GROWTH AND THE POTENTIAL FOR CONJUGATION BY <em>BACILLUS SPHAERICUS </em>IN MOSQUITO LARVAL CADAVERS

GROWTH AND THE POTENTIAL FOR CONJUGATION BY BACILLUS SPHAERICUS IN MOSQUITO LARVAL CADAVERS

Margarita Correa and Allan A.Yousten

Microbiology and Immunology Section, Biology Department, Virginia Polytechnic Institute and State University, Blacksburg, VA 24061, fax (540)231-9307, Yousten@vt.edu

SUMMARY

Spores of binary toxin-producing strains of the mosquito pathogen, Bacillus sphaericus, germinated in high percentage in larval cadavers. Low toxicity strains producing the 100 kDa toxin germinated in low percentage. Normally non-toxic strains supplied with the binary or 100 kDa toxins germinated poorly. Thus, recycling (spore germination, vegetative growth, and resporulation) seems limited to the high toxicity strains. One high toxicity strain that successfully recycles (strain 2362) was shown to be able to transfer conjugative plasmid pAMß1 to several related strains of B. sphaericus. However, strain 22362 [pAMß1] was not able to conjugate with less related B. sphaericus strains or with some other members of Bacillus. We did not detect conjugation among strains that successfully conjugated on membranes when the conjugations were carried out in larval cadavers. There was no indication that B. sphaericus 2362 could conjugate utilizing the large cryptic plasmid resident in that strain.

Key words: Bacillus sphaericus, spore germination, bacterial conjugation

INTRODUCTION

Bacillus sphaericus is a heterogeneous species of bacteria that contains strains belonging to at least five different DNA homology groups (Krych et al., 1980). The bacteria in these homology groups are sufficiently phenotypically similar that it has not been practical to establish each as a new distinct species. Homology group IIA does differ from the other genospecies in that it contains strains that are pathogenic for mosquito larvae. Pathogenicity is caused by the production of one or both of two unrelated protein toxins. One toxin of 100 kDa is produced by vegetative cells and bacteria producing this toxin alone are of low toxicity. The second toxin is produced at the time of sporulation and it accumulates in the sporangium as a parasporal body in much the same way as that found in B. thuringiensis. This is a binary toxin consisting of two distinct proteins of 51.4 and 41.9 kDa. Strains producing this toxin are of high toxicity. Some strains produce both toxins and are of high toxicity. These protein toxins have been sequenced and are unrelated to each other or to the B. thuringiensis toxins (Porter et al., 1993).

B. sphaericus has proven effective against a variety of mosquito larvae and one isolate (strain 2362) will be available commercially as a larvicide in the near future. Despite its attractive features as a biological control agent, efforts are being made to modify this bacterium to improve its toxicity and manufacturing characteristics. For example, one of the B. thuringiensis israelensis dipteran toxins has been inserted into B. sphaericus to enhance activity against Aedes aegypti larvae (Poncet et al., 1994). Initially, these genetic modifications are being done by inserting genes on plasmid vectors that also carry antibiotic resistance genes. The delivery of very large numbers of the spores of these bacteria into the environment might be undesirable if the spores are capable of germinating and if the resulting vegetative cells can transfer the antibiotic resistance genes to environmental microflora. We believe that if this can occur, the most likely site is in the mosquito larval cadaver where spores, the form delivered with toxin, may germinate to produce vegetative cells. Spore germination, outgrowth, and resporulation (referred to as "recycling") has been shown for strain 2362 (Charles and Nicolas, 1986). The vegetative cells resulting from spore germination would be in close proximity to other bacteria in a nutrient rich environment, the larval cadaver, where conjugation might be possible.

The goals of this research were i. to determine if B. sphaericus spores can germinate in the larval cadaver and ii. to determine if B. sphaericus is capable of conjugation.

MATERIALS AND METHODS

Bacteria and Bacteriophages. Bacterial strains (with exceptions noted below) were obtained from the Virginia Tech culture collection. Bacteriophages were previously described by Yousten et al. (1980). B. sphaericus NRS718 [pUE382] and B. subtilis DB104 [pUE382] were obtained from P. Baumann. B. sphaericus NRS1693 [pC35] was obtained from A. Porter. B. sphaericus 2362 carrying plasmid pAMß1 was produced by membrane mating strain 2362 with Enterococcus faecalis JH2.2 [pAMß1]. Rifampicin resistant mutants were obtained by growing bacteria in increasing concentrations of the antibiotic until all were resistant to at least 25 µg/ml.

Recycling of Bacteria in Larval Cadavers. Spore suspensions of rifampicin resistant strains were prepared by bacterial growth at 30°C in NYSM broth (Myers and Yousten, 1978) with shaking at 150 rpm until maximum sporulation was achieved. Spores were washed three times in sterile distilled water, suspended in distilled water and stored at 4°C. Three hundred, third instar Culex quinquefasciatuslarvae were allowed to feed for 15 min at 25°C in 200 ml tap water containing 1-5 x 10⁶ spores/ml. After feeding, larvae were removed, rinsed with sterile tap water and placed in 300 ml of tap water. In some experiments, soluble binary toxin prepared from B. sphaericus 2362 was added at a level of 12 g protein/ml. At intervals, 25 larvae were removed, rinsed, suspended in 5 ml sterile water and homogenized with a glass tissue grinder. Spore counts (cells resistant to 80°C for 12 min) and total viable B. sphaericus counts were made in NYSM agar containing 25 g/ml rifampicin. Total bacterial counts (B. sphaericus and all other bacteria growing under the incubation conditions) were made in Difco plate count agar.

Conjugal Transfer of Plasmid pAMß1. The procedure was a modification of that described by Koehler and Thorne (1987). Donor and recipient cultures were grown separately with shaking for 5 hr in 5 ml of NY broth (Difco nutrient broth supplemented with 0.05% yeast extract). Equal volumes of donor and recipient were mixed and 0.1 ml spread onto a 45 mm, 0.45 µm membrane filter disc resting on NYSM agar. The filter was incubated for 20 hr at 37°C (time and temperature determined in preliminary experiments). After incubation, bacteria were scraped off with a glass rod, suspended in 2 ml NY broth, and plated on medium containing 10 µg/ml erythromycin (resistance determined by pAMß1 in the donor strain), 25 µg/ml rifampicin (resistance in the recipients), or both antibiotics to detect transconjugants. Plates were incubated at 30°C for 48 h.

Transconjugants were verified by extracting plasmids (Kado and Liu, 1981) and analyzing by agarose gel electrophoresis. Identity of donors and recipients was verified by the use of auxotrophic markers or by sensitivity to bacteriophages. Bacterial surface layer protein was extracted and analyzed by SDS-PAGE as described by Lewis et al. (1987).

Conjugation in larval cadavers was tested by feeding a mixture of spores of the donor and recipient bacteria to third instar Culex quinquefasciatus mosquito larvae. Larvae were homogenized at intervals and the homogenate plated on the appropriate antibiotic-containing media. Spore counts were determined by heating homogenate at 80°C for 12 min prior to plating.

RESULTS

Spore Germination in Larval Cadavers. Larvae fed spore suspensions containing the binary toxin died within 12 - 24 hr after toxin ingestion whereas larvae fed the 100 kDa toxin died more slowly, usually 24 - 48 hr following ingestion. This death rate was observed whether the toxin was produced by the bacteria or whether the toxin was a soluble preparation from a different strain. A high percentage of spores from high toxicity strains 2362 (serotype 5a5b), IAB59 (serotype 6), and 2297 (serotype 25) germinated rapidly during the 12 hr following their ingestion. Although the number of heat resistant spores dropped during this period, the total number of B. sphaericus cells remained high indicating that the loss of heat resistant forms was due to germination and not to spore destruction or defecation. The total number of bacteria associated with larvae (B. sphaericus and other bacteria growing under the cultivation conditions used) also increased as the larvae died (Figure 1). This is a pattern typical of "recycling". In contrast, the number of spores of a low toxicity strain, 1883, producing only the 100 kDa toxin decreased slowly and the total number of B. sphaericus also decreased (Figure 2). The addition of soluble binary toxin from strain 2362 to the spores of strain 1883 produced a more rapid kill of larvae and a more rapid germination of spores, however, the percentage of spores germinating was about the same as without the addition of binary toxin.

To determine if either the binary or the 100 kDa toxin might have a direct germinative effect on spores and promote recycling, spores of normally nontoxic bacteria were fed along with toxin. The toxin was either produced by the bacteria themselves (encoded on recombinant plasmids) or supplied as sterile, soluble binary toxin from strain 2362. B. sphaericus NRS 718 [pUE382] and B. subtilis DB104 [pUE382] produced binary toxin as parasporal inclusions, and spores of B. sphaericus ATCC 14577 were fed along with soluble toxin. Although larvae died rapidly from ingesting binary toxin produced by NRS 718 [pUE382], there was less spore germination than observed with the other pathogens and little of no increase in spore number in cadavers (Figure 3). Similar results were obtained with the other nontoxic strains. The effect of the 100 kDa toxin on germination of spores of a normally nontoxic bacterium, was tested with B. sphaericus 1693 [pC35]. The spores of this recombinant germinated poorly in larval cadavers.

Conjugation. The ability of B. sphaericus to act as a donor in conjugation was examined by transferring the promiscuous, conjugative plasmid pAMß1 into B. sphaericus 2362 by conjugation with Enterococcus faecalis [pAMß1]. The recovered transconjugant, B. sphaericus 2362 [pAMß1], was erythromycin resistant and was used as donor in conjugation experiments with rifampicin-resistant mutants of a number of other B. sphaericus strains as well as with other bacteria. Proof that recovered transconjugants were actually the recipient strains was provided by using auxotrophic mutants or recipients whose response to certain lytic bacteriophages could be tested.

B. sphaericus 2362 was able to transfer pAMß1 to other pathogenic B. sphaericus strains of serotypes 5a5b, 25, and 9a9c. Conjugation was much more efficient when carried out on membranes than in broth (frequency of about 10^-4 to 10^-5 per recipient on membranes vs. about 10^-7 in broth). Conjugation frequency was about 10^-3 using strain 1593-P51, a bacteriophage resistant mutant of strain 1593. This mutant has a surface layer protein of lower molecular weight than the parent strain. No transfer was detected using low toxicity pathogens of serotypes 1 or 2 or to a high toxicity pathogen of serotype 6 as recipients. Also, no transfer was detected using nonpathogens of DNA homology groups I or V as recipients. No transfer was detected with B. thuringiensis serovar. israelensis or B. cereus var. mycoides (isolated from field collected mosquito larvae) as recipients.

B. sphaericus 2362 and several other pathogenic strains carry a large (approximately 180 kb) cryptic plasmid. It seemed possible that this plasmid might be conjugative, but it lacked known markers to follow its transfer. In B. thuringiensis, cryptic, conjugative plasmids have been detected by their ability to mobilize transfer of smaller, nonconjugative plasmids carrying antibiotic resistance genes. To use this tactic, B. sphaericus 2362 was transformed with the small nonconjugative plasmid pUB110 carrying neomycin resistance. No transfer of pUB110 was detected between the donor and any of the other B. sphaericus strains tested. This included the strains that had been successful recipients for pAMß1.

To test the suitability of the mosquito larval cadaver as a site for B. sphaericus conjugation, spores of strain 2362 [pAMß1] and strains that had been successful recipients in membrane matings were fed to C. quinquefasciatus larvae. Larvae were homogenized and plated at 48 and 72 hr to recover transconjugants. Under these conditions the larvae died rapidly, spores germinated, but no transconjugants were recovered. The larval cadaver may not be a suitable site for conjugation, however, mixing larval homogenate with strains used in filter matings did not inhibit mating on filters.

DISCUSSION

Spore germination is a critical event in the life cycle of B. sphaericus mosquito pathogens. It is significant that a high percentage of spores from three different serotypes capable of producing the binary toxin were able to germinate in larval cadavers and produce vegetative cells. These vegetative cells could grow at the expense of nutrient liberated from tissue and carry out the typical growth cycle resulting in sporulation. It appears that the ability of a high percentage of ingested spores to germinate in cadavers may have evolved in these strains to take advantage of the nutrients provided by the lethal effect of toxin. A smaller percentage of the spores of low toxicity strains 1883 and of recombinant 1693 [pC35] germinated than of the high toxicity strains. During the period of spore germination, there was a concomitant decrease in total B. sphaericus cells in the larvae. Since the larvae that consumed low toxicity strains died more slowly than larvae consuming binary toxin, it is possible that the loss of spores might be due not only to spore germination but also at least partially to defecation of spores. Since the 100 kDa toxin is produced during vegetative growth, there may have been less advantage to these strains to evolve spores that germinated well in cadavers.

It was possible that the binary toxin had some direct germinative effect on spores. However, the failure of normally nontoxic recombinants producing the toxin to germinate well argues against this hypothesis. Also, the addition of soluble binary toxin to the normally non-toxic strain ATCC 14577 failed to induce a high percentage of spores to germinate in cadavers.

The ability of the spores of high toxicity strains to germinate in cadavers provides the vegetative cells that would be necessary for conjugation. The ability of a binary toxin producing strain, 2362, to conjugate was demonstrated by introducing the conjugative plasmid pAMß1 into this strain. A high frequency of conjugation was found when this bacterium was mixed with a limited range of other B. sphaericus strains on membrane filters. A mutant having a surface layer protein of lower than normal molecular weight had a conjugation frequency about 1 log higher than the parent strain. This suggests a role for the surface layer protein in the conjugation process in these bacteria. The frequency of conjugation in liquid was at least three logs lower than on membranes.

The ability of the larval cadaver to serve as a suitable milieu for conjugation was tested by feeding a mixture of spores from strains that had been shown to conjugate on membranes. No transconjugants were recovered. However, if the conjugation frequency in larvae is more similar to that shown in liquid (about 10^-7) than to that found on membranes, the presence of approximately 10⁵ bacteria in each larva would then require at least 100 larvae to be examined to recover even one transconjugant. In our experiments, 50 or 75 larvae were homogenized and plated at each time point. It may be that the limit of detection in this experiment was too low to detect the rare transconjugants. There is probably no substance in cadavers that is specifically inhibitory to conjugation because the addition of larval homogenate to filter matings did not decrease the frequency of mating.

ACKNOWLEDGEMENT

This research was supported by cooperative research agreement CR819744-01 from the U.S. Environmental Protection Agency Environmental Research Laboratory (Duluth, MN).

REFERENCES

Charles J, Nicolas L (1986) Recycling of Bacillus sphaericus 2362 in mosquito larvae: a laboratory study. Ann Inst Pasteur/Microbiol 137B:101-111.

Kado C, Liu S-T (1981) Rapid procedure for detection and isolation of large and small plasmids. J Bacteriol 145: 1365-1373.

Koehler T, Thorne C (1987) Bacillus subtilis (natto) plasmid pLS20 mediates interspecies plasmid transfer. J Bacteriol 169:5271-5278.

Krych V, Johnson J, Yousten A (1980) Deoxyribonucleic acid homologies among strains of Bacillus sphaericus. Int J Syst Bacteriol 30: 476-484.

Lewis L, Yousten, Murray R (1987) Characterization of the surface protein layers of the mosquito-pathogenic strains of Bacillus sphaericus. J Bacteriol 169: 72-79.

Myers P, Yousten A (1978) Toxic activity of Bacillus sphaericus SSII-1 for mosquito larvae. Infect Immun 19:1047-1053.

Poncet S, Delecluse A, Anello G, Klier A, Rapoport G (1994) Transfer and expression of the cryIVB and cryIVD genes of Bacillus thuringiensis subsp. israelensis in Bacillus sphaericus 2297. FEMS Microbiol Lett 117: 91-96.

Porter A, Davidson E, Liu J-W (1993) Mosquitocidal toxins of bacilli and their genetic manipulation for effective biological control of mosquitoes. Microbiol. Rev. 57: 838-861.

Yousten A, deBarjac H, Hedrick J, Cosmao Dumanoir V, Myers P (1980) Comparison between bacteriophage typing and serotyping for the differentiation of Bacillus sphaericus strains. Ann Microbiol (Inst Pasteur) 131B: 297-308.

Figure 1. Recycling of B. sphaericus 2362 in larval cadavers. (l) B. sphaericus spores, (m) Total B. sphaericus cells (spores, germinated spores, and vegetative cells), () Total bacteria associated with cadavers.

Figure 2. Recycling of B. sphaericus 1883 in larval cadavers. (l) B. sphaericus spores, (m) Total B. sphaericus cells (spores, germinated spores, and vegetative cells), () Total bacteria associated with cadavers.

Figure 3. Recycling of B. sphaericus NRS 718 [pUE382] in larval cadavers. (l) B. sphaericus spores, (m) Total B. sphaericus cells (spores, germinated spores, and vegetative cells), () Total bacteria associated with cadavers. The normally nontoxic strain produced binary toxin encoded on recombinant plasmid pUE382.

Figure 4. Recycling of B. sphaericus 1693 [pC35] in larval cadavers. (l) B. sphaericus spores, (m) Total B. sphaericus cells (spores, germinated spores, and vegetative cells), () Total bacteria associated with cadavers. The normally nontoxic strain produced 100 kDa toxin encoded on recombinant plasmid pC35.