, conjugative plasmid pAM beta1 was used to determine favorable conditions for plasmid donation by B. sphaericus 2362. With the same strain as recipient, conjugation frequency increased both with increase in temperature and time of incubation. The frequency was much higher when mating was done on a membrane surface than when done in broth, a common observation among gram positive bacteria. Transfer of pAM beta1 was successful when strains of the same serotype (5a5b) as the 2362 donor were used as recipients. pAM beta1 was also transferred to two other pathogens, strains 2297 and 31-2, of different serotypes. However, not all strains of DNA homology group IIA (mosquito pathogens) functioned as recipients. Strains IAB460, SSII-1, and Q yielded no transcipients. Also, strains outside homology group IIA as well as three other species of Bacillus did not yield transcipients. Successful recipients of serotype 5a5b are known to have restriction endonucleases, but these are likely to have DNA modified in the same way as the donor (2362) of that serotype. Strains 2297 and 31-2 of different serotypes than the donor lack restriction endonuclease activity and this may explain their success as recipients. Kellen K and SSII-1 have restriction endonuclease of different specificity than 2362 and this may explain their failure as recipients (Zahner & Priest, pers. comm.). Also, B. sphaericus possesses a surface protein layer (S layer) that is presumably the point of contact between cells at the initiation of conjugation. The protein of this layer is known to differ immunologically and in peptide maps between serotypes. It also differs in arrangement of subunits between homology groups (Lewis et al. 1987). It is possible that differences in S-layer protein create physical barriers that prevent plasmid transfer. The observation that the conjugation frequency was higher with a mutant strain (1593-5-1) having a modified S-layer protein than with the parent strain, suggests that the S layer may be involved in conjugation in this species. The existence of cryptic, conjugal plasmids in B. thuringiensis was demonstrated by detecting the mobilization of pUB110 and pBC16 (Battisti et al. 1985, Reddy et al. 1987). However, no transfer of pUB110 was detected from B. sphaericus 2362 to strains that had been shown to be recipients for transfer of pAM beta1. This indicates that the large, cryptic plasmid of B. sphaericus 2362 is not capable of mobilizing smaller plasmids. Jarrett and Stephenson (1990) demonstrated plasmid transfer between strains of B. thuringiensis in cadavers of lepidopteran larvae killed by the toxins of that bacterium. No transcipients were recovered from mosquito larvae killed by the mixture of B. sphaericus spores fed to them although the same strains were capable of transferring pAM beta1 on membranes. It is possible that the conditions in the larval cadavers are not suitable for conjugation or that the failure to detect transcipients is related to the frequency of conjugation and the number of bacteria available in the cadavers. If the frequency of conjugation was similar to that found on membranes (about 10^-3 to 10^-4 for the recipients tested in larvae), a few transcipients should have been recovered from among the recipient cells present in each cadaver. However, conjugation frequency was much lower in broth than on membranes, and the conditions in the decomposing cadaver may have resembled broth more than membrane. A lower conjugation frequency similar to that in broth would have produced too few transcipients to detect unless several hundred or thousand larval cadavers had been examined. The number of B. thuringiensis cells in the lepidopteran larvae used by Jarrett and Stephenson (1990) was higher (about 10^6 to 10^7 per larva) than the number present in smaller mosquito larvae. Different conditions in the lepidopteran cadaver and a higher number of bacteria may have been responsible for conjugation in that site. If conjugation occurred in mosquito larval cadavers, it was a rare event below the level of our detection sensitivity in these experiments. The failure of B. sphaericus 2362 to mobilize transfer of pUB110 and the apparent low (undetectable) conjugation frequency of a known conjugative plasmid in larval cadavers, indicates that it is unlikely that genetically engineered B. sphaericus would pass recombinant plasmids to other bacteria in the larval cadaver. ACKNOWLEDGEMENT To Viviane Zahner and Fergus Priest for sharing unpublished data on restriction endonuclease activity. This research was supported by cooperative research agreement CR819744-01 from the U.S. Environmental Protection Agency, Environmental Research Laboratory (Duluth, MN). Mention of commercial products or company names does not imply endorsement by the U.S. Environmental Protection Agency. REFERENCES
Copyright 1997 Fundacao Oswaldo Cruz The following images related to this document are available:Line drawing images[oc97076a.gif] |
| |||||||||
