BIOPESTICIDES

Abstract

The present invention relates to Bacillus thuringiensis strains which are phenotypically stable and have increased virulence compared to the wild-type strains, whereby the increased virulence has been achieved by the exposure of the strain to a mutagen during one or more passages. Such strains are particularly useful as biopesticides. Methods for increasing virulence in microbial pesticides are also described.

Claims

1. A bacterial strain comprising: a B. thuringiensis strain which is phenotypically stable and has increased virulence compared to the wild-type strain, whereby the increased virulence has been achieved by the exposure of the strain to a mutagen during one or more passages.

2. The B. thuringiensis strain of claim 1, wherein the mutagen comprises ethyl methanesulfonate (EMS).

3. The B. thuringiensis strain of claim 1, e wherein the strain is a pesticide.

4. The B. thuringiensis strain of claim 3, wherein the pesticide is toxic to one or more of the following: aphids, other Hemipteran pests, thrips, Lepidopteran larvae, Dipteran larvae, Coleopteran larvae, spider mites, locusts, crickets, ants, cockroaches, flies, wasps, termites woodworm, wood ants, bookworms, silverfish, carpet beetles, clothes moths, gypsy moths, chiggers, Sarcoptes scabiei, ticks, mites, lice, fleas, bed bugs, mosquitoes, tsetse flies, kissing bugs, root-knot nematode, soybean cyst nematode, potato cyst nematode, slugs, snails, and the diamondback moth Plutella xylostella and its larvae.

5. (canceled)

6. The B. thuringiensis strain of claim 3, wherein the pesticide is applied to, near to, or in the soil of, cruciferous plants.

7. The B. thuringiensis strain of claim 1, wherein the B. thuringiensis strain comprises Bacillus thuringiensis entomocidus, Bacillus thuringiensis aizawai or Bacillus thuringiensis kurstaki.

8. A composition comprising the B. thuringiensis strain according to claim 1.

9. The composition according to claim 8, wherein the B. thuringiensis strain is in the form of spores or crystals.

10. The composition according to claim 8, wherein the composition is in the form of a solid or a liquid.

11. The composition according to claim 8, wherein the composition further comprises one or more of the following ingredients: wetting agents, thickeners, viscosity modifying agents, stabilisers and surfactant and a dispersing medium such as water or edible oils.

12. The composition according to claim 8, wherein the composition is a pesticide.

13. The composition according to claim 8, wherein the composition is applied to plants, soil surface and/or into the soil itself by spraying, spreading or sprinkling.

14. A method for improving virulence of microbial pesticides comprising: (a) splitting a population of pests into subpopulations; (b) exposing the subpopulations of pests to one or more microbial pesticides; (c) selecting pest cadavers from the subpopulations with highest mortality; (d) inoculating sporulation media with the pest cadavers so as to provide sufficient readily quantifiable inoculum for a subsequent round of infection; (e) purifying and quantifying microbial spores from the selected cadavers so as to identify one or more microbial pesticides having improved virulence; and wherein the microbial pesticide is exposed to a mutagen.

15. The method according to claim 14, further comprising using microbial spores from step (e) to infect one or more second pest populations.

16. The method according to claim 14, wherein steps (a) to (e) are repeated at least once, and optionally, the microbial pesticide is repeatedly exposed to the mutagen.

17. The method according to claim 14, wherein the cadavers are selected during step (c) on the basis of total mortality of the pest population reaching in the range of 25-30%.

18. The method according to claim 14, wherein a single clone is selected from the purified microbial spores of step (e) to infect further pest populations.

19. The method according to claim 18, wherein a single clone is selected for further infection at least twice per 5 passages.

20-24. (canceled)

25. The method according to claim 14, wherein the mutagen comprises ethyl methanesulfonate (EMS).

26-29. (canceled)

30. The method according to claim 14 wherein the microbial pesticide comprises a bacterium, and the bacterium is mutated by inactivation or disruption of a mutS gene.

31-36. (canceled)

Description

DETAILED DESCRIPTION OF THE INVENTION

[0059] Embodiments of the present invention will now be described, by way of example only, in which:

[0060] FIG. 1 is a schematic diagram showing the steps involved in infecting insect populations with bacterial spores according to group selection and pool selection methods;

[0061] FIG. 2 is a bar chart showing early larval mortality of Cry1Ac resistant larvae of P. xylostella exposed to ancestral (wt anc), unevolved B. thuringiensis kurstaki and evolved treatments. Data are means and SEs pooled over 2 replicate bioassays, at three doses for four replicated lineages per treatment;

[0062] FIG. 3 is a group of 6 graphs showing the overall mortality of Cry1Ac resistant larvae of P. xylostella exposed to ancestral (wt anc) unevolved B. thuringiensis kurstaki and evolved treatments. Data are proportional mortality per dose (n=50-60 larvae) with different symbols for replicate lineages within treatments. Lines are fitted statistical models (back-transformed into probability) fitted for each lineage and each bioassay. Line types indicate (dashed or solid) independent bioassays. Doses are viable spores per microlitre;

[0063] FIG. 4 is two graphs showing the efficacy of the best performing wild type clone (p3c2) against B. thuringiensis kurstaki resistant and susceptible P. xylostella in comparison to its ancestor and an alternative commercial Bt isolate (XenTari) used to target resistant larvae. Data are raw 5 day mortality from bioassays (n=60 larvae per dose);

[0064] FIG. 5 is two graphs showing the relative fitness of selected evolved clones, p3c1 (Bacillus thuringiensis kurstaki NCTC 17080301)) and p3c2 (Bacillus thuringiensis kurstaki NCTC 17080302), with increased virulence relative to ancestor across a range of initial frequencies. Data are based on the ratio of estimated Malthusian parameter; dashed line indicates fitted linear models with SE in grey. N=40-50 larvae per treatment;

[0065] FIG. 6 is two graphs showing the data of the Cry1Ac-resistant and -susceptible lavae;

[0066] FIG. 7 is a graph showing the alternating background clone bioassay in Cry1Ac-susceptible larvae;

[0067] FIG. 8 is three graphs showing the bioassay of resistant and co-evolved passage treatments for co-evolved larvae, Cry1Ac-resistant larvae and Cry1Ac-susceptible larvae; and

[0068] FIG. 9 is a graph showing the results of a bioassay assessing the toxin class for the initial mix, HCO and HCO+EMS in the toxin library selection experiment.

EXAMPLE

[0069] The first experiment sought to increase virulence of B. thuringiensis kurstaki to an insect host genotype of diamondback moth (DBM) Plutella xylostella, which was highly resistant (LC5010.sup.5 spores) to infection. The supply of mutations was manipulated by evolving wild type and mutator (mut) bacteria. In addition, group level selection was imposed in two ways: the first treatment used pooling of larvae from sub-populations at each round of infection, a regime that should select on high bacterial population size within insects and a protocol that has produced modest increases in virulence previously (11). The second selection treatment imposed group-level competition on mortality rates in a meta-population: in each round of selection pathogens were only propagated from groups with high levels of mortality. In the second selection experiment a Bt biopestidal isolate (B. thuringiensis entomocidus i) was evolved with already quite effective virulence in P. xylostella (LC5010.sup.2 spores); in this experiment we used a mutagen (EMS) to supply additional mutations in lieu of a mutator strain, and varied the insect genetic background used to impose selectin on virulence, in order to either improve evolution of virulence or produce mutants capable of infecting more than one genotype with higher efficacy.

Strains Used and Construction

[0070] The wild type B. thuringiensis kurstaki strain was obtained by transforming 7.1.o strain with pHT315-gfp plasmid containing the ermB gene conferring erythromycin resistance (21). It was maintained in vitro with 10 ug/mL erythromycin. When grown from insect cadavers, 300 g/mL erythromycin was used to inhibit growth of Gram-negative bacteria. The mut (mutator) strain was obtained by inactivating the mutS gene in 7.1.o. A fragment of mutS coding sequence containing an insertion of the cat gene was synthesised and cloned within BamHI sites of the thermosensitive vector pRN1501 (29). The resulting plasmid was transformed into 7.1.o strain, and clones with recombination of the interrupted mutS sequence into the chromosome were obtained after growth at 42 C. by screening for chloramphenicol (Cm) resistance and loss of erythromycin resistance (29). Insertion was confirmed by PCR and sequencing of mutS locus. The mutator phenotype of the resulting mut strain was confirmed with fluctuation tests showing an increase in mutation rate of 40-fold compared to the wt strain. The mut strain was grown in the presence of 5 ug/mL Cm. The strain of B. thuringiensis entomocidus in passage experiment 2 was recovered from a commercial the Bacillus Genetic Stock Centre (BGSC 414) and transformed with pHT315-gfp plasmid containing the ermB, as above.

Passage Experiment 1; Scale of Competition and Mutation Supply

[0071] Diamondback moth larvae used for selection were reared on an autoclaved artificial diet (30) without supplementary antibiotics until 3.sup.rd instar. Bt strains used for the selection experiment were first grown from single colonies into 100 mL of sporulating media (HCO) with appropriate antibiotic, then pasteurised (65 C., 20 min). Selection was then performed with 4 replicate lineages per treatment. The Cry1Ac resistant line of Plutella xylostella used for experiments was derived from a cross of NOQA-GE (from David Heckel, Max Planck, Jena) with the diet adapted susceptible line Vero Beach (from Oxitec UK), this line was selected for resistance to Cry1Ac and denoted VL-FR. A population that was fixed for resistance using a PCR screen of resistance alleles in the parental population (31) was established. An insect strain with similar genetic background (derived from NOQA-GE X Vero Beach) was established from F2 offspring but in this case fixed for susceptibility to Cry1Ac, this line was denoted VL-SS.

[0072] For infection, sterile diet was cut into quarters in 55 mL dishes, and each quarter was inoculated with 90 uL of spore suspension containing 510.sup.4 to 10.sup.5 spores/L (chosen to produce 25-30% mortality over 2-4 days for resistant diamondback moth) and dried. 10 third instar larvae were then added per 55 mL dish. Insect death was recorded from two days after infection. When total mortality reached 25-30%, the earliest cadavers were transferred to micro-centrifuge tubes with sterile toothpicks. After two days at room temperature (20-22 C.), 0.85% NaCl solution (saline) was added to the tubes. Cadavers were homogenised with pellet pestles, then the tubes were briefly vortexed and the suspension was transferred to wells containing 1 mL of sporulation medium in 24-well growth plates, for spore amplification in vitro. Sporulation medium was HCO (32) containing polymyxin (100 IU/mL) (Oxoid) and an additional relevant antibiotic: erythromycin 300 g/mL for wt lineages containing pHT315:gfp or chloramphenicol 5 g/mL for mut lineages containing the chromosomal cat marker. Rows of inoculated wells from the same line of selection were alternated with empty rows to prevent and control for contamination between lines. The plates were then sealed with tape to limit evaporation, and placed at 30 C. with 150 rpm agitation for three days.

[0073] Selection of cadavers and inoculation into sporulation medium depended on experimental treatment (as shown in FIG. 1). This experiment compared the efficacy of selecting for virulence by imposing group level competition on mortality between subpopulations in an insect host meta-population (group selection) or by imposing selection on efficient replication of pathogens in pools of cadavers (pool selection). For the group selection treatment, each lineage was split into 12 populations maintained separately. The 6 populations with highest larval mortality were selected; 6 other populations were discarded. If the same mortality was observed for several populations, the ones with earliest observed deaths were selected. Cadavers from selected populations were transferred into separate tubes (6 total). 50 L saline was added in each tube, and the suspension from each tube was inoculated into separate wells. After spore purification (described below), spores from each well were used to inoculate 2 populations in the next selection round. For the pooled selection treatment, each lineage was split into 6 populations, which were pooled together at each round of selection. All cadavers from 6 populations in the pooled treatment were transferred to the same microfuge tube. 300 L saline was then added and the suspension was distributed into 6 wells for inoculation. To maintain comparable population sizes, at inoculation, across treatments, only a subset of cadavers was selected from replicate lineages that had more than 30% death at time of selection.

[0074] After 3 days of growth, 200 L spore aliquots were transferred to 96-well PCR plates with aluminium sealing film for quantification, infection in next passage and storage. Plates were centrifuged at 3000 g for 15 min, the supernatant was removed and spores were resuspended in 100 L 0.85% NaCl. Plates containing spores to be used in next infection step were pasteurised (20 min, 65 C.) and kept at 10 C. until use, the others were stored at 20 C. Spore density was estimated by diluting 10 L of spores into 90 L saline and measuring OD600 nm in a 96-well microplate reader. A subset of the samples was plated at appropriate dilutions on LBA plates, in order to visually check for contamination and calibrate OD600 nm measurements.

Phenotypic and Genotypic Characterization of Strains

Bioassays

[0075] After five rounds of selection, the evolved lineages were frozen at 80 C. (LB 20% glycerol) without pasteurisation. These frozen stocks were used to inoculate sporulation medium for assays on whole populations. Clones were isolated from population samples by streaking 3 times (successively single colonies on LB agar plates supplemented with 20 g/mL erythromycin (wt clones) or 5 g/mL chloramphenicol (mut clones).

[0076] For bioassays, mixed populations or clones were inoculated into 1 mL HCO (without antibiotics) and grown for 3 days at 30 C., with 150 rpm agitation. Spore concentration was measured by plating on LBA after pasteurisation.

[0077] Lineage-level assays were carried out twice with independent spore stocks. Each assay was performed with at least three doses with 50-60 insects per dose. Initial screens of clone level variation examined 6 clones per lineage; these used 15-30 insects per dose at two doses. Data shown are deaths after 5 days. In passage experiment 1 clonal screens were analysed in terms of viable spores (counted as colony forming units on LBA) but also in terms of dilution factor of inocula relative to purified spore stocks as some clones displayed variable germination rates on LBA. Clones of interest were further characterized in additional bioassays, with methods that followed lineage assays, these assays were repeated at least twice.

Genotypic and Phenotypic Characterization

[0078] Ten evolved clones from both the wild type and mutator backgrounds, in addition to their respective ancestors were selected for whole genome sequencing on an Illumina HiSeq 2500 at a minimum of 30 coverage by 150 bp paired end sequencing in a single multiplexed run. DNA was extracted using Qiagen genomic extraction kits, after appropriate lysis for Gram-positive bacteria. Genomes were assembled using Velvet (33) and mapped against draft assemblies of ancestors and the completed assembly of B. thuringiensis kurstaki HD-1 using PAGIT (34). Measurements of relative fitness within insects used a marked strain of the 7.1.o wild type ancestor transformed with a pHT315-rfp plasmid containing the tetR gene conferring tetracycline resistance (21). This marked ancestor therefore has a plasmid with a similar backbone to the evolved wild type strains but is distinguishable on the basis of its distinct antibiotic resistance. Relative fitness is calculated from an estimate of relative reproductive rates of two genotypes within each insect, as described previously (21).

ResultsPassage Experiment 1

[0079] After five rounds of selection clear differences in virulence between selection treatments (pooled and group) and between the two genetic backgrounds (wild type and mut) relative to the ancestral wild type were seen. FIG. 2 shows the change in early larval mortality of Cry1Ac resistant larvae of P. xylostella exposed to ancestral (wt anc) unevolved B. thuringiensis kurstaki and evolved treatments. For early mortality, both group level selection (F.sub.1,44=6.54, P=0.014) and the mutator background increased the effectiveness of evolved Bt lines in killing resistant insects (F.sub.1,45=4.10, P=0.049) (see FIG. 2). The data from overall (day 5) mortality broadly confirms these simple conclusions. Mutators clearly performed better than wild type evolved strainsdoses as low as 25,000 spores were regularly producing mortality in excess of 75%, while 150,000 spores of the ancestor could not reliably produce this level of mortality. Group selection was also more effective than pooled selection, although the gains for wild type lineages were weaker than for mutators (as shown in FIG. 3). The full analysis of overall mortality was more complex as the effects of genetic background and selection treatment interacted with bioassay dose. However, pooling cadavers as a selection treatment was clearly inferior to group selection (generalized linear mixed model (glmm), z=2.37, P=0.0178) while the mutators were more efficient at killing insects because mortality responded more rapidly with dose (glmm, z=10.67, P<0.0001) (see FIG. 3). FIG. 3 shows the change overall of Cry1Ac resistant larvae of P. xylostella exposed to ancestral (wt anc), unevolved B. thuringiensis kurstaki and evolved treatments.

[0080] However, it is important to note that assays conducted with evolved lineages are potentially diverse bacterial assemblages and that lineages with increased virulence may contain clonal variants with substantially higher gains in killing power toward previously resistant insects. To investigate this clonal level variation 6 clones for each evolved lineage from passage 5 were streaked out and characterized as well as 6 clones from a subpopulation of a wild type group selected lineage that showed enhanced virulence at passage 3. The analysis of passage 5 clones broadly confirmed the lineage level assays, in that mutator and group selection treatments identified more clones with putative increases in virulence (as shown in Table 1 below). This analysis also showed that evolved lineages (especially in the pooled treatment) contain a substantial number of clones that appeared to have reduced virulence relative to ancestors (as shown in Table 1).

TABLE-US-00001 TABLE 1 (Clone level endpoint assays from passage 5. Six clones per independent lineage (48 overall) were screened for increases or decreases in virulence relative to their ancestor in two- dose discriminating assays. These data indicate the number of evolved clones with significant (P < 0.05) changes in virulence in logistic regression models with binomial errors.) Clones with Clones with increased decreased Treatment virulence virulence mut Group 14 5 mut Pool 8 6 wild type Group 9 9 wild type Pool 3 15

[0081] Sequencing of a number of these low virulence clones confirms that they are Cry toxin cheaters, strains which have lost one or both of the large bacterial plasmid on which the majority of Cry toxin genes are encoded (as shown in Table 2 below).

TABLE-US-00002 TABLE 2 (Genetic characterization of evolved clones (by short-read whole genome sequencings) revealed different patterns of loss for the B. t. kurstaki Cry toxin bearing plasmids. 24 clones from each treatment group were screened with PCR.) loss of both Cry plasmids (pKur2/ loss of Cry pBMB299 and (pKur6/ megaplasmid Treatment pBMB65) (pKur2) only mut Group 0 0 mut Pool 1 5 wildtype Group 1 3 wildtype Pool 7 8

[0082] In addition to this broad characterization of a large number of clones, a subset of clones of interest was characterized in more detail in Table 3 below. Notably, several mutants were identified with increases in killing power (i.e. a reduction in LC50) of greater than an order of magnitude. For some mutators (mgd6), and indeed for the lineage data in FIG. 4, there was substantial variation in virulence between assays, which is likely to be related to instability of mutator genotypes. FIG. 5 shows the efficacy of best performing wild type clone (p3c2) against B. thuringiensis kurstaki resistant and susceptible P. xylostella in comparison to its ancestor and an alternative commercial Bt isolate (XenTari) used to target resistant larvae. Data are raw 5 day mortality from bioassays (n=60 larvae per dose). Detailed assays of the best performing clone, p3c2, showed that it suffered some reduction in virulence against the Bt susceptible line VL-SS. However, it performed as well, or better, than the currently available Bt strain (XenTari) used to target Cry1Ac resistant diamondback moth (see FIG. 4).

TABLE-US-00003 TABLE 3 Detailed clone level variation in virulence (LC50s) - measured as log.sub.10 transformed spore dose/colony forming units l.sup.1 for a selection of high performing clones with increased virulence relative to their ancestor. Increase in virulence Genetic relative to Clone background LC50 (SE) ancestor ancestral B. t. wild-type 4.98 (0.056) kurstaki 7.1.o p3c1 wild-type 3.93 (0.02) 11 p3c2 wild-type 2.67 (0.024) 203 p3c3 wild-type 4.19 (0.026) 6.1 mga3 mutator 3.73 (0.055) 17.8 mgab2 mutator 4.32 (0.04) 4.6 mgd6 mutator 3.6 (0.58) 24.0 mgc5 mutator 4.4 (0.11) 3.8

[0083] In addition to sequencing the relative fitness of two of the high virulence clones was investigated with respect to their evolved ancestor (p3c1 and p3c2 (Bacillus thuringiensis kurstaki NCTC (provisional accession number) 17080301 and Bacillus thuringiensis kurstaki NCTC (provisional accession number) 17080302 respectively)). Since group selection has the potential to increase the frequency of traits that are costly to individuals it was expected that the protocols may have allowed identification of clones that are more effective at killing hosts than their ancestors but which have reduced competitive fitness in insects, potentially because they have evolved greater levels of investment in production of virulence factors. As predicted, clone p3c2 was substantially less fit in terms of competitive fitness with respect to its ancestor (t=4.2, P<0.0001); while both clones showed fitness that declined as their initial frequencies increased (t=2.3, P=0.022) (see FIG. 5), a pattern consistent with social interactions and the costs of producing secreted virulence factors in B. thuringiensis (21, 28). FIG. 5 shows the relative fitness of selected evolved clones with increased virulence relative to ancestor across a range of initial frequencies. Data are based on the ratio of estimated Malthusian parameters.

Passage Experiment 2

[0084] The second passage experiment had several additional aims. First, to repeat the group selection experimental protocol with a different bacterial genotype (here B. thuringiensis entomocidus) and second to test whether the virulence could be improved of a pathogen against a relatively susceptible host. We chose B. thuringiensis entomocidus (Bacillus Genetic Stock Centre strain 414) as a target, as it is as effective against Cry 1Ac diamondback moth larvae (LC50490 cfu/1) (FIG. 6) although not as pathogenic as B. thuringiensis kurstaki. Resistance to Cry1Ac also provides some cross-resistance to this B. thuringiensis entomocidus strain (conferring an of LC50 7900 cfu/1). Here we also set out to test whether a mutagenic agent (0.02% ethyl methanesulfonate (EMS)) could be used in lieu of mutator strains in selection to provide addition genetic variation for selection to act upon (see below). Finally, we had two specific evolutionary hypotheses to evaluate: we tested whether diversification of the host genotypes in selection lines could be used to generally improve virulence against multiple insect genotypes; this was tested by alternately passaging Bt lineages in Cry1Ac susceptible and resistant populations. Finally, an additional treatment tested whether increased responsiveness to selection could be gained by co-evolving both insect hosts and bacterial pathogen, as opposed to evolving pathogens against a host with a constant genetic background.

[0085] There were therefore four treatments in this selection experiment, all of which used a chemical mutagen: [0086] 1. Selection in Cry1Ac-susceptible larvae; [0087] 2. Selection in Cry1Ac resistant larvae; [0088] 3. Alternating selection in susceptible and resistant larvae; [0089] 4. Coevolution treatment, here larval populations that were heterogeneous for the Cry1Ac-resistance allele (at 10% initial frequency) were used to initiate passage. In each independent replicate, larvae that survived an infection from Bt under selection (ca. 400 per replicate) were reared on to adult stages to provide larvae for the next passage.

[0090] The overall structure of this selection experiment followed the group selection design in experiment 1 (FIG. 1). For each passage round there were 4 treatments, each of 4 replicates of 10 subpopulations. Once mortality had reached >20%, cadavers from the 5 subpopulations with the highest/earliest mortality were kept and spores recovered from them for the next passage round. Spore inocula were adjusted between rounds to try and maintain 20% mortality.

Mutagenic Protocol

[0091] The protocol differed from previous ones as during an in-media growth phase a chemical mutagen was added, with the aim of increasing variation for selection. In practice, selected cadavers were mashed and incubated in 100 l of LB at 30 C. for 3 hrs, then 0.2% ethyl methanesulfonate (EMS) added and incubated for a further 3 hrs. The EMS mutagen was then removed via centrifuging and resuspension, and spores the produced by adding the resuspension to 1 ml HCO per subpopulation (with selective erythromycin) and incubating at 30 C. for 6 days. Preliminary work with EMS identified a concentration that produced the greatest number of rifampicin resistant mutants without significant cell death, and the chosen 0.2% concentration was then used as described above.

[0092] After the final (5.sup.th) round of selection, spores from each selected subpopulation in each replicate were bio-assayed against resistant and susceptible larvae (and the subsequent generation of co-evolved larvae for the resistant and co-evolution treatments) at a range of spore dilutions. Spore dilutions were varied depending on treatment and larvae in the bioassay to cover from 0% to 100% mortality, and dilutions were refined in a 2.sup.nd round of bioassays to better achieve this. The ancestral Bt entomocidus strain was also included in all assays for comparison. Spores were grown in HCO for all endpoint bioassays.

ResultsPassage Experiment 2.

[0093] Here we investigated whether our use of a mutagen could also produce significant increases in virulence; whether specialized adaptation to one Cry1Ac resistant insect genotype could be offset by selecting pathogens in alternating insect genotypes and finally whether co-evolving insects and pathogens could accelerate the evolution of resistance.

[0094] B thuringiensis entomocidus in the susceptible larvae and alternating treatment evolved significantly increased virulence to both Cry1Ac-resistant and susceptible hosts (mixed model binomial errors, .sup.2=6.67, df=2, P=0.036) (FIG. 6). However, gains in virulence tended to be relatively modest and resulted only in the halving of LC50s; with the exception of replicate 3 in the alternating treatment (shaded triangles in FIG. 6), this a had 10-fold reduction in LC50 in susceptible insects (LC50=53.7 cfu/l) (FIG. 6). Multiple individual clones from this lineage shared similar increases in virulence in Cry1Ac susceptible larvae relative to their ancestor (FIG. 7). While this demonstrated that use of a chemical mutagenic could facilitate improvement of virulence, the alternating exposure to different genetic background did not lead to avoidance of trade-offs: gains in virulence in targeting Cry1Ac susceptible larvae were accompanied by reduction in virulence against Cry1Ac resistant larvae (FIG. 6; LC50 of replicate 3 in resistant insects 26,900 cfu/l). Note that other evolved lines showed a consistent halving of LC50s for both resistant and susceptible larvae.

[0095] B thuringiensis entomocidus was also selected to increase virulence against a constant Cry1Ac resistant insect genetic background, and in a co-evolving insect background, which increased resistance as selection for virulence proceeded. Both these selection treatments produced bacteria that were significantly more virulent than their ancestor (FIG. 8: resistant larvae treatmentestimate=1.36, SE=0.40, t=3.44, P=0.00787; co-evolved larvae treatmentestimate=1.23, SE=0.41, t=2.99, P=0.0034). However, both these selection regimes produced similar gains in virulence (FIG. 8) so there was no additional benefit to be gained from co-evolving larvae with pathogen in this experiment. Notably at the end of the experiment the co-evolved larvae had similar levels of resistance to B. thuringiensis entomocidus as the Cry1Ac resistant larvae (estimate 0.3, SE=0.26, t=1.18, P=0.24). In selection treatments with resistant and co-evolving larvae, improvements in killing these genetic backgrounds did not appear to result in a trade-off; evolved lineages did not acquire reduced virulence in Cry1Ac-susceptible larvae (FIG. 8).

Toxin Library Selection Experiment.

[0096] Here we set out to test whether group selection could isolate mutants with high virulence from a library of clones with pre-existing variation in the Bt toxin Cry1Ac. Selection was started with a mix of 16 clones from a library of toxin variants; each variant contained one to seven amino-acid changes in Cry1Ac. Each variant was present on a pHT315 plasmid containing the ermB gene conferring erythromycin resistance, transformed in Bt 407 Cry-background. Plasmids were maintained in vitro with 10 ug/mL erythromycin. When grown from insect cadavers, 300 g/mL erythromycin was used to inhibit growth of Gram-negative bacteria.

[0097] The passage experiment was performed according to the group selection regime as described in Passage experiment 1 (FIG. 1), using the insect strain VL-SS susceptible to Cry1Ac and an infective dose of around 30 spores/L (chosen to produce 25-30% mortality over 2-4 days for susceptible diamondback moth). In this experiment we applied two treatments, one in which spores were produced in HCO media, as per passage experiment 1, while the second treatment used additional chemical mutagenesis as per the protocol set out in passage experiment 2 (HCO+EMS). We carried out three rounds of selection. After selection 12 clones from each selection replicate were isolated from cadavers and Cry variants were identified using PCR and Sanger sequencing (48 clones per treatment in all).

ResultsToxin Library Selection Experiment

[0098] Initial bioassays indicated that 6 mutants had virulence that was indistinguishable from negative controls, i.e. these genotypes were null mutants that had no Cry1Ac activity (N135Q; EP8; EP26; EP23; EP14; EP15). In addition 3 mutants had suffered some moderation in virulence as a result of amino acid changes: EP13 (LC50 52 spores/l); EP27 (32.7 spores/l) and EP24 (23 spores/l); these mutants are classified as having very low, low and moderate virulence respectively (see shading in FIG. 9). All other mutants had amino acid changes that did not appear to affect mortalitythese were classed as having high virulence and shared an LC50 of 12 spores/l. Three rounds of group selection decreased the representation of Cry1Ac null mutants in both selection treatments. However, the HCO only treatment also increasing the representation of the high virulence mutants, whereas in HCO+EMS (mutagen) treatment, mutants with low and moderate virulence increased frequencies at the expense of highly virulent mutants.

Discussion

[0099] These experiments demonstrate a number of novel and effective steps that could be used to improve a biological pesticide based on B. thuringiensis, or other insect pathogens. First, it demonstrates that a combination of in vivo selection on mortality and in vitro bulking up of infectious material can be used to produce rapid changes in virulence phenotypes. Each round of infection and sporulation took under two weeks, and therefore an experienced researcher could conduct five rounds of selection in less than three months. These methods appear to be repeatable for different bacterial genetic backgrounds with different applications in mind. Notably experiments could consistently increase virulence in both highly and moderately resistant insects, although one replicate in the alternating selection regime in the B. thuringiensis entomocidus passage experiment did gain significantly increased virulence against susceptible insects.

[0100] Importantly, the details of the selection protocol are critical to producing good results. While previous selection experiments in both Bt and other pathogens have shown that pooling groups of bacteria and selecting for high population size and effective use of resources can maintain investment in virulence and produce strains with shorter time to death (7, 11), here it is shown that group level competition in a meta-population produced lineages with greater increases in virulence (see FIGS. 2, 3; Table 1) and was more effective at maintaining production of Cry toxins (see Table 2). The consequences of modifying selection design to including diverse genetic material in insect targets are not so clear-cut. Nevertheless, the improvement of 5/8 replicates in the alternating and co-evolution treatments in passage experiment 2 suggests that these modifications can further refine the selective process. In these experiments co-evolving target insects with pathogens did not improve response to selection for virulence. However, this could be because resistance to Cry1Ac (in the NA-QAGE derived insects) in these experiments is monogenic and largely recessive. Individuals that are heterozygous for resistance may not have the intermediate levels of resistance that can facilitate pathogen adaptation and may appear only fleetingly in selection experiments. Coevolution regimes in target hosts with more complex modes of resistance could prove more successful. Characterization of evolved clones with increased virulence has not yet revealed any changes in sequence of Cry toxins; evolved clones also vary in their level of virulence (see Table 3) and in their reproductive rate in insects (as shown in FIG. 5), suggesting that improvement has occurred via a range of mechanisms.

[0101] In addition, it has been shown that group-selection can facilitate isolation of clones with improved biocontrol potential, but which suffer some individual level cost, i.e. they have a reduced fitness in competition with their ancestor (see FIG. 5). Thus this selection regime has the power to create and identify mutants that would not rise to high frequency in conventional sequential passage experiments. Since investment in virulence factors and secondary metabolites is expected to be costly for a range of pathogens, and not just Bt, group selection based protocols could therefore have broad potential for improving virulence in range of biological control agents. Other entomopathogenic microbes, such as fungi, which also invest in a wide range of secondary metabolites used to enhance virulence might also respond to group level selection techniques. Previous research has indicated that cheating and conflict might account for the dynamics of virulence in entomopathogenic nematodes (35), another important group of biological control agents. As with Bt these organisms, and their symbionts, secreted a wide range of protein based virulence factors important for invading, killing and digesting their insect hosts (36). Moreover, selection experiments indicate that treatments that limit opportunities for cheating and within host competition are better at maintaining virulence. In addition, group-level competition based on maximizing reproductive rate in cadavers did not lead to better maintenance of virulence in that study, since there appeared to be a trade-off between virulence and maximizing the production of progeny in hosts (35).

[0102] While the methods presented here provide a framework for the general improvement of virulence against a particular insect host, the combination of in vivo and in vitro selection could also be applied to adapted strains with good virulence characteristics to cheaper or alternative growth media, while retaining pesticidal efficacy. We have shown that protocols that rely on generating genetic variation via fixed, higher mutation rates or by chemically-induced mutagenesis can facilitate selection for increased virulence. Of the two protocols, the use of a chemical mutagen has two advantages; first it does not require the use of a genetically modified strain and second any clones produced during selection are phenotypically more stable.

[0103] We have also demonstrated proof of principle for the application of these selection techniques to pools of mutants with pre-existing genetic variation, here derived from a library of clones in which proteins of interest have been mutagenized. Notably in this experiment, there was no benefit to supplementing mutation supply with a mutagen during the course of the selection experiment.

[0104] The forgoing embodiments are not intended to limit the scope of the protection afforded by the claims, but rather to describe examples of how the invention may be put into practice.

Biological Deposits

[0105] The application refers to the following indications of deposited biological material comprising two (2) deposits: [0106] Name: European Collection of Cell Cultures [0107] Address: Public Health England Porton Down, [0108] National Collection of Type Cultures, [0109] PHE Culture Collections, Microbiological Services, [0110] Porton Down, [0111] Sailsbury, [0112] SP4 0JG [0113] United Kingdom [0114] Date: 3 Aug. 2017 [0115] Accession Number: NCTC 17080301 (provisional number) [0116] Descriptor: Bacillus thuringiensis kurstaki (p3c1) [0117] Depositor: University of Exeter

[0118] -and- [0119] Date: 3 Aug. 2017 [0120] Accession Number: NCTC 17080302 (provisional number) [0121] Descriptor: Bacillus thuringiensis kurstaki (p3c2) [0122] Depositor: University of Exeter

REFERENCES

[0123] 1. M. Kruger, J. Van Rensburg, J. Van den Berg, Environ. Entomol. 40, 477 (2011). [0124] 2. B. E. Tabashnik, J. B. J. Van Rensburg, Y. Carrire, J. Econ. Entomol. 102, 2011 (2009). [0125] 3. T. Vaughn et al., Crop Science 45, 931 (2005). [0126] 4. M. Soberon et al., Science 318, 1640 (2007). [0127] 5. A. H. Badran et al., Nature 533, 58 (2016). [0128] 6. S. P. Diggle, A. S. Griffin, G. S. Campbell, S. A. West, Nature 450, 411 (2007). [0129] 7. A. S. Griffin, S. A. West, A. Buckling, Nature 430, 1024 (2004). [0130] 8. T. F. Cooper, PloS Biol. 5, e225 (2007). [0131] 9. R. E. Lenski, Evolution 42, 425 (1988). [0132] 10. R. D. Schulte, C. Makus, B. Hasert, N. K. Michiels, H. Schulenburg, Proceedings of the National Academy of Sciences 107, 7359 (2010). [0133] 11. J. Garbutt, M. B. Bonsall, D. J. Wright, B. Raymond, Ecol Letts 14, 765 (2011). [0134] 12. M. Berling et al., Applied and Environmental Microbiology 75, 925 (2009). [0135] 13. R. Maddamsetti, P. J. Hatcher, S. Cruveiller, arXiv.org, (2012). [0136] 14. M. E. Watson, J. L. Burns, A. L. Smith, Microbiology(UK) 150, 2947 (2004). [0137] 15. C. D. Bayliss et al., Infection and Immunity 76, 5038 (2008). [0138] 16. C. Pal, M. D. Macia, A. Oliver, I. Schachar, A. Buckling, Nature 450, 1079 (2007). [0139] 17. D. Racey, R. F. Inglis, F. Harrison, A. Oliver, A. Buckling, Evolution 642, 515 (2009). [0140] 18. J. A. G. M. de Visser, Microbiology (Reading, Engl) 148, 1247 (2002). [0141] 19. A. C. Shaver et al., Genetics 162, 557 (2002). [0142] 20. E. van Leeuwen, S. O. A. Neill, A. Matthews, B. Raymond, ISME J 9, 2328 (2015). [0143] 21. L. Zhou, L. Slamti, C. Nielsen-Leroux, D. Lereclus, B. Raymond, Current Biology 24, 2417 (2014). [0144] 22. F. Ginetti, M. Perego, A. M. Albertini, A. Galizzi, Microbiology 142, 2021 (1996). [0145] 23. N. Colegrave, S. Collins, Heredity 100, 464 (2008). [0146] 24. S. West, S. Diggle, A. Buckling, A. Gardner, A. Griffin, Annu. Rev. Ecol. Syst 38, 53 (2007). [0147] 25. S. A. West, A. Buckling, Proc. R. Soc. Lond. Ser. B-Biol. Sci. 270, 37 (2003). [0148] 26. J. B. Xavier, K. R. Foster, Proc. Natl. Acad. Sci. U.S.A 104, 876 (2007). [0149] 27. G. J. Velicer, L. Kroos, R. E. Lenski, Nature 404, 598 (2000). [0150] 28. B. Raymond, S. A. West, A. S. Griffin, M. B. Bonsall, Science 337, 85 (2012). [0151] 29. D. Lereclus, M. Vallade, J. Chaufaux, O. Arantes, S. Rambaud, Biotechnology 10, 418 (1992). [0152] 30. F. Hunter, N. E. Crook, P. F. Entwistle, in Viruses as pathogens for the control of insects., J. M. L. Grainger, Ed. (Academic Press, New York, 1984), pp. 323-347. [0153] 31. S. W. Baxter et al., Genetics 189, 675 (2011). [0154] 32. M. M. Lecadet, M. O. Blondel, J. Ribier, J Gen Microbiol 121, 203 (1980). [0155] 33. D. Zerbino, E. Birney, Genome Research 18, 821 (2008). [0156] 34. M. Swain et al., Nature Protocols 7, 1260 (2012). [0157] 35. D. Shapiro-Ilan, B. Raymond, Evol Appl 9, 462 (2016). [0158] 36. I. Eleftherianos, S. Joyce, R. H. ffrench-Constant, D. J. Clarke, S. E. Reynolds, Parasitology 137, 1695 (2010).