- Email: [email protected]
Quantifying the reproduction of Bacillus thuringiensis HD1 in cadavers and live larvae of Plutella xylostella
Available online at www.sciencedirect.com Journal of
INVERTEBRATE PATHOLOGY Journal of Invertebrate Pathology 98 (2008) 307–313 www.elsevier.com/locate/yjipa
Quantifying the reproduction of Bacillus thuringiensis HD1 in cadavers and live larvae of Plutella xylostella Ben Raymond a,*, Simon L. Elliot b, Richard J. Ellis c b
a Department of Biological Science, Imperial College, Silwood Park, Ascot, Berks SL5 7PY, UK Setor de Entomologia, Departamento de Biologia Animal, Universidade Federal de Vicßosa, UFV, Brazil c Centre for Population Biology, Imperial College, Silwood Park, Ascot, Berks SL5 7PY, UK
Received 24 October 2007; accepted 29 January 2008 Available online 5 February 2008
Abstract The Bacillus cereus group comprises a range of micro-organisms with diverse habits, including gut commensals, opportunistic pathogens and soil saprophytes. Using quantitative microbiological methods we tested whether Bacillus thuringiensis (Bt) could reproduce in cadavers of Plutella xylostella killed by Bt, or in the gut of live insects, or be transmitted vertically from females to their offspring. We also tested whether diverse Bt strains could grow in high nutrient broth at a pH similar to that in the larval midgut. Low levels of reproduction were found in insect cadavers but there was no evidence of vertical transmission, or of significant reproduction in live insects. Four strains of B. thuringiensis var. kurstaki and one of B. thuringiensis var. tenebrionis were found to be capable of growth at high pH. Greater spore recovery rates in frass were found in hosts that were resistant or tolerant of infection. We concluded that that spores recovered in frass represent, in general, an ungerminated fraction of ingested inoculum and that germination rates are reduced in unsuitable hosts. Ó 2008 Elsevier Inc. All rights reserved. Keywords: Diamondback moth; Transmission; Reproductive strategy; Sub-lethal; Vertical transmission
1. Introduction There is a wide diversity of niches occupied by the Bacillus cereus complex (Jensen et al., 2003). The entomopathogen Bacillus thuringiensis (Bt) is considered to be part of this group, together with B. cereus sensu stricto and Bacillus anthracis, the causative agent of anthrax. Despite high levels of pathogenicity, the ability of B. thuringiensis strains to grow and sporulate effectively within insect cadavers is highly variable (Prasertphon et al., 1973; Suzuki et al., 2004). Effective transmission of Bt between larvae has been difficult to demonstrate experimentally (Takatsuka and Kunimi, 1998) and can require a high density of hosts and/or cannibalism (Knell et al., 1998). These low rates of transmission and reproduction may, in part, explain
*
Corresponding author. Fax: +44 1865 271281. E-mail address: [email protected] (B. Raymond).
0022-2011/$ - see front matter Ó 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.jip.2008.01.005
the lack of secondary infection from Bt microbial sprays (van Frankenhuyzen, 1993). The variable success of Bt as a pathogen has led to much speculation regarding possible alternative reproductive strategies. This includes the suggestion that Bt is a soil micro-organism with incidental insecticidal activity (Martin and Travers, 1989); that Bt is part of the phylloplane microflora and has evolved to provide protection against plants (Elliot et al., 2000; Smith and Couche, 1991) and that Bt may be part of the commensal gut flora of many insects without causing overt disease (Jensen et al., 2003). All of these hypotheses remain largely untested, leaving the ecological niche of Bt a matter of continuing contention. Bt spores can germinate and grow vegetatively in the insect midgut in the process of lethal infections (Chiang et al., 1986) and vegetative cells of Bt have been isolated from the guts of soil dwelling insects (Hendriksen and Hansen, 2002). Bacillus cereus sensu stricto has also been described as a gut commensal in Lepidoptera and other
308
B. Raymond et al. / Journal of Invertebrate Pathology 98 (2008) 307–313
invertebrates (Jung and Kim, 2006; Margulis et al., 1998). Nevertheless, it is not currently known whether Bt can proliferate in non-lethal infections in the larval gut and this question is the main focus of this work. The high pH of the Lepidopteran midgut is a potential physiological barrier to the proliferation of Bt in the larval intestine. Early reports of Bt (Kreig, 1964)—cited in Burges and Hurst (1977) and B. cereus (Raevuori and Genigeorgis, 1975) have claimed that bacteria in this group are not capable of growth at high pH. Bacillus thuringiensis israelensis cannot grow at high pH in buffered media, however, pathogenesis in the midgut lowers midgut pH readily (Walther et al., 1986) and vegetative growth of Bt can occur 40 min after ingestion of inoculum (Chiang et al., 1986). In nonlethal infections, a functioning gut may retain a high pH and it is not clear, in general, to what extent Lepidopteran adapted strains of Bt can tolerate these conditions. If sub-lethal reproduction does occur and commensal reproduction is a primary or major niche for Bt this begs the question of what it is the purpose of investing in a large quantity of protein-based toxin? Jensen et al. (2003) speculate that toxin receptors provide germination cues and increased germination improves the ability of Bt to proliferate in the gut. Toxin may of course lower pH by reducing gut function. An additional possibility is that sub-lethal pore-formation caused by low doses of endotoxin and increases the flow of haemolymph and nutrients into the midgut. Using the diamondback moth, Plutella xylostella, as a host we compared the extent to which Bt could replicate in lethal infections, and in live hosts before death or sublethally infected hosts. We also tested whether a range of B. thuringiensis isolates were capable of growth at high pH. We hypothesized that if Bt can propagate infections sub-lethally in the gut then spores should either be released in the frass or that non-overt infection should be transmitted from mother to offspring. If spores are released in frass, reproduction should be indicated by the ratio of spores released to those ingested by the host. Finally, if toxin binding or pore-formation assists growth in the intestine we hypothesize that insects which are resistant to or tolerant of infection should be less suitable for sub-lethal proliferation. It follows that if toxin binding is important for sub-lethal growth the proportion of spores in frass to spores ingested should increase in insects that are more susceptible to toxins. 2. Materials and methods 2.1. Bacterial strains, preparation of inocula and enumeration In order to be able to conclusively identify experimental inoculum from commensal B. cereus and other contaminants we used an antibiotic resistant strains of in all experiments. Bacillus thuringiensis var. kurstaki HD1 (Btk) was isolated from a commercial biopesticide preparation, DiPel
(32,000 IU mg 1; supplied by Biowise, UK) by culturing spores on B. cereus specific agar (BcSA; Oxoid, UK). A spontaneous rifampicin resistant mutant, Btk rifR, was produced by repeated sub-culture with this antibiotic on BcSA. This was repeated until the strain could grow at 100 lg ml 1 rifampicin. The stability of this resistance was ensured by sub-culturing the strain without selection a further 10 times in Luria broth (LB; Sigma–Aldrich, UK). Bacillus thuringiensis var. tenebrionis (Btt) was a gift from DJI Thomas (HRI, Wellesbourne, UK). A spontaneous spectinomycin derivative, designated Btt specR, was selected using the methods described above. Additional strains of B. t. kurstaki were sourced from the Bacillus Genetic Stock Centre (BGSC 4A4, 4D4); one strain was isolated from wild Brassica oleracea in the UK (Dorset 7.1.o) and has been identified as B. t. kurstaki by the presence of parasporal inclusions and multilocus sequence typing (ST8, sequence data available at http://pubmlst.org/ bcereus). Inocula were prepared by streaking glycerol stocks onto BcSA plates. After incubation at 30 °C for 5 days bacterial biomass was removed and resuspended in sterile saline (0.85% NaCl). The cell/spore/toxin suspension was washed by 3 rounds of centrifugation and resuspension in fresh sterile saline. Spore density was then assessed using a haemocytometer and appropriate dilutions made in sterile saline containing 0.005% Triton X-100. The total number of bacteria in experimental samples (cadavers, frass, eggs, or adult abdomens) was calculated by selective plating procedures. Samples were homogenized by grinding in sterile saline. Tenfold serial dilutions were performed and triplicate 20 ll aliquots of each dilution were dropped onto BcSA containing the appropriate selective antibiotic (100 lg ml 1 rifampicin for Btk rifR, or 100 lg ml 1 spectinomycin for Btt specR) together with 20 lg ml 1 cycloheximide to suppress fungal growth. All dilutions were then pasteurized (70 °C for 45 min) before plating out as above to generate spore counts. Plates were incubated at 30 °C for 48 h and the number of colonies counted. 2.2. Insect strains A Bt susceptible population (Lab-UK) of P. xylostella was obtained from the Institute of Arable Crops Research, Rothamsted (Harpenden, Hertfordshire, United Kingdom). A second population of P. xylostella (Karak) with high levels of resistance to Btk was also used. Mechanisms of resistance, maintenance of resistance and insect culture methods have been described previously (Sayyed et al., 2004). 2.3. Growth at neutral and high pH Autoclaved Luria-Bertani (LB) broth (Sigma) was adjusted to pH 10 using concentrated sterile NaOH in a laminar flow cabinet. The pH of standard LB broth and
B. Raymond et al. / Journal of Invertebrate Pathology 98 (2008) 307–313
alkaline LB was confirmed with a Mettler Toledo MP220 pH meter. Sterile 30 ml universals were either filled with 9 ml aliquots of high pH LB or standard LB broth (pH 7.3). In the first experiment, we compared the growth of Btk rifR at high pH in cultures initiated with 106 cells from either pasteurized spore preparations or overnight broth culture (predominantly vegetative cells). The experiment was cultured at 30 °C with continuous shaking (100 rpm). Cell density was measured five times over 24 h using spectrophotometry (OD at 600 nm) and pH readings taken at the same timepoint. In the second experiment, we compared the growth of a range of Bt strains in standard LB or LB adjusted to pH 10 as above. Spore suspensions of the strains Btk rifR, Btt specR, Dorset 7.1.o, BGSC 4A4 and BGSC 4D4 were prepared as above and 106 pasteurized spores used to inoculate each universal. Bacteria were cultured for 48 h, as above, and final densities measured by diluted plating two samples from each universal onto LB agar, and incubating plates at 30 °C. Means for each universal were used in subsequent analysis to avoid pseudoreplication. All treatments were replicated four times. 2.4. Reproduction within cadavers and larval guts Second and third instar larvae were exposed to Bt on leaf discs dipped in spore suspension at three standard concentrations (2 106, 2 105, 2 104 spores ml 1). Three larvae were confined on single leaf discs in 50 mm Petri dishes, with ten dishes per dose per instar (60 replicates in all). Leaf discs were suspended on 30 mm Petri dish lids within the larger 50 mm dishes so larval frass could be collected without contamination from leaf discs. We also set up five additional replicates (=Petri dishes) for Btk-resistant insects, and ten replicates of larvae exposed to Btt specR, a strain not pathogenic to P. xylostella, using the highest spore concentration. After 7 days frass was recovered from the bottom of the Petri dish and the consumption of each leaf disc estimated by calculating the size of holes in leaf discs. The average quantity of liquid inoculum retained by each leaf disc was calculated by weighing a subsample of leaf discs (n = 5) before and after dipping. This varied little between leaves and so the total number of bacteria on each disc was estimated at mean volume inoculum retained bacterial concentration. Inoculum concentration and proportion of leaf consumed were used to calculate the number of spores consumed in each Petri dish. Pasteurized and unpasteurized homogenates of cadavers were plated out, whereas all frass samples were pasteurized. 2.5. Vertical transmission from sub-lethally infected insects We exposed 150+ third instar Lab-UK larvae to Btk rifR on leaf discs dipped in a spore suspension (2 105 spores ml 1) as above. Insects which survived on these leaf discs were transferred to fresh sterile 90 mm Petri dishes (6 pupae in each of 10 dishes). Adults were
309
allowed to emerge and laid eggs on Parafilm that had been dipped in boiled cabbage extract within each dish. Eggs were washed from the Parafilm in sterile saline with sterilized paintbrushes, homogenized and plated out. Two abdomens were dissected from females within each Petri dish, homogenized and plated out on BcSA media as above. 2.6. Statistical analysis To measure the growth of bacteria in cadavers and frass, we used a standard metric of reproductive rates loge [(colony forming units recovered + 1)/(spores consumed)]. This measure controls for the variation in spores consumed associated with differences between instars, genotypes or replicates. Unpasteurized cadaver samples, containing vegetative cells and spores, very rarely had higher counts than pasteurized samples, indicating that bacteria had almost entirely sporulated within cadavers; the maximum value was used in analysis of reproduction, as this was most representative of reproduction at the time of sampling. All data were analysed in R v.2.1.0 (www.r-project.org/) using generalized linear modelling to carry out one-way analysis of variance, two-way analysis of covariance. Mixed model maximum likelihood anova was used to analyse time course data, with replicate incorporated as a random factor. Significance testing was carried out by sequential removal of non-significant terms (P > 0.05) in order to develop minimal adequate models. Model assumptions (homoscedasticity, normality) were checked with graphical analyses. 3. Results and discussion The reproductive rate of Btk rifR in cadavers increased with spore concentration (ANCOVA: F1,58 = 35.9, P < 0.0001; Fig. 1) but was not affected by larval instar (F1,57 = 0.28, P = 0.60) and there was no interaction between concentration and instar (F1,56 = 2.54, P = 0.12). However, the reproductive rate in cadavers only exceeded zero (i.e. more spores were recovered than were ingested) in a small proportion of replicates in the two treatments with the highest concentration of spores (Fig. 1), even though the mortality of second and third instars at the highest concentration of spores approached 100% (Mortality was 29/30 for both second and third instar larvae at the top spore concentration; 20/30 and 15/30, respectively, at the intermediate concentration and 1/30 and 6/30 at the low concentration). The cause of the poor reproduction of Btk rifR cadavers killed by this highly pathogenic isolate requires further investigation. However, since cadavers infected with this isolate, when plated out on less selective media (LB) routinely contain high counts of other bacterial species (>106 CFU per cadaver) it is likely that heterospecific competition is severely limiting growth in this strain (Takatsuka and Kunimi, 2000).
310
B. Raymond et al. / Journal of Invertebrate Pathology 98 (2008) 307–313
Fig. 1. The reproductive rate of Btk in cadavers [ln (CFU from cadavers/spores eaten)] in second and third instar diamondback moth over a range of concentrations of leaf-dip inoculum. A reproductive rate of zero indicates that the number of spores recovered is equal to the number of spores ingested. The raw data (open circles), with means for each dose (solid squares) have been plotted to show that reproduction exceeded 0 in some replicates. There are multiple overlapping data points at the minimum values for each dose.
While absolute counts of spores in frass are likely to reflect the quantity of material ingested we hypothesized that if spores recovered from frass were the product of active growth, we would recover more spores from frass than were ingested. In other words, the reproductive rate or the log ratio of spores recovered to spores ingested should exceed zero. The reproductive rate in frass in over a range of doses (both lethal and non-lethal) was always negative by a com-
fortable margin (Fig. 2). We also failed to detect any evidence of vertical transmission of Bt from insects surviving infection in eggs or adult females. An analysis of covariance showed that the proportion of spores was affected by dose (ANCOVA: log spore concentration F1,58 = 19.8, P 0.001, Fig. 2) and instar (F1,57 = 6.47, P = 0.014). Cultures initiated with spores and overnight broth culture (predominantly vegetative cells) could grow in broth
Fig. 2. The reproductive rate of Btk in frass [ln (CFU from frass/spores eaten)] in second and third instar diamondback moth over a range of concentrations of leaf-dip inoculum. A reproductive rate of zero indicates that the number of spores recovered is equal to the number of spores ingested. The raw data (open circles), with means for each dose (solid squares) have been plotted to show that reproduction never exceeded 0. There are multiple overlapping data points at the minimum values for each dose.
B. Raymond et al. / Journal of Invertebrate Pathology 98 (2008) 307–313
adjusted to a pH of 10.0 (Fig. 3a), although the growth in the treatment initiated with spores was faster than in treatment initiated with overnight cell cultures (mixed model ANOVA; time treatment interaction, df = 1, likelihood ratio = 39.5, P < 0.0001; Fig. 3a) This experiment used alkali broth that was not buffered, and pH declined as cultures grew. This does, however, mimic the situation in live hosts since toxins and replication of Bt in the gut disrupts its buffering capacity and lowers pH (Angus and Heimpel, 1956; Walther et al., 1986). Nevertheless bacteria were replicating while the pH was between 9.5 and 10.0 (Fig. 3a). All four strains of B. t. kurstaki and the one strain of B. thuringiensis var. tenebrionis proved capable of growth at both neutral and high pH in high nutrient media. High pH did retard the growth of all strains (F1,38 = 11.1, P = 0.002, Fig. 3b). While results with some strains were more variable at high pH there was no significant main
311
effect of strain (F4,34 = 2.08, P = 0.1) or significant interaction between strain and pH (F4,30 = 1.75, P = 0.17). This variability is best accounted for by the anomalous behaviour of a few replicates rather than consistent differences between strains. Whilst the ability of Bt to grow at high pH contradicts early reports, the ability to tolerate high pH seems consistent with a ecological niche that involves the invasion of a host from the gut. The data presented here, together with the observation that the high pH of Lepidopteran midguts can decline in the posterior midgut, and approach neutral values in the hindgut (Dow, 1984) indicates that pH is not a major barrier to the proliferation of Bt in the Lepidopteran intestine. An unexpected finding was the ability of Btt specR to replicate at high pH, given the lower pH of Coleopteran guts (Egert et al., 2005). Non-pathogenic strains of Bt can, however, reproduce in mixed infections with pathogenic isolates (Raymond
Fig. 3. (a) The growth and concomitant changes in pH for Btk rifR in alkaline LB broth. Open data points are cultures initiated with overnight cultures (predominantly vegetative cells), filled data points are cultures initiated with pasteurized spore/toxin suspensions. Square points represent optical density measurements, and lines are the fitted minimal adequate statistical model of bacterial growth. Triangular points are pH measurements. Data are means + SE, error bars are commonly smaller than the symbols. (b) The density of diverse Bt strains after 48 h growth in liquid media: initially adjusted to pH 10.0 (shaded bars) or control broth at pH 7.3 (open bars). Data are mean counts per replicate ± 1SE.
312
B. Raymond et al. / Journal of Invertebrate Pathology 98 (2008) 307–313
Fig. 4. The log ratio of spores recovered in frass: spores ingested from Btk resistant and susceptible larvae that had consumed Btk spores (shaded bars), and from susceptible larvae that had consumed B. t. tenebrionis spores (open bars), a subspecies of Bt not infectious to Lepidoptera. All leaf discs were inoculated with a standard concentration of spores (2 106 spores ml 1). Data are mean log ratios ± 1SE.
et al., 2007). Given the high local genetic diversity of Bt isolates (Collier et al., 2005), mixed infections may be common in the field and would maintain selective pressure for the tolerance of high pH. We hypothesized that if pore-formation by crystal toxins was involved in reproduction within the gut in live hosts, insects that were tolerant or resistant to Bt infection should have lower proportion of spores in frass to spores ingested relative to susceptible insects that were exposed to the same dose. In contrast, we found the opposite pattern: when hosts were resistant to Bt, a higher proportion of spores were recovered from frass relative to the quantity ingested (one-way ANOVA F2,22 = 20.4, P < 0.0001, Fig. 4). Model simplification revealed that there was no significant difference between the insects exposed to Btt specR and the Btk resistant insects exposed to Btk (ANOVA F1,23 = 0.13, P = 0.73). No insects exposed to Btt specR died, under half of the Btk resistant insects died (7/15), the majority (29/30) of the Btk susceptible insects also died. Since the resistant insects in this experiment have a loss of binding resistance mechanism (Sayyed et al., 2004) the absence of suitable toxin receptors in the host midgut, therefore, increased the proportion of viable spores found in frass. Given that there was no evidence for reproduction in the midgut, our interpretation of this data is that the spores recovered in frass are a fraction of the original inoculum. We conclude, therefore, that a higher proportion of spores in frass indicates higher spore survival, or lower germination rates. The inference of lower germination rates in a resistant host is consistent with the finding that toxin receptors within the insect midgut act as germination cues for Bt (Du and Nickerson, 1996). In summary, in this study Bt was able to tolerate high pH, but there was no evidence that it can reproduce in diamondback moth larvae without killing the host, at least for the strain used in this study. Currently, there is no evidence
that vegetative cells of Bt have any specialized mechanism for adhering to the gut, demonstrably contrasting with the attachments seen in Arthromitus form of B. cereus (Margulis et al., 1998). On the other hand, there are several lines of evidence that suggest vegetative growth of Bt in the gut may be important in lethal infections. The importance of live spores in synergizing the mortality caused by endotoxins (Li et al., 1987) indicates that vegetatively growing Bt contributes to the lethality of Bt toxins. Toxins produced by vegetative cells lyse midgut epithelial cells (Yu et al., 1997). The tolerance of high pH and the use of toxin-receptors as germination cues (Du and Nickerson, 1996) also suggest that Bt is adapted for growth in the gut. However, since ingestion of lethal doses of Bt endotoxins can cause paralysis very quickly (Endo and Nishiitsutsujiuwo, 1980) we would not necessarily expect to be able to detect vegetative growth in lethal infections in the frass excreted prior to death. Acknowledgments Thanks to BBSRC and NERC for financial support, and for the comments of anonymous referees who helped improve this manuscript. References Angus, T.A., Heimpel, A.M., 1956. An effect of Bacillus sotto on the larvae of Bombyx mori. Can. Entomol. 88, 138–139. Burges, H.D., Hurst, J.A., 1977. Ecology of Bacillus thuringiensis in storage moths. J. Invertebr. Pathol. 30, 131–139. Chiang, A.S., Yen, D.F., Peng, W.K., 1986. Germination and proliferation of Bacillus thuringiensis in the gut of rice moth larva, Corcyra cephalonica. J. Invertebr. Pathol. 48, 96–99. Collier, F.A., Elliot, S.L., Ellis, R.J., 2005. Spatial variation in Bacillus thuringiensis/cereus populations within the phyllosphere of broadleaved dock (Rumex obtusifolius) and surrounding habitats. Fems Microbiol. Ecol. 54, 417–425.
B. Raymond et al. / Journal of Invertebrate Pathology 98 (2008) 307–313 Dow, J.A., 1984. Extremely high pH in biological systems: a model for carbonate transport. Am. J. Physiol. Regul. Integr. Comp. Physiol. 246, R633–R636. Du, C., Nickerson, K.W., 1996. Bacillus thuringiensis HD-73 spores have surface-localized Cry1Ac toxin: physiological and pathogenic consequences. Appl. Environ. Microbiol. 62, 3722–3726. Egert, M., Stingl, U., Bruun, L.D., Pommerenke, B., Brune, A., Friedrich, M.W., 2005. Structure and topology of microbial communities in the major gut compartments of Melolontha melolontha larvae (Coleoptera: Scarabaeidae). Appl. Environ. Microbiol. 71, 4556–4566. Elliot, S.L., Sabelis, M.W., Janssen, A., van der Geest, L.P.S., Beerling, E.A.M., Fransen, J., 2000. Can plants use entomopathogens as bodyguards?. Ecol. Lett. 3, 228–235. Endo, Y., Nishiitsutsujiuwo, J., 1980. Mode of action of Bacillus thuringiensis delta-endotoxin-histopathological changes in the silkworm midgut. J. Invertebr. Pathol. 36, 90–103. Hendriksen, N.B., Hansen, B.M., 2002. Long-term survival and germination of Bacillus thuringiensis var. kurstaki in a field trial. Can. J. Microbiol. 48, 256–261. Jensen, G.B., Hansen, B.M., Eilenberg, J., Mahillon, J., 2003. The hidden lifestyles of Bacillus cereus and relatives. Environ. Microbiol. 5, 631– 640. Jung, S.C., Kim, Y., 2006. Synergistic effect of entomopathogenic bacteria (Xenorhabdus sp. and Photorhabdus temperata ssp. temperata) on the pathogenicity of Bacillus thuringiensis ssp. aizawai against Spodoptera exigua (Lepidoptera: Noctuidae). Environ. Entomol. 35, 1584–1589. Knell, R.J., Begon, M., Thompson, D.J., 1998. Host–pathogen population dynamics, basic reproductive rates and threshold densities. Oikos 81, 299–308. Kreig, A., 1964. Bacillus thuringiensis (Bacillaceae). Vegetative Vermehrung und Sporenbildung. In: Wolf, G. (Ed.), Encylocpaedia Cinematographica E555/1963. Institut fu¨r den Wissenschaftlichen Film, Gottingen, Germany, pp. 1–9. Li, R.S., Jarrett, P., Burges, H.D., 1987. Importance of spores, crystals, and delta-endotoxins in the pathogenicity of different varieties of Bacillus thuringiensis in Galleria mellonella and Pieris brassicae. J. Invertebr. Pathol. 50, 277–284. Margulis, L., Jorgensen, J.Z., Dolan, S., Kolchinsky, R., Rainey, F.A., Lo, S.-C., 1998. The Arthromitus stage of Bacillus cereus: intestinal symbionts of animals. Proc. Nat. Acad. Sci. USA 95, 1236–1241.
313
Martin, P.A.W., Travers, R.S., 1989. Worldwide abundance and distribution of Bacillus thuringiensis isolates. Appl. Environ. Microb. 55, 2437–2442. Prasertphon, S., Areekul, P., Tanada, Y., 1973. Sporulation of Bacillus thuringiensis in cadavers. J. Invert. Pathol. 21, 205–207. Raevuori, M., Genigeorgis, C., 1975. Effect of pH and sodium chloride on growth of Bacillus cereus in laboratory media and certain foods. Appl. Microbiol. 29, 68–73. Raymond, B., Davis, D., Bonsall, M.B., 2007. Competition and reproduction in mixed infections of pathogenic and non-pathogenic Bacillus spp.. J. Invertebr. Pathol. 96, 151–155. Sayyed, A.H., Raymond, B., Ibiza-Palacios, M.S., Escriche, B., Wright, D.J., 2004. Genetic and biochemical characterization of field evolved resistance to Bacillus thuringiensis toxin Cry1Ac in diamondback moth, Plutella xylostella. Appl. Environ. Microbiol. 70, 7010–7017. Smith, R.A., Couche, G.A., 1991. The phylloplane as a source of Bacillus thuringiensis variants. Appl. Environ. Microbiol. 57, 311–315. Suzuki, M.T., Lereclus, D., Arantes, O.M.N., 2004. Fate of Bacillus thuringiensis strains in different insect larvae. Can. J. Microbiol. 50, 973–975. Takatsuka, J., Kunimi, Y., 1998. Replication of Bacillus thuringiensis in larvae of the Mediterranean flour moth, Ephestia kuehniella (Lepidoptera: Pyralidae): growth, sporulation and insecticidal activity of parasporal crystals. Appl. Entomol. Zool. 33, 479–486. Takatsuka, J., Kunimi, Y., 2000. Intestinal bacteria affect growth of Bacillus thuringiensis in larvae of the oriental tea tortrix, Homona magnanima Diakonoff (Lepidoptera: Tortricidae). J. Invertebr. Pathol. 76, 222–226. van Frankenhuyzen, K., 1993. The challenge of Bacillus thuringiensis. In: Entwistle, P.F. et al. (Eds.), Bacillus thuringiensis; an Environmental Biopesticide: Theory and Practice. John Wiley, Chichester, UK, pp. 311–330. Walther, C.J., Couche, G.A., Pfannenstiel, M.A., Egan, S.E., Bivin, L.A., Nickerson, K.W., 1986. Analysis of mosquito larvicidal potential exhibited by vegetative cells of Bacillus thuringiensis subsp. israelensis. Appl. Environ. Microbiol. 52, 650–653. Yu, C.G., Mullins, M.A., Warren, G.W., Koziel, M.G., Estruch, J.J., 1997. The Bacillus thuringiensis vegetative insecticidal protein Vip3A lyses midgut epithelium cells of susceptible insects. Appl. Environ. Microbiol. 63, 532–536.
INVERTEBRATE PATHOLOGY Journal of Invertebrate Pathology 98 (2008) 307–313 www.elsevier.com/locate/yjipa
Quantifying the reproduction of Bacillus thuringiensis HD1 in cadavers and live larvae of Plutella xylostella Ben Raymond a,*, Simon L. Elliot b, Richard J. Ellis c b
a Department of Biological Science, Imperial College, Silwood Park, Ascot, Berks SL5 7PY, UK Setor de Entomologia, Departamento de Biologia Animal, Universidade Federal de Vicßosa, UFV, Brazil c Centre for Population Biology, Imperial College, Silwood Park, Ascot, Berks SL5 7PY, UK
Received 24 October 2007; accepted 29 January 2008 Available online 5 February 2008
Abstract The Bacillus cereus group comprises a range of micro-organisms with diverse habits, including gut commensals, opportunistic pathogens and soil saprophytes. Using quantitative microbiological methods we tested whether Bacillus thuringiensis (Bt) could reproduce in cadavers of Plutella xylostella killed by Bt, or in the gut of live insects, or be transmitted vertically from females to their offspring. We also tested whether diverse Bt strains could grow in high nutrient broth at a pH similar to that in the larval midgut. Low levels of reproduction were found in insect cadavers but there was no evidence of vertical transmission, or of significant reproduction in live insects. Four strains of B. thuringiensis var. kurstaki and one of B. thuringiensis var. tenebrionis were found to be capable of growth at high pH. Greater spore recovery rates in frass were found in hosts that were resistant or tolerant of infection. We concluded that that spores recovered in frass represent, in general, an ungerminated fraction of ingested inoculum and that germination rates are reduced in unsuitable hosts. Ó 2008 Elsevier Inc. All rights reserved. Keywords: Diamondback moth; Transmission; Reproductive strategy; Sub-lethal; Vertical transmission
1. Introduction There is a wide diversity of niches occupied by the Bacillus cereus complex (Jensen et al., 2003). The entomopathogen Bacillus thuringiensis (Bt) is considered to be part of this group, together with B. cereus sensu stricto and Bacillus anthracis, the causative agent of anthrax. Despite high levels of pathogenicity, the ability of B. thuringiensis strains to grow and sporulate effectively within insect cadavers is highly variable (Prasertphon et al., 1973; Suzuki et al., 2004). Effective transmission of Bt between larvae has been difficult to demonstrate experimentally (Takatsuka and Kunimi, 1998) and can require a high density of hosts and/or cannibalism (Knell et al., 1998). These low rates of transmission and reproduction may, in part, explain
*
Corresponding author. Fax: +44 1865 271281. E-mail address: [email protected] (B. Raymond).
0022-2011/$ - see front matter Ó 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.jip.2008.01.005
the lack of secondary infection from Bt microbial sprays (van Frankenhuyzen, 1993). The variable success of Bt as a pathogen has led to much speculation regarding possible alternative reproductive strategies. This includes the suggestion that Bt is a soil micro-organism with incidental insecticidal activity (Martin and Travers, 1989); that Bt is part of the phylloplane microflora and has evolved to provide protection against plants (Elliot et al., 2000; Smith and Couche, 1991) and that Bt may be part of the commensal gut flora of many insects without causing overt disease (Jensen et al., 2003). All of these hypotheses remain largely untested, leaving the ecological niche of Bt a matter of continuing contention. Bt spores can germinate and grow vegetatively in the insect midgut in the process of lethal infections (Chiang et al., 1986) and vegetative cells of Bt have been isolated from the guts of soil dwelling insects (Hendriksen and Hansen, 2002). Bacillus cereus sensu stricto has also been described as a gut commensal in Lepidoptera and other
308
B. Raymond et al. / Journal of Invertebrate Pathology 98 (2008) 307–313
invertebrates (Jung and Kim, 2006; Margulis et al., 1998). Nevertheless, it is not currently known whether Bt can proliferate in non-lethal infections in the larval gut and this question is the main focus of this work. The high pH of the Lepidopteran midgut is a potential physiological barrier to the proliferation of Bt in the larval intestine. Early reports of Bt (Kreig, 1964)—cited in Burges and Hurst (1977) and B. cereus (Raevuori and Genigeorgis, 1975) have claimed that bacteria in this group are not capable of growth at high pH. Bacillus thuringiensis israelensis cannot grow at high pH in buffered media, however, pathogenesis in the midgut lowers midgut pH readily (Walther et al., 1986) and vegetative growth of Bt can occur 40 min after ingestion of inoculum (Chiang et al., 1986). In nonlethal infections, a functioning gut may retain a high pH and it is not clear, in general, to what extent Lepidopteran adapted strains of Bt can tolerate these conditions. If sub-lethal reproduction does occur and commensal reproduction is a primary or major niche for Bt this begs the question of what it is the purpose of investing in a large quantity of protein-based toxin? Jensen et al. (2003) speculate that toxin receptors provide germination cues and increased germination improves the ability of Bt to proliferate in the gut. Toxin may of course lower pH by reducing gut function. An additional possibility is that sub-lethal pore-formation caused by low doses of endotoxin and increases the flow of haemolymph and nutrients into the midgut. Using the diamondback moth, Plutella xylostella, as a host we compared the extent to which Bt could replicate in lethal infections, and in live hosts before death or sublethally infected hosts. We also tested whether a range of B. thuringiensis isolates were capable of growth at high pH. We hypothesized that if Bt can propagate infections sub-lethally in the gut then spores should either be released in the frass or that non-overt infection should be transmitted from mother to offspring. If spores are released in frass, reproduction should be indicated by the ratio of spores released to those ingested by the host. Finally, if toxin binding or pore-formation assists growth in the intestine we hypothesize that insects which are resistant to or tolerant of infection should be less suitable for sub-lethal proliferation. It follows that if toxin binding is important for sub-lethal growth the proportion of spores in frass to spores ingested should increase in insects that are more susceptible to toxins. 2. Materials and methods 2.1. Bacterial strains, preparation of inocula and enumeration In order to be able to conclusively identify experimental inoculum from commensal B. cereus and other contaminants we used an antibiotic resistant strains of in all experiments. Bacillus thuringiensis var. kurstaki HD1 (Btk) was isolated from a commercial biopesticide preparation, DiPel
(32,000 IU mg 1; supplied by Biowise, UK) by culturing spores on B. cereus specific agar (BcSA; Oxoid, UK). A spontaneous rifampicin resistant mutant, Btk rifR, was produced by repeated sub-culture with this antibiotic on BcSA. This was repeated until the strain could grow at 100 lg ml 1 rifampicin. The stability of this resistance was ensured by sub-culturing the strain without selection a further 10 times in Luria broth (LB; Sigma–Aldrich, UK). Bacillus thuringiensis var. tenebrionis (Btt) was a gift from DJI Thomas (HRI, Wellesbourne, UK). A spontaneous spectinomycin derivative, designated Btt specR, was selected using the methods described above. Additional strains of B. t. kurstaki were sourced from the Bacillus Genetic Stock Centre (BGSC 4A4, 4D4); one strain was isolated from wild Brassica oleracea in the UK (Dorset 7.1.o) and has been identified as B. t. kurstaki by the presence of parasporal inclusions and multilocus sequence typing (ST8, sequence data available at http://pubmlst.org/ bcereus). Inocula were prepared by streaking glycerol stocks onto BcSA plates. After incubation at 30 °C for 5 days bacterial biomass was removed and resuspended in sterile saline (0.85% NaCl). The cell/spore/toxin suspension was washed by 3 rounds of centrifugation and resuspension in fresh sterile saline. Spore density was then assessed using a haemocytometer and appropriate dilutions made in sterile saline containing 0.005% Triton X-100. The total number of bacteria in experimental samples (cadavers, frass, eggs, or adult abdomens) was calculated by selective plating procedures. Samples were homogenized by grinding in sterile saline. Tenfold serial dilutions were performed and triplicate 20 ll aliquots of each dilution were dropped onto BcSA containing the appropriate selective antibiotic (100 lg ml 1 rifampicin for Btk rifR, or 100 lg ml 1 spectinomycin for Btt specR) together with 20 lg ml 1 cycloheximide to suppress fungal growth. All dilutions were then pasteurized (70 °C for 45 min) before plating out as above to generate spore counts. Plates were incubated at 30 °C for 48 h and the number of colonies counted. 2.2. Insect strains A Bt susceptible population (Lab-UK) of P. xylostella was obtained from the Institute of Arable Crops Research, Rothamsted (Harpenden, Hertfordshire, United Kingdom). A second population of P. xylostella (Karak) with high levels of resistance to Btk was also used. Mechanisms of resistance, maintenance of resistance and insect culture methods have been described previously (Sayyed et al., 2004). 2.3. Growth at neutral and high pH Autoclaved Luria-Bertani (LB) broth (Sigma) was adjusted to pH 10 using concentrated sterile NaOH in a laminar flow cabinet. The pH of standard LB broth and
B. Raymond et al. / Journal of Invertebrate Pathology 98 (2008) 307–313
alkaline LB was confirmed with a Mettler Toledo MP220 pH meter. Sterile 30 ml universals were either filled with 9 ml aliquots of high pH LB or standard LB broth (pH 7.3). In the first experiment, we compared the growth of Btk rifR at high pH in cultures initiated with 106 cells from either pasteurized spore preparations or overnight broth culture (predominantly vegetative cells). The experiment was cultured at 30 °C with continuous shaking (100 rpm). Cell density was measured five times over 24 h using spectrophotometry (OD at 600 nm) and pH readings taken at the same timepoint. In the second experiment, we compared the growth of a range of Bt strains in standard LB or LB adjusted to pH 10 as above. Spore suspensions of the strains Btk rifR, Btt specR, Dorset 7.1.o, BGSC 4A4 and BGSC 4D4 were prepared as above and 106 pasteurized spores used to inoculate each universal. Bacteria were cultured for 48 h, as above, and final densities measured by diluted plating two samples from each universal onto LB agar, and incubating plates at 30 °C. Means for each universal were used in subsequent analysis to avoid pseudoreplication. All treatments were replicated four times. 2.4. Reproduction within cadavers and larval guts Second and third instar larvae were exposed to Bt on leaf discs dipped in spore suspension at three standard concentrations (2 106, 2 105, 2 104 spores ml 1). Three larvae were confined on single leaf discs in 50 mm Petri dishes, with ten dishes per dose per instar (60 replicates in all). Leaf discs were suspended on 30 mm Petri dish lids within the larger 50 mm dishes so larval frass could be collected without contamination from leaf discs. We also set up five additional replicates (=Petri dishes) for Btk-resistant insects, and ten replicates of larvae exposed to Btt specR, a strain not pathogenic to P. xylostella, using the highest spore concentration. After 7 days frass was recovered from the bottom of the Petri dish and the consumption of each leaf disc estimated by calculating the size of holes in leaf discs. The average quantity of liquid inoculum retained by each leaf disc was calculated by weighing a subsample of leaf discs (n = 5) before and after dipping. This varied little between leaves and so the total number of bacteria on each disc was estimated at mean volume inoculum retained bacterial concentration. Inoculum concentration and proportion of leaf consumed were used to calculate the number of spores consumed in each Petri dish. Pasteurized and unpasteurized homogenates of cadavers were plated out, whereas all frass samples were pasteurized. 2.5. Vertical transmission from sub-lethally infected insects We exposed 150+ third instar Lab-UK larvae to Btk rifR on leaf discs dipped in a spore suspension (2 105 spores ml 1) as above. Insects which survived on these leaf discs were transferred to fresh sterile 90 mm Petri dishes (6 pupae in each of 10 dishes). Adults were
309
allowed to emerge and laid eggs on Parafilm that had been dipped in boiled cabbage extract within each dish. Eggs were washed from the Parafilm in sterile saline with sterilized paintbrushes, homogenized and plated out. Two abdomens were dissected from females within each Petri dish, homogenized and plated out on BcSA media as above. 2.6. Statistical analysis To measure the growth of bacteria in cadavers and frass, we used a standard metric of reproductive rates loge [(colony forming units recovered + 1)/(spores consumed)]. This measure controls for the variation in spores consumed associated with differences between instars, genotypes or replicates. Unpasteurized cadaver samples, containing vegetative cells and spores, very rarely had higher counts than pasteurized samples, indicating that bacteria had almost entirely sporulated within cadavers; the maximum value was used in analysis of reproduction, as this was most representative of reproduction at the time of sampling. All data were analysed in R v.2.1.0 (www.r-project.org/) using generalized linear modelling to carry out one-way analysis of variance, two-way analysis of covariance. Mixed model maximum likelihood anova was used to analyse time course data, with replicate incorporated as a random factor. Significance testing was carried out by sequential removal of non-significant terms (P > 0.05) in order to develop minimal adequate models. Model assumptions (homoscedasticity, normality) were checked with graphical analyses. 3. Results and discussion The reproductive rate of Btk rifR in cadavers increased with spore concentration (ANCOVA: F1,58 = 35.9, P < 0.0001; Fig. 1) but was not affected by larval instar (F1,57 = 0.28, P = 0.60) and there was no interaction between concentration and instar (F1,56 = 2.54, P = 0.12). However, the reproductive rate in cadavers only exceeded zero (i.e. more spores were recovered than were ingested) in a small proportion of replicates in the two treatments with the highest concentration of spores (Fig. 1), even though the mortality of second and third instars at the highest concentration of spores approached 100% (Mortality was 29/30 for both second and third instar larvae at the top spore concentration; 20/30 and 15/30, respectively, at the intermediate concentration and 1/30 and 6/30 at the low concentration). The cause of the poor reproduction of Btk rifR cadavers killed by this highly pathogenic isolate requires further investigation. However, since cadavers infected with this isolate, when plated out on less selective media (LB) routinely contain high counts of other bacterial species (>106 CFU per cadaver) it is likely that heterospecific competition is severely limiting growth in this strain (Takatsuka and Kunimi, 2000).
310
B. Raymond et al. / Journal of Invertebrate Pathology 98 (2008) 307–313
Fig. 1. The reproductive rate of Btk in cadavers [ln (CFU from cadavers/spores eaten)] in second and third instar diamondback moth over a range of concentrations of leaf-dip inoculum. A reproductive rate of zero indicates that the number of spores recovered is equal to the number of spores ingested. The raw data (open circles), with means for each dose (solid squares) have been plotted to show that reproduction exceeded 0 in some replicates. There are multiple overlapping data points at the minimum values for each dose.
While absolute counts of spores in frass are likely to reflect the quantity of material ingested we hypothesized that if spores recovered from frass were the product of active growth, we would recover more spores from frass than were ingested. In other words, the reproductive rate or the log ratio of spores recovered to spores ingested should exceed zero. The reproductive rate in frass in over a range of doses (both lethal and non-lethal) was always negative by a com-
fortable margin (Fig. 2). We also failed to detect any evidence of vertical transmission of Bt from insects surviving infection in eggs or adult females. An analysis of covariance showed that the proportion of spores was affected by dose (ANCOVA: log spore concentration F1,58 = 19.8, P 0.001, Fig. 2) and instar (F1,57 = 6.47, P = 0.014). Cultures initiated with spores and overnight broth culture (predominantly vegetative cells) could grow in broth
Fig. 2. The reproductive rate of Btk in frass [ln (CFU from frass/spores eaten)] in second and third instar diamondback moth over a range of concentrations of leaf-dip inoculum. A reproductive rate of zero indicates that the number of spores recovered is equal to the number of spores ingested. The raw data (open circles), with means for each dose (solid squares) have been plotted to show that reproduction never exceeded 0. There are multiple overlapping data points at the minimum values for each dose.
B. Raymond et al. / Journal of Invertebrate Pathology 98 (2008) 307–313
adjusted to a pH of 10.0 (Fig. 3a), although the growth in the treatment initiated with spores was faster than in treatment initiated with overnight cell cultures (mixed model ANOVA; time treatment interaction, df = 1, likelihood ratio = 39.5, P < 0.0001; Fig. 3a) This experiment used alkali broth that was not buffered, and pH declined as cultures grew. This does, however, mimic the situation in live hosts since toxins and replication of Bt in the gut disrupts its buffering capacity and lowers pH (Angus and Heimpel, 1956; Walther et al., 1986). Nevertheless bacteria were replicating while the pH was between 9.5 and 10.0 (Fig. 3a). All four strains of B. t. kurstaki and the one strain of B. thuringiensis var. tenebrionis proved capable of growth at both neutral and high pH in high nutrient media. High pH did retard the growth of all strains (F1,38 = 11.1, P = 0.002, Fig. 3b). While results with some strains were more variable at high pH there was no significant main
311
effect of strain (F4,34 = 2.08, P = 0.1) or significant interaction between strain and pH (F4,30 = 1.75, P = 0.17). This variability is best accounted for by the anomalous behaviour of a few replicates rather than consistent differences between strains. Whilst the ability of Bt to grow at high pH contradicts early reports, the ability to tolerate high pH seems consistent with a ecological niche that involves the invasion of a host from the gut. The data presented here, together with the observation that the high pH of Lepidopteran midguts can decline in the posterior midgut, and approach neutral values in the hindgut (Dow, 1984) indicates that pH is not a major barrier to the proliferation of Bt in the Lepidopteran intestine. An unexpected finding was the ability of Btt specR to replicate at high pH, given the lower pH of Coleopteran guts (Egert et al., 2005). Non-pathogenic strains of Bt can, however, reproduce in mixed infections with pathogenic isolates (Raymond
Fig. 3. (a) The growth and concomitant changes in pH for Btk rifR in alkaline LB broth. Open data points are cultures initiated with overnight cultures (predominantly vegetative cells), filled data points are cultures initiated with pasteurized spore/toxin suspensions. Square points represent optical density measurements, and lines are the fitted minimal adequate statistical model of bacterial growth. Triangular points are pH measurements. Data are means + SE, error bars are commonly smaller than the symbols. (b) The density of diverse Bt strains after 48 h growth in liquid media: initially adjusted to pH 10.0 (shaded bars) or control broth at pH 7.3 (open bars). Data are mean counts per replicate ± 1SE.
312
B. Raymond et al. / Journal of Invertebrate Pathology 98 (2008) 307–313
Fig. 4. The log ratio of spores recovered in frass: spores ingested from Btk resistant and susceptible larvae that had consumed Btk spores (shaded bars), and from susceptible larvae that had consumed B. t. tenebrionis spores (open bars), a subspecies of Bt not infectious to Lepidoptera. All leaf discs were inoculated with a standard concentration of spores (2 106 spores ml 1). Data are mean log ratios ± 1SE.
et al., 2007). Given the high local genetic diversity of Bt isolates (Collier et al., 2005), mixed infections may be common in the field and would maintain selective pressure for the tolerance of high pH. We hypothesized that if pore-formation by crystal toxins was involved in reproduction within the gut in live hosts, insects that were tolerant or resistant to Bt infection should have lower proportion of spores in frass to spores ingested relative to susceptible insects that were exposed to the same dose. In contrast, we found the opposite pattern: when hosts were resistant to Bt, a higher proportion of spores were recovered from frass relative to the quantity ingested (one-way ANOVA F2,22 = 20.4, P < 0.0001, Fig. 4). Model simplification revealed that there was no significant difference between the insects exposed to Btt specR and the Btk resistant insects exposed to Btk (ANOVA F1,23 = 0.13, P = 0.73). No insects exposed to Btt specR died, under half of the Btk resistant insects died (7/15), the majority (29/30) of the Btk susceptible insects also died. Since the resistant insects in this experiment have a loss of binding resistance mechanism (Sayyed et al., 2004) the absence of suitable toxin receptors in the host midgut, therefore, increased the proportion of viable spores found in frass. Given that there was no evidence for reproduction in the midgut, our interpretation of this data is that the spores recovered in frass are a fraction of the original inoculum. We conclude, therefore, that a higher proportion of spores in frass indicates higher spore survival, or lower germination rates. The inference of lower germination rates in a resistant host is consistent with the finding that toxin receptors within the insect midgut act as germination cues for Bt (Du and Nickerson, 1996). In summary, in this study Bt was able to tolerate high pH, but there was no evidence that it can reproduce in diamondback moth larvae without killing the host, at least for the strain used in this study. Currently, there is no evidence
that vegetative cells of Bt have any specialized mechanism for adhering to the gut, demonstrably contrasting with the attachments seen in Arthromitus form of B. cereus (Margulis et al., 1998). On the other hand, there are several lines of evidence that suggest vegetative growth of Bt in the gut may be important in lethal infections. The importance of live spores in synergizing the mortality caused by endotoxins (Li et al., 1987) indicates that vegetatively growing Bt contributes to the lethality of Bt toxins. Toxins produced by vegetative cells lyse midgut epithelial cells (Yu et al., 1997). The tolerance of high pH and the use of toxin-receptors as germination cues (Du and Nickerson, 1996) also suggest that Bt is adapted for growth in the gut. However, since ingestion of lethal doses of Bt endotoxins can cause paralysis very quickly (Endo and Nishiitsutsujiuwo, 1980) we would not necessarily expect to be able to detect vegetative growth in lethal infections in the frass excreted prior to death. Acknowledgments Thanks to BBSRC and NERC for financial support, and for the comments of anonymous referees who helped improve this manuscript. References Angus, T.A., Heimpel, A.M., 1956. An effect of Bacillus sotto on the larvae of Bombyx mori. Can. Entomol. 88, 138–139. Burges, H.D., Hurst, J.A., 1977. Ecology of Bacillus thuringiensis in storage moths. J. Invertebr. Pathol. 30, 131–139. Chiang, A.S., Yen, D.F., Peng, W.K., 1986. Germination and proliferation of Bacillus thuringiensis in the gut of rice moth larva, Corcyra cephalonica. J. Invertebr. Pathol. 48, 96–99. Collier, F.A., Elliot, S.L., Ellis, R.J., 2005. Spatial variation in Bacillus thuringiensis/cereus populations within the phyllosphere of broadleaved dock (Rumex obtusifolius) and surrounding habitats. Fems Microbiol. Ecol. 54, 417–425.
B. Raymond et al. / Journal of Invertebrate Pathology 98 (2008) 307–313 Dow, J.A., 1984. Extremely high pH in biological systems: a model for carbonate transport. Am. J. Physiol. Regul. Integr. Comp. Physiol. 246, R633–R636. Du, C., Nickerson, K.W., 1996. Bacillus thuringiensis HD-73 spores have surface-localized Cry1Ac toxin: physiological and pathogenic consequences. Appl. Environ. Microbiol. 62, 3722–3726. Egert, M., Stingl, U., Bruun, L.D., Pommerenke, B., Brune, A., Friedrich, M.W., 2005. Structure and topology of microbial communities in the major gut compartments of Melolontha melolontha larvae (Coleoptera: Scarabaeidae). Appl. Environ. Microbiol. 71, 4556–4566. Elliot, S.L., Sabelis, M.W., Janssen, A., van der Geest, L.P.S., Beerling, E.A.M., Fransen, J., 2000. Can plants use entomopathogens as bodyguards?. Ecol. Lett. 3, 228–235. Endo, Y., Nishiitsutsujiuwo, J., 1980. Mode of action of Bacillus thuringiensis delta-endotoxin-histopathological changes in the silkworm midgut. J. Invertebr. Pathol. 36, 90–103. Hendriksen, N.B., Hansen, B.M., 2002. Long-term survival and germination of Bacillus thuringiensis var. kurstaki in a field trial. Can. J. Microbiol. 48, 256–261. Jensen, G.B., Hansen, B.M., Eilenberg, J., Mahillon, J., 2003. The hidden lifestyles of Bacillus cereus and relatives. Environ. Microbiol. 5, 631– 640. Jung, S.C., Kim, Y., 2006. Synergistic effect of entomopathogenic bacteria (Xenorhabdus sp. and Photorhabdus temperata ssp. temperata) on the pathogenicity of Bacillus thuringiensis ssp. aizawai against Spodoptera exigua (Lepidoptera: Noctuidae). Environ. Entomol. 35, 1584–1589. Knell, R.J., Begon, M., Thompson, D.J., 1998. Host–pathogen population dynamics, basic reproductive rates and threshold densities. Oikos 81, 299–308. Kreig, A., 1964. Bacillus thuringiensis (Bacillaceae). Vegetative Vermehrung und Sporenbildung. In: Wolf, G. (Ed.), Encylocpaedia Cinematographica E555/1963. Institut fu¨r den Wissenschaftlichen Film, Gottingen, Germany, pp. 1–9. Li, R.S., Jarrett, P., Burges, H.D., 1987. Importance of spores, crystals, and delta-endotoxins in the pathogenicity of different varieties of Bacillus thuringiensis in Galleria mellonella and Pieris brassicae. J. Invertebr. Pathol. 50, 277–284. Margulis, L., Jorgensen, J.Z., Dolan, S., Kolchinsky, R., Rainey, F.A., Lo, S.-C., 1998. The Arthromitus stage of Bacillus cereus: intestinal symbionts of animals. Proc. Nat. Acad. Sci. USA 95, 1236–1241.
313
Martin, P.A.W., Travers, R.S., 1989. Worldwide abundance and distribution of Bacillus thuringiensis isolates. Appl. Environ. Microb. 55, 2437–2442. Prasertphon, S., Areekul, P., Tanada, Y., 1973. Sporulation of Bacillus thuringiensis in cadavers. J. Invert. Pathol. 21, 205–207. Raevuori, M., Genigeorgis, C., 1975. Effect of pH and sodium chloride on growth of Bacillus cereus in laboratory media and certain foods. Appl. Microbiol. 29, 68–73. Raymond, B., Davis, D., Bonsall, M.B., 2007. Competition and reproduction in mixed infections of pathogenic and non-pathogenic Bacillus spp.. J. Invertebr. Pathol. 96, 151–155. Sayyed, A.H., Raymond, B., Ibiza-Palacios, M.S., Escriche, B., Wright, D.J., 2004. Genetic and biochemical characterization of field evolved resistance to Bacillus thuringiensis toxin Cry1Ac in diamondback moth, Plutella xylostella. Appl. Environ. Microbiol. 70, 7010–7017. Smith, R.A., Couche, G.A., 1991. The phylloplane as a source of Bacillus thuringiensis variants. Appl. Environ. Microbiol. 57, 311–315. Suzuki, M.T., Lereclus, D., Arantes, O.M.N., 2004. Fate of Bacillus thuringiensis strains in different insect larvae. Can. J. Microbiol. 50, 973–975. Takatsuka, J., Kunimi, Y., 1998. Replication of Bacillus thuringiensis in larvae of the Mediterranean flour moth, Ephestia kuehniella (Lepidoptera: Pyralidae): growth, sporulation and insecticidal activity of parasporal crystals. Appl. Entomol. Zool. 33, 479–486. Takatsuka, J., Kunimi, Y., 2000. Intestinal bacteria affect growth of Bacillus thuringiensis in larvae of the oriental tea tortrix, Homona magnanima Diakonoff (Lepidoptera: Tortricidae). J. Invertebr. Pathol. 76, 222–226. van Frankenhuyzen, K., 1993. The challenge of Bacillus thuringiensis. In: Entwistle, P.F. et al. (Eds.), Bacillus thuringiensis; an Environmental Biopesticide: Theory and Practice. John Wiley, Chichester, UK, pp. 311–330. Walther, C.J., Couche, G.A., Pfannenstiel, M.A., Egan, S.E., Bivin, L.A., Nickerson, K.W., 1986. Analysis of mosquito larvicidal potential exhibited by vegetative cells of Bacillus thuringiensis subsp. israelensis. Appl. Environ. Microbiol. 52, 650–653. Yu, C.G., Mullins, M.A., Warren, G.W., Koziel, M.G., Estruch, J.J., 1997. The Bacillus thuringiensis vegetative insecticidal protein Vip3A lyses midgut epithelium cells of susceptible insects. Appl. Environ. Microbiol. 63, 532–536.