Antimicrobial compounds from rotifers

Abstract

The present invention relates to methods comprising exposure of a rotifer to a microbe. The invention also relates to compounds, polynucleotides, methods of therapy and methods of controlling microbial disease in plants.

Description

[0001] METHOD Field of the Invention The present invention relates to methods comprising exposure of a rotifer to a microbe. The invention also relates to compounds, polynucleotides, methods of therapy and methods of controlling microbial disease in plants. Cross reference to related application This application claims priority from GB 2317905.4 filed on 23 November 2023, the contents of which are hereby incorporated by reference. Background of the invention Antagonistic interactions among species are strong, ubiquitous and relentless sources of selection in natural populations. Examples include arms-races between pathogen virulence factors and host immune systems, and coevolution between antimicrobial compounds or toxins and pathways to resist them. These dynamics are central to various global challenges, including emerging infectious diseases, management of crop pathogens, and antimicrobial resistance. According to theory, selection arising from antagonistic coevolution can favour adaptations to shuffle existing combinations of genes or to acquire genes bringing new functions. These processes help explain the especially rapid and intense adaptive evolution seen at loci encoding the molecular mediators of conflict. Different domains of life typically address the challenge of ongoing genetic mixing in distinct ways. In bacteria and archaea, horizontal gene transfer (HGT) occurs by various mechanisms and is well known as a route for the spread of antimicrobial resistance, as well as the genes encoding antibiotic production themselves. For example, non-ribosomal peptide synthetases (NRPS) and polyketide synthetases (PKS) are large, multi-modular enzymes that catalyse assembly of a range of natural products, including antimicrobial compounds. These can be encoded as biosynthetic gene clusters on plasmids as well as chromosomes, and their mobility and modular structure facilitates diversification to produce a vast array of secondary metabolites via recombination within and between genomes. The natural ‘reservoir’ of mobile genetic diversity for antibiotic synthesis and resistance is thought to reflect a history of coevolution between the producers and targets of antimicrobial compounds. In contrast, among eukaryotes, the most important mechanism of genetic exchange is meiotic sex, by which whole genomes are shuffled every generation though recombination, segregation and outcrossing. Although meiotic shuffling has different effects to HGT, sex too has been linked to biotic conflict because it can speed up host adaptation against pathogens by generating new combinations of resistance alleles. When coevolving pathogens are common and virulent, the theoretical benefits of genetic exchange are so substantial that they may outweigh the inherent costs of sexual reproduction compared with parthenogenesis. Antagonistic coevolution may therefore help explain why obligately asexual plant and animal lineages are typically rare and short-lived despite major advantages. This so-called ‘Red Queen Hypothesis’ (RQH) draws support from associations between recombination and immunity, from host-pathogen dynamics in mixed sexual and asexual populations, and from the susceptibility of asexually propagated lineages to pathogens. Herein is revealed the links between genetic transfer and biotic conflict in a group of animals that challenge typical distinctions between domains described above. Bdelloid rotifers are a class of microscopic invertebrates that live in freshwater and limno-terrestrial habitats worldwide. Reproduction is only known by parthenogenetic eggs and neither males nor sperm have been reported despite centuries of microscopic observation and the description of hundreds of species, leading to the hypothesis that the class Bdelloidea has diversified for tens of millions of years in the absence of sexual reproduction. Genetic evidence to either confirm or refute obligate asexuality in bdelloids has proved complicated, and seemingly definitive evidence both for and against this hypothesis has been overturned or reinterpreted by later work. In contrast, repeated studies have demonstrated that bdelloid genomes encode extraordinarily high proportions of genes acquired horizontally from non-metazoan taxa. Approximately 10% of genes appear to have been captured from bacteria, fungi, plants and other sources, rather than sharing recent common ancestry with metazoan orthologs. This estimate is an order of magnitude greater than for other animals, holds for all bdelloid genomes so far examined, and is consistent across various methods for detecting HGT. Comparisons among bdelloid species indicate that most HGT events are ancient, with ongoing acquisition rates estimated to be on the order of one gene per 100,000 years. At these rates, the phenomenon would be too slow to equate with the sexual shuffling seen in typical eukaryotes, or the rapid dynamics of bacterial accessory genomes. Nevertheless, HGT has been hypothesised to introduce novel biochemical functions that help bdelloids adapt to environmental challenges, as it does in bacteria. Acquired genes are expressed and incorporated into metabolic pathways, some of which are not shared by other metazoans. Putative functions identified to date include desiccation tolerance, nutrient exploitation and repair of DNA damage. However, the deep associations between genetic transfer and coevolution raise the hypothesis that HGT may particularly help rotifers deal with biotic antagonism; for instance by acquiring genes with pathogen resistance functions. Isolated examples of horizontally acquired genes contributing to immunity are known from invertebrates, but the massive scale of HGT in bdelloids and the prevalence of asexual reproduction might especially favour the co-option of unusual pathways to resist microbial enemies, with parallels to the role of HGT in bacterial conflict. If so, this could compensate in part for the challenge that an asexual lineage theoretically faces from pathogens. This hypothesis is investigated herein by testing whether horizontally acquired genes are disproportionately involved in the response of bdelloid rotifers to infection. Like all animals, bdelloid rotifers are exploited by a range of natural enemies, including over 60 species of virulent fungal and oomycete pathogens. These can exterminate cultured populations in a few weeks, and significantly depress the abundance of hosts in natural habitats. However, almost nothing is known about variation in susceptibility among rotifers, how this compares with observations in other invertebrate pathosystems, whether its genetic basis matches models of coevolution, or how the underlying mechanisms evolve and remain effective if sex is rare or absent. Summary of the invention Coevolutionary antagonism generates relentless selection that can favour genetic exchange, including transfer of antibiotic synthesis and resistance genes among bacteria, and sexual recombination of disease resistance alleles in eukaryotes. Herein is described an unusual link between biological conflict and DNA transfer in microbe-resistant rotifers, microscopic animals whose genomes show elevated levels of horizontal gene transfer from non-metazoan taxa. When such rotifers are challenged with a microbe (e.g. a fungal pathogen), horizontally acquired genes are observed to be over twice as likely to be upregulated — a far stronger enrichment than observed under abiotic stress. Among hundreds of differentially expressed non-metazoan genes, the most markedly overrepresented are genetic clusters resembling bacterial polyketide and non-ribosomal peptide synthetases that produce antibiotics. Upregulation of these clusters in a pathogen-resistant rotifer species is shown to be nearly ten times stronger than in a susceptible species. By acquiring, domesticating, and expressing non-metazoan biosynthetic pathways, bdelloids may have evolved to resist natural enemies using antimicrobial mechanisms absent from other animals. Accordingly, the inventors have devised a novel approach of producing compounds having anti-microbial activity. This has been achieved through the discovery that by exposing a microbe- resistant rotifer to a microbe, the rotifer produces a compound having anti-microbial activity. Also identified is that genes horizontally acquired from non-metazoan taxa in microbe-resistant rotifers become overexpressed upon exposure to a microbe, wherein such horizontally acquired genes may comprise genes associated with antimicrobial compound production. The present invention thereby provides a novel method for the production of compounds wherein the method has the potential to yield novel compounds suitable for clinical application, and particularly compounds which may address the unmet need for compounds to target pathogenic microbes possessing resistance to the antibiotics currently approved for use in the clinic. The present invention thereby also provides a novel method for the production of compounds wherein the method has the potential to yield novel compounds suitable for agricultural application, and particularly compounds which may control microbial disease in plants. The invention therefore provides a method comprising providing a microbe-resistant species of rotifer in culture, exposing the rotifer to a microbe, and isolating a compound produced by the rotifer following exposure to the microbe. The invention also provides a method comprising assaying anti-microbial activity of a compound isolated according to the method of the invention. The invention also provides a compound produced and isolated by the method of the invention. The invention also provides a method of treating a microbial infection, wherein the method comprises administering the compound of the invention to a subject. The invention also provides a polynucleotide comprising one or more genes which are comprised within the genome of a microbe-resistant rotifer and have been acquired by the rotifer via HGT from non-metazoan taxa, preferably wherein said one or more genes have increased expression upon exposure of the rotifer to a microbe, preferably wherein the microbe is a fungus. The invention also provides a method of controlling microbial disease in a plant, the method comprising contacting the plant with a compound according to the invention, preferably wherein the microbial disease is a fungal disease. Brief Description of Figures Figure 1 shows the response of bdelloid rotifers to experimental inoculation with the fungal pathogen R. globospora. a Active and contracted A. vaga (left panel); healthy A. ricciae (middle panel); composite image of three A. ricciae corpses with fully developed R. globospora infections, with hyphae differentiating into irregular resting spores and long conidiophores bearing spherical infectious conidia (scale bars 100µm). b Proportion of rotifers killed by infection 72 hours after initial exposure to R. globospora. Points indicate replicate laboratory populations of A. vaga (red) and A. ricciae (blue) with approximately 10 individuals (range 6 to 20) exposed to 1000 conidia. c Dynamics of gene upregulation (solid bars) and downregulation (hatched bars) at 7- and 24-hours post-inoculation (timepoints T7 and T24 respectively), relative to control populations inoculated with UV-inactivated spores. At T7, 541 A. vaga genes and 962 A. ricciae genes showed significant DE (absolute fold change in expression > 4 and FDR < 1e–3). Most were upregulated in the pathogen treatment group relative to controls (452 in A. vaga, ~84%; 709 in A. ricciae, ~74%); the remainder were downregulated. At T24, the number of DE genes rose to 1767 in A. vaga (1093 upregulated, ~62%), and 2388 in A. ricciae (1590 upregulated, ~67%). Thus, the number of genes showing significant DE increased between timepoints in both species, but more genes were DE (and especially upregulated) in A. ricciae than in A. vaga. At T7, the relative magnitude of DE among the upregulated subset did not differ significantly between A. vaga and A. ricciae (estimate = 0.076, SE = 0.053, t = 1.44, P = 0.15). However, at T24 the relative magnitude of DE was significantly higher for the upregulated subset in A. ricciae than in A. vaga (three-way interaction of time, species and DE set: estimate = 0.23, SE = 0.058, t = 4.02, P <0.0001). Significant differences were not detected between species in the magnitude of DE in the downregulated sets at either timepoint (see Fig.2 and Table 2). Error bars show 95% confidence intervals from 100 bootstrap replicates, sampling across all genes with replacement. Figure 2 shows gene expression in response to pathogen exposure in two species of bdelloid rotifer. Differential gene expression data for a A. vaga (red) and b A. ricciae (blue) at timepoints 7 (T7) and 24 hours (T24) post-inoculation. Each point on the ‘volcano’ plot represents a gene plotted by log2 fold-change in expression level on the X-axis, and significance (expressed as – log10 FDR) on the Y-axis. Positive X-axis values indicate genes that were upregulated in treatment groups (i.e., those exposed to the live pathogen) relative to control groups. Genes with significant changes in expression level (defined as absolute fold-change > 4 and FDR < 1e–3, dashed lines) are shown in colour, with HGTC indicated by darker shading. Genes with non-significant expression changes are shown in grey (>95% of genes are non-significant at these thresholds; see legends for counts). Bar plots show the proportion (%) of HGTCper DE subset: significantly downregulated (left-hand bar), no significant change (middle bar, ‘NS’), and significantly upregulated (right-hand bar). P-values refer to tests of non-association between HGTC and the corresponding up- or downregulated subset (Fisher’s exact tests, see Table 1). Results are robust to the threshold of fold- change used to define the upregulated gene set (1.5-, 2-, 8-, and 16-fold absolute differences yielded the same results as 4-fold) and to reanalysis using three alternative packages to infer differential expression (Figs.9–14; see Methods). Figure 3 shows gene expression in response to desiccation stress in A. vaga. Reanalysis of gene expression data from Hecox-Lea and Mark Welch (2018)47, showing the proportion of HGTCin significantly DE gene subsets when animals are a entering and b recovering from desiccation. Plots arranged as in Fig.2. c Overlap in DE genes between all treatment groups from pathogen (T7 and T24) and desiccation (Entering and Recovery) experiments. Values in bold on the diagonal indicate the number of genes in each category (e.g., 452 genes significantly upregulated at timepoint T7 in pathogen experiment). Values off the diagonal show the proportion of each category that are shared between treatment groups, e.g., of the 1807 genes upregulated during recovery from desiccation, only 228 (12.6%) were also upregulated in the pathogen response at T24. Overall, the proportion of upregulated genes shared between experiments is low (mean = 13%) compared to within experiments (mean = 52%). Values just for HGTCshowed a similar pattern: of the 285 HGTC upregulated during recovery from desiccation, only 33 (~12%, T7) and 63 (~24%, T24) were also upregulated in response to pathogens (see Table 6). Even fewer genes were shared between downregulated subsets (none and ~10% for T7 and T24, respectively). Cells are shaded in proportion to the value of gene sharing (upregulated in green, downregulated in red). Figure 4 shows enriched Gene Ontology (GO) terms for upregulated HGTC. Significantly enriched GO terms relating to the molecular function of HGTCupregulated by a A. ricciae and b A. vaga in response to the pathogen (FDR < 0.001). The area of the rectangle corresponds to the relative magnitude of enrichment of each term; related sub-terms are grouped under a single colour. Terms associated with NRP / PKS functions are highlighted with a yellow border. ‘Catalytic activity’ and ‘oxidoreductase activity’ are associated with NRP / PKS functions too, but these are more generic terms. As a control, we applied the same functional enrichment analysis to HGTC that were differentially expressed (FDR < 0.05) by A. vaga c entering and d recovering from desiccation (Supplementary Data 3). If enriched gene categories reflect a generalised stress response rather than putative adaptations to a fungal attack, or arise from biases in GO or HGTC annotation, we would expect some of the same terms to emerge. We found no enrichment of GO terms relating to NRP / PKS, RNA ligase or glucan-binding (licheninase) functions among upregulated genes in either of the desiccation conditions, even with relaxed thresholds (P < 0.05; P < 0.1), indicating little functional overlap between the pathogen and desiccation responses. Instead, NRP / PKS-associated terms were significantly enriched among genes downregulated by A. vaga when entering or recovering from desiccation (e.g. ‘phosphopantetheine binding’, FDR = 1.45e–5; ‘antibiotic biosynthetic process’, FDR = 0.00289; Supplementary Data 3). In response to desiccation, therefore, rotifers seem more likely to downregulate NRP / PKS genes that had been constitutively expressed in hydrated control animals, rather than upregulating those that were not previously active. Figure 5 shows diversity, structure, genomic location, and expression dynamics of putative NRP / PKS encoded in bdelloid genomes. a Phylogeny of bdelloid NRP / PKS coding sequences (CDS) based on alignment of the predicted condensation domain to bacterial and fungal sequences. CDS with significant upregulation (absolute fold-change > 4 and FDR < 1e–3) are highlighted with a black border and darker shade. Selected bacterial and fungal proteins involved in antimicrobial compound production are shown for reference (EntF, involved in the biosynthesis of enterobactin in Escherichia coli; PksJ, involved in the biosynthesis of the antibiotic polyketide bacillaene in Bacillus subtilis; AcvA, part of the penicillin G biosynthesis pathway in the fungus Aspergillus nidulans). The ‘upregulated clade’ in blue dashed lines comprises a single A. vaga and multiple closely related A. ricciae CDS. The other upregulated A. vaga CDS (marked with an asterisk) groups more closely with bacterial homologs, and may be a more recent acquisition. b Domain arrangement cartoon for two putative NRPS-PKS hybrid clusters upregulated by bdelloids, based on InterProScan analysis of predicted proteins. Multiple partially-assembled A. ricciae CDS have been aligned to the putatively orthologous NRPS-PKS cluster in A. vaga, which is more completely assembled. The condensation domain aligned for the phylogeny is highlighted in yellow. The sequence for ARIC|g51138 was not aligned (double asterisk) as it lacks the focal C domain, which is encoded instead in the adjacent gene model (ARIC|g51137). c Locations of ca.40 putative biosynthetic gene clusters (black ticks) in a haploid, chromosome-scale A. vaga assembly (Av20), with blue-shaded sectors showing subtelomeric regions, and an orange track showing density of HGTC. Clusters corresponding to AVAG|g23567 and AVAG|g48151 CDS are demarked with black and blue arrows respectively. d Expression dynamics for 60 and 36 NRPS / PKS CDS identified in A. vaga and A. ricciae respectively, at timepoints T7 and T24 after pathogen exposure. Grey boxplots show the distribution for all upregulated NRP / PKS CDS (significant or not). The set showing substantially elevated expression at T24 in A. ricciae corresponds to the upregulated clade. Figure 6 shows predicted products of putative NRP / PKS encoded in bdelloid genomes. a The NRPS-PKS hybrid cluster corresponding to gene model AVAG|g48151 is predicted by SeMPI 2.0 to synthesise a cyclic heptadepsipeptide. b Maximum common substructures are shown for the two named compounds with the highest similarity to the predicted product: tyrocidine A, a cyclic decapeptide with potent antifungal activity produced by Bacillus brevis, and hymenamide J, isolated from extracts of the marine sponge Hymeniacidon sp. and its microbiota. Figure 7 shows correlation in DE between timepoints. Each point represents a single gene, and its log2 fold change in expression in treatment groups versus control groups at timepoint T7 (X-axis) versus timepoint T24 (Y-axis) for a A. vaga and b A. ricciae. Positive values represent upregulation in the treatment groups relative to control groups; negative values represent downregulation. Genes with significant DE in both timepoints are shown in red (A. vaga) and blue (A. ricciae); genes significant in one timepoint but not the other are shown with diamond and circle symbols (see legends). Genes with non-significant DE in both timepoints are plotted in grey. Solid black lines show the linear relationship for all upregulated genes (i.e., genes with log2 fold change > 0 in both timepoints; Pearson’s correlation R = 0.84 and 0.82 for A. vaga and A. ricciae, respectively; P < 2e–16 in both cases). Dashed black lines show the linear relationship for downregulated genes (log2fold change < 0 in both timepoints; Pearson’s correlation R = 0.67 and 0.66 for A. vaga and A. ricciae, respectively; P < 2e–16 in both cases). Figure 8 shows correlation in DE between gene copies within and between genomes. Plots show the log2fold change in expression in treatment groups versus control groups for a gene copies within A. vaga at timepoint T7, b gene copies within A. vaga at timepoint T24, c gene copies within A. ricciae at T7, d gene copies within A. ricciae at T24, e gene copies between species at T7, and finally f gene copies between species at T24. Note that each point represents a relationship between a pair of genes, not a gene itself (i.e., putative homologs, homoeologs or paralogs for within-genome comparisons, or orthologs for between-genome comparisons). Solid black lines show the linear relationship for all upregulated genes (i.e., genes with log2fold change > 0 in both copies; Pearson’s correlation R = 0.43, 0.43, 0.49, 0.60, 0.35 and 0.37 for A. vaga and A. ricciae, timepoints T7 and T24 respectively; P < 2e–16 in all cases). Dashed black lines the linear relationship for downregulated genes (log2fold change < 0 at both timepoints; Pearson’s correlation R = 0.50, 0.50, 0.39, 0.42, 0.26 and 0.40 for A. vaga and A. ricciae, timepoints T7 and T24 respectively; P < 2e–16 in all cases). For example, of the 1093 genes that were significantly upregulated in A. vaga at T24, 552 (50.5%) shared an ortholog that was also significantly upregulated in A. ricciae, and the magnitude of DE between these orthologs was significantly correlated (Pearson’s correlation R = 0.62, P < 2e–16). Figure 9 shows effect of varying DE significance parameters on HGTC enrichment, A. vaga. Dashed lines show ± log2fold change equivalent to a 1.5-fold, b 2-fold, c 8-fold, and d 16- fold change in expression value to demark up- and downregulated subsets. FDR threshold < 1e–3 in all cases. P-values in blue show the probability of observing these data given the null hypothesis of no enrichment of HGTC in corresponding subset (Fisher exact tests). Figure 10 shows effect of varying DE significance parameters on HGTC enrichment, A. ricciae. Dashed lines show ± log2 fold change equivalent to a 1.5-fold, b 2-fold, c 8-fold, and d 16- fold change in expression value to demark up- and downregulated subsets. FDR threshold < 1e–3 in all cases. P-values in blue show the probability of observing these data given the null hypothesis of no enrichment of HGTC in corresponding subset (Fisher exact tests). Figure 11 shows DE results using the edgeR package, A. vaga. Plots arranged as above. Figure 12 shows DE results using the voom package, A. vaga. Plots arranged as above. Figure 13 shows DE results using the edgeR package, A. ricciae. Plots arranged as above. Figure 14 shows DE results using the voom package, A. ricciae. Plots arranged as above. Figure 15 shows genomic context of putative NRP / PKS clusters. The 50 kb flanking regions surrounding putative NRP / PKS gene models in both A. vaga and A. ricciae show significantly a fewer genes, b more TEs, and c more telomeric repeats (‘TGTGGG’), compared to BUSCO genes (see also Table 7). Figure 16 shows phylogenetic placement of horizontally acquired genes with putative roles in pathogen resistance in bdelloids. Phylogenies are shown for the RNA ligase domains a 2_5_RNA_ligase2, b RNA_ligase and c RNA_lig_T4_1; the glycosyl hydrolase domains d Glyco_hydro_16 and e Glyco_hydro_64; and f the caspase domain Peptidase_C14. Sequences from A. vaga and A. ricciae shown in red and blue, respectively; genes that were upregulated on exposure to the fungal pathogen are highlighted with a black circle. Non-rotifer sequences were taken from the Pfam seed alignments for each domain, and thus represent the known diversity across the tree of life. Tip symbols indicate taxonomic classifications (see legend). Metazoan sequences are indicated with a blue ‘M’ if present. Figure 17 shows circos plot illustrating locations of transposable elements (TEs), genes, HGTC and putative secondary metabolite synthesis clusters in the Av20 haploid genome assembly. Profiles for chromosomes 1-6 show seven tracks (from inside to outside): biosynthesis clusters (highlighted black ticks), HGTC (red histogram), core metazoan genes (blue histogram), TE-like elements (green histogram), Av13-TE consensus annotations (black histogram), PacBio DNA sequencing coverage (blue line) and Illumina DNA sequencing coverage (red line). Transposable element (TE) repeats described by Simion et al.2021 were predominantly located in subtelomeric regions (green histogram), with distribution confirmed here by mapping compiled A. vaga TE consensus sequences (Flot el al.2013) to the Av20 assembly (black histogram). Gene-rich regions (blue histogram) are mainly located outside of subtelomeric regions. Chromosome ideograms (outer layer) are plotted as color bars (homologous pairs showing similar color), with labels (1-6) showing chromosome numbers from the source genome assembly (GCA_021613535.1). External label ticks are spaced at 2.4 Mb apart. Coverage and histogram layers were calculated with a 100-kb sliding window. Tracks 1-2 and 4-6 are drawn after Simion et al.202144; tracks 3 and 7 are new. Of approximately 40 putative biosynthesis clusters, two appear to be 3’-incomplete and 11 appear to contain frameshifts and / or stop codons; nevertheless, several of these are substantially transcribed. Figure 18 shows expression of foreign genes in control groups. a Proportion of native (light shade) and foreign (dark shade) genes expressed in control groups (i.e., trimmed mean of M- values [TMM] > 0 in any control replicate). Fisher exact test for count data, for A. vaga odds ratio = 1.34 (95% CI = 1.32–1.38), for A. ricciae odds ratio = 1.22 (95% CI = 1.19–1.27); P < 0.001 in both cases. b Level of expression (log TMM) of native and foreign genes expressed in control replicates. Mean (log) TMM Av native = 3.11 ± 5.47 SD; Av HGTC= 2.38 ± 4.54 SD; Welch two sample t-test, t = 80 (49,628 d.f.), P < 0.001; Mean (log) TMM Ar native = 2.99 ± 5.31 SD; Ar HGTC = 2.58 ± 4.78 SD; Welch two sample t-test, t = 60 (31,226 d.f.), P < 0.001. Figure 19 shows gene expression in control replicates of A. vaga in response to drying. a Proportion of native (light shade) and foreign (dark shade) genes expressed in control groups (i.e., trimmed mean of M-values [TMM] > 0 in any control replicate). Fisher exact test for count data, odds ratio = 1.43 (95% CI = 1.38–1.47), P < 0.001. b Level of expression (log TMM) of native and foreign genes expressed in control replicates. Mean (log) TMM Av native = 3.05 ± 5.91 SD; Av foreign = 2.54 ± 5.22 SD; Welch two sample t-test, t = 56 (25,801 d.f.), P < 0.001. Figure 20 shows evidence for antifungal activity of lysate from a rotifer (Adineta ricciae) primed by fungal challenge. A. Petri dishes with lawns of R. globospora fungus and paper test discs inoculated with lysates. B. Examples of dishes showing fungal growth after incubation. C. Examples of microscope images of assay discs treated with water (left) and with disinfectant. After 4 days of fungal growth, the water disc is highly occluded, the disinfectant disc shows a microscopic zone of fungal inhibition and higher light intensity, as analysed by ImageJ. Figure 21. A. Results of ImageJ light intensity analysis of replicate paper discs with different treatments. Higher relative light intensity (arbitrary units) reflects stronger inhibition of the growth of the fungal pathogen Rotiferophthora globospora. Relative to distilled water (negative control), inhibition of fungal growth is highly significant for the antimicrobial positive control standard (20% UniSafe® disinfectant, P < 0.001). Minute (30μL) quantities of lysate from resistant rotifers (A. ricciae) show significantly stronger fungal inhibition than water (P = 0.0251**), and only when the rotifers are challenge-primed by exposure to the pathogen (A_ricc_path). Non- challenge-primed lysate from A. ricciae (A_ricc_ctrl) is statistically indistinguishable from water. Lysate from a susceptible rotifer species (A. vaga) is also indistinguishable from water, but challenge-primed A. vaga lysate shows slightly higher inhibition than the unprimed lysate. These patterns are consistent with an antifungal compound in the lysate that is expressed most strongly by A. ricciae after challenge, and less strongly by A. vaga after challenge, but is not expressed in unprimed rotifers. This matches the expression patterns seen for BGC48151 and its hypothetical antimicrobial product. B. Statistical analysis of results in A. Figure 22 shows a heat map showing proportions of rotifers killed by fungal pathogens 72h after exposure. Lighter squares indicate matches where the rotifer is more resistant to the pathogen; darker squares indicate matches where the rotifer is more susceptible. Trees to the right and top of the plot show the phylogenetic relationships among rotifers and fungi respectively. Figure 23. A. Proportions of bdelloid rotifers from different species succumbing to infection by two fungal pathogen species within 72h of challenge. The first two boxes on the left encompass data that are also presented elsewhere, with respect to A. ricciae and A. vaga versus R. globospora 8995. Other datapoints correspond to newer experiments. B. Statistical analysis indicates resistance of A. steineri to the fungal pathogen R. sp.10274. C. Reads from six transcriptome libraries were mapped to a focal candidate biosynthetic gene cluster extracted from the A. steineri genome (“As_AS23R|g36239”). For each library, we counted the number of reads mapped to the cluster, divided by the total number of reads in the library (‘Mapped reads per million’). We then compared this value between the libraries corresponding to pathogen-challenged A. steineri (“R. sp. 10274”, red datapoints) and UV-inactivated conidia (“CTRL”, blue datapoints), using parametric and nonparametric statistical tests. This reveals a significant (approximately twofold) upregulation of the BGC in the populations exposed to live pathogens. Figure 24. Predicted products of putative NRP / PKS encoded in bdelloid genomes. A. The NRPS-PKS hybrid cluster corresponding to gene model AVAG|g48151 is predicted by SeMPI 2.0 to synthesise a cyclic heptadepsipeptide. B-C. Maximum common substructures are shown for the two named compounds with the highest similarity to the predicted product: tyrocidine A, a cyclic decapeptide with potent antifungal activity produced by Bacillus brevis, and hymenamide J, isolated from extracts of the marine sponge Hymeniacidon sp. and its microbiota. D-E. Maximum common substructures between two named antimicrobial compounds and hypothetical products predicted by SEMPI 2.0 from the A. steineri biosynthetic gene cluster “As_AS23R|g36239”, which is upregulated in response to a fungal challenge. Figure 25. A. Resistance to fungal pathogens bybdelloid rotifers of the genus Habrotrocha: proportions of bdelloid rotifers from different species succumbing to infection by two fungal pathogen species within 72h of challenge. Statistical analysis indicates that none of the three rotifer species shows any susceptibility to R. globospora, but H. ligula shows a nearly eightfold higher resistance (lower mortality) against R. denticulospora as compared with either of the other two host species. B. Reads from six transcriptome libraries were mapped to a focal candidate biosynthetic gene cluster extracted from the H. ligula genome (“HB003|g97330”). For each library, we counted the number of reads mapped to this cluster, divided by the total number of reads in the library (‘Mapped reads per million’). We compared this value between the libraries corresponding to pathogen-challenged H. ligula (‘R. dent 12628’, red datapoints) and UV-inactivated conidia (“CTRL”, blue datapoints), using parametric and nonparametric statistical tests. This reveals a significant (about twofold) upregulation of the BGC in the rotifers exposed to live pathogens. Figure 26. GC content surrounding the upregulated NRPS / PKS cluster corresponding to AVAG|g23567. GC content was calculated in a 20 kb sliding window along (A): contig AVAG00146 (Av13, 36.5 kb), and (B): contig CP075492 (Av20, Chromosome 3, 20.35 Mb). In each case the position of the focal annotation (shown in red) corresponds to a global peak in GC content, as predicted if this cluster had recently been acquired from a genome with a substantially higher baseline GC content than A. vaga. Figure 27. Rotifers of the species A. ricciae and A. vaga were exposed to a standardised inoculum of spores from the basidiomycete pathogen Triacutus subcuticularis and survival was measured as an estimate of differential resistance. The data are compared to survival rates for the ascomycete pathogen Rotiferophthora globospora, demonstrating that A. ricciae shows increased resistance against both the ascomycete and basidiomycete pathogen as compared with A. vaga. Figure 28. Summary of outcomes for 23 rotifer-fungus matches. Resist: grey shading indicates evidence that the rotifers resist the fungal pathogen (N=23; see Figure 22 for further details of resistance in respect of Rotiferophthora); RNA: molecular and bioinformatic evidence has been obtained of the transcriptional basis of the resistance (N=11); BGC: green tick indicates evidence from RNA-seq confirming upregulation of horizontally acquired putatively antimicrobial BGCs (N=11). Figure 29. Differential expression of multiple BGCs in A. Adineta steineri and B. Adineta vaga (clone ‘AD002’) in response to challenge with either R. globospora or R. sp. ARSEF 10274. Table shows results of ANOVA, with P-values adjusted using the Benjamini-Hochberg method to control the false-discovery rate (FDR). Direction and percentage change in expression refer to the most strongly significant pairwise difference between the UV-irradiated control treatment and the spores of the two live pathogens, adjusted using Tukey’s HSD. Rows in red highlight upregulated clusters, blue indicates downregulation. Cluster numbers refer to the position of the most closely orthologous BGC in the Av20 reference genome for A. vaga (‘AD008’). Figure 30. Differential expression of horizontally acquired biosynthetic gene clusters by the bdelloid rotifers A. ricciae (A) and A. vaga (B) in response to the ascomycete fungal pathogen from an earlier Example (R. globospora, first column), and to three further tested fungi from different ascomycete and basidiomycete genera (right columns). Red indicates upregulation, blue indicates downregulation. A black outline indicates that the differential expression is statistically significant after using the Benjamini-Hochberg method to control the false-discovery rate (FDR). Horizontal arrows highlight two clusters uniquely upregulated by each bdelloid species in response to the basidiomycete but not in response to any ascomycete. The vertical arrow in (B) highlights a single ascomycete fungal pathogen of undescribed species and genus (Gen. nov. cf. Harposporium) that induced significant expression of 11 BGCs in a single experiment. Figure 31. SEMPI 2.0 metabolite predictions for two newly identified BGCs in A. ricciae, upregulated in response to attack by the basidiomycete pathogen Triacutus subcuticularis. These predictions are novel compounds but they show highest similarities to surfactins (A) and fengycins (B), with lower similarities to a range of other secondary metabolites with antimicrobial activity. Brief Description of the Sequences All sequences are presented in the accompanying sequence listing compliant with ST.26. SEQ ID NO: 1 - ‘BGC48151’ as sequenced in Adineta vaga, from assembly Av13, upregulated vs e.g. R. globospora (23,409bp) SEQ ID NO: 2 - Ortholog to ‘BGC48151’ as sequenced in Adineta ricciae, from assembly Ar18, upregulated vs. e.g. R. globospora (22,734bp) SEQ ID NO: 4 - BGC 'AVAG|g23567' (corresponding to Cluster 3.4) as sequenced in Adineta vaga, from assembly Av20 (21637 bp) SEQ ID NO: 5 - BGC 'HB003g97330', as sequenced from Habrotrocha ligula, from ‘filtered_AS23R.primary_ctg 2023’, upregulated versus e.g. R. sp. 'ARSEF 10274' (73695 bp) SEQ ID NO: 8 - BGC 'ptg000020l' (corresponding to AD2_6.1) as sequenced in Adineta vaga 'AD002' and upregulated versus e.g. R. sp. 'ARSEF 10274' (14494 bp) SEQ ID NO: 9 - 'BGC 'ARICg51902' as sequenced in Adineta ricciae, and upregulated versus e.g. Triacutus subcuticularis, from assembly ARIC003 (35196 bp) SEQ ID NO: 10 - BGC 'ARICg46825' as sequenced in Adineta ricciae, and upregulated versus e.g. Triacutus subcuticularis, from assembly ARIC003 (11040 bp) SEQ ID NO: 11 - BGC 'AVAG_NRPS_region_3.6' as sequenced in Adineta vaga from assembly Av20 and upregulated versus e.g. H. sp 'ARSEF 14199' (13015 bp) SEQ ID NO: 12 - BGC 'AVAG_NRPS_region_1.8' as sequenced in Adineta vaga from assembly Av20 and upregulated versus e.g. H. sp 'ARSEF 14199' (32749 bp) SEQ ID NO: 13 - BGC 'AVAG_NRPS_region_3.2' as sequenced in Adineta vaga from assembly Av20 and upregulated versus e.g. H. sp 'ARSEF 14199' (24730 bp) SEQ ID NO: 14 - BGC 'AVAG_NRPS_region_5.1' as sequenced in Adineta vaga from assembly Av20 and upregulated versus e.g. H. sp 'ARSEF 14199' (12638 bp) SEQ ID NO: 15 - BGC 'AVAG_NRPS_region_6.2' as sequenced in Adineta vaga from assembly Av20 and upregulated versus e.g. H. sp 'ARSEF 14199' (7175 bp) SEQ ID NO: 16 - BGC 'AVAG_NRPS_SDH_region_1.5' as sequenced in Adineta vaga from assembly Av20 and upregulated versus e.g. T. subcuticularis (3123 bp) SEQ ID NO: 17 - BGC 'AVAG_NRPS_SDH_region_2.3' as sequenced in Adineta vaga from assembly Av20 and upregulated versus e.g. H. sp 'ARSEF 14199' (3050 bp) ‘filtered_AS23R.primary_ctg 2023’ and upregulated versus e.g. R. sp. 'ARSEF 10274' (47370 bp) Detailed Description of the Invention The present invention will be described with respect to particular embodiments and with reference to certain drawings but the disclosure is not limited thereto but only by the claims. Of course, it is to be understood that not necessarily all aspects or advantages may be achieved in accordance with any particular embodiment. Thus, for example those skilled in the art will recognize that the disclosed embodiments may be embodied or carried out in a manner that achieves or optimizes one advantage or group of advantages as taught herein without necessarily achieving other aspects or advantages as may be taught or suggested herein. The disclosure, both as to organization and method of operation, together with features and advantages thereof, may best be understood by reference to the following detailed description when read in conjunction with the accompanying drawings. The aspects and advantages of the invention will be apparent from and elucidated with reference to the embodiment(s) described hereinafter. Reference throughout this specification to "one embodiment" or "an embodiment" means that a particular feature, structure or characteristic described in connection with the embodiment is included in at least one disclosed embodiment. Thus, appearances of the phrases "in one embodiment" or "in an embodiment" in various places throughout this specification are not necessarily all referring to the same embodiment, but may. Similarly, it should be appreciated that in the description of exemplary disclosed embodiments, various features are sometimes grouped together in a single embodiment, figure, or description thereof for the purpose of streamlining the disclosure and aiding in the understanding of one or more of the various inventive aspects. This method of disclosure, however, is not to be interpreted as reflecting an intention that the claimed invention requires more features than are expressly recited in each claim. Rather, as the following claims reflect, inventive aspects lie in less than all features of a single foregoing disclosed embodiment. It should be appreciated that “embodiments” of the disclosure can be specifically combined together unless the context indicates otherwise. The specific combinations of all disclosed embodiments (unless implied otherwise by the context) are further disclosed embodiments of the claimed invention. In addition, as used in this specification and the appended claims, the singular forms “a”, “an”, and “the” include plural referents unless the content clearly dictates otherwise. Thus, for example, reference to “a polynucleotide” includes two or more polynucleotides, reference to “a protein” includes two or more proteins, and the like. "About" as used herein when referring to a measurable value such as an amount, a temporal duration, and the like, is meant to encompass variations of ± 20 % or ± 10 %, more preferably ± 5 %, even more preferably ± 1 %, and still more preferably ± 0.1 % from the specified value, as such variations are appropriate to perform the disclosed methods. “Polynucleotide”, “nucleotide sequence”, “DNA sequence”, or “nucleic acid molecule(s)” as used herein refers to a polymeric form of nucleotides of any length, which comprises ribonucleotides or deoxyribonucleotides. This term refers only to the primary structure of the molecule. Thus, this term includes double- and single-stranded DNA, and RNA. The term “polynucleotide” as used herein, may be a single or double stranded covalently-linked sequence of nucleotides in which the 3' and 5' ends on each nucleotide are joined by phosphodiester bonds. The polynucleotide may be made up of deoxyribonucleotide bases or ribonucleotide bases. Polynucleotides may be manufactured synthetically in vitro or isolated from natural sources. Polynucleotides may further include modified DNA or RNA, for example DNA or RNA that has been methylated, or RNA that has been subject to post-translational modification, for example 5’- capping with 7-methylguanosine, 3’-processing such as cleavage and polyadenylation, and splicing. Polynucleotides may also include synthetic nucleic acids (XNA), such as hexitol nucleic acid (HNA), cyclohexene nucleic acid (CeNA), threose nucleic acid (TNA), glycerol nucleic acid (GNA), locked nucleic acid (LNA) and peptide nucleic acid (PNA). The term “wild-type” refers to a gene or gene product isolated from a naturally occurring source. A wild-type gene is that which is most frequently observed in a population and is thus arbitrarily designed the “normal” or “wild-type” form of the gene. In contrast, the term “modified”, “mutant” or “variant” refers to a gene or gene product that displays modifications in sequence (e.g., substitutions, truncations, or insertions), post-translational modifications and / or functional properties (e.g., altered characteristics) when compared to the wild-type gene or gene product. Sequence identity can also be to a fragment or portion of the full-length polynucleotide or polypeptide. Hence, a sequence may have only 50% overall sequence identity with a full-length reference sequence, but a sequence of a particular region, domain or subunit could share 70%, 80%, 90%, or as much as 99% sequence identity with the reference sequence. Features, integers, characteristics, compounds, chemical moieties or groups described in conjunction with a particular aspect, embodiment or example of the invention are to be understood to be applicable to any other aspect, embodiment or example described herein unless incompatible therewith. All of the features disclosed in this specification (including any accompanying claims, abstract and drawings), and / or all of the steps of any method or process so disclosed, may be combined in any combination, except combinations where at least some of such features and / or steps are mutually exclusive. The invention is not restricted to the details of any foregoing embodiments. The invention extends to any single feature, or any novel combination, of the features disclosed in this specification (including any accompanying claims, abstract and drawings), or to any novel single feature, or any novel combination, of the steps of any method or process so disclosed. The contents of any document cited herein is incorporated by reference. Method Provided herein is a novel method for producing and isolating a compound from a microbe- resistant species of rotifer. This is achieved through the provision herein of a method comprising providing a microbe-resistant species of rotifer in culture, exposing the rotifer to a microbe, and isolating a compound produced by the rotifer following exposure to the microbe. The method may advantageously lead to the generation of novel antimicrobial compounds suitable for clinical or agricultural applications. The method herein comprises the use of a microbe-resistant rotifer. The rotifer may be any rotifer, preferably wherein the rotifer has acquired genetic material via horizontal gene transfer (HGT), and most preferably wherein the genetic material acquired has been horizontally transferred from one or more non-metazoan taxa. The rotifer is preferably a bdelloid rotifer, i.e. a rotifer comprised in the Bdelloidea class of rotifer. The bdelloid rotifer may be any species of microbe- resistant bdelloid rotifer, preferably wherein the species is Adineta ricciae, Adineta vaga, Adineta steineri, Habrotrocha ligula, Habrotrocha constricta, Habrotrocha elusa or Habrotrocha bidens. Preferably the bdelloid rotifer is Adineta ricciae, Adineta vaga, Adineta steineri or Habrotrocha ligula, and most preferably the bdelloid rotifer is Adineta ricciae. Methods of culturing rotifer species are known in the art. Preferably rotifer species are cultured in water, e.g. sterilised distilled water. Preferably the water is replaced on a regular basis with fresh water. Rotifer species may be preferably cultured in an atmosphere having a temperature of at least about 15oC and no more than about 23oC, although most preferably the atmosphere has a temperature of about 20oC. Rotifer species may be preferably cultured in normal atmospheric CO2. Methods of culturing rotifer species may comprise feeding the rotifer species with a microbe that is not toxic to the rotifer, such as Escherichia coli and / or Saccharomyces cerevisiae. Methods of culturing rotifer species may also comprise feeding the rotifer species with a dilute lettuce extract. The method herein may comprise the use of a rotifer that is resistant to any type of microbe. The microbe is preferably a fungus or bacterial microorganism, wherein most preferably the microbe is a fungus. The same microbe, preferably a fungus, may be the microbe to which the rotifer is exposed in the method herein. The fungus is preferably a fungus of the Ascomycota or Basidiomycota phylum, and more preferably a fungus of the Hypocreales. More preferably, when the fungus is of the Ascomycota phylum, the fungus is of the Rotiferophthora genus or Harposporium genus. More preferably, when the fungus is of the Basidiomycota phylum, the fungus is of the Triacutus genus. The fungus may preferably be a conidium. The fungus species may be Rotiferophthora globospora, Rotiferophthora sp. ‘ARSEF 10274’, Rotiferophthora denticulospora, Rotiferophthora lacrima, Rotiferophthora minutispora, Rotiferophthora tagenophora, Rotiferophthora barronii, Rotiferophthora amamiensis, Rotiferophthora cylindrospora, Rotiferophthora japonica, Rotiferophthora angustispora or Rotiferophthora brevipes. Preferably the fungus is Rotiferophthora globospora, Rotiferophthora denticulospora or Rotiferophthora sp. ‘ARSEF 10274’, and most preferably the fungus is Rotiferophthora globospora. The fungus species may preferably be Harposporium spirosporum or Harposporium sp. ‘ARSEF 14199’. The fungus species may preferably be Triacutus subcuticularis. The rotifer used in the method of the invention may be resistant to any microbe, preferably wherein the microbe is a microbe described herein. Microbe-resistance comprises the ability an organism to resist the effects of infection of the organism with a microbe and / or the toxic effects of exposure of the organism to a microbe. Microbe-resistance preferably comprises the ability to the organism to survive following exposure to a microbe. Methods for determining microbe resistance in an organism, for example in rotifers, are known in the art. Preferably, wherein the organism is a rotifer, the exposing of the rotifer to a microbe comprises contacting the rotifer with the microbe, and more preferably wherein a rotifer culture is inoculated with the microbe. The methods of determining microbe-resistance may preferably for example comprise culturing one or more rotifers, exposing the one or more rotifers to a microbe, and determining the proportion of the one or more rotifers that survive the exposure. More preferably, a rotifer culture is inoculated with a conidial suspension at a density of about 125 spores uL-1and mean infection mortality is determined after at least 48 hours, more preferably at least 72 hours. The method described herein comprises isolating a compound produced by the rotifer following exposure to a microbe, preferably wherein the compound is a secondary metabolite. As will be appreciated by the skilled person, a secondary metabolite is an organic compound produced by an organism that - whilst not directly involved in the growth of said organism – may provide the organism with a selective advantage in its ecosystem due to its effects on other organisms present in said ecosystem. Methods for determining whether a compound has been produced by an organism in culture following exposure to a form of challenge are known in the art. Those skilled in the art will therefore appreciate how it can be determined whether exposure of a rotifer in culture to a microbe leads to the production of compounds. Typically, identification of a compound not present prior to exposure with the microbe, yet present after exposure with microbe, indicates that the compound has been produced by the rotifer following exposure to the microbe. As a person skilled in the art will appreciate, control assays can be used to validate whether said compound is produced by the rotifer following exposure to the microbe. Means for isolating a compound from bioactive mixtures such as a culture or culture extract are known in the art. Preferably the rotifer is lysed prior to the isolation of the compound, and wherein the compound is isolated from the rotifer lysate. Lysis may be performed by any suitable means, for example sonication. The isolation preferably comprises one or more of the steps: (i) pelleting the rotifer following exposure to the microbe, (ii) sonication of the rotifers to yield a lysate, (iii) concentration of the lysate, for example at least fourfold using a vacuum concentrator, and (iv) filtration of the lysate using a 0.2 µm syringe filter. The isolation may further comprise fractionation by any suitable means known in the art. For example, the fractionation may comprise one or more of the steps: (i) solvent extraction (water, ethanol, methanol, chloroform, dichloromethane, diethylether, acetone), (ii) normal- and reverse- phase liquid chromatography, (iii) anion- and cation-exchange chromatography, and (iv) high- performance liquid chromatography with identification by mass spectrometry characterisation of bioactive fractions in comparison to equivalently fractionated lysate from non-challenged rotifers. In the method provided herein, the genome of the rotifer may comprise genetic material acquired via horizontal gene transfer (HGT) from non-metazoan taxa, preferably wherein said genetic material has increased expression following the exposure of the rotifer to the microbe. The skilled person would understand how gene expression can be determined through use of known techniques in the art, e.g. by unbiased or targeted assays. Increased expression of said genetic material may refer to increased transcription of open reading frames comprised in said genetic material. Means for determining whether genetic material has been horizontally acquired are known to the skilled person. The horizontally acquired genetic material may comprise portions of genes, intact genes, and gene clusters. Genetic material acquired by horizontal gene transfer may be determined using one or more of the steps (i) protein sequence alignment to the UniProt90 sequence database and determining alignment to Metazoan and non-Metazoan sequence, and (ii) protein sequence searched against translated nucleotide sequences (tBLASTn) alignment to a known rotifer genome assembly. Preferably, the genetic material acquired via horizontal gene transfer (HGT) from non-metazoan taxa that has increased expression following the exposure of the rotifer to the microbe comprises genetic sequences encoding non-ribosomal peptide synthetases, polyketide synthetases, hybrid non-ribosomal peptide and polyketide synthetases, wherein the non-ribosomal peptide synthetases, polyketide synthetases, hybrid non-ribosomal peptide and polyketide synthetases may produce antibiotics. The genetic material may preferably comprise any portion of, or all of, a sequence defined according to SEQ ID NO 1, SEQ ID NO 2 or SEQ ID NO 3, or any one of the sequences defined according to SEQ ID NOs 1 to SEQ ID NO: 19. The portion of said sequence may be at least about 1%, at least about 2%, at least about 3%, at least about 4%, at least about 5%, at least about 6%, at least about 7%, at least about 8%, at least about 8%, at least about 9%, at least about 10%, at least about 20%, at least about 30%, at least about 40%, at least about 50%, at least about 60%, at least about 70%, at least about 80%, or at least about 90% of the sequence defined according to SEQ ID NO 1, SEQ ID NO 2 or SEQ ID NO 3, or any one of the sequences defined according to SEQ ID NOs 1 to SEQ ID NO: 19. The genetic material may share at least about 50%, at least about 60%, at least about 70%, at least about 80%, at least about 90%, at least about 95%, at least about 96%, at least about 97%, at least about 98%, at least about 99%, or 100% sequence identity with a portion of the sequence defined according to SEQ ID NO 1, SEQ ID NO 2 or SEQ ID NO 3, or any one of the sequences defined according to SEQ ID NOs 1 to SEQ ID NO: 19. The genetic material may preferably comprise a sequence of at least 1500bp nucleotides in length and wherein the 1500bp sequence shares at least 70% sequence identity with any 1500bp nucleotide sequence of SEQ ID NO 1, SEQ ID NO 2 or SEQ ID NO 3, or any one of the sequences defined according to SEQ ID NOs 1 to SEQ ID NO: 19, preferably wherein the 1500bp shares at least 70% sequence identity with any 1500bp nucleotide sequence of SEQ ID NO 1 or SEQ ID NO 2, and most preferably with any 1500bp nucleotide sequence of SEQ ID NO 1. The genetic material may preferably share at least about 70% sequence identity with the entire nucleotide sequence defined according to SEQ ID NO 1, SEQ ID NO 2 or SEQ ID NO 3, or any one of the sequences defined according to SEQ ID NOs 1 to SEQ ID NO: 19, more preferably according to SEQ ID NO 1 or SEQ ID NO 2, and most preferably according to SEQ ID NO 1. The genetic material may preferably share at least about 80%, at least about 90%, at least about 95%, at least about 96%, at least about 97% with the entire nucleotide sequence defined according to SEQ ID NO 1, SEQ ID NO 2 or SEQ ID NO 3, or any one of the sequences defined according to SEQ ID NOs 1 to SEQ ID NO: 19, more preferably according to SEQ ID NO 1 or SEQ ID NO 2, and most preferably according to SEQ ID NO 1. Method of assaying anti-microbial activity Provided herein is a method comprising assaying anti-microbial activity of a compound isolated according to the method of the invention described herein. The method of assaying for anti-microbial activity may therefore comprise assaying the activity of a compound, wherein the compound is obtained through a method comprising providing a microbe-resistant species of rotifer in culture, exposing the rotifer to a microbe, and isolating a compound produced by the rotifer following exposure to the microbe. The anti-microbial activity may be against any microbe, preferably wherein the microbe is described herein, and most preferably wherein the microbe is a fungus. The fungus may for example be Rotiferophthora globospora, Rotiferophthora denticulospora Rotiferophthora sp. ‘ARSEF 10274’, Rotiferophthora lacrima, Rotiferophthora minutispora, Rotiferophthora tagenophora, Rotiferophthora barronii, Rotiferophthora anamiensis, Rotiferophthora cylindrospora, Rotiferophthora japonica, Rotiferophthora angustispora or Rotiferophthora brevipes, preferably wherein the fungus is Rotiferophthora globospora, Rotiferophthora denticulospora or Rotiferophthora sp. ‘ARSEF 10274’, and most preferably the fungus is Rotiferophthora globospora. Anti-microbial activity of the compound can be determined by any known methods for determining anti-microbial activity. Preferably, the assay comprises culturing the microbe and exposing the microbe to the compound, and subsequently assessing the viability of the microbe. In the event that that the compound leads to decreased rate of survival in the microbe, the compound may be deemed to possess anti-microbial activity. Compound Provided herein is a compound produced and isolated by the method according to the method of the invention described herein. The compound of the invention may therefore be obtained through a method comprising providing a microbe-resistant species of rotifer in culture, exposing the rotifer to a microbe, and isolating a compound produced by the rotifer following exposure to the microbe. The compound may preferably be produced by virtue of the increased expression of genetic material acquired via horizontal gene transfer (HGT) from non-metazoan taxa when a microbe-resistant rotifer comprising said genetic material is exposed to a microbe. The compound preferably has anti-microbial activity, and most preferably wherein the anti-microbial activity is anti-fungal activity. The compounds may be produced by the activity of one or more non-ribosomal peptide synthetases and / or polyketide synthetases and / or hybrid non-ribosomal peptide and polyketide synthetases. The compound may be a cyclic compound similar to known antibiotics, and is preferably a cyclic heptadepsipeptide. The compound may be an analog of, consist of or comprise a compound defined according to Fig 24A-E. Method of treatment Provided herein is a method of treating a medical disorder wherein the method comprises administering the compound of the invention to a subject. The compound administered to the subject may therefore be a compound obtained through a method comprising providing a microbe- resistant species of rotifer in culture, exposing the rotifer to a microbe, and isolating a compound produced by the rotifer following exposure to the microbe. The compound administered to the subject may therefore also be a compound produced by a host cell according to the invention. Those of skill in the art will appreciate that the method of treatment disclosed herein has many different clinical applications as a potential treatment of various medical disorders, however preferably the medical disorder is microbial infection, and most preferably the microbial infection is a fungal infection. The compound may be administered to a subject wherein the compound is a composition formulated for any suitable administration route such as oral, topical or injectable. The compound may be formulated for administration with a pharmaceutically acceptable carrier or diluent. For example, solid oral forms may contain, together with the active compound, diluents, e.g. lactose, dextrose, saccharose, cellulose, corn starch or potato starch; lubricants, e.g. silica, talc, stearic acid, magnesium or calcium stearate, and / or polyethylene glycols; binding agents; e.g. starches, arabic gums, gelatin, methylcellulose, carboxymethylcellulose or polyvinyl pyrrolidone; disaggregating agents, e.g. starch, alginic acid, alginates or sodium starch glycolate; effervescing mixtures; dyestuffs; sweeteners; wetting agents, such as lecithin, polysorbates, laurylsulphates; and, in general, non-toxic and pharmacologically inactive substances used in pharmaceutical formulations. Such pharmaceutical preparations may be manufactured in known manner, for example, by means of mixing, granulating, tableting, sugar coating, or film coating processes. As explained herein, the compounds and compositions provided herein are useful in treating or preventing various medical disorders. The present invention therefore provides a compounds and compositions as provided herein for use in medicine, preferably wherein the compositions and compounds herein are for use in the treatment of microbial infection, preferably wherein the microbial infection is a fungal infection. The invention also provides the use of compounds and compositions as provided herein in the manufacture of a medicament, wherein the medicament may be for the treatment of microbial infection, and preferably wherein the microbial infection is a fungal infection. Also provided is a method of treating a subject in need of such treatment, said method comprising administering to the subject a compound or composition provided herein. In some embodiments the subject suffers from or is at risk of suffering from one of the disorders disclosed herein, preferably wherein the disorder is a microbial infection, and preferably wherein the microbial infection is a fungal infection. A subject is typically an animal subject, and is preferably a human patient. The patient may be male or female. The age of the patient is typically at least 18 years, for instance from 30 to 70 years or from 40 to 60 years. The animal may be in any animal in a veterinary setting, for example a dog, cat, horse, or a livestock animal, for example a cow, pig, chicken, sheep, lamb, turkey or goat. A therapeutically or prophylactically effective amount of the compound or composition may be administered to a subject. The dose may be determined according to various parameters, especially according to the compound used; the age, weight and condition of the subject to be treated; the route of administration; and the required regimen. Again, a physician will be able to determine the required route of administration and dosage for any particular subject. Polynucleotide, vector and host cell Provided herein is a polynucleotide comprising one or more genes which are comprised within the genome of a microbe-resistant rotifer and have been acquired by the rotifer via HGT from non-metazoan taxa, preferably wherein said one or more genes have increased expression upon exposure of the rotifer to a microbe, preferably wherein the microbe is a fungus. The one or more genes may be defined according to the definition provided herein of the genetic material acquired via horizontal gene transfer (HGT) from non-metazoan taxa that has increased expression following the exposure of the rotifer to the microbe, wherein for example the one or more genes comprise any portion of, or all of, a sequence defined according to SEQ ID NO 1, SEQ ID NO 2 or SEQ ID NO 3, or any one of the sequences defined according to SEQ ID NOs 1 to SEQ ID NO: 19; and optionally wherein the portion of said sequence may be at least about 1%, at least about 2%, at least about 3%, at least about 4%, at least about 5%, at least about 6%, at least about 7%, at least about 8%, at least about 8%, at least about 9%, at least about 10%, at least about 20%, at least about 30%, at least about 40%, at least about 50%, at least about 60%, at least about 70%, at least about 80%, or at least about 90% of the sequence defined according to SEQ ID NO 1, SEQ ID NO 2 or SEQ ID NO 3, or any one of the sequences defined according to SEQ ID NOs 1 to SEQ ID NO: 19; and further optionally wherein said one or more genes may share at least about 50%, at least about 60%, at least about 70%, at least about 80%, at least about 90%, at least about 95%, at least about 96%, at least about 97%, at least about 98%, at least about 99%, or 100% sequence identity with a portion of the sequence defined according to SEQ ID NO 1, SEQ ID NO 2 or SEQ ID NO 3, or any one of the sequences defined according to SEQ ID NOs 1 to SEQ ID NO: 19. The one or more genes preferably comprise a sequence of at least 1500bp nucleotides in length and wherein the 1500bp sequence shares at least 70% sequence identity with any 1500bp nucleotide sequence of SEQ ID NO 1, SEQ ID NO 2 or SEQ ID NO 3, or any one of the sequences defined according to SEQ ID NOs 1 to SEQ ID NO: 19, preferably wherein the 1500bp shares at least 70% sequence identity with any 1500bp nucleotide sequence of SEQ ID NO 1 or SEQ ID NO 2, and most preferably with any 1500bp nucleotide sequence of SEQ ID NO 1. The one or more genes preferably share at least about 70% sequence identity with the entire nucleotide sequence defined according to SEQ ID NO 1, SEQ ID NO 2 or SEQ ID NO 3, or any one of the sequences defined according to SEQ ID NOs 1 to SEQ ID NO: 19, more preferably according to SEQ ID NO 1 or SEQ ID NO 2, and most preferably according to SEQ ID NO 1. The one or more genes preferably share at least about 80%, at least about 90%, at least about 95%, at least about 96%, at least about 97% with the entire nucleotide sequence defined according to SEQ ID NO 1, SEQ ID NO 2 or SEQ ID NO 3, or any one of the sequences defined according to SEQ ID NOs 1 to SEQ ID NO: 19, more preferably according to SEQ ID NO 1 or SEQ ID NO 2, and most preferably according to SEQ ID NO 1. Preferably the one or more genes encode one or more non-ribosomal peptide synthetases, polyketide synthetases, hybrid non-ribosomal peptide or polyketide synthetases. Provided herein is a vector comprising the polynucleotide of the invention. Such expression vectors are routinely constructed in the art of molecular biology and may for example involve the use of plasmid DNA and appropriate initiators, promoters, enhancers and other elements, such as for example polyadenylation signals which may be necessary, and which are positioned in the correct orientation, in order to allow for expression of a peptide of the invention. The vector may also encode a further fluorescent protein aside from any fluorescent protein expressed in frame with the water-forming oxidase, wherein said further fluorescent protein is separated from the nucleotide sequence of the polynucleotide of the invention by an internal ribosomal entry site (IRES) or under the control of an additional promoter. The vector preferably comprises an origin of replication, and a T-DNA right border repeat of a Ti or Ri plasmid, optionally further comprising a left border repeat of a Ti or Ri plasmid, and at least one bacterial selectable marker. The vector may comprise one or more of: i. an enhancer; ii. a plant selectable marker; iii. a multicloning site; and iv. a recombination site. Suitable vectors for this purpose are well known to those of skill in the art and include, without limitation, plasmids, cosmids, vectors, artificial chromosomes, modified viruses, and mobile genetic elements. Other suitable vectors would be apparent to persons skilled in the art. By way of further example in this regard we refer to Sambrook et al. (Sambrook, J., Fritsch, E. R., & Maniatis, T. (1989). Molecular Cloning: A Laboratory Manual. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press). The invention also includes cells that have been modified to express a polypeptide of the invention. Host cells may comprise the polynucleotide of the invention and / or the vector of the invention. Such host cells typically include prokaryotic cells such as bacteria, for example E. coli, or mammalian cells. Such bacteria or cells may be cultured using routine methods to produce a compound when exposed to a microbe in accordance with the method of the invention. Method of controlling microbial disease Provided herein is a method of controlling microbial disease in a plant, the method comprising contacting the plant with a compound according to the invention. Preferably the microbial disease is a fungal disease. The compound contacted with the plant may therefore be a compound obtained through a method comprising providing a microbe-resistant species of rotifer in culture, exposing the rotifer to a microbe, and isolating a compound produced by the rotifer following exposure to the microbe. The compound contacted with the subject may therefore also be a compound produced by a host cell according to the invention. Those of skill in the art will appreciate that the method of controlling microbial disease disclosed herein may be applied to any form of plant in order to control any form of microbial disease. The microbe may be any microbe known in the art to be pathogenic towards a plant. The microbial disease is preferably a fungal disease. The method of controlling a fungal disease disclosed herein may be applied to any form of plant in order to control any form of fungal disease. The fungus may be any fungus known in the art to be pathogenic towards a plant. The fungal disease may preferably comprise any one or more of anthracnose, leaf spot, rust, wilt, blight, coils, scab, gall, canker, damping-off, root rot, mildew, and dieback. Those of skill in the art will appreciate that “controlling” refers to any effect on a microbe that results in limiting the microbial disease that it causes in a plant, preferably wherein the microbe is a fungus and preferably wherein the microbial disease is a fungal disease. Assays for measuring the controlling of a microbial disease in a plant are known in the art, preferably wherein the assays are for measuring the controlling of a fungal disease in a plant. The microbe is preferably in contact with the plant and the controlling may preferably comprise killing the microbe or inhibiting the growth of the microbe, preferably wherein the microbe is a fungus and the controlling may preferably comprise killing the fungus or inhibiting the growth of the fungus. The compound may therefore control fungal disease through its fungicidal activity. The plant may preferably be any plant susceptible to microbial disease, and is preferably susceptible to fungal disease. The plant is preferably a crop plant. The plant may preferably be a monocot plant and may, for example, be selected from the families Arecaceae, Amaryllidaceae or Poaceae. For example, the plant may be a cereal crop, such as wheat, rice, barley, oat, triticale, rye, buckwheat, or a non-cereal monocot crop such as garlic, onion, leek, yam, oil palm, or banana. The plant may be a dicot plant and may, for example, be selected from the families Asteraceae, Brassicaceae (e.g. Brassica napus), Chenopodiaceae, Cucurbitaceae, Leguminosae (Caesalpiniaceae, Aesalpiniaceae Mimosaceae, Papilionaceae or Fabaceae), Malvaceae, Rosaceae or Solanaceae. For example, the plant may be selected from lettuce, sunflower, broccoli, spinach, water melon, squash, cabbage, tomato, potato, sweet potato, capsicum, tobacco, cotton, okra, apple, rose, strawberry, alfalfa, bean, soybean, field (fava) bean, pea, lentil, peanut, chickpea, apricots, pears, peach, grape vine, bell pepper, chilli, citrus or coffee species. The plant may be a biofuel or bioenergy crop such as rape / canola, linseed, lupin and willow, poplar, poplar hybrids, or gymnosperms, such as loblolly pine. Also included are crops for silage, grazing or fodder (grasses, clover, sanfoin, alfalfa), fibres (e.g. hemp, cotton, flax), building materials (e.g. pine, oak, teak, rosewood), pulping (e.g. poplar), feeder stocks for the chemical industry (e.g. high erucic acid oil seed rape, linseed, jute, oil palm). The compound contacted with the plant is preferably in an amount sufficient to control microbial disease, and preferably in an amount sufficient to kill or inhibit the growth of the microbe that is in contact with the plant. Preferably the compound contacted with the plant is preferably in an amount sufficient to control fungal disease, and preferably in an amount sufficient to kill or inhibit the growth of the fungus that is in contact with the plant. The compound may be contacted with the plant in any suitable form, such as in the form of a spray or powder. EXAMPLES The present invention will now be described with reference to specific Examples, which should not be construed as in any way limiting. Example 1 Abstract Coevolutionary antagonism generates relentless selection that can favour genetic exchange, including transfer of antibiotic synthesis and resistance genes among bacteria, and sexual recombination of disease resistance alleles in eukaryotes. We report an unusual link between biological conflict and DNA transfer in bdelloid rotifers, microscopic animals whose genomes show elevated levels of horizontal gene transfer from non-metazoan taxa. When rotifers were challenged with a fungal pathogen, horizontally acquired genes were over twice as likely to be upregulated — a far stronger enrichment than observed under abiotic stress. Among hundreds of differentially expressed non-metazoan genes, the most markedly overrepresented were clusters resembling non- ribosomal peptide synthetases, polyketide synthetases, hybrid non-ribosomal peptide and polyketide synthetases that produce antibiotics. Upregulation of these clusters in a pathogen-resistant rotifer species was nearly ten times stronger than in a susceptible species. By acquiring, domesticating, and expressing non-metazoan biosynthetic pathways, bdelloids may have evolved to resist natural enemies using antimicrobial mechanisms absent from other animals. Introduction Antagonistic interactions among species are strong, ubiquitous and relentless sources of selection in natural populations1–3. Examples include arms-races between pathogen virulence factors and host immune systems, and coevolution between antimicrobial compounds or toxins and pathways to resist them. These dynamics are central to various global challenges, including emerging infectious diseases, management of crop pathogens, and antimicrobial resistance. According to theory, selection arising from antagonistic coevolution can favour adaptations to shuffle existing combinations of genes or to acquire genes bringing new functions. These processes help explain the especially rapid and intense adaptive evolution4seen at loci encoding the molecular mediators of conflict5. Different domains of life typically address the challenge of ongoing genetic mixing in distinct ways. In bacteria and archaea, horizontal gene transfer (HGT) occurs by various mechanisms6,7and is well known as a route for the spread of antimicrobial resistance8, as well as the genes encoding antibiotic production themselves. For example, non-ribosomal peptide synthetases (NRPS) and polyketide synthetases (PKS) are large, multimodular enzymes that catalyse assembly of a range of natural products, including antimicrobial compounds. These can be encoded as biosynthetic gene clusters on plasmids9as well as chromosomes, and their mobility and modular structure facilitates diversification to produce a vast array of secondary metabolites via recombination within and between genomes10. The natural ‘reservoir’ of mobile genetic diversity for antibiotic synthesis and resistance is thought to reflect a history of coevolution between the producers and targets of antimicrobial compounds11. In contrast, among eukaryotes, the most important mechanism of genetic exchange is meiotic sex, by which whole genomes are shuffled every generation though recombination, segregation and outcrossing. Although meiotic shuffling has different effects to HGT, sex too has been linked to biotic conflict because it can speed up host adaptation against pathogens by generating new combinations of resistance alleles12,13. When coevolving pathogens are common and virulent14, the theoretical benefits of genetic exchange are so substantial that they may outweigh the inherent costs15of sexual reproduction compared with parthenogenesis16–18. Antagonistic coevolution may therefore help explain why obligately asexual plant and animal lineages are typically rare and short-lived despite major advantages19,20. This so-called ‘Red Queen Hypothesis’3(RQH) draws support from associations between recombination and immunity, from host-pathogen dynamics in mixed sexual and asexual populations21,22, and from the susceptibility of asexually propagated lineages to pathogens23. Here, we investigate the links between genetic transfer and biotic conflict in a group of animals that challenge typical distinctions between domains described above. Bdelloid rotifers are a class of microscopic invertebrates that live in freshwater and limno-terrestrial habitats worldwide. Reproduction is only known by parthenogenetic eggs and neither males nor sperm have been reported despite centuries of microscopic observation and the description of hundreds of species24, leading to the hypothesis that the class Bdelloidea has diversified for tens of millions of years in the absence of sexual reproduction25,26. Genetic evidence to either confirm or refute obligate asexuality in bdelloids has proved complicated, and seemingly definitive evidence both for and against this hypothesis27–31has been overturned or reinterpreted by later work32–38. In contrast, repeated studies have demonstrated that bdelloid genomes encode extraordinarily high proportions of genes acquired horizontally from non-metazoan taxa. Approximately 10% of genes appear to have been captured from bacteria, fungi, plants and other sources, rather than sharing recent common ancestry with metazoan orthologs29,39,40. This estimate is an order of magnitude greater than for other animals41, holds for all bdelloid genomes so far examined, and is consistent across various methods for detecting HGT33,42–44. Comparisons among bdelloid species indicate that most HGT events are ancient, with ongoing acquisition rates estimated to be on the order of one gene per 100,000 years45. At these rates, the phenomenon would be too slow to equate with the sexual shuffling seen in typical eukaryotes, or the rapid dynamics of bacterial accessory genomes. Nevertheless, HGT has been hypothesised to introduce novel biochemical functions that help bdelloids adapt to environmental challenges, as it does in bacteria. Acquired genes are expressed and incorporated into metabolic pathways40,42, some of which are not shared by other metazoans46. Putative functions identified to date include desiccation tolerance, nutrient exploitation and repair of DNA damage40,42,47. However, the deep associations between genetic transfer and coevolution raise the hypothesis that HGT may particularly help rotifers deal with biotic antagonism; for instance by acquiring genes with pathogen resistance functions. Isolated examples of horizontally acquired genes contributing to immunity are known from invertebrates48–50, but the massive scale of HGT in bdelloids and the prevalence of asexual reproduction might especially favour the co-option of unusual pathways to resist microbial enemies, with parallels to the role of HGT in bacterial conflict51,52. If so, this could compensate in part for the challenge that an asexual lineage theoretically faces from pathogens. We investigated this hypothesis by testing whether horizontally acquired genes are disproportionately involved in the response of bdelloid rotifers to infection. Like all animals, bdelloid rotifers are exploited by a range of natural enemies, including over 60 species of virulent fungal and oomycete pathogens53. These can exterminate cultured populations in a few weeks54,55, and significantly depress the abundance of hosts in natural habitats56. However, almost nothing is known about variation in susceptibility among rotifers, how this compares with observations in other invertebrate pathosystems57–62, whether its genetic basis matches models of coevolution12,63, or how the underlying mechanisms evolve and remain effective if sex is rare or absent. We used RNA-seq to identify genes that are differentially expressed when bdelloid rotifers are attacked by a natural fungal pathogen in the genus Rotiferophthora64(Clavicipitaceae, Hypocreales), which preys specifically upon them (Fig.1a). We assessed variation by comparing two host species, Adineta ricciae and A. vaga, which differ by a factor of four in resistance to the pathogen (Fig.1b). We compared the scale and speed of the transcriptomic response and asked whether horizontally transferred genes were especially likely to be differentially expressed compared to regular metazoan genes in each species. We compared our results with RNA-seq data obtained when bdelloids were exposed to desiccation47, to test whether genes of non-metazoan origin contribute disproportionately when responding to a biotic as opposed to an abiotic stressor. Functional enrichment analysis identified the most strongly upregulated classes of horizontally transferred genes, and revealed groups whose expression profiles and genomic representations differed between the resistant and susceptible rotifer species. Susceptibility to a fungal pathogen differs markedly between two Adineta species We inoculated clonal populations of the bdelloid rotifers A. ricciae and A. vaga with spores of the fungal pathogen Rotiferophthora globospora64and monitored them over 3–4 days. In both species, >95% of animals ingested spores and contracted within minutes of exposure (Fig.1a). The proportion of animals killed by fungal infection at 72h differed markedly between the species: 18% for A. ricciae and 71% for A. vaga (Fig.1b, relative risk = 3.74, 95% CI: 2.8–5.0, z = 8.758, P < 0.001). Given that these species are morphologically and ecologically33,65similar, were reared clonally and exposed under standardised conditions, the difference in mortality seems to indicate variation in the physiological response and genetic resistance of each host to the fungus. Rapid and large-scale transcriptional response to pathogen attack in both species To investigate the genetic basis of the response to pathogens, we harvested total RNA from exposed and control populations of each species in triplicate at 7 and 24 hours after initiation of fungal attack and quantified gene expression across species and timepoints, defining significant differential expression (DE) as an absolute fold-change >4 and a False Discovery Rate-adjusted P- value < 0.001. Even by these stringent criteria, a rapid and pronounced transcriptional response was detected in both species. The greater magnitude of upregulation in A. ricciae and the larger number of responding genes suggest a stronger and broader defensive response than A. vaga (Fig.1c). For both species, the direction and magnitude of DE at 7h post-inoculation (timepoint T7) correlated significantly with that seen at 24h post-inoculation (timepoint T24; Pearson’s correlation R = 0.66– 0.84; P < 2e–16 in all cases; Fig.7), indicating that many of the transcriptional changes initiated at T7 were ongoing and consistent at T24, even as new genes joined the response. Because our transcripts were mapped to annotated reference genomes for each species (denoted ‘Av13’ and ‘Ar18’ for A. vaga and A. ricciae, respectively29,33), we were also able to analyse orthology. The two species showed substantially overlapping profiles of differentially expressed genes, as expected if defensive or recognition pathways are conserved (Fig.8; Table 3). Both species therefore mounted rapid, extensive, and partially overlapping transcriptional responses to the pathogen, beginning less than 7 hours after exposure. Twofold overrepresentation of non-metazoan genes in the response to pathogens Based on the predicted proteins in each reference genome, the background frequency of high-confidence HGT candidate genes (denoted ‘HGTC’) is ~11.5% for both species44. Among the significantly upregulated subset, however, 23–32% were HGTC, depending on species and timepoint (Table 1). This two- to three-fold enrichment for HGTC was highly significant (binomial GLM, estimate = 1.18, SE = 0.07, P < 0.0001; Fig.2; Table 4) in both species at both timepoints. The proportional enrichment was significantly stronger for the early response (estimate = –0.22, SE = 0.08, P = 0.0048), though the absolute numbers of upregulated HGTC genes were higher at T24 (Fig.2 and Table 1; Table 4). We observed a similar enrichment of HGTC among significantly downregulated genes at both experimental timepoints (estimate = 1.09, SE = 0.12, P < 0.0001; Fig. 2 and Table 1; Table 4), which again did not differ between the two species (see Supplementary Note to Table 8). Horizontally acquired genes were therefore markedly over-represented among the genes that are differentially expressed in the pathogen treatments in both species, accounting for about 33% and 25% of the most strongly responding genes at T7 and T24, respectively. Enrichment of non-metazoan genes is stronger when responding to pathogens than to desiccation We considered whether the apparent enrichment of HGTC was a feature of the response to fungal attack specifically, or might result from more generic properties of such genes and their expression. For example, HGTCmight tend to show differential expression between any given pair of conditions or timepoints if they are overrepresented among effectors at the ends of gene regulatory networks, or if their expression level is less precisely regulated than core metazoan genes66,67. Alternatively, HGTCmight be overrepresented among general stress-response genes, rather than those responding to pathogens in particular40,42,47. To test these possibilities, we repeated the analyses above using RNA-seq data from a prior study47that investigated gene expression when the same cultured strain of A. vaga was exposed to desiccation stress. Most bdelloid species can tolerate complete loss of cellular water by entering a state of physiological dormancy known as anhydrobiosis68,69. Hecox-Lea and Mark Welch (2018) compared gene expression in hydrated animals to those that were either entering or recovering from anhydrobiosis, and we tested for enrichment of HGTCin these responses. The proportion of HGTCthat were significantly upregulated in response to desiccation was substantially lower than in the pathogen experiment (13.8% for ‘entering’, 14.7% for ‘recovering’, compared to ~23–32% in the pathogen experiments; Fig.3 and Table 1). There was no significant HGTC enrichment when animals were entering desiccation (estimate = 0.14, SE = 0.12, P = 0.26; Fig 3a and Table 1; Table 5), but significant enrichment was seen during recovery (estimate = 0.24, SE = 0.07, P = 0.0004; Fig.3b and Table 1; Table 5), especially for downregulated genes (estimate = 0.64, SE = 0.07, P < 0.0001). The markedly lower enrichment of HGTCis unlikely to reflect a less extensive transcriptional response overall because we detected nearly twice as many differentially expressed genes during recovery from desiccation (n = 3129) versus pathogen challenge (n = 1767). While the datasets from different experiments should be compared with caution, both stressors clearly induced strong differential expression in hundreds of genes, but the enrichment of HGTCgenes was between two and three times stronger when responding to pathogens compared with desiccation. To test how far the transcriptional responses are specific to each challenge or reflect general stress response pathways, we checked the identities of genes that were differentially expressed in the two experiments. Overall, the proportion of upregulated genes shared between any two treatment conditions was low (mean = ~13%; Fig 3c and Table 6), indicating that the large majority of significantly differentially expressed genes in A. vaga, including HGTC, are specific to either the infection or desiccation experiment. A striking difference between susceptible and resistant species in expression of horizontally transferred genes with predicted biosynthetic functions To investigate putative functions of these genes we performed functional enrichment analyses for Gene Ontology (GO) terms70overrepresented among upregulated HGTC, relative to all genes (Fig.4; Supplementary Data 1). We focused first on A. ricciae at T24, taking this as a mature expression profile that ought to include genes mediating resistance, given that over 70% of these hosts eventually survived. Applying stringent statistical criteria (FDR < 0.001), we identified 29 significantly enriched GO terms among upregulated HGTCgenes (summarised in Fig.4a). The most highly overrepresented term was ‘phosphopantetheine binding’ (GO:0031177). Phosphopantetheine is a key cofactor in the activity of non-ribosomal peptide, polyketide and hybrid synthetases (NRP / PKS), acting as a ‘swinging arm’ to bring activated fatty acid or amino acid groups into contact with sequential catalytic centres during biosynthesis71,72. Three more of the top five most strongly enriched GO terms were similarly associated with NRP / PKS functions or processes. Phosphopantetheine binding remained the most strongly enriched GO term among all upregulated genes (i.e., before filtering for HGTC), indicating that enrichment is not simply because genes with this term are especially likely to be classed as HGTC, but because their contribution to the pathogen response is disproportionate (Supplementary Data 5). The more susceptible species shows a markedly different pattern. Applying the same analysis to A. vaga at T24, we found no significant enrichment among upregulated HGTCfor any of eight NRP / PKS-associated GO terms that were enriched in A. ricciae (Fig.4b; Supplementary Data 1), even with relaxed statistical stringency (FDR < 0.05). This was especially notable because 78% of the enriched (FDR < 0.001) GO terms in A. vaga were otherwise shared with A. ricciae (14 / 18 terms, Supplementary Data 1), rising to 100% overlap at FDR < 0.05. The divergent pattern with respect to NRP / PKS biosynthesis-associated terms therefore appears to be the main difference between the resistant and the susceptible species in the predicted functional profiles of upregulated HGTC. The detailed output of the GO analysis indicated that these results were driven by 27 predicted proteins in A. ricciae, each associated with two or more NRP / PKS-related GO terms (Supplementary Data 2). BLAST queries of these upregulated HGTCproducts against the UniProt protein database returned top matches (24–35% identity) to NRP / PKS encoded by fungi (n = 6) or, more commonly, bacteria (n = 21). Fungal matches were NRP / PKS enzymes that synthesise a range of secondary metabolites (malpicyclins, pyranterreones, oxaleimides, andrastatins), some of which are cytotoxic. Bacterial matches were all NRPS proteins that catalyse the production of antimicrobial compounds, including bacitracin, surfactin, mycosubtilin, tyrocidines and gramicidins (Supplementary Data 2). These bacterial products show broad-spectrum activity, including against fungal hyphae73–75and spores76. This result raises the hypothesis that A. ricciae resists fungal pathogens in part by synthesising antifungal secondary metabolites using highly modified non- ribosomal peptide synthetases originally acquired from bacteria. Both species contain multiple divergent NRP / PKS clusters, with stronger differential expression by the resistant species in response to fungal infection To survey putative NRP / PKS genes in the two species, we searched the predicted proteomes for matches to three key canonical NRPS-related domains (see Methods for details). This returned positive matches in both species: 60 in A. vaga, 36 in A. ricciae (Supplementary Data 2). The diversity of these predicted NRP / PKS proteins within and between species is revealed by aligning the rotifer sequences with their nearest fungal or bacterial counterparts based on the condensation domain (Fig.5a, b). Both species appear to encode multiple NRP / PKS-like proteins that differ substantially from known bacterial or fungal counterparts. Counting and interpreting these NRP / PKS-like coding sequences (CDS) is challenging, because NRP / PKS are large multimodular proteins that can be encoded as single genes or clusters. They often contain duplications, recombination or fusion of copies or modules77that are not easily resolved in short-read assemblies such as Av13 and Ar18. To check our ‘three-domain’ NRP / PKS CDS inventory, we used a combination of manual and automated methods to locate putative biosynthetic gene clusters in a recently available chromosome-scale assembly for A. vaga (Av20), independent of our other annotation pipelines. We identified approximately 40 clusters (Fig.5c; Supplementary Data 8). Like many HGTC, these are disproportionately located in highly dynamic subtelomeric regions that are hard to assemble (the same holds for A. ricciae; Fig.15 and Table 7). Nevertheless, 59 of 60 (98%) of our NRP / PKS gene models from Av13 mapped to independently identified biosynthetic clusters in Av20, and over 94% of Av20 clusters with a canonical condensation domain were represented at least partially in our putative NRP / PKS gene set. This suggests that the ‘three domain’ prediction set from the Av13 and Ar18 assemblies gives an accurate, if conservative estimate of NRP / PKS diversity. Expression profiles of putative NRP / PKS CDS show that the more resistant A. ricciae upregulates two at T7, rising to 11 at T24, whereas in the more susceptible A. vaga, no NRP / PKS CDS are significantly upregulated at T7, rising to two at T24 (Fig.5d; Supplementary Data 2). Moreover, the fold-change for upregulated CDS was over an order of magnitude higher for A. ricciae than A. vaga at T24, driven by this group of 11. These genes were barely expressed under control conditions: of those with measurable expression, the mean normalised read count (TMM = 0.015) was lower than 96.8% of genes, reflecting a rapid switch from almost no transcription to large-scale production. The resistant species shows duplication of a shared, upregulated NRP / PKS cluster that is predicted to produce a cyclic compound similar to known antibiotics The condensation phylogeny and expression profile for the set of predicted NRP / PKS CDS shows a closely related (aLRT > 95%) cluster of 11 CDS from A. ricciae and 1 from A. vaga that are all significantly upregulated (‘Upregulated clade’ in Fig 5a), suggesting duplications in A. ricciae of an ortholog that is present in both species. To circumvent assembly issues discussed above, we manually investigated copy number and orthology of this focal cluster using alternative assemblies for both A. vaga and A. ricciae (see Methods for further details). This revealed that A. vaga encodes two nearly identical homologous copies of this NRPS-PKS hybrid cluster (represented by AVAG|g48151 in Fig 5b), whereas A. ricciae appears to encode at least six copies of an orthologous NRPS-PKS cluster that shares all the key modules. Furthermore, the clusters in A. ricciae are flanked by multiple copies of genes encoding the two major classes of metazoan and bacterial metabolite transporter proteins (ABCB1 and MFS-family), associated with resistance to antibiotics and export of toxic compounds from cells. These, too, are dramatically upregulated on exposure to the pathogen (Supplementary Data 6 and Table 16). As a final step to explore potential function, we predicted the structure of the secondary metabolite synthesised by the NRPS-PKS hybrid cluster represented by AVAG|g48151, using SeMPI v2.078. The predicted product is a cyclic heptadepsipeptide (Fig.6a) with no highly similar hits (metabolite score >0.7) to known compounds in the screened databases. However, the named bacterial metabolite with the highest similarity (score 0.64) is tyrocidine A (Fig.6b), a cyclic decapeptide with broad-spectrum antimicrobial activity that disrupts fungal membranes73, retards conidial germination and inhibits extension of hyphae79. A second match (score 0.61) is to hymenamide J, an octapeptide isolated from the microbiota associated with a marine sponge80, with cytotoxic properties81but to our knowledge unscreened for antimicrobial activity (although a related compound, hymenamide E, shows potent antifungal activity82). Horizontally acquired genes with matches to RNA ligases and glucanases are also disproportionately upregulated in both hosts Whereas expression of NRP / PKS CDS differed between the host species, two other prominent HGTC-encoded elements of the pathogen response were shared. First, terms related to RNA repair are at or near the top of enrichment analyses for both species and timepoints (e.g. GO:0008452, ‘RNA ligase activity’; Fig.4a and b). Only about 15 genes in either species have this term, but 50–70% of these are consistently, significantly and strongly upregulated in response to the fungus (Supplementary Data 1). This includes the two most strongly upregulated genes at both time points in A. vaga. The rotifer-encoded RNA ligase domains (Pfam accession PF013563) cluster phylogenetically with non-metazoan lineages (including bacteria, fungi, viruses, and non-metazoan eukaryotes; Fig.16a–c). The top BLAST matches are to RNA ligases from bacteria or bacteriophages, where they function in tail fibre assembly (Supplementary Data 6). Second, both species show upregulation of HGTCwith putative glucanase activity at both timepoints (e.g. GO:0042972, ‘licheninase activity’, active on β-1,4 or β-1,3 glycosidic bonds; 6 genes in A. ricciae; 3 in A. vaga at T24; Fig.16d and e; Supplementary Data 6). These genes had top BLAST matches to glycoside hydrolase (GH) family enzymes in bacteria, and are interesting because of possible roles in recognising or attacking fungal cell walls. While present in both species, the more resistant A. ricciae seems to encode and express a marginally expanded repertoire of these enzymes. Neither RNA ligase nor licheninase functions were enriched in the responses to desiccation (Fig.4c, and d). A final category that was consistently enriched was the generic term ‘catalytic activity’ (GO:0003824). Closer examination of individual genes revealed at least 6 with matches to components or regulators of plant immunity, including phenylpropanoid and shikimate biosynthetic pathways (Supplementary Data 6). Other genes of interest included dioxygenases and a gene of fungal origin linked to programmed cell death (Supplementary Data 6). Horizontally acquired pathogen-induced genes were largely inherited from a common ancestor Functional similarities between the responses of A. ricciae and A. vaga might either reflect shared inheritance of an ancient HGTC portfolio common to many bdelloid rotifers, or independent gains of genes with common functions but different and potentially recent non-metazoan origins. We tested these alternative hypotheses by examining patterns of orthology for all upregulated HGTC across A. vaga, A. ricciae, and other sequenced bdelloid genomes (see Methods). Of 372 HGTC upregulated by A. ricciae at T24, 351 (94%) showed evidence (BLAST hit with E-value < 1e–5) of orthologs in distant bdelloid genera (Rotaria or Didymodactylos), consistent with acquisition millions of years ago83(Supplementary Data 4). All of the remaining 21 genes shared orthologs within Adineta, indicating gains at least as old as congeneric divergences. Similarity between the two focal species therefore seems to reflect inheritance of horizontally acquired genes from a common ancestor. In A. vaga, at T24, just one upregulated HGTCout of 285 (0.35%) had no identified orthologous counterparts and represents a putative recent acquisition. This gene (AVAG|g23567) appears in Fig.5a (marked with an asterisk) because it appears to encode an NRP / PKS hybrid synthetase, with closest BLAST matches to bacterial PKSs that synthesise antibiotic precursors (narbonolide and 6-deoxyerythronolide B), and to fungal NRP-PKS hybrids synthesising cytotoxic compounds (pyrophen, swainsonine). The SeMPI prediction for this cluster is a small linear polyketide whose closest named match is alpiniamide (metabolite score 0.66), a product of Streptomyces sp. whose activity is unknown (Supplementary Data 2 and Supplementary Data 7). Example 2 - Materials and methods Rotifer and pathogen isolates Animals belonging to the species Adineta ricciae65and A. vaga118,119were isolated respectively in 1998 from mud in Australia65and ca.1984 from moss in Italy120. The fungal pathogen Rotiferopthora globospora64was found attacking co-occurring rotifers of the genus Adineta in soil in northern New York55. A pure culture on potato dextrose agar (PDA) was obtained in December 2008 using methods described elsewhere121, and deposited in April 2009 with the USDA Agricultural Research Service Collection of Entomopathogenic Fungal Cultures (ARSEF) for long-term cryogenic storage under the accession code ARSEF 8995. Infection assays Adult rotifers were transferred to 96-well plates (Thermo-Fisher), with approximately 11 animals per well (mean: 11.0, SD: 3.5) in 60µL of sterilised, distilled water. Wells were inoculated with 8µL of freshly prepared R. globospora conidial suspension at a density of 125 spores µL-1. Negative control wells received 8µL of distilled water or inactivated spore suspension (see Methods). The final density of spores in each well was high (ca.15 conidia µL-1) to ensure every animal was exposed to the pathogen in a synchronised pulse. Rotifers were counted after 48 and 72 hours and classed as active, contracted, killed by infection (if at least one hypha had emerged through the integument from the interior) or otherwise dead. RNA-seq experimental design Rotifer populations were reared in eight replicate Petri dishes per species, with ca.50 founders per dish, fed only with E. coli (OP50, 5 x 108cells per dish) in distilled water. Rotifers were counted and harvested after 4 weeks, when the mean population size was about 3000 per dish for A. ricciae and 2000 for A. vaga. Populations were then subdivided to yield 16 replicate 1.5mL tubes for each species, with approximately 1000 animals per tube for A. ricciae and 600 for A. vaga. Tubes were then randomly allocated to receive either live or irradiated pathogen spores, and to have RNA extracted either 7 or 24 hours later, with each combination replicated four times. Irradiated spore suspensions (see Methods) were used as a control treatment to account for physical, chemical, or nutritional effects of ingesting fungal cells, so that all else was equal except for pathogen viability. Tubes were inoculated with 20µL of live or irradiated spore suspension at a density of 500 spores µL-1, for a total of 10,000 spores and a final density of 80 conidia µL-1, to ensure every animal was exposed as synchronously as possible, and then incubated upright at 20oC in a blocked layout until RNA extraction. RNA extraction and sequencing Total RNA was extracted from each tube at the appropriate timepoint using a column-based RNeasy Mini kit (Qiagen #74104), following the manufacturer’s protocol for animal tissues. Extracted RNA was eluted in 32µL of RNase-free water and 1.5µL aliquots were analysed using a Nanodrop 2000 (ThermoFisher). Spectrophotometric measurements were used to select the three replicates with the highest RNA concentrations from each treatment group for further analysis and sequencing (24 tubes total). Libraries were prepared using the TruSeq stranded mRNA kit (Illumina) and sequenced on an Illumina NovaSeq 6000 at Edinburgh Genomics (Edinburgh, UK), using an SP flow cell to generate 50-base paired-end reads (~200 bp inserts). Data filtering and quality control Raw sequencing reads were quality- and adapter-trimmed using BBTools ‘bbduk’ v38.73 and error-corrected using BBTools ‘tadpole’ (https: / / sourceforge.net / projects / bbmap / ). Unwanted ribosomal RNA (rRNA) reads were removed by mapping to the SILVA rRNA database122using BBTools ‘bbmap’. Contaminant reads derived from either the bacterial rotifer food present in the dishes (E. coli strain OP50) or from the fungal pathogen itself were removed using a similar approach, by mapping to the OP50 genome (NCBI accession GCF_009496595.1) or to all available genomes of fungi in the family Clavicipitaceae (NCBI taxid 34397; see Methods for further details). An average of 78.5 million reads were retained per library after filtering (94.2 Gb total data; Table 9), with >99% of filtered data mapping to the A. vaga (Av13) and A. ricciae (Ar18) reference genomes (Table 15). Differential expression and functional enrichment analyses Transcript quantification was performed using Salmon ‘quant’ v0.14.1123, using the gene models of Nowell et al. (2018)33as the target transcriptomes. Short transcripts (<150 bases) were removed prior to analysis, and genomic scaffolds were appended to each transcriptome data as ‘decoys’ prior to quantification as recommended in the Salmon documentation. Relationships between biological replicates within and between samples were checked visually using utility scripts from the Trinity software124. Statistical analysis of the resulting count matrix was performed with DESeq2 v1.26.0125, which uses negative binomial generalized linear models to test for differential expression. P-values were adjusted for multiple testing using the Benjamini-Hochberg method126to control the false discovery rate (FDR). Stringent thresholds of FDR < 1e–3 and absolute log2fold-change > 2 (i.e.4-fold difference in expression) were used to define an initial set of differentially expressed genes for downstream analysis. To test if our results were affected by the choice of software, we repeated the DE analysis using two alternative DE packages, with consistent results (Figs.11–14; see Methods for further details). Functional annotations for all genes were assimilated using the Trinotate v3.2.0127pipeline, while putative HGT candidate genes (HGTC) were classified based on the recent analysis of Jaron et al. (2021) (see Methods for further details). Functional enrichment analysis was performed using GOseq v1.38.070, based on gene ontology (GO) terms identified during functional annotation. For each timepoint, the test set was defined as HGTC genes that were significantly up- or downregulated (based on absolute log2 fold-change > 4 and FDR < 1e–3) versus the background set of the whole genome. Significant GO terms were visualised using Revigo128(default parameters). To test if our results were affected by the choice of genome-based transcriptome targets used, we repeated the above analyses using transcriptomes that were assembled de novo from the RNA-seq data (see Methods for further details), independent of reference assemblies or HGTCfiltering (Supplementary Data 5). We found ‘phosphopantetheine binding’ (GO:0031177) was still the most highly enriched term for A. ricciae (FDR = 3.72e–15), with ‘antibiotic biosynthetic process’ and ‘isomerase activity’ also significantly enriched (FDR = 0.0001), whereas none of these NRP / PKS-associated terms was enriched at any level for A. vaga. The same conclusion is therefore reached whether transcriptomes are assembled and annotated de novo from RNA-seq data, or mapped to gene models predicted from the reference genomes. To account for the lower RNA-seq coverage and power to detect functional enrichment in the desiccation dataset47, we relaxed the FDR threshold for GO term enrichment to P < 0.05, then to P < 0.1 (Supplementary Data 3). This ensured we would not miss weak signals of functional overlap between HGTCupregulated in response to biotic and abiotic stress. Statistical analyses To assess differences in transcriptional response to infection between species, we fitted a linear mixed effects model to predict DE (log2fold-change) including species and timepoint as two- level fixed factors, DE category (‘upregulated’, ‘downregulated’ and ‘non-significant’ according to the thresholds outlined above) as a three-level fixed factor, and gene ID as a random intercept term using the ‘lmer’ function from the lme4129package in R v4.0.2130. To assess the significance of differences in the proportion of HGTC in up- and downregulated subsets of genes, we first fitted generalized linear models for HGTC as a binary response variable (1 = is HGTC; 0 = is not HGTC) against species as a two-level fixed factor and DE category as a three-level fixed factor using the ‘glm’ function in base R. Separate models were run for T7 and T24 timepoints. A reduced model was then run using ‘glmer’ from lme4, including DE category as a three-level fixed factor and gene ID as a random intercept term. Phylogenetic analyses Phylogenetic trees of rotifer-encoded NRP / PKS proteins with bacterial and fungal homologs were constructed based on the alignment of the condensation domain (PF00668) with Pfam ‘seed’ representatives. Alignments were built using HMMER ‘hmmalign’, and phylogenetic analysis performed using IQ-TREE v1.6.12131(see Methods for further details). Phylogenies of other domains of interest were constructed using the same approach. Putative NRP / PKS screen and genomic validation in A. vaga An automated screen was conducted on the predicted proteomes of A. vaga and A. ricciae for putative NRP / PKS genes based on the presence of the canonical adenylation (AMP-binding, Pfam accession PF00501), thiolation and peptide carrier protein (PP-binding, PF00550) and condensation (PF00668) domains, using HMMER ‘hmmsearch’ v3.3 (http: / / hmmer.org / ). Only proteins with significant matches (E-value ≤ 1e–5) to all three domains were classified as putative NRP / PKS in this ‘three-domain’ set of gene models. This approach was validated by performing a manual survey in A. vaga, where a chromosome-scale genome assembly (Av20) has recently become available36(see Methods for further details). To test if the identified NRP / PKS set were located in telomeric regions, we counted the frequency of (a) genes, (b) transposable elements (TEs)34, and (c) the telomere-associated repeat motif ‘TGTGGG’132in (max) 50 kb windows surrounding each putative NRP / PKS gene model using BEDTools v2.29.2. The span of each feature was converted to a proportion by dividing by the actual window size for each flanking region, to correct for variation in window size. The genomic context of core eukaryote BUSCO genes (n = 303) was also evaluated for comparison (Fig.15 and Table 7). For A. vaga, the genomic locations of NRP / PKS clusters were also ascertained directly based on the Av20 genome assembly. The Av20 assembly (along with an additional, alternative assembly for A. ricciae) was also used to determine copy number and orthology of the set of upregulated putative NRP / PKS gene models highlighted in Fig.5 (see Methods for further details). Secondary metabolite prediction Secondary metabolite products of focal NRP / PKS gene models were predicted using the SeMPI v2.0 web server78, with maximum cluster distance set to 25 kb and all metabolite databases selected, but otherwise default settings (see Methods for further details). Data availability All raw sequencing data generated by this study have been deposited in the relevant International Nucleotide Sequence Database Collaboration (INSDC) database with the BioProject ID PRJEB39927, with the SRA run accessions ERR4469891, ERR4469902–8, ERR4471099–102, ERR4471104–11, ERR4471113–6 (see Table 9). This study also analysed publicly available data from BioProjects PRJEB1171, PRJEB23547, and PRJEB43248 (genome assemblies, gene annotations, TE annotations), and PRJNA494578 (RNA-seq data for the desiccation test). Code availability Custom scripts used in this study are available at GitHub (https: / / github.com / reubwn / bdelloid-immunity). Rotifer and pathogen isolates Animals belonging to the species Adineta ricciae1and A. vaga2,3were isolated respectively in 1998 from mud in Australia1and ca.1984 from moss in Italy4and propagated clonally in continuous long-term culture across several laboratories5–8. We use the code ‘AD001’ for this lineage of A. ricciae, and ‘AD008’ for A. vaga. Annotated genome assemblies are available for both species, denoted ‘Ar18’ for A. ricciae and ‘Av13’ for A. vaga7,9. Cultures were maintained in 60mm plastic Petri dishes in sterilised distilled water, fed with Escherichia coli (OP50) and Saccharomyces cerevisiae (S288c) and subcultured approximately once per month. They were stored at 20°C in an illuminated incubator (LMS, Kent, UK) with a 12:12 hour light:dark cycle. The fungal pathogen Rotiferopthora globospora10was found attacking co-occurring rotifers of the genus Adineta in soil in northern New York11. A pure culture on potato dextrose agar (PDA) was obtained in December 2008 using methods described elsewhere12, and deposited in April 2009 with the USDA Agricultural Research Service Collection of Entomopathogenic Fungal Cultures (ARSEF) for long-term cryogenic storage under the accession code ARSEF 8995. Frozen mycelium was retrieved and revived from this collection in 2013, and since maintained in serial subculture on PDA at 20°C with a 12:12 hour light:dark cycle and with transfers every 4 months. This pathogen strain has no recent history of laboratory co-passage with either of the two rotifer hosts tested here and all three were isolated on different continents. However, R. globospora appears to be globally distributed—it was originally described from New Zealand10and has also been recorded in Japan (CGW, pers. obs.), in both cases attacking Adineta. Therefore, both host species are likely to encounter this pathogen in nature, especially given the high dispersal capacity and global distribution of bdelloid rotifers13–16. Infection by R. globospora is initiated when a host ingests infectious spores (conidia), which are spherical and approximately 3.5µm in diameter10. These lodge in the mouth or oesophagus, at which point the animal stops feeding and contracts within a few minutes (Fig.1a, main text). Over the next 6–8 hours, the conidium produces a thin germ tube that penetrates the gut wall and swells into assimilative hyphae, which begin to invade and digest the surrounding host tissue after about 12–24 hours, as the infection becomes established. By 36–48 hours, fungal hyphae have filled most of the body cavity and the host is dead. Between 48–72 hours, hyphae begin to emerge through the host integument, and will eventually differentiate to form two types of spores: a handful of thick- walled resting spores, and hundreds of fresh infectious conidia (Fig.1a, main text). Because Rotiferophthora can only complete its life cycle by killing its host, these fungi are inherently highly virulent pathogens, and have been described as “devastating” to populations of Adineta10. R. globospora (ARSEF 8995) has been shown to exterminate laboratory populations of a sympatric Adineta clone within 28 days of initial exposure to a low density of conidia11. However, even if rotifer individuals are inevitably killed once the pathogen becomes established, an individual host can resist the initial attack and prevent an ingested spore from establishing a successful infection, or at least delay its progression. Even partial resistance at an early stage could dramatically slow the spread of an epidemic17, giving clonal relatives time to escape the infested habitat16,18before the local population is exterminated. Infection assays To quantify resistance to R. globospora under standard conditions, rotifers were transferred by pipette from stock populations to a 2mL droplet of sterile distilled water, where eggs, corpses and food from the cultures were washed away. Adult individuals were transferred to 96-well plates (Thermo-Fisher), with approximately 11 animals per well (mean: 11.0, SD: 3.5) in 60µL of sterilised, distilled water. Rotifers were counted using a compound microscope (Nikon Eclipse E400), noting whether each animal was active (feeding or locomoting, Fig.1a), contracted (with head withdrawn) or dead. Animals that died during the transfer (<1.5% of the total) were physically removed where possible, or recorded so they could be excluded from later counts. To obtain pure suspensions of conidia, approximately 5x5mm of freshly subcultured sporulating mycelium on PDA was moved to 500µL of sterile distilled water in a 1.5mL Eppendorf tube. After vortexing, 400µL of conidial suspension was transferred to a fresh sterile tube and conidial density was measured using a haemocytometer and lactophenol cotton blue stain. An inactivated inoculum was prepared simultaneously as a control, by exposing an aliquot of conidia in water to a germicidal ultraviolet lamp (25W, 253.7nm) at a distance of 10cm for 45 minutes, with regular vortexing to ensure all spores were irradiated. This treatment was successful, as none of the rotifers in control wells treated with irradiated spores became infected. Wells were inoculated with 8µL of freshly prepared conidial suspension at a density of 125 spores µL-1. Negative control wells received 8µL of distilled water or inactivated spore suspension. The final density of spores in each well was high (ca.15 conidia µL-1) to ensure every animal was exposed to the pathogen in a synchronised pulse. This appeared to work, because 94% of animals in experimental wells were contracted 8 hours after exposure, versus a baseline of 3% of animals contracted in control wells. Plates were then stored in incubators (LMS 300NP) at 20oC, with a 12:12 hour light:dark cycle. After 48 and 72 hours, rotifers were counted again, and classed as active, contracted, killed by infection or otherwise dead. A rotifer was considered to be killed by infection if at least one hypha had emerged through the integument from the interior. This criterion was unambiguous, in contrast with the difficulty of determining whether a fungal infection had established inside a contracted animal. By 72h, the proportion of infected animals in experimental wells appeared to have stabilised. In a subset of wells recounted at 96h, only a small fraction of animals (~6.5%) had newly developed visible infections. The relative risk of infection for A. vaga versus A. ricciae at 96h (RR 2.94, 95% CI 1.93–4.46; Z = 5.06, P < 0.0001) had not narrowed significantly since 72h (ratio of RR 0.78, 95% CI 0.47–1.31, Z = –0.926, P = 0.354), but the survivors had reproduced, making further tracking of the originally exposed cohort difficult. We therefore took the 72h timepoint as the standard measure of infection mortality. Individuals in negative control wells never became infected, whether they received water or sterilized spores, and the background death rate was ~2%. Inoculation trials with both A. ricciae and A. vaga were replicated on multiple occasions using at least three separate source dishes for each species prepared at different times, and separately prepared cultures and inocula of fungi, with each set of trials replicated in at least three wells with accompanying negative controls in a blocked design. The total number of animals exposed across counted trials was 216 for A. ricciae and 189 for A. vaga (Fig.1b, main text). Rates of mortality from R. globospora infection were consistent across these trials, as was the difference in susceptibility between the species (Fig.1b, main text). One set of well trials was run simultaneously and in close linkage with the scaled-up RNA-seq experiment described below, using rotifers from the same source populations, inoculated with the same live and irradiated pathogen suspensions. This enabled the timings and presumed final outcomes of infections in the RNA tubes to be inferred, even though those animals could not be directly counted and were sacrificed for RNA by 24h. The results of the RNA-linked well experiments were similar to earlier trials: mean infection mortality at 72h was 18% for A. ricciae and 79% for A. vaga. RNA-seq experimental design Rotifers for the RNA-seq experiment were reared at scale in eight replicate Petri dishes per species, with ca.50 founders per dish, all from the same clonal laboratory line of each species. These were fed only with E. coli (OP50, 5 x 108cells per dish) in distilled water. S. cerevisiae was omitted to avoid introducing further eukaryotic transcripts, and because gene expression by rotifers metabolising a fungal food source might complicate inferences about transcription in response to a fungal pathogen. The omission of S. cerevisiae as a food did not affect infection outcomes, because results of RNA-linked well assays were almost identical to earlier trials. When the A. ricciae dishes were initially established, each was inoculated with 50µL of 5µm-filtered water from the A. vaga source population, and vice-versa, to homogenise any co-cultured bacterial communities whose composition might differ between the cultured lines of the two species, and limit this as a potential source of gene expression differences. Dishes of the two species were stored in evenly interspersed blocks while the populations were growing. Rotifers were counted and harvested after 4 weeks, when the mean population size was about 3000 per dish for A. ricciae and 2000 for A. vaga, which reproduces more slowly19. Each dish was gently washed, using several water changes to float and pour off eggs, bacterial cells and corpses, while active animals adhered to the plastic. Cleaned rotifers were detached from the plastic by dropping distilled water onto them from a 10mL syringe at 40cm height, followed by repeated aspiration and forcible expulsion of medium using a P1000 pipette and a 1mL tip. Physical detachment was preferred over a salt or cold shock approach7, to reduce the chance of inducing transcriptomic, immunological or behavioural changes. Suspended rotifers were poured into a 50mL centrifuge tube and pelleted at 3000 x g for 5 minutes using a swing-bucket cradle. All but 1.5mL of supernatant was removed, then a vortex shaker was used to resuspend the rotifer pellets and 1mL was transferred immediately by pipette to a 1.5mL Eppendorf tube. This process was repeated to yield 16 replicate 1.5mL tubes for each species. Each tube was centrifuged at 17000 x g for 3 minutes, and all but 100µL of water was removed from the rotifer pellet. The pelleted rotifers were left overnight to recover and redistribute themselves around the submerged interior of the tube. We estimate that the efficiency of rotifer recovery via this method is about 70%, based on numbers of animals left over in plates or tubes, giving approximately 1000 animals per tube for A. ricciae and 600 for A. vaga. The 16 tubes for each species were randomly allocated to receive either live or irradiated pathogen spores, and to have RNA extracted either 7 or 24 hours later, with each combination replicated four times. Irradiated spore suspensions, as described above, were used as a control treatment to account for physical, chemical, or nutritional effects of ingesting fungal cells, so that all else was equal except for pathogen viability. RNA sampling times were chosen to correspond to the early stages of infection12—by 7h, germ tubes from ingested spores have attempted to infiltrate the host, but assimilative hyphae would not yet be established. By 24h, assimilative hyphae would be present if the germ tube had been successful, but these would not yet have colonised the host extensively or killed it. Each tube was inoculated with 20µL of live or irradiated spore suspension at a density of 500 spores µL-1, for a total of 10,000 spores and a final density of 80 conidia µL-1, to ensure every animal was exposed as synchronously as possible. Tubes were incubated upright at 20oC in a blocked layout until RNA extraction. RNA extraction and sequencing Total RNA was extracted from each tube at the appropriate timepoint using a column-based RNeasy Mini kit (Qiagen #74104), following the manufacturer’s protocol for animal tissues. To lyse and homogenise the rotifers, tubes were centrifuged at 17,000 x g for 3 minutes and 100µL of excess water was removed, leaving rotifers pelleted in approximately 20µL of water. After adding 50µL of Buffer RLT, the pellet was immediately disrupted and homogenised for 1 minute by pulsed application of a pellet pestle attached to a cordless motor (Kimble). A further 250µL of Buffer RLT was added to stabilise the lysate and rinse residue from the pestle, before proceeding with the manufacturer’s protocol, including an on-column DNase I digestion step. Lysis and stabilisation were completed for all tubes within 30 minutes of the target timepoint, in a balanced order with respect to treatment and species. RNA was eluted in 32µL of RNase-free water and 1.5µL aliquots were analysed using a Nanodrop 2000 (ThermoFisher). Spectrophotometric measurements were used to select the three replicates with the highest RNA concentrations from each treatment group for further analysis and sequencing, so that downstream steps had three biological replicates. These 24 tubes were frozen at -80oC and shipped on dry ice to the University of Edinburgh, where further quality control was undertaken by Edinburgh Genomics, including RNA quantitation with a Qubit 2.0 fluorometer in duplicate for each sample, and RNA ScreenTape analysis with an Agilent 2200 TapeStation system to assess RNA integrity. All 24 samples proceeded to cDNA library preparation, using the TruSeq stranded mRNA kit (Illumina) to enrich for polyadenylated transcripts. The indexed ~200 bp insert libraries were sequenced in multiplex on an Illumina NovaSeq 6000 at Edinburgh Genomics, using an SP flow cell to generate 50-base paired-end reads. Data filtering and quality control The 24 libraries yielded 102.9 Gb of raw sequencing data. Raw sequencing reads were quality- and adapter-trimmed using BBTools ‘bbduk’ v38.73 (parameters ‘ktrim=r k=23 mink=11 hdist=1 tpe tbo’), and error-corrected using BBTools ‘tadpole’ (parameters ‘mode=correct tossjunk tossuncorrectable’) Unwanted reads derived from ribosomal RNA (rRNA) were removed by mapping the data to the SILVA rRNA database20using BBTools ‘bbmap’ with parameters ‘local=t outu=filtered_R#.fq.gz’ (i.e. retaining only unmapped pairs of reads). Sequences derived from the fungal pathogen were removed using the same approach, mapping to sequenced genomes of fungi in the family Clavicipitaceae (NCBI taxid 34397). The proportion of remaining reads mapping to the Ar18 or Av13 target genomes was assessed using STAR v2.7.3a21with the parameter ‘--twoPassMode Basic’. Final data quality was assessed visually using FastQC v0.11.622and MultiQC v1.923. All raw sequencing data have been deposited in the relevant International Nucleotide Sequence Database Collaboration (INSDC) database with the Study ID PRJEB39927 (see Table 9 for run accessions). Before using the data to quantify expression of foreign genes encoded by rotifers, it was important to account for reads from contaminating microorganisms in the non-axenic experimental tubes. For example, bacterial contamination might complicate expression measurements for rotifer HGTCgenes that share partial homology with true bacterial sequences. Rotifer-encoded HGTCgenes also preclude simply filtering out all reads mapping to known bacterial sequences. These potential issues were instead assessed by focusing on the E. coli that was fed to the rotifers (strain OP50), whose genome sequence is available (GenBank accession GCF_009496595.1). Despite efforts to wash the animals and allow time to clear the gut, OP50 was still by far the largest potential source of contaminating transcripts—approximately 3.56% (± 2.8 SD) of unfiltered reads mapped to the OP50 genome (Table 10). In filtered reads, however, this was reduced to 0.016% (± 0.015 SD; Table 11), indicating that the vast majority of sequenced OP50 reads were derived from 16S and 23S rRNA, as confirmed by further direct mapping to OP50 rRNA (Table 12). E. coli therefore made a negligible contribution to the mRNA pool, probably owing to massive enrichment for eukaryotic sequences via the poly-A selection step. Nevertheless, to exclude even a marginal effect of mis-mapped bacterial reads on estimated HGTC expression, each significantly upregulated HGTCgene was examined for any contribution from OP50-derived reads. For none of these genes did more than one or two reads from OP50 contribute to expression calculations, whereas the corresponding mean number of total reads was 13655 (minimum 153) (Tables 13 and 14). Given that E. coli OP50 was the most common contaminant, its negligible contributions to mRNA-derived reads and HGTCexpression calculations indicate that the data and results reflect genes encoded and expressed by rotifers, and not RNA originating in bacterial contaminants. The hypothesis that HGTC sequences arise from contamination is further refuted by the presence of corresponding paired copies on genomic scaffolds for A. vaga and A. ricciae whose divergence matches the rest of the genomes, by the fact that investigated HGTCgenes often show a high degree of divergence from the nearest sequenced non-metazoan copies, and by the presence of introns in genes encoding proteins whose closest homology is otherwise to bacteria. Many upregulated gene copies were closely related to each other and showed patterns of homologous or homoeologous relationships consistent with the ancestral tetraploidy of bdelloid species24,25. After data filtering and quality-control, 78.5 million reads were retained per library (94.2 Gb total data, Table 9). Over 99% of filtered reads mapped to the Av13 and Ar18 reference genomes (Table 15). Filtered data were de novo assembled using the Trinity software v2.8.926,27, resulting in 118,860 and 63,749 transcripts representing 41,527 and 35,079 ‘genes’ for A. ricciae and A. vaga, respectively (Table 15). Transcriptomes showed G + C proportions that were consistent with their respective genomes (A. ricciae 37.6%; A. vaga 33.0%) and high ‘completeness’ scores as measured by BUSCO analysis (>97% of core eukaryote genes completely recovered in both cases). Transcript-to-genome mapping rates were >99% in both species, with 87.1% (A. ricciae) and 81.8% (A. vaga) of introns correctly called based on previously published gene models. Differential expression analysis Transcript quantification was performed using Salmon ‘quant’ v0.14.128, using the gene models of Nowell et al. (2018)9as the target transcriptomes. Short transcripts <150 bases were removed prior to analysis, and genomic scaffolds were appended to each transcriptome data as ‘decoys’ prior to quantification as recommended in the Salmon documentation. Relationships between biological replicates within and between samples were checked visually using utility scripts from the Trinity software26,27. Statistical analysis of the resulting count matrix was performed with DESeq2 v1.26.029, which uses negative binomial generalized linear models to test for differential expression. P-values were adjusted for multiple testing using the Benjamini- Hochberg method30to control the false discovery rate (FDR). Stringent thresholds of FDR < 1e–3 and log2 fold-change > 2 (i.e.4-fold difference in expression) were used to define an initial set of differentially expressed genes for downstream analysis. Comparison of control populations (Figs. 18 and 19) showed that HGTCgenes are more likely to be expressed than native genes under control conditions, but at significantly lower levels. The enrichment for HGTC we found among differentially expressed genes is thus unlikely to be explained by a known false-positive bias in some RNA-seq analyses toward genes with higher expression levels31,32. However, to test for this effect and to check for consistency in our overall results, we also calculated differential expression using two other software packages, edgeR v3.28.133–35and limma / voom v3.42.236. These packages handle known biases in RNA-seq analysis differently32, allowing us to establish that differential expression results are not dependent on the software used (Figs.11–14). To assess the quality of the A. vaga and A. ricciae gene predictions9, a transcriptome was assembled de novo from the filtered RNA-seq data using Trinity v2.8.926with default parameters, and mapped to the genomes of A. ricciae9and A. vaga7using Minimap237v2.17 with the parameters ‘-ax splice -C5’. The degree of correspondence between the predicted gene models and the assembled transcriptome was then assessed using ‘paftools.js junceval’, which provides a count of the number of correctly predicted introns based on the exact agreement between the location of the intron-exon junction in the annotation file and gaps in the transcriptome alignments. Gene completeness of assembled transcriptomes was also measured by Benchmarking Universal Single- Copy Orthologs (BUSCO) analysis38, using the core eukaryote and metazoa gene sets (n = 303 and 978 query genes, respectively). Transcriptome quality metrics are shown in Table 15. HGT classification Putative HGT candidate genes (denoted HGTC) were determined from the recent analysis of Jaron et al.39. Briefly, proteins were aligned to the UniProt90 sequence database40using Diamond ‘blastp’ v0.9.2141with the parameters ‘--sensitive -k 500 -e 1e-10’. For each protein, a HGT ‘index’ (hU) was computed based on the alignment bitscores to the best-matching hits from the Metazoa (BIN) and non-Metazoa (BOUT) with the formula: hU= BOUT– BIN42. A ‘consensus hit support’ (CHS) score was also calculated as the proportion of secondary hits in agreement with the result based on the hU9,43. An initial set of HGTC were defined if hU > 30 and CHS > 90%. This initial set was further filtered based on tBLASTn alignment to the recently published, near-chromosomal level A. vaga assembly44(diploid version), retaining only those with a good match (E-value ≤ 1e–5) to this highly curated genome. Phylogenetic trees of selected HGTC and their putative co-orthologs from the UniRef90 database were constructed using IQ-TREE v1.6.12 with the parameters ‘-alrt 1000 - bb 5000 - Functional annotation Protein sequences from the A. ricciae and A. vaga reference genomes were aligned to the SwissProt40and Pfam48sequence databases using BLAST v2.10.1+49and HMMER v3.3 (http: / / hmmer.org / ) respectively. Signal peptide cleavage sites and transmembrane helices were identified using SignalP v4.150and TMHMM v2.051respectively. Protein domains were further identified using InterProScan v552. Functional annotations were assimilated using the Trinotate v3.2.053pipeline. Functional enrichment analyses Functional enrichment analysis was performed using GOseq54, based on gene ontology (GO) terms identified during functional annotation above. For each timepoint, the test set was defined as HGTCthat were significantly up- or downregulated (based on absolute fold-change > 4 and FDR < 1e–3) versus the background set of the whole genome. GOseq was run using the ‘run_GOseq.pl’ utility script in the Trinity software package. At the most stringent threshold (FDR < 0.001), 29 GO terms were significantly enriched for genes upregulated by A. ricciae at T24 (rising to 43 and 68 for FDR < 0.01 and < 0.05 respectively). Significant GO terms were visualised using Revigo55(default parameters). To assist in interpreting enriched GO terms, each unique HGTCassociated with significantly enriched terms was individually translated and processed as a BLASTp query against the UniProt database, with rotifers (NCBI taxid 10190) excluded. For further clarification in selected cases, we also ran a tBLASTn query against the NCBI nucleotide database. The list of significantly enriched GO terms that appeared to point to NRP / PKS functions was: GO:0031177 (MF) phosphopantetheine binding; GO:0017000 (BP) antibiotic biosynthetic process; GO:0016999 (BP) antibiotic metabolic process; GO:0072341(MF) modified amino acid binding; GO:0017144 (BP) drug metabolic process; GO:0019842 (MF) vitamin binding; GO:0033218 (MF) amide binding; GO:0016853 (MF) isomerase activity. We tested whether the apparent pattern of enrichment for NRP / PKS-related terms in A. ricciae versus A. vaga was affected by the reference genomes used for functional enrichment and HGTCanalyses, which were assembled and annotated using different methods for the two species. We repeated the functional enrichment analysis as above, using transcriptomes assembled and annotated de novo from RNA-seq data, independent of reference assemblies or HGTC filtering (Supplementary Data 5). ‘Phosphopantetheine binding’ (GO:0031177) was still the most highly enriched term for A. ricciae (FDR = 3.72e–15), with ‘antibiotic biosynthetic process’ and ‘isomerase activity’ also significantly enriched (FDR = 0.0001), whereas none of these NRP / PKS- associated terms was enriched at any level for A. vaga. The same conclusion is therefore reached whether transcriptomes are assembled and annotated de novo from RNA-seq data, or mapped to the predicted proteome of existing reference genomes. To account for the lower RNA-seq coverage and power to detect functional enrichment in the desiccation dataset56, we relaxed the FDR threshold for GO term enrichment to P < 0.05 (Fig.4, main text), then to P < 0.1 (Supplementary Data 3). This ensured we would not miss weak signals of functional overlap between HGTCgenes upregulated in response to biotic and abiotic stress. Even at this lowest threshold, however, only five of the 54 significantly (FDR < 0.05) enriched GO terms for pathogen-exposed A. vaga at T24 (Supplementary Data 1) overlapped with desiccation stress, all unrelated to the putative defensive functions discussed elsewhere: three redundant GO terms for oxidoreductase activity, ‘trimethyllysine dioxygenase activity’, ‘iron ion binding’ and ‘carnitine biosynthetic process’ (which is not mediated by NRP / PKS). Gene orthology Orthologous relationships between A. ricciae and A. vaga genes was determined using OrthoFinder v2.3.1257with default parameters. Proteomes from nine other rotifer species9,58,59were included in the analysis to aid orthology inference. Phylogenetic analyses Phylogenetic trees of rotifer-encoded NRP / PKS proteins with bacterial and fungal homologs were constructed based on the alignment of the condensation domain (PF00668) with Pfam ‘seed’ representatives. Alignments were built using HMMER ‘hmmalign’, and phylogenetic analysis performed using IQ-TREE v1.6.1245, using ModelFinder for automatic model selection46and the ultrafast bootstrap60and SH-like approximate likelihood ratio test61for inferring branch support (parameters ‘-m TEST -bb 5000 -alrt 1000’). Resultant trees were plotted in R using the ‘phytools’ v0.7-7062and ‘ape’ v5.4.163packages. Aligned rotifer proteins generally show low similarity to even the best-matching bacterial hits (e.g. pairwise BLSM62 residue matches are < 50%, even where long sequences spanning multiple domains can be aligned). Putative NRP / PKS screen and genomic validation in A. vaga An automated screen was conducted on the predicted proteomes of A. vaga and A. ricciae for putative NRP / PKS genes using a conservative approach based on the presence of the canonical adenylation (AMP-binding, Pfam accession PF00501), thiolation and peptide carrier protein (PP- binding, PF00550) and condensation (PF00668) domains, using HMMER ‘hmmsearch’ v3.3 (http: / / hmmer.org / ). Only proteins with significant matches (E-value ≤ 1e–5) to all three domains were classified as putative NRP / PKS in this ‘three-domain’ set of gene models. To check this set against the genomic repertoire of NRP / PKS clusters, a more inclusive manual survey was conducted in A. vaga, where a chromosome-scale haploid genome assembly (Av20) is available44alongside the diploid reference assembly used elsewhere (Av13). First, Av13 was screened for putative biosynthetic clusters based on the presence of adenylation (A-) and condensation (C-) domains using tBLASTn, agnostic to prior annotations. The number of individual domains occurring once per module was counted, to obtain an independent estimate of the number of modules while accounting for partially assembled NRPS genes that may be distributed between scaffolds. This yielded a total of 306 hits to A- domains and 276 hits to similarly-sized C-domains. Since NRPS are organized as iterative modules, each scaffold represented by a set of A-domains and / or C-domains was manually curated to obtain the putative multi-domain cluster. A total of 110 gene models corresponding to putative biosynthetic clusters were identified, most of which exhibit higher similarity to bacterial than to fungal NRPS, alongside 12 annotated polyketide synthases (PKS), with some of the annotated clusters representing NRPS-PKS hybrids. This list recovers all members of the conservative three-domain set (n = 60), but also many fragmented or potentially non-biosynthetic annotations, lacking one or more canonical domains. Cross-referencing to Av20 was performed using a multi-step approach. First, the software AntiSMASH v6.1.164was used for automated detection; this revealed potential NRPS-PKS cluster locations, but gene annotations were not optimal, especially intron detection, since the method is optimized for fungal or bacterial genomes. Further searches used tBLASTn with translated CDS of the 122 NRP / PKS-related annotations from the maximal Av13 set. Additionally, putative NRPS A- domains were confirmed with NRPSpredictor265, which uses support vector machines to predict substrate specificity based on the configuration of the amino acid residues in the active site of an A- domain, with Av20 haploid gene annotations showing a total of 51 with substrate predictions consistent with A-domains. In combination, these lines of evidence identified approximately 40 putative biosynthetic gene clusters in the haploid Av20 assembly (Supplementary Data 8), distributed across the six chromosomes (Fig.5c, main text). Approximately 10 appeared defective in some way (3’ incomplete; apparent frameshifts in key modules), but several of these CDS were highly expressed, so these apparent defects may reflect ongoing assembly and annotation challenges even in a high-quality genome. Clusters (especially those consisting of more than one module) are preferentially located in subtelomeric regions, which agrees with the known enrichment of HGTC in telomeric and subtelomeric regions in the A. vaga genome. All of the Av13 annotations in the three-domain NRP / PKS set (n = 60) could be matched by BLAST to one of 20 clusters in Av20. Each matched cluster attracted a median of 2 annotations, as expected when mapping to a haploid assembly, but with some variability. For example, one especially large subtelomeric Av20 cluster (3.2; 108.5 kb) had 8 matching Av13 annotations, perhaps reflecting scaffold fragmentation in the Av13 assembly, while two focal annotations (AVAG|g23567 and AVAG|g48151) mapped uniquely to a single Av20 cluster each (3.4 and 5.2 respectively), suggesting unresolved alleles in Av13. The remaining Av20 clusters (n = 19) did not attract matching annotations from the three-domain set, largely because they lacked the canonical C-domains required in the automated survey. Of the 17 Av20 clusters with annotated C-domains, 16 (94%) were captured in the three-domain set. The genomic context of identified clusters was ascertained directly for A. vaga using the Av20 assembly. Of 37 putative NRP / PKS clusters shown in Fig.5c, 24 (81%) are inside (or within 1 Mb of) a subtelomeric region, as defined by a mean relative density of >10 telomeric repeats assessed in a 1 Mb sliding window (Fig.17; see also Figure S10 of Simion et al.2021). For both species, a statistical approach was also used to confirm subtelomeric localisation, by counting the frequency of three telomere-associated genome features (the telomeric repeat hexamer “TGTGGG”66, low gene density and elevated transposable element repeat density) in a (maximum) 50 kb window surrounding each putative NRP / PKS gene model using BEDTools v2.29.2. The span of each feature was converted to a proportion by dividing by the actual window size for each flanking region, to correct for variation in window size. The genomic context of eukaryote BUSCO genes (n = 303) was also evaluated for comparison. Relative to BUSCO genes, NRP / PKS CDS in both rotifer species demonstrated significant associations with increased telomeric repeats (estimate = 0.72, SE = 0.05, P < 0.001) and transposable element repeats (estimate = 1.33, SE = 0.11, P < 0.001), and decreased gene density (estimate = –0.09, SE = 0.01, P < 0.001; Fig.15 and Table 7). Phylogenies of rotifer-encoded NRP / PKS gene models with bacterial and fungal homologs were constructed based on the alignment of the condensation domain (PF00668) with Pfam ‘seed’ representatives. Alignments were built using HMMER ‘hmmalign’, and phylogenetic analysis performed using IQTREE v1.6.12 with the parameters ‘-alrt 1000 -bb 5000 -m TEST’. Phylogenies of other domains of interest were constructed using the same approach (Fig.16). Orthology and copy number estimate for the upregulated NRP / PKS cluster To test for orthology between AVAG|g48151 and the upregulated, putatively fragmented A. ricciae CDS (Fig.5a and b, main text), the corresponding regions were examined in Av20, and in an alternative short-read genome assembly for A. ricciae (denoted Ar2159). Although internal sections of the putative NRPS-PKS hybrid cluster may have been collapsed where modules are identical, or fragmented into separately-assembled contigs where alleles diverge, the number of unique flanking sequences at the 5’- and 3’-termini of the gene model can be used to estimate the copy number for each cluster, even if the assembly is less reliable in the middle. We therefore inspected the adjacent flanking regions in the A. vaga and A. ricciae genomic contigs encoding the N-terminal and C-terminal regions of upregulated clusters, looking for local microsynteny in the 5’- and 3’-flanks. The single-copy AVAG|g23567 (scaffold AVAG00146 in Av13; Chr5: 665120..721107 in haploid Av20) does not have a recognizable ortholog in A. ricciae assemblies. The single-copy AVAG|g48151 in A. vaga (scaffold AVAG00591in Av13; Chr5: 665120..721107 in haploid Av20) has one true ortholog in A. ricciae: ARIC|g35898, as judged by the presence of a cytochrome P450 CDS in its immediate 3’-flank on the corresponding Ar21 genomic contig ARIC00373, and an ABC transporter B1-like CDS at the 5’-flank. It also has at least two potentially intact paralogs, indicated by upregulated CDS with alternative 5’ and 3’ flanks encoding different products: ARIC|g15363, with amiloride-sensitive sodium channel at the 5’-flank in contig ARIC00114; ARIC|g49019 with RNA-recognition motif RRM1_CPEB2-like protein at the 3’-flank in the Ar21 assembly contig ARIC003_00110, and ARIC|g51257, with ABC transporter B1-like protein at the 3’-flank in the Ar21 assembly contig ARIC003_01015; and AVAG|g51868 with MFS transporter-like abhydrolase domain-containing protein at the 5’-flank in the Ar21 assembly contig ARIC003_02601. An additional distinct 3’-flank was detected on a contig harbouring an NRPS C- terminal fragment, however it was 5’-truncated and the copy was deemed non-functional. The remaining CDS from Fig.5b belonging to the middle modules from the same cluster could not be confidently matched with contigs harbouring flanking genes, and can be assumed to originate from collapsed or improperly assembled middle modules from one of the paralogs. Note that some copies may have undergone further tandem duplications, thus the current estimate of 3 (or 6, if both alleles are counted) represents the minimal number, which may be revised upwards when a chromosome- quality assembly becomes available. The difficulty in assembling and resolving so many divergent copies of a multimodular cluster (and potential accessory transporter genes) helps explain the apparent fragmentation into multiple gene models for A. ricciae (Fig.5b; Supplementary Data 2). Nevertheless, even where multiple gene models overlap (e.g. ARIC|g35898; AVAG|g51868), each feature was assessed as being strongly upregulated in its own right, in analyses where all 11 CDS were available to attract RNA reads (Fig.5d). Therefore, automated and manually curated annotations across multiple alternative genome assemblies support the conclusion that A. ricciae both encodes more copies of the focal NRPS-PKS cluster than A. vaga and upregulates these copies more strongly. Secondary metabolite prediction Secondary metabolite products of focal NRP / PKS gene models were predicted using the SeMPI v2.0 web server67, with maximum cluster distance set to 25 kb and all metabolite databases selected, but otherwise default settings. Two different sequences were submitted: an extraction of the region spanning the predicted gene model for AVAG|g48151 (23.4 kb including introns), and the full Av13 genomic scaffold on which this gene model is encoded (AVAG00591, 98.7 kb). The automatic reannotation performed as part of the SeMPI pipeline gave slightly different module and product predictions (gene model region: C70H86N8O26; full scaffold: C64H77N7O23), though both predictions had highest similarity to tyrocidines. Fig.6a (main text) reports the prediction for the full scaffold input because this is agnostic with respect to prior annotations, but both outputs are provided in Supplementary Data 7. For AVAG|g23567, the full scaffold (AVAG00146, 365.1 kb) was used. The predicted metabolite for AVAG|g48151 was initially rendered by SeMPI with chemical structural errors (pentavalent carbon atoms), so was re-exported in SMILES format to ChemDraw JS (version 19.0.0-CDJS-19.0.x.9+da9bec968, PerkinElmer Informatics), cyclicised with appropriate valences and re-rendered using RDKit.js (https: / / github.com / rdkit / rdkit). References 1. Van Valen, L. A new evolutionary law. Evol Theory 1, 1–30 (1973). 2. Hamilton, W. D. Sex versus Non-Sex versus Parasite. Oikos 35, 282–290 (1980). 3. Brockhurst, M. A. et al. Running with the Red Queen: the role of biotic conflicts in evolution. Proc. Biol. Sci. 281, (2014). 4. Schulte, R. D., Makus, C. & Schulenburg, H. Host-parasite coevolution favours parasite genetic diversity and horizontal gene transfer. J. Evol. Biol.26, 1836–1840 (2013). 5. Choi, K. et al. Recombination rate heterogeneity within Arabidopsis disease resistance genes. PLoS Genet.12, e1006179 (2016). 6. Colnaghi, M., Lane, N. & Pomiankowski, A. Genome expansion in early eukaryotes drove the transition from lateral gene transfer to meiotic sex. Elife 9, (2020). 7. Ambur, O. H., Engelstädter, J., Johnsen, P. J., Miller, E. L. & Rozen, D. E. Steady at the wheel: conservative sex and the benefits of bacterial transformation. Philos. Trans. R. Soc. Lond. B Biol. Sci.371, (2016). 8. von Wintersdorff, C. J. H. et al. Dissemination of antimicrobial resistance in microbial ecosystems through horizontal gene transfer. Front. Microbiol.7, 173 (2016). 9. Sit, C. S. et al. Variable genetic architectures produce virtually identical molecules in bacterial symbionts of fungus-growing ants. Proc. Natl. Acad. Sci. U. S. A.112, 13150–13154 (2015). 10. Baunach, M., Chowdhury, S., Stallforth, P. & Dittmann, E. The landscape of recombination events that create nonribosomal peptide diversity. Mol. Biol. Evol.38, 2116–2130 (2021). 11. Caldera, E. J., Chevrette, M. G., McDonald, B. R. & Currie, C. R. Local adaptation of bacterial symbionts within a geographic mosaic of antibiotic coevolution. Appl. Environ. Microbiol.85, (2019). 12. Hamilton, W. D., Axelrod, R. & Tanese, R. Sexual reproduction as an adaptation to resist parasites (a review). Proc. Natl. Acad. Sci. U. S. A. 87, 3566–3573 (1990). 13. Agrawal, A. F. Similarity selection and the evolution of sex: revisiting the red queen. PLoS Biol.4, e265 (2006). 14. Otto, S. P. & Nuismer, S. L. Species interactions and the evolution of sex. Science 304, 1018–1020 (2004). 15. Lehtonen, J., Jennions, M. D. & Kokko, H. The many costs of sex. Trends Ecol. Evol. 27, 172–178 (2012). 16. Howard, R. S. & Lively, C. M. Parasitism, mutation accumulation and the maintenance of sex. Nature 367, 554– 557 (1994). 17. West, S. A., Lively, C. M. & Read, A. F. A pluralist approach to sex and recombination. J. Evol. Biol.12, 1003– 1012 (1999). 18. Hodgson, E. E. & Otto, S. P. The red queen coupled with directional selection favours the evolution of sex. J. Evol. Biol.25, 797–802 (2012). 19. Neiman, M., Meirmans, S. & Meirmans, P. G. What can asexual lineage age tell us about the maintenance of sex? Ann. N. Y. Acad. Sci.1168, 185–200 (2009). 20. Moreira, M. O., Fonseca, C. & Rojas, D. Parthenogenesis is self-destructive for scaled reptiles. Biol. Lett.17, 20210006 (2021). 21. Decaestecker, E. et al. Host-parasite “Red Queen” dynamics archived in pond sediment. Nature 450, 870–873 (2007). 22. Ordonez, N. et al. Worse comes to worst: Bananas and Panama Disease-when plant and pathogen clones meet. PLoS Pathog. 11, e1005197 (2015). 23. Morran, L. T., Schmidt, O. G., Gelarden, I. A., Parrish, R. C., 2nd & Lively, C. M. Running with the Red Queen: host-parasite coevolution selects for biparental sex. Science 333, 216–218 (2011).24. Birky, C. W., Jr. Positively negative evidence for asexuality. J. Hered.101 Suppl 1, S42-5 (2010). 25. Donner, J. Ordnung Bdelloidea.297 (Akademie Verlag, Berlin, 1965). 26. Segers, H. Annotated checklist of the rotifers (Phylum Rotifera), with notes on nomenclature, taxonomy and distribution. Zootaxa (2007). 27. Arkhipova, I. & Meselson, M. Transposable elements in sexual and ancient asexual taxa. Proc. Natl. Acad. Sci. U. S. A. 97, 14473–14477 (2000). 28. Mark Welch, D. & Meselson, M. Evidence for the evolution of bdelloid rotifers without sexual reproduction or genetic exchange. Science 288, 1211–1215 (2000). 29. Flot, J.-F. et al. Genomic evidence for ameiotic evolution in the bdelloid rotifer Adineta vaga. Nature 500, 453– 457 (2013).30. Signorovitch, A., Hur, J., Gladyshev, E. & Meselson, M. Allele sharing and evidence for sexuality in a mitochondrial clade of bdelloid rotifers. Genetics 200, 581–590 (2015). 31. Debortoli, N. et al. Genetic exchange among bdelloid rotifers is more likely due to horizontal gene transfer than to meiotic sex. Curr. Biol. 26, 723–732 (2016). 32. Mark Welch, D. B., Mark Welch, J. L. & Meselson, M. Evidence for degenerate tetraploidy in bdelloid rotifers. Proc. Natl. Acad. Sci. U. S. A. 105, 5145–5149 (2008). 33. Nowell, R. W. et al. Comparative genomics of bdelloid rotifers: Insights from desiccating and nondesiccating species. PLoS Biol. 16, e2004830 (2018). 34. Nowell, R. W. et al. Evolutionary dynamics of transposable elements in bdelloid rotifers. Elife 10, e63194 (2021).35. Wilson, C. G., Nowell, R. W. & Barraclough, T. G. Cross-contamination explains “Inter and intraspecifichorizontal genetic transfers” between asexual bdelloid rotifers. Curr. Biol.28, 2436–2444 (2018). 36. Simion, P. et al. Chromosome-level genome assembly reveals homologous chromosomes and recombination in asexual rotifer Adineta vaga. Sci Adv 7, eabg4216 (2021). 37. Laine, V. N., Sackton, T. B. & Meselson, M. Genomic signature of sexual reproduction in the bdelloid rotifer Macrotrachella quadricornifera. Genetics 220, (2022). 38. Vakhrusheva, O. A. et al. Genomic signatures of recombination in a natural population of the bdelloid rotifer Adineta vaga. Nat. Commun.11, 6421 (2020). 39. Gladyshev, E. A., Meselson, M. & Arkhipova, I. R. Massive horizontal gene transfer in bdelloid rotifers. Science 320, 1210–1213 (2008).40. Boschetti, C., Pouchkina-Stantcheva, N., Hoffmann, P. & Tunnacliffe, A. Foreign genes and novel hydrophilicprotein genes participate in the desiccation response of the bdelloid rotifer Adineta ricciae. J. Exp. Biol. 214, 59– 68 (2011). 41. Yoshida, Y., Nowell, R. W., Arakawa, K. & Blaxter, M. Horizontal Gene Transfer in Metazoa: Examples and Methods. in Horizontal Gene Transfer: Breaking Borders Between Living Kingdoms (eds. Villa, T. G. & Viñas, M.) 203–226 (Springer International Publishing, 2019). doi:10.1007 / 978-3-030-21862-1_7. 42. Boschetti, C. et al. Biochemical diversification through foreign gene expression in bdelloid rotifers. PLoS Genet. 8, e1003035 (2012). 43. Richards, T. A. & Monier, A. A tale of two tardigrades. Proc. Natl. Acad. Sci. U. S. A. 113, 4892–4894 (2016). 44. Jaron, K. S. et al. Genomic features of parthenogenetic animals. J. Hered.112, 19–33 (2021).45. Eyres, I. et al. Horizontal gene transfer in bdelloid rotifers is ancient, ongoing and more frequent in species fromdesiccating habitats. BMC Biol.13, 90 (2015). 46. Rodriguez, F., Yushenova, I. A., DiCorpo, D. & Arkhipova, I. R. Bacterial N4-methylcytosine as an epigenetic mark in eukaryotic DNA. Nat. Commun.13, 1072 (2022). 47. Hecox-Lea, B. J. & Mark Welch, D. B. Evolutionary diversity and novelty of DNA repair genes in asexual bdelloid rotifers. BMC Evol. Biol.18, 177 (2018). 48. Chou, S. et al. Transferred interbacterial antagonism genes augment eukaryotic innate immune function. Nature 518, 98–101 (2015). 49. Di Lelio, I. et al. Evolution of an insect immune barrier through horizontal gene transfer mediated by a parasitic wasp. PLoS Genet.15, e1007998 (2019).50. Verster, K. I. et al. Horizontal transfer of bacterial cytolethal distending Toxin B genes to insects. Mol. Biol. Evol.36, 2105–2110 (2019). 51. Harrison, E. & Brockhurst, M. A. Ecological and evolutionary benefits of temperate phage: What does or doesn’t kill you makes you stronger. Bioessays 39, (2017). 52. Hall, R. J., Whelan, F. J., McInerney, J. O., Ou, Y. & Domingo-Sananes, M. R. Horizontal gene transfer as a source of conflict and cooperation in prokaryotes. Front. Microbiol. 11, 1569 (2020). 53. Barron, G. L. Fungal parasites and predators of rotifers, nematodes, and other invertebrates. in Biodiversity of Fungi (eds. Mueller, G. M., Bills, G. F. & Foster, M. S.) 435–450 (Academic Press, 2004). doi:10.1016 / B978- 012509551-8 / 50022-2. 54. Wilson, C. G. & Sherman, P. W. Anciently asexual bdelloid rotifers escape lethal fungal parasites by drying up and blowing away. Science 327, 574–576 (2010).55. Wilson, C. G. Desiccation-tolerance in bdelloid rotifers facilitates spatiotemporal escape from multiple species ofparasitic fungi. Biol. J. Linn. Soc. Lond.104, 564–574 (2011). 56. Wilson, C. G. & Sherman, P. W. Spatial and temporal escape from fungal parasitism in natural communities of anciently asexual bdelloid rotifers. Proc. Biol. Sci.280, 20131255 (2013). 57. Irazoqui, J. E., Urbach, J. M. & Ausubel, F. M. Evolution of host innate defence: insights from Caenorhabditis elegans and primitive invertebrates. Nat. Rev. Immunol.10, 47–58 (2010). 58. Cogni, R., Cao, C., Day, J. P., Bridson, C. & Jiggins, F. M. The genetic architecture of resistance to virus infection in Drosophila. Mol. Ecol. 25, 5228–5241 (2016). 59. Bento, G. et al. The genetic basis of resistance and matching-allele interactions of a host-parasite system: The Daphnia magna-Pasteuria ramosa model. PLoS Genet.13, e1006596 (2017).60. Hudson, A. I., Fleming-Davies, A. E., Páez, D. J. & Dwyer, G. Genotype-by-genotype interactions between aninsect and its pathogen. J. Evol. Biol. 29, 2480–2490 (2016). 61. Ekroth, A. K. E., Gerth, M., Stevens, E. J., Ford, S. A. & King, K. C. Host genotype and genetic diversity shape the evolution of a novel bacterial infection. ISME J.15, 2146–2157 (2021). 62. Fredericksen, M. et al. Infection phenotypes of a coevolving parasite are highly diverse, structured, and specific. Evolution (2021) doi:10.1111 / evo.14323. 63. Engelstädter, J. Host-parasite coevolutionary dynamics with generalized success / failure infection genetics. Am. Nat.185, E117-29 (2015). 64. Barron, G. L. A new genus, Rotiferophthora, to accommodate the Diheterospora-like endoparasites of rotifers. Can. J. Bot.69, 494–502 (1991).65. Segers, H. & Shiel, R. Tale of a sleeping beauty: a new and easily cultured model organism for experimentalstudies on bdelloid rotifers. Rotifera X 141–145 (2005). 66. Jain, R., Rivera, M. C. & Lake, J. A. Horizontal gene transfer among genomes: the complexity hypothesis. Proc. Natl. Acad. Sci. U. S. A.96, 3801–3806 (1999). 67. Gluck-Thaler, E. & Slot, J. C. Dimensions of horizontal gene transfer in eukaryotic microbial pathogens. PLoS Pathog.11, e1005156 (2015). 68. Ricci, C. & Caprioli, M. Anhydrobiosis in bdelloid species, populations and individuals. Integr. Comp. Biol. 45, 759–763 (2005). 69. Mark Welch, D. B., Ricci, C. & Meselson, M. Bdelloid Rotifers: Progress in Understanding the Success of an Evolutionary Scandal. in Lost Sex 259–279 (Springer, Dordrecht, 2009). doi:10.1007 / 978-90-481-2770-2_13.70. Young, M. D., Wakefield, M. J., Smyth, G. K. & Oshlack, A. Gene ontology analysis for RNA-seq: accountingfor selection bias. Genome Biol. 11, R14 (2010). 71. Finking, R. & Marahiel, M. A. Biosynthesis of nonribosomal peptides. Annu. Rev. Microbiol. 58, 453–488 (2004). 72. Sieber, S. A. & Marahiel, M. A. Molecular mechanisms underlying nonribosomal peptide synthesis: approaches to new antibiotics. Chem. Rev. 105, 715–738 (2005). 73. Rautenbach, M., Troskie, A. M., Vosloo, J. A. & Dathe, M. E. Antifungal membranolytic activity of the tyrocidines against filamentous plant fungi. Biochimie 130, 122–131 (2016). 74. Jiang, J. et al. Identification of novel surfactin derivatives from NRPS modification of Bacillus subtilis and its antifungal activity against Fusarium moniliforme. BMC Microbiol.16, 31 (2016).75. Leclère, V. et al. Mycosubtilin overproduction by Bacillus subtilis BBG100 enhances the organism’s antagonisticand biocontrol activities. Appl. Environ. Microbiol.71, 4577–4584 (2005). 76. Murray, T., Leighton, F. C. & Seddon, B. Inhibition of fungal spore germination by gramicidin S and its potential use as a biocontrol against fungal plant pathogens. Lett. Appl. Microbiol.3, 5–7 (1986). 77. Medema, M. H., Cimermancic, P., Sali, A., Takano, E. & Fischbach, M. A. A systematic computational analysis of biosynthetic gene cluster evolution: lessons for engineering biosynthesis. PLoS Comput. Biol. 10, e1004016 (2014). 78. Zierep, P. F., Ceci, A. T., Dobrusin, I., Rockwell-Kollmann, S. C. & Günther, S. SeMPI 2.0-A web server for PKS and NRPS predictions combined with metabolite screening in natural product databases. Metabolites 11, (2020). 79. Troskie, A. M., de Beer, A., Vosloo, J. A., Jacobs, K. & Rautenbach, M. Inhibition of agronomically relevant fungal phytopathogens by tyrocidines, cyclic antimicrobial peptides isolated from Bacillus aneurinolyticus. Microbiology 160, 2089–2101 (2014). 80. Schirmer, A. et al. Metagenomic analysis reveals diverse polyketide synthase gene clusters in microorganisms associated with the marine sponge Discodermia dissoluta. Appl. Environ. Microbiol. 71, 4840–4849 (2005). 81. Tsuda, M., Sasaki, T. & Kobayashi, J. Hymenamides G, H, J, and K, four new cyclic octapeptides from the Okinawan marine sponge Hymeniacidon sp. Tetrahedron 50, 4667–4680 (1994). 82. Dahiya, R., Pathak, D., Himaja, M. & Bhatt, S. First total synthesis and biological screening of hymenamide E. Acta Pharm. 56, 399–415 (2006).83. Tang, C. Q., Obertegger, U., Fontaneto, D. & Barraclough, T. G. Sexual species are separated by larger geneticgaps than asexual species in rotifers. Evolution 68, 2901–2916 (2014). 84. Richards, T. A., Leonard, G., Soanes, D. M. & Talbot, N. J. Gene transfer into the fungi. Fungal Biol. Rev.25, 98–110 (2011). 85. Husnik, F. & McCutcheon, J. P. Functional horizontal gene transfer from bacteria to eukaryotes. Nat. Rev. Microbiol.16, 67–79 (2018). 86. Thiessen, K. D. et al. Zebrafish otolith biomineralization requires polyketide synthase. Mech. Dev.157, 1–9 (2019). 87. Shou, Q. et al. A hybrid polyketide-nonribosomal peptide in nematodes that promotes larval survival. Nat. Chem. Biol.12, 770–772 (2016).88. Izoré, T. et al. Drosophila melanogaster nonribosomal peptide synthetase Ebony encodes an atypicalcondensation domain. Proc. Natl. Acad. Sci. U. S. A. 116, 2913–2918 (2019). 89. Faddeeva-Vakhrusheva, A. et al. Coping with living in the soil: the genome of the parthenogenetic springtail Folsomia candida. BMC Genomics 18, 493 (2017). 90. Du, D. et al. Multidrug efflux pumps: structure, function and regulation. Nat. Rev. Microbiol. 16, 523–539 (2018). 91. Sung, G.-H. et al. Phylogenetic classification of Cordyceps and the clavicipitaceous fungi. Stud. Mycol.57, 5–59 (2007). 92. Herrero-Galán, E. et al. The insecticidal protein hirsutellin A from the mite fungal pathogen Hirsutella thompsonii is a ribotoxin. Proteins 72, 217–228 (2008). 93. Olombrada, M. et al. Fungal ribotoxins: Natural protein-based weapons against insects. Toxicon 83, 69–74 (2014).94. Olombrada, M. et al. Characterization of a new toxin from the entomopathogenic fungus Metarhizium anisopliae: the ribotoxin anisoplin. Biol. Chem. 398, 135–142 (2017). 95. Maxwell Burroughs, A. & Aravind, L. RNA damage in biological conflicts and the diversity of responding RNA repair systems. Nucleic Acids Res.44, 8525–8555 (2016). 96. Feng, H., Zhou, D., Daly, P., Wang, X. & Wei, L. Characterization and functional importance of two glycoside hydrolase family 16 genes from the rice white tip nematode Aphelenchoides besseyi. Animals 11, (2021). 97. Brown, G. D. & Gordon, S. Immune recognition of fungal beta-glucans. Cell. Microbiol.7, 471–479 (2005). 98. Hamilton, C. & Bulmer, M. S. Molecular antifungal defenses in subterranean termites: RNA interference reveals in vivo roles of termicins and GNBPs against a naturally encountered pathogen. Dev. Comp. Immunol.36, 372– 377 (2012).99. Liu, T. et al. Structural and biochemical insights into an insect gut-specific chitinase with antifungal activity. Insect Biochem. Mol. Biol.119, 103326 (2020). 100. Boller, T. Antimicrobial Functions of the Plant Hydrolases, Chitinase and ß-1,3-Glucanase. in Mechanisms of Plant Defense Responses (eds. Fritig, B. & Legrand, M.) 391–400 (Springer Netherlands, 1993). doi:10.1007 / 978-94-011-1737-1_124. 101. Aktuganov, G. et al. Wide-range antifungal antagonism of Paenibacillus ehimensis IB-X-b and its dependence on chitinase and beta-1,3-glucanase production. Can. J. Microbiol.54, 577–587 (2008). 102. Benítez, T., Rincón, A. M., Limón, M. C. & Codón, A. C. Biocontrol mechanisms of Trichoderma strains. Int. Microbiol.7, 249–260 (2004).103. Jaenike, J. A hypothesis to account for the maintenance of sex within populations. Evol. Theory 3, 191–194(1978). 104. Judson, O. P. & Normark, B. B. Ancient asexual scandals. Trends Ecol. Evol.11, 41–46 (1996). 105. Normark, B. B., Judson, O. P. & Moran, N. A. Genomic signatures of ancient asexual lineages. Biol. J. Linn. Soc. Lond.79, 69–84 (2003). 106. Peters, A. D. & Lively, C. M. The Red Queen and fluctuating epistasis: A population genetic analysis of antagonistic coevolution. Am. Nat. 154, 393–405 (1999). 107. Sasaki, A., Hamilton, W. D. & Ubeda, F. Clone mixtures and a pacemaker: new facets of Red-Queen theory and ecology. Proc. Biol. Sci.269, 761–772 (2002). 108. Gandon, S. & Otto, S. P. The evolution of sex and recombination in response to abiotic or coevolutionary fluctuations in epistasis. Genetics 175, 1835–1853 (2007). 109. Salathé, M., Kouyos, R. D., Regoes, R. R. & Bonhoeffer, S. Rapid parasite adaptation drives selection for high recombination rates. Evolution 62, 295–300 (2008). 110. Salathé, M., Kouyos, R. D. & Bonhoeffer, S. The state of affairs in the kingdom of the Red Queen. Trends Ecol. Evol.23, 439–445 (2008). 111. Lively, C. M. A review of Red Queen models for the persistence of obligate sexual reproduction. J. Hered.101 Suppl 1, S13-20 (2010). 112. Arkhipova, I. R., Yushenova, I. A. & Rodriguez, F. Giant reverse transcriptase-encoding transposable elements at telomeres. Mol. Biol. Evol. (2017) doi:10.1093 / molbev / msx159. 113. Arkhipova, I. R., Yushenova, I. A. & Rodriguez, F. Endonuclease-containing Penelope retrotransposons in the bdelloid rotifer Adineta vaga exhibit unusual structural features and play a role in expansion of host gene families.Mob. DNA 4, 19 (2013). 114. Linardopoulou, E. V. et al. Human subtelomeres are hot spots of interchromosomal recombination and segmental duplication. Nature 437, 94–100 (2005). 115. Barry, J. D., Ginger, M. L., Burton, P. & McCulloch, R. Why are parasite contingency genes often associated with telomeres? Int. J. Parasitol.33, 29–45 (2003). 116. Miethke, M. et al. Towards the sustainable discovery and development of new antibiotics. Nat Rev Chem 5, 726– 749 (2021). 117. Gao, J. et al. A rotifer-derived paralytic compound prevents transmission of schistosomiasis to a mammalian host. PLoS Biol. 17, e3000485 (2019).118. Davis, H. A new callidina: With the result of experiments on the desiccation of rotifers. Mon. Microsc. J. 9, 201–209 (1873). 119. Örstan, A. The trouble with Adineta vaga (Davis, 1873): a common rotifer that cannot be identified (Rotifera: Bdelloidea: Adinetidae). Zootaxa 4830, 597–600 (2020). 120. Ricci, C. Culturing of some bdelloid rotifers. Hydrobiologia 112, 45–51 (1984). 121. Barron, G. L. Fungal parasites of bdelloid rotifers: Diheterospora. Can. J. Bot.63, 211–222 (1985). 122. Quast, C. et al. The SILVA ribosomal RNA gene database project: improved data processing and web-based tools. Nucleic Acids Res. 41, D590-6 (2013). 123. Patro, R., Duggal, G., Love, M. I., Irizarry, R. A. & Kingsford, C. Salmon provides fast and bias-aware quantification of transcript expression. Nat. Methods 14, 417–419 (2017).124. Haas, B. J. et al. De novo transcript sequence reconstruction from RNA-seq using the Trinity platform forreference generation and analysis. Nat. Protoc. 8, 1494–1512 (2013). 125. Love, M. I., Huber, W. & Anders, S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol.15, 550 (2014). 126. Benjamini, Y. & Hochberg, Y. Controlling the False Discovery Rate: A practical and powerful approach to multiple testing. J. R. Stat. Soc. Series B Stat. Methodol. 57, 289–300 (1995). 127. Bryant, D. M. et al. A tissue-mapped axolotl de novo transcriptome enables identification of limb regeneration factors. Cell Rep. 18, 762–776 (2017). 128. Supek, F., Bošnjak, M., Škunca, N. & Šmuc, T. REVIGO summarizes and visualizes long lists of gene ontology terms. PLoS One 6, e21800 (2011).129. Bates, D., Mächler, M., Bolker, B. & Walker, S. Fitting linear mixed-effects models using lme4. Journal ofStatistical Software, Articles 67, 1–48 (2015). 130. R Core Team. R: A language and environment for statistical computing. (2016). 131. Nguyen, L.-T., Schmidt, H. A., von Haeseler, A. & Minh, B. Q. IQ-TREE: a fast and effective stochastic algorithm for estimating maximum-likelihood phylogenies. Mol. Biol. Evol.32, 268–274 (2015). 132. Gladyshev, E. A. & Arkhipova, I. R. Telomere-associated endonuclease-deficient Penelope-like retroelements in diverse eukaryotes. Proc. Natl. Acad. Sci. U. S. A. 104, 9352–9357 (2007). Example 3 - Evidence for upregulation of a biosynthetic gene cluster by the bdelloid rotifer Adineta steineri when challenged with the fungal pathogen Rotiferophthora sp. “ARSEF 10274” Elsewhere, we demonstrate that the bdelloid rotifers Adineta vaga and A. ricciae both strongly upregulate at least one biosynthetic gene cluster (‘BGC 48151’) when challenged with a fungal pathogen, Rotiferophthora globospora, to which they are at least partially resistant. The upregulation is stronger in A. ricciae, which is more resistant than A. vaga, but the clusters in the two species are elsewhere demonstrated to be orthologous. The predicted product of the enzyme encoded by BGC 48151 is a novel secondary metabolite with chemical similarity to known antibiotics (e.g. Tyrocidine A). We here consider a third bdelloid rotifer species, Adineta steineri, and a second pathogen species, Rotiferophthora sp. (an undescribed species known by its strain code, “ARSEF 10274”). Replicate populations of all three rotifer species were exposed to each fungal pathogen using the methods described elsewhere for well assays. Again, the spore density was 125 spores µL-1and the volume of inoculum was 8 µL. We measured the proportion of rotifers in each population visibly infected by the fungus at 72h post exposure, as described elsewhere (Fig 23a). The bdelloid rotifer Adineta steineri (orange boxes) is found to be significantly more resistant to the fungal pathogen R. sp.10274 than A. vaga, and marginally more resistant than A. ricciae (Fig 23b), whereas Adineta steineri shows no resistance to R. globospora 8995. In a separate experiment, total RNA was extracted from six replicate populations of A. steineri, according to the methods described elsewhere (“RNA-seq experimental design”). Three populations were exposed to live conidia of R. sp.10274; three were exposed to UV-inactivated spores as a non-challenged control. RNA was harvested from all populations at 18 hours and processed as described elsewhere to generate six transcript libraries with a mean of 250 million sequencing reads per library. Using methods described elsewhere (“Putative NRP / PKS screen and genomic validation in A. vaga”), we screened the genome of A. steineri for putative NRP / PKS biosynthetic gene clusters. We identified a predicted cluster of horizontal origin (“As_AS23R|g36239”) that has some areas of homology to ‘BGC 48151’, but is at least 20% divergent from the other species at the nucleotide level. We mapped reads from each of the six transcript libraries to this predicted cluster and compared the number of mapped reads from populations exposed to live pathogens versus UV- inactivated spores. This enables a test for upregulation of the focal cluster in response to attack by pathogens (Fig 23c). We conclude that A. steineri significantly upregulates expression of at least one biosynthetic gene cluster approximately twofold in response to challenge by R. sp. ARSEF 10274. Finally, we used methods described elsewhere (“Secondary metabolite prediction”) to predict the metabolite that might be synthesised by the cluster corresponding to “As_AS23R|g36239”. When considering the most highly expressed region of this cluster, the top hit by metabolite score was to the cyclic peptide antibiotic cairomycin B (Shimi et al.1977), while the top hit to the full cluster was the cyclic octapeptide champacyclin (Pesic et al.2013), which has activity against the plant pathogen Erwinia amylovora, the agent of fire blight in fruit trees. In both cases, the predicted rotifer product was novel and showed several differences of chemical structure from the named antibiotics (Fig 24D-E). References: Pesic A, Baumann HI, Kleinschmidt K, Ensle P, Wiese J, Süssmuth RD, Imhoff JF.2013. Champacyclin, a new cyclic octapeptide from Streptomyces strain C42 isolated from the Baltic Sea. Mar Drugs.11: 4834-57. doi: 10.3390 / md11124834. Shimi IR, Abedallah N, Fathy S.1977. Cairomycin B, a new antibiotic. Antimicrob Agents Chemother., 11: 373-5. doi: 10.1128 / AAC.11.3.373.

[0002] Example 4 - Evidence for resistance to fungal pathogens by bdelloid rotifers of the genus Habrotrocha (Family: Habrotrochidae, Order: Philodinida), and for upregulation of a biosynthetic gene cluster by the rotifer Habrotrocha ligula when challenged with the fungal pathogen Rotiferophthora denticulospora Elsewhere, we demonstrate that three bdelloid rotifer species of the genus Adineta strongly upregulate biosynthetic gene clusters with homology to ‘BGC 48151’ when challenged with fungal pathogens in the genus Rotiferophthora to which they are partially resistant. We here show evidence that resistance to fungal pathogens and the mechanism of upregulating horizontally acquired biosynthetic gene clusters in response to fungal attack can be generalised to bdelloid rotifers in a different order, family and genus, namely Habrotrocha. Under controlled conditions, we exposed multiple species and clonal lineages of bdelloid rotifers from two different genera (Habrotrocha and Adineta) to a diverse set of fungal pathogens of the genus Rotiferophthora. Replicate populations of fifteen different rotifer clones were exposed to 20 fungal pathogen isolates using the methods described elsewhere for well assays (“Infection assays”). Again, the spore density was 125 spores µL-1and the volume of inoculum was 8 µL. We measured the proportion of rotifers in each population killed and visibly infected by the fungus at 72h post exposure, as described elsewhere. Each of the pairings was replicated between 3 and 21 times, using over 20,000 individual rotifers in total. Results were plotted as a heat-map organised by the phylogenetic relatedness of the hosts and pathogens (Fig.22), and analysed using a Bayesian co-phylogenetic generalised linear model approach (Hadfield et al.2014). The results reveal substantial and significant variation in resistance between rotifer species and between fungal pathogens (Table 17), such that resistance is a function of host and pathogen identities and their interaction as well as their reciprocal and individual evolutionary histories. This method can be used to predict (and indicates directly) which combinations of rotifers and pathogens will show phenotypic resistance. The results demonstrate that resistance to fungal pathogens as initially exemplified for the genus Adineta also extends to the genus Habrotrocha. To further establish this generality, we examined the data for three species: Habrotrocha constricta, H. elusa and H. ligula, when challenged with two fungal pathogens selected from the panel in Figure 22: R. globospora and R. denticulospora (‘ARSEF 12628’). Statistical analysis of mortality patterns shows that all three are unaffected by R. globospora (no infections observed), suggesting these species may be beyond the host range of this pathogen (Figure 25A). Meanwhile, all three species were infected by R. denticulospora, but H. ligula showed approximately eightfold higher resistance to this fungal pathogen than either of the other two hosts, a highly significant difference (Figure 25A). Given the strong evidence of fungal resistance in H. ligula, we tested whether fungal challenge induces this species to upregulate biosynthetic gene clusters as we demonstrated for the distantly related genus Adineta. Total RNA was extracted from six replicate populations of H. ligula, according to the methods described elsewhere (“RNA-seq experimental design”). Three populations were exposed to live conidia of R. denticulospora; three were exposed to UV- inactivated spores as a non-challenged control. RNA was harvested from all populations at 24 hours and processed as described elsewhere to generate six transcript libraries with a mean of 251 million sequencing reads per library. We used tblastn to query a draft genome assembly for H. ligula using the protein translation of the biosynthetic gene cluster ‘BGC 48151’ from A. vaga. We identified a biosynthetic gene cluster ‘HB003_g97330’ with homology to the polyketide synthetase modules of the hybrid PKS- NRPS encoded by ‘BGC 48151’. We mapped reads from each of the six transcript libraries to this cluster and compared the number of mapped reads from populations exposed to live pathogens versus UV-inactivated spores (Figure 25B), normalising by millions of total reads in each library. Statistical tests indicate significant upregulation of the assayed cluster by a factor of 2.16 in response to the live fungus as compared with inactivated spores. To investigate the origin and potential identity of the cluster, we used methods described elsewhere (“Functional enrichment analyses”) to query non-metazoan sequences in the SwissProt database using the HB003_g97330 cluster. The top hits include polyketide synthetases used by the fungus Sordaria araneosa to produce the antifungal secondary metabolites sordarin and hypoxysordarin (Kudo et al.2016). According to a recent review, “sordarins were perceived as one of the more promising anti-fungal agents to fight invasive fungal infections. There has been significant development of a vast number of sordarin derivatives displaying extraordinary in vitro and in vivo efficacy, with high specificity toward numerous fungal species and very low toxicity, thus making sordarins much safer than the other drugs used nowadays. Importantly, sordarins display a unique modus operandi targeting the fungal translational machinery exclusively…” (Shao et al.2022 p.1126). We conclude that in response to challenge by R. denticulospora, the rotifer H. ligula significantly upregulates expression (by approximately twofold) of at least one biosynthetic gene cluster linked to antifungal bioactivity. The genera Adineta and Habrotrocha belong to different bdelloid orders, but both show resistance to fungal pathogens and upregulation of biosynthetic gene clusters, indicating that these traits generalise across bdelloid rotifers. We also found that although Habrotrocha did not show evidence of differential resistance (or indeed susceptibility at all) to R. globospora, we were able to select a different pathogen (R. denticulospora) from our challenge panel that did so strongly. This demonstrates the advantage of combining multiple rotifer and pathogen species in controlled cross-inoculation assays, which help to identify microbe-resistant species of rotifers in culture together with suitable pathogens to which they may be exposed in order to induce production of compounds. References: Hadfield, J. D., B. R. Krasnov, R. Poulin, S. Nakagawa, A tale of two phylogenies: comparative analyses of ecological interactions. American. Naturalist.183, 174–87 (2014). Kudo,F., Matsuura,Y., Hayashi,T., Fukushima,M. and Eguchi,T. Genome mining of the sordarin biosynthetic gene cluster from Sordaria araneosa Cain ATCC 36386: characterization of cycloaraneosene synthase and GDP-6-deoxyaltrose transferase J. Antibiot.69, 541-548 (2016) Shao, Y., Molestak, E., Su, W., Stankevič, M., & Tchórzewski, M. (2022). Sordarin—An anti‐ fungal antibiotic with a unique modus operandi. British Journal of Pharmacology, 179(6), 1125- 1145. Example 5 – Extraction of secondary metabolite enriched lysates from challenge-primed rotifers and in-vitro activity assays against pathogenic fungi Elsewhere, we have demonstrated that applying a controlled challenge by the fungal pathogen Rotiferophthora globospora to populations of the bdelloid rotifers A. vaga and A. ricciae induces strong upregulation of orthologs of the biosynthetic gene cluster ‘BGC 48151’. We further found marked and significant differences in the extent of upregulation and the number of encoded copies, with A. ricciae showing substantially higher expression and copy number than A. vaga. We further showed that differential expression of this cluster is the single most marked functional difference in the transcriptomic response to fungal attack between the two species. BGC 48151 is elsewhere characterised as encoding a hybrid PKS-NRPS, and its biosynthetic product is predicted as a secondary metabolite with substantial similarity to compounds with known anti-microbial activity. As demonstrated elsewhere, A. ricciae shows threefold higher resistance to the fungal pathogen than A. vaga. Together, these results demonstrate that exposing the rotifers to the fungus causes overexpression of a gene cluster encoding a biosynthetic enzyme, and further indicate that this enzyme enables rotifer cells to produce one or more secondary metabolite compounds with anti-microbial activity, which in turn underpins the strong resistance of A. ricciae to the fungus, and the weaker resistance of A. vaga to the same fungus. Here we describe experiments and methods to induce production of the target compound in both A. ricciae and A. vaga populations through pathogen challenge, then to lyse the primed, challenge-enriched rotifers and their cells to release the compound, then to separate the compound from rotifer tissue, bacteria and other insoluble fractions, then to assay and quantify the anti- microbial activity of the resulting challenge-enriched clarified lysate against the original fungal pathogen in vitro, through comparison with rotifer lysates that had not undergone challenge-primed enrichment for biosynthetic enzymes and secondary compounds. According to methods described elsewhere (‘RNA-Seq Experimental Design’), hundreds of thousands of rotifers of each species were reared in Petri dishes or tissue-culture flasks, then counted, washed and combined into centrifuge tubes. Rotifers were gathered into dense pellets by centrifugation, then inoculated either with live or UV-inactivated R. globospora spores, with a spore density equivalent to ~100 spores per animal. After inoculation, the primed rotifers were incubated at 20°C for thirty hours to allow the challenge to initiate upregulation of the BGC and synthesis of its product according to the timeline established previously (Figure 1c). Tubes were then centrifuged (Eppendorf 5810R) at 4000 rpm for 5 minutes to create a pellet; excess liquid was removed to leave approximately 0.8mL. To release compounds from the rotifers, tubes were placed inside a frozen aluminium block and 24 rounds of 23Khz ultrasonication were applied in pulses of 10 seconds. Lysates were clarified of insoluble debris by centrifugation and retrieval of the supernatant; products within the lysate were then concentrated by pooling multiple tubes and applying a vacuum concentrator (SpeedVac, ThermoFisher Scientific). Concentrated lysate was further clarified and sterilised using a 0.2µm syringe-mounted microfilter. An agar disc-diffusion assay (e.g. as reviewed by Balouiri et al., 2016) was employed to test and compare bioactivity of rotifer extracts against a lawn of the fungal pathogen R. globospora. Mycelium was grown according to methods described elsewhere (“Rotifer and pathogen isolates”), then fragmented by vortexing in water in association with stainless steel beads and spread as an even lawn across replicate 90mm PDA plates. The plates were allowed to dry, and sterile blank 5mm filter discs were positioned on the surface. To each disc was applied 20µL of sterile lysate from a single rotifer treatment, such that each treatment combination (rotifer species, live versus sham challenge) was applied to each plate in a blocked design, together with a positive control (20% UniSafe® disinfectant, a biocidal preparation of alkyldimethylbenzylammonium chlorides) and negative control (sterile water) on each plate (Figure 20). Immediately following preparation of the disc assay plates (T=0), a SPOT Insight camera was used in combination with a Nikon Olympus SMZ1500 dissecting microscope to take standardised images of each disc individually. Plates were then incubated for 4 days at 20°C to give the fungus an opportunity to grow before discs were re-imaged (Figure 20B). The analysis software ImageJ was used to measure the light intensity in each image relative to T0, and thereby quantify the extent of fungal growth inhibition around each treated disc (e.g. Alonso et al.2017). Higher intensity indicates less occlusion by fungal mycelium and provides evidence of bioactivity of the corresponding disc treatment, inhibiting growth of the fungus relative to the negative control disc. Results from two replicate assays (Fig 21A) demonstrate the efficacy of this method for asaying antimicrobial activity. When analysed using a linear mixed-effects model (Fig 21B), highly significant inhibition was recorded for the biocidal preparation relative to the negative control (P = 1.75 x 10-7). Significant inhibition of fungal growth also was demonstrated for the challenge-primed extract from the bdelloid rotifer A. ricciae, and not for any of the other extracts, including A. ricciae where no live pathogen challenge had been applied, which was statistically indistinguishable from water. Lysate from the less resistant rotifer species (A. vaga) also is indistinguishable from water, but challenge-primed A. vaga lysate shows a higher median inhibition than the unprimed lysate (Fig.21A). These patterns are consistent with successful extraction from the rotifers into the lysate of a compound with antimicrobial activity against R. globospora, which is produced most strongly by A. ricciae after challenge, and less strongly by A. vaga after challenge, but is not expressed in unprimed rotifers. This pattern closely matches the expression results obtained elsewhere for BGC48151 and its hypothetical antimicrobial product as well as the relative resistances of the two rotifers to attack in vivo by the fungal pathogen that was here assayed in vitro. References: Alonso, C.A. et al. (2017) ‘Antibiogramj: A tool for analysing images from disk diffusion tests’, Computer Methods and Programs in Biomedicine, 143, pp.159–169. Balouiri, M., Sadiki, M. and Ibnsouda, S.K. (2016) ‘Methods for in vitro evaluating antimicrobial activity: A review’, Journal of Pharmaceutical Analysis, 6(2), pp.71–79. Example 6 – Resistance of bdelloid rotifers and upregulation of horizontally-acquired biosynthetic gene clusters in response to challenge by fungal pathogens from three genera in the phylum Ascomycota and one in the phylum Basidiomycota Various features and advantages of the present invention have been illustrated in evidence and examples elsewhere. For instance, animal-encoded sources of antimicrobial secondary metabolites are vanishingly rare; these are expected to have advantageous safety and toxicity profiles given their production within metazoan cells. Bdelloid rotifers have conducted tens of millions of years of natural ‘bioprospecting’ via HGT, acquiring biosynthetic gene clusters from diverse organisms and environmental sources, adapting them for expression in their own cells against microbial challenges, then ‘screening’ the resulting natural products for antimicrobial activity. Given this history and their taxonomic diversity, bdelloid rotifers offer a large pool of novel natural products to be developed for clinical and agricultural applications. Elsewhere we show that at least 40 biosynthetic gene clusters may be identified by looking at just one species (A. vaga). At least one BGC in A. vaga (AVAG|g23567) appears to have been captured so recently that it is not shared by the related species A. ricciae and is thus a potentially unique acquisition. Figure 26 shows further evidence of the recent capture of this biosynthetic gene cluster. AVAG|g23567 was not located in a subtelomeric region of Av20, but approximately 6.8 Mb along Chromosome 3 (Fig.5c; black arrow). It was also unusual among upregulated annotations in apparently lacking orthologous copies in other bdelloids. Contamination can be excluded, as it was assembled almost identically in Av13 and the Hi-C-scaffolded Av20, with consistent linkage to flanking metazoan genes. The cluster also has few or no introns. Together, this evidence suggests a recent acquisition. We further tested this hypothesis by measuring the guanosine-cytosine (GC) content of the cluster relative to the local genomic context. Elevated GC content is predicted for recently acquired sequences that have yet to converge with the genomic background. We tested this for AVAG|g23567 by calculating %GC in a 20kb sliding window along its Av13 contig (AVAG00146). The focal annotation sits at the global maximum of the resulting landscape (Figure 26); no equivalent window in the 365 kb contig shows such high GC content (38%). We used the same method to assess %GC in Av20, where AVAG|g23567 corresponds to cluster 3.4 on Chromosome 3 (the longest chromosome, 20.35 Mb). Again, the maximum GC content (38.1%) for this contig occurs within the focal NRP / PKS cluster. The next-highest 20kb window has a substantially lower value (37.2%). Cluster 3.4 thus underpins the most marked peak in GC content seen on Chromosome 3, consistent with other evidence that it has recently been acquired from a bacterial genome. Other bdelloid species are likely to yield similarly unique biosynthetic gene clusters as a result of recent acquisitions. We have shown that bdelloid rotifers of different species and genera encode and can be induced to upregulate various biosynthetic gene clusters upon challenge with pathogens of the ascomycete genus Rotiferophthora, according to methods described elsewhere. Here we show that the same mechanisms of rotifer resistance can be induced in response to multiple fungal pathogens from different groups of agricultural concern, including those in the distantly related phylum Basidiomycota. Basidiomyetes rank highly among fungal pathogen taxa of concern in relation to economically important crops (Savary et al.2019), such as rusts of wheat (Puccinia spp.) and soybeans (Phakospora pachyrhizi). We used methods described elsewhere (“Infection assays”) to test multiple combinations of bdelloid rotifers against fungi of different genera, first assessing phenotypic resistance to the pathogen and then extracting RNA from populations of partially resistant rotifers under experimentally controlled conditions (“RNA-seq experimental design”), and mapping reads as described elsewhere against panels of candidate BGCs retrieved by BLAST from draft genomes (“Putative NRP / PKS screen and genomic validation”). In every case where we found rotifers with differential resistance against the relevant pathogen (e.g. Fig.27 for the basidiomycete pathogen Triacutus subcuticularis), and wherever we extracted and analysed RNA, we found evidence of upregulation of horizontally acquired BGCs (Figure 28). This is further evidence that the results generalise across multiple bdelloid rotifer species and genera, but it also extends the findings to multiple fungal genera (including a basidiomycete pathogen) and demonstrates that the methods can be applied reliably and repeatedly on demand to reveal and induce expression of biosynthetic clusters with products linked to antimicrobial activity. In many of the cases summarised in Figure 28, evidence of resistance and BGC upregulation has been discussed in previous Examples (Adineta vaga ‘AD008’ and Adineta ricciae versus R. globospora in Examples 1 and 2; Adineta steineri versus Rotiferophthora sp. ‘ARSEF 10274’ in Example 3; Habrotrocha ligula versus Rotiferophthora denticulospora in Example 4. For A. vaga (clone ‘AD002’) and for A. steineri (clone ‘AS23R’), lists of BGCs that were assessed as differentially expressed in response to challenge with either R. globospora or R. sp. ‘ARSEF 10274’ are shown in Figure 29. These BGCs are a subset of a larger panel named and ordered according to their closest orthology with the A. vaga clusters whose genomic locations are shown in Figure 5c. Differential expression was measured according to the reads-mapped-per- million (RPM) metric described elsewhere, and is reported for all orthologs where RPM differed significantly between rotifers exposed to live versus UV-inactivated pathogen spores, regardless of the size of the significant difference, or which pathogen induced it (where both did, the larger difference is shown). To compare read counts between treatments, linear models were fitted with treatment (live versus inactivated pathogen) as the fixed factor. To account for testing multiple clusters, we adjusted P-values using the Benjamini-Hochberg method (Benjamini and Hochberg, 1995), and then used a Tukey-HSD test for post-hoc pairwise comparisons between relative gene expression levels, given that each line represents two contrasts between the same rotifer and two different pathogens. Only BGCs that were significantly differentially expressed according to both adjusted tests are shown in Figure 29. There is evidence that 6 BGCs are upregulated for A. steineri and 7 for A. vaga (clone ‘AD002’). In both cases, the putative ortholog of ‘BGC 48151’ (cluster 5.2 in A. vaga ‘AD008’) is significantly upregulated. As already shown elsewhere for A. steineri, putative partial orthology of a BGC to A. vaga is compatible with substantial divergence in both sequence and predicted metabolite product of a given BGC. Our experiments also showed that fungal genera other than Rotiferophthora can induce upregulation of horizontally acquired biosynthetic gene clusters. A. vaga was challenged with two fungal pathogens in different ascomycete genera to which it shows partial resistance: Harposporium spirosporum (Ophiocordycipitaceae, Hypocreales), and an undescribed pathogen species belonging to an undescribed genus that superficially resembles Harposporium but is unrelated: Gen. nov. cf. Harposporium (incertae sedis, Hypocreales). We identify this fungus by its strain code: “ARSEF 14199”. Differential expression was assessed across the panel of BGCs using the RPM metric described above, and the Benjamini-Hochberg correction for multiple tests. Figure 30 shows a heat- map of relative upregulation or downregulation of BGCs in response to challenges from different genera, alongside differential expression versus R. globospora. Some BGCs respond to different pathogens in similar ways, but other clusters respond in opposite ways or are only induced in response to certain fungi. This is consistent with a model in which some anti-microbial mechanisms show broad activation and activity, with others induced more specifically against certain fungal taxa. This indicates a further advantage when using the methods described here to induce and derive novel natural products from bdelloid rotifers, in that the search may be tailored according to whether a broad-acting or targeted spectrum of anti-microbial (in this case anti-fungal) activity is desired. The undescribed ascomycete fungal pathogen (N. sp.14199) was unusually effective in inducing BGC upregulation in A. vaga (11 clusters significantly upregulated, including 8 apparently unique to this challenge). Together, this evidence further supports the robustness and advantages of the innovation. We investigated the responses of A. ricciae and A. vaga to a basidiomycete fungal pathogen, Triacutus subcuticularis. A panel of genomically encoded BGCs was identified for each of A. vaga and A. ricciae according to methods described elsewhere (“Putative NRP / PKS screen and genomic validation”), and their expression was investigated when the two species were challenged with the pathogen (Figure 30A,B). We found at least one BGC in A. ricciae that was downregulated in response to the original ascomycete fungus and therefore would not have been amenable to further study, but was upregulated in response to challenge with the basidiomycete. A. vaga had a BGC that was upregulated versus the basidiomycete but not three ascomycetes, further demonstrating that different challenges can induce and reveal different clusters and compounds. We investigated predicted chemical products of two novel BGCs upregulated significantly by A. ricciae versus the basodiomycete T. subcuticularis using methods described elsewhere (“Secondary metabolite prediction”). The BGC corresponding to ARIC|g46825 was upregulated exclusively in response to the basidiomycete, whereas the BGC corresponding to ARIC|g51902 was upregulated in response to both tested pathogens. These BGCs are preducted to synthesise novel secondary metabolites with chemical similarity to surfactins (Fig.31A) and fengycins (Fig.31B). Fengycins and surfactins are lipopeptides synthesised by bacteria (Bacillus spp.) with powerful antifungal activity. Fengycin production is a key mechanism of biocontrol by Bacillus against Monilinia spp., one of the priority pathogens of stone fruit crops (Yánez-Mendizábal et al., 2012). Fengycin and surfactin induce plant systemic defenses (Ongena et al., 2007) and act synergistically to inhibit the growth of Phytophthora infestans (Wang et al., 2020). This is not a fungus but an oomycete pest of potato. This evidence suggests that antimicrobial effects of rotifer-synthesised metabolites could extend to economically important oomycete pathogens as well as ascomycete and basidiomycete fungi. Surfactin itself is toxic to animal cells, but its promising antimicrobial and anticancer properties have inspired ongoing research to mitigate toxicity by artificially altering its structure or formulation (Wu et al., 2017). Deriving a related molecule from rotifers that is evolution-tested for reduced toxicity in animal cells would have advantages in this area. References: Savary, S., Willocquet, L., Pethybridge, S.J. et al. The global burden of pathogens and pests on major food crops. Nat Ecol Evol 3, 430–439 (2019). https: / / doi.org / 10.1038 / s41559-018-0793-y Benjamini, Y., & Hochberg, Y. (1995). Controlling the false discovery rate: a practical and powerful approach to multiple testing. Journal of the Royal statistical society: series B (Methodological), 57(1), 289-300. Ongena, M., Jourdan, E., Adam, A., Paquot, M., Brans, A., Joris, B., Arpigny, J. L., & Thonart, P. (2007). Surfactin and fengycin lipopeptides of Bacillus subtilis as elicitors of induced systemic resistance in plants. Environmental Microbiology, 9(4), 1084–1090. https: / / doi.org / 10.1111 / J.1462- 2920.2006.01202.X Wang, Y., Zhang, C., Liang, J., Wang, L., Gao, W., Jiang, J., & Chang, R. (2020). Surfactin and fengycin B extracted from Bacillus pumilus W-7 provide protection against potato late blight via distinct and synergistic mechanisms. Applied Microbiology and Biotechnology, 104(17), 7467– 7481. https: / / doi.org / 10.1007 / S00253-020-10773-Y / FIGURES / 9 Wu, Y. S., Ngai, S. C., Goh, B. H., Chan, K. G., Lee, L. H., & Chuah, L. H. (2017). Anticancer activities of surfactin potential application of nanotechnology assisted surfactin delivery. Frontiers in Pharmacology, 8(OCT), 296621. https: / / doi.org / 10.3389 / FPHAR.2017.00761 / BIBTEX Yánez-Mendizábal, V., Zeriouh, H., Viñas, I., Torres, R., Usall, J., de Vicente, A., Pérez-García, A., & Teixidó, N. (2012). Biological control of peach brown rot (Monilinia spp.) by Bacillus subtilis CPA-8 is based on production of fengycin-like lipopeptides. European Journal of Plant Pathology, 132(4), 609–619. https: / / doi.org / 10.1007 / S10658- TABLES Table 1 Number of HGTCin significantly up- and downregulated gene subsets in response to biotic (pathogen exposure) and abiotic (desiccation) stress. TestaSpecies Contrast DE HGTCcin: Odds ratio P-valuedsubsetbDE0 DE1 (95% CI) Pathogen A. ricciae T7 Up 6041 (11%) 196 (28%) 3.3 (2.7–3.8) <0.001 Down 68 (27%) 3.1 (2.3–4.2) <0.001 T24 Up 5738 (10%) 372 (23%) 2.7 (2.4–3.0) <0.001 Down 195 (24%) 2.8 (2.4–3.3) <0.001 A. vaga T7 Up 7918 (12%) 145 (32%) 3.4 (2.8–4.2) <0.001 Down 26 (29%) 3.0 (1.8–4.8) <0.001 T24 Up 7629 (12%) 285 (26%) 2.6 (2.3–3.0) <0.001 Down 175 (26%) 2.6 (2.1–3.0) <0.001 Desiccation A. vaga Entering Up 7975 (12%) 77 (14%) 1.2 (0.9–1.5) 0.24 Down 37 (10%) 0.8 (0.6–1.2) 0.29 Recovering Up 7553 (12%) 266 (14%) 1.3 (1.1–1.5) <0.001 Down 270 (20%) 1.9 (1.6–2.2) <0.001aTest: ‘Pathogen’, treatment groups represent animals exposed to live pathogen spores, versus controls exposed to inactivated spores, at 7h (T7) and 24h (T24) post exposure; ‘Desiccation’, treatment groups represent animals either entering into or recovering from desiccation, versus control animals that remained hydrated47.bDE subset: ‘Up’, genes that are significantly upregulated in the treatment group; ‘Down’, genes that are significantly downregulated in the treatment group. Significance is defined as genes with absolute fold change in expression > 4 and FDR < 1e–3.cThe number of HGTC in different subsets as follows: ‘DE0’, genes with no significant DE; ‘DE1’, genes either significantly up- or downregulated.dP-value for Fisher’s exact test (two tailed) for an association between classifications; null hypothesis: true odds ratio is equal to 1.

[0003] Table 2 Fixed and random effects for a linear mixed effects model testing for significant differences in magnitude of DE (normalised log2FC) between species. The reference levels for fixed factors were as follows: ‘Species’ = ‘Av’ (A. vaga), ‘DE category’ = ‘NS’ (i.e. genes with no significant change in expression, defined by the thresholds stated in the main text), ‘Timepoint’ = ‘T7’. Results are based on 249,392 observations from 124,696 genes. Random effects Groups Variance Std. Dev. Gene ID 0.2700 0.5197 Residual 0.5679 0.7536 Fixed effects Estimate Std. Error T-value Pr(>|t|) Signif. code (Intercept) 1.95e-02 3.57e-03 5.47 4.50e-08 *** speciesAr -1.44e-01 5.23e-03 -27.461 < 2e-16 *** DE_categoryDOWN -2.43E+00 9.22e-02 -26.379 < 2e-16 *** DE_categoryUP 2.35E+00 4.16e-02 56.403 < 2e-16 *** timepointT24 1.19e-01 4.18e-03 28.47 < 2e-16 *** speciesAr:DE_DOWN 8.57e-02 1.07e-01 0.798 0.4247 speciesAr:DE_UP 7.64e-02 5.33e-02 1.433 0.1519 speciesAr:timeT24 -1.18e-01 6.13e-03 -19.174 < 2e-16 *** DE_DOWN:timeT24 -1.12e-01 9.57e-02 -1.171 0.2417 DE_UP:timepointT24 -2.22e-01 4.52e-02 -4.909 9.18e-07 *** speciesAr:DE_DOWN:timeT2 . 4 1.90e-01 1.13e-01 1.678 0.0934 speciesAr:DE_UP:timeT24 2.34e-01 5.81e-02 4.02 5.82e-05 ***

[0004] Table 3 Pearson’s product-moment correlations in log2 fold-change expression. TestaSpecies TimepointbDE typecPearson’s R (95%-CI) P-value Timepoints A. vaga Both Up 0.84 (0.84–0.85) <0.001 Down 0.67 (0.66–0.68) <0.001 A. ricciae Both Up 0.82 (0.82–0.83) <0.001 Down 0.66 (0.65–0.67) <0.001 Within genomes A. vaga T7 Up 0.43 (0.42–0.44) <0.001 Down 0.50 (0.49–0.52) <0.001 T24 Up 0.43 (0.42–0.44) <0.001 Down 0.50 (0.49–0.51) <0.001 A. ricciae T7 Up 0.49 (0.48–0.50) <0.001 Down 0.39 (0.38–0.40) <0.001 T24 Up 0.60 (0.60–0.61) <0.001 Down 0.42 (0.41–0.43) <0.001 Between genomes Both T7 Up 0.35 (0.34–0.36) <0.001 Down 0.26 (0.25–0.27) <0.001 T24 Up 0.37 (0.36–0.37) <0.001 Down 0.40 (0.39–0.40) <0.001aExplanation of tests: “Timepoints”, correlation in differential expression (DE) in the same gene across timepoints T7 and T24 (Supplementary Fig.1); “Within genomes”, correlation in DE between inferred gene copies within genomes (including putative homologs, homoeologs and other gene copies) (Supplementary Fig.2a–d); “Between genomes”; correlation in DE between inferred orthologs across species (Supplementary Fig.2e–f).bTimepoints: “T7”, 7h post-infection; “T24”, 24h post-infection.cDE type: “Up”, genes that are upregulated in the treatment group (i.e., positive DE); “Down”, genes that are downregulated in the treatment group (negative DE). Note that all genes are included in correlations regardless of significance. Table 4 A Summary of a generalised (binomial) linear model testing for enrichment of HGTC in up- and downregulated subsets of genes following exposure to a fungal pathogen. The reference levels for fixed factors were as follows: ‘DE category’ = ‘NS’ (i.e. genes with no significant change in expression, defined by the thresholds stated in the main text). Results are based on 249,392 observations of 124,696 genes. Estimate Std. Error Z-value Pr(>|z|) Signif. code (Intercept) -2.057368 0.008989 -228.886 <2e-16 *** DE_categoryDOWN 1.087234 0.121459 8.951 <2e-16 *** DE_categoryUP 1.179946 0.065063 18.136 <2e-16 *** timepointT24 -0.024297 0.012843 -1.892 0.0585 . DE categoryDOWN:timeT24 -0.096948 0.135819 -0.714 0.4753 DE categoryUP:timeT24 -0.224416 0.079581 -2.82 0.0048 ** B Fixed effects for generalized linear models run separately for timepoints T7 and T24. The reference levels for fixed factors were as follows: ‘Species’ = ‘Av’ (A. vaga), ‘DE category’ = ‘NS’ (i.e. genes with no significant change in expression, defined by the thresholds stated in the main text). Timepoint T7 Estimate Std. Error Z-value Pr(>|z|) Signif. code (Intercept) -1.98809 0.01198 -165.911 < 2e-16 *** DE_category.DOWN 1.10305 0.23341 4.726 2.29e-06 *** DE_category.UP 1.23798 0.10148 12.200 < 2e-16 *** Species.Ar -0.15337 0.01813 -8.461 < 2e-16 *** DE_category.DOWN:species.Ar 0.03756 0.27345 0.137 0.891 DE_category.UP:species.Ar -0.05868 0.13242 -0.443 0.658 Timepoint T24 (Intercept) -2.00893 0.01219 -164.768 < 2e-16 *** DE_category.DOWN 0.96111 0.08870 10.836 < 2e-16 *** DE_category.UP 0.96686 0.06996 13.819 < 2e-16 *** Species.Ar -0.16190 0.01851 -8.744 < 2e-16 *** DE_category.DOWN:species.Ar 0.08080 0.12185 0.663 0.507 DE_category.UP:species.Ar 0.01790 0.09273 0.193 0.847 Table 5 Fixed effects for a generalized linear model testing for significant enrichment of HGTC in up- and downregulated subsets of genes while entering and recovering from desiccation. The reference levels for the fixed factor ‘DE category’ = ‘NS’ (i.e. genes with no significant change in expression, defined by the thresholds stated in the main text). Entering desiccation Estimate Std. Error Z-value Pr(>|z|) Signif. code (Intercept) -1.97337 0.01195 -165.126 < 2e-16 *** DE_category.DOWN -0.19646 0.17393 -1.130 0.259 DE_category.UP 0.14130 0.12332 1.146 0.252 Recovering from desiccation (Intercept) -1.99608 0.01226 -162.771 < 2e-16 *** DE_category.DOWN 0.63605 0.06932 9.176 < 2e-16 *** DE_category.UP 0.23939 0.06752 3.546 0.000392 ***

[0005] Table 6 Proportion of A. vaga DE genes shared between pathogen and desiccation conditions. Values in bold on diagonals indicate the total number of genes in each category (e.g., 136 HGTC genes upregulated at T7). Percentages in each cell then indicate the proportion of shared genes relative to the row total (e.g., 89% of 145 T7 upregulated HGTC genes are shared with T24 upregulated set). Green and red shading is to highlight the relative size of the shared fraction for up- and downregulated subsets, respectively. Gene type DE type Pathogen (T7) Pathogen (T24) Entering Recovering All DE genes Up T7 452 89.6% 7.1% 21.9% All DE genes Up T24 37.1% 1093 9.3% 20.9% All DE genes Up Ent 5.7% 18.3% 558 62.7% All DE genes Up Rec 5.5% 12.6% 19.4% 1807 All DE genes Down T7 89 66.3% 2.2% 0.0% All DE genes Down T24 8.8% 674 2.1% 11.3% All DE genes Down Ent 0.6% 3.9% 361 26.3% All DE genes Down Rec 0.0% 5.7% 7.2% 1322 HGTCUp T7 145 89.0% 9.7% 22.8% HGTCUp T24 45.3% 285 10.9% 22.1% HGTCUp Ent 18.2% 40.3% 77 57.1% HGTCUp Rec 12.4% 23.7% 16.5% 266 HGTCDown T7 26 76.9% 3.8% 0.0% HGTCDown T24 11.4% 175 2.3% 14.9% HGTCDown Ent 2.7% 10.8% 37 40.5% HGTCDown Rec 0.0% 9.6% 5.6% 270

[0006] Table 7 Fixed and random effects for a linear mixed effects model testing for an association between putative NRP / PKS genes and three genomic features of interest (a gene density, b transposable elements, and c telomeric repeats) as measured by the proportion of base pairs occupied by these features in 25 kb flanking regions surrounding putative NRP / PKS genes compared to BUSCO genes. A Gene density model output. The reference levels for the fixed factor ‘Type’ = ‘BUSCO’ Random effects Groups Variance Std. Dev. Gene ID 0.0001711 0.01308 Residual 0.0329041 0.18139 Fixed effects Estimate Std. Error T-value Pr(>|t|) Signif. code (Intercept) 0.53400 0.01002 53.313 0.0113 * Type.NRPS -0.09479 0.01327 -7.142 1.24e-12 *** B Transposable elements model output. The reference levels for the fixed factor ‘Type’ = ‘BUSCO’. Random effects Groups Variance Std. Dev. Gene ID 3.37e-05 0.005805 Residual 1.93e-02 0.138912 Fixed effects Estimate Std. Error T-value Pr(>|t|) Signif. (Intercept) 3.984e-02 5.314e-03 7.497 0.0771 . Type.NRPS 7.657e-02 1.127e-02 6.794 1.54e-11 *** C Telomeric repeats model output. The reference levels for the fixed factor ‘Type’ = ‘BUSCO’. Random effects Groups Variance Std. Dev. Gene ID 0.03866 0.1966 Residual 0.42313 0.6505 Fixed effects Estimate Std. Error T-value Pr(>|t|) Signif. (Intercept) -6.98263 0.13973 -49.97 0.0127 * Type.NRPS 0.71584 0.04617 15.50 <2e-16 *** Table 8 Evidence for the hypothesis that upregulation of HGTC-enriched effectors might be paired with downregulation of HGTC-enriched regulators or repressors. Occurrence of the word “regulation” in the text of GO terms that are significantly enriched among differentially expressed HGTC genes. This word appears only in terms enriched among downregulated genes, and at a significantly higher rate in the pathogen response than in the desiccation response, where HGTCare not significantly enriched (e.g. Fisher’s exact test for A. vaga T24 downregulated versus desiccation recovery downregulated: P < 0.0001). A. ricciae A. vagaPathogen Pathogen Pathogen Pathogen Desiccation T7 T24 T7 T24 recovery UP DOWN UP DOWN UP DOWN UP DOWN UP DOWN “regulation” in term 0 14 0 13 0 18 0 18 0 1 Total GO terms 15 28 29 33 8 41 18 59 4 55 FDR<0.001 Proportion of terms 0 0.5 0 0.39 0 0.44 0 0.31 0 0.02 Supplementary Note: The hypothesis that HGTC expression is particularly important in pathogen defence might initially seem to predict stronger enrichment among upregulated versus downregulated genes. Although upregulated HGTCgenes outnumber downregulated in every condition, HGTC enrichment was proportionally the same in both directions. One explanation—that HGTC are more loosely regulated and fluctuate in expression more than metazoan genes—can be rejected because rotifers entering desiccation showed no such pattern, and HGTC enrichment in recovering rotifers was asymmetric in favour of downregulated genes. A second possibility is that rotifers often regulate HGTC products using HGTC regulatory proteins, either because these were acquired together or became ‘wired’ together in their new metazoan context. If so, then upregulation of HGTC-enriched effectors might be paired with downregulation of a smaller set of HGTC-enriched negative regulators or repressors. This explanation has some support: the word “regulation” appears in 40% of GO terms enriched among HGTCthat are downregulated in response to pathogens, but not at all among upregulated genes (Supplementary Table 7), Moreover, it only appears once among the 55 GO terms (2%) enriched among HGTC that are downregulated during desiccation recovery, where there was only very weak enrichment of upregulated HGTC(Fig.3b). Thirdly, perhaps a set of HGTC genes performing other functions is downregulated to divert resources to the immediate threat posed by the fungus. This could include other defensive genes, either constitutively expressed against microbial attack in general, or perhaps responding to bacteria present in the cultures. In nematodes, for instance, induction of antifungal defences is correlated with repression of antibacterial immune response genes68, perhaps to focus resources on a more pressing threat or balance biochemical trade-offs in defence mechanisms69. This hypothesis could help explain why, at T24, A. ricciae significantly downregulated 5 HGTC NRP / PKS CDS (Fig.5d) that had been substantially expressed under control conditions, and which BLAST to prokaryotic antimicrobial synthetase clusters. The magnitude of this downregulation was higher than the single significantly downregulated NRP / PKS CDS in A. vaga (Fig.5d, Supplementary Data 2), consistent with the hypothesis that A. ricciae requires stronger downregulation of constitutively expressed NRP / PKS to match its stronger upregulation of the focal cluster. HGTCenrichment among downregulated genes might therefore reflect a combination of pairing between HGTCregulatory and effector proteins, and trade-offs among pathogen-specific defensive HGTCsets.

[0007] noiss20 1 9 0 4 1 0 2 9 0 3 1 0 80 50 70 19 99 60 00 60 01 90 51 41 31 80 40e 9 9 9 9 8 0 9 1 1 1 1 1 1 1 1 1c 96 17 96 17 96 96 96 96 96 17 96 17 1 1 1 1 1 1 1 1c 4 4 4 4 44 44 44 4 7 7 7 7 7 7 7 7a4 4 4 4 4 44 44 44 44 44 44 44 44 44 44 44 44A R R R R R R R R R R R R R R R R R R R RR S R E R E R E R E R E R E R E R E R E R E R E R E R E R E R E R E R E R E R E R E ) % a6%1%2%4%5%3%0%8%3%8%8%7%4%6%1%3% % % %wr. . . . . . . . . . . .2 4 3 2(19 19 59 29 29 29 98 19 39 19 37 6.8 3.9 0.9 2.9 3.9 6.8 3.9 3.9 49 %)d 00 63 06 89 46 44 89 64 4 4 6 0 0 2 6 4 8 0 0 8e 2 4 5 96 1 3 8 3 8 2 1 3 8 9 3r , , , ,3,7,2,6,9,3,9,4,2,9,2,0,7,4 7 4et61 34 72 01 88 37 14 67 3 4 9 3 2 2 7 3 1,8,7,1li7 2 7 7 1 3 6 361082770524505103021552f( ,0,0,0,4,3,2,4,2,0,2, , , , , , , , ,6,se1s30 ,7189 8 1 0 4 4 0 08 49 11 02 25 73 82 75 99 91 3,4,03,55,43,64,23,44,43,94,5,6,4,7,0,0,8,2,6,a 2 4 3 4 3 5 4 3 3 3 B ) d 43 67 08 47 67 27 00 65 05 47 64 03 67 27 80 46 8 0 2 4er6 0 9 9 9 0 3 8 4 7 7 18 4 3 3e, , ,5, , , , , ,8 3 7 8 8 6 7t7 4 9,4 7 4 7 4 6 3,6,9,1,0,6,3,9,9,5,8li222022197662016818702 9 4 3 6 6 7 5 0 0f( ,s6, , , , , , , , ,6,8,2,4,0,7,5,1,0,4,6 49 67 10 1 1 7 88 27 48 86 88 95 19 27 88 57 00 0 7 6 2 1 8 7 6 7da e R ) d 4 0 4 6 6 8 8 2 0 2 2 6 0 4 6 6 2 6 0 8e 55089592320833080025118 6 7 4 2 6 4 4 8 ,, , , , , ,3 7 3 4 1 9 2 8 5m 6 7 5, , , , , , , , , , , , ,3 9 3 52 84 37 61 49 52 64 15 75 18 30 69 69 44 4 7 8mir3t,2(2,55,05,75,04,90,02,1 4 3 7 6 0 5 4 06865676,5,1,9,0,3,9,7,5,2,4,2,9s75904954383799553637793 2 2 2 2 2 7 8 8es ,3,5,3,5,3,4,3,4,3,4,23,85,83,04,34,65,04,44,73,3a B ) d 85 44 44 87 00 25 04 25 8 2 8 8 2 8 4 2 8 8 2 8e3 0 7 21 7 6 1 1 3 3 2 5 5 9 8 ,, , ,2,0,1,0,9,4,1,5,0,1,6,8,0,6 7 5m 1 7 6 2 0 0 4 0 6 7 3 5 1, , ,6 2 2 1 2 7 4 1 4 0 2 9 4 2 5 6 2mir4t,9 921 7 6 8 1 7 6 63694066535641566181,1,8,9,6,4,9,1,2,4,9,4, , , , , , , ,(7 0 7 0 7 9 7 6 6 0 6 2 1 9 5s1 1 9 7 9 7 01 7 9 8 01 9 8 6 7da e R ) 28 20 29 62 24 27 69 85 0 8 0 6 2 2 6 0 8 6 8 2war 2 8 2 5 6 4 8 7927117115508529088160219( ,0,2,5,7,1,9,2,8,9,8, , , , , , , , , ,0 5 3 2 0 4 1 4 4 2 7 5 4 2 6 9s3 19 5 7 9 0 3 2 7 9 8 0 5 6 2d, ,0,2,5,6,9,4,7,8,8,0,3,5,4,9 5 5 7 8a 2 0 7 5 0 2 3 5 0 7, , , , ,e 7 30 8 01 7 9 6 7 1 7 3 2 0 6 1 1 8 9 7 9 8 0 7 9 8 0 9 8 7 7 R 1 1 lo l r o l r orl l l l l l t t t t t t oro o l l l t t t o o o n n n n n nr ro o o n n nt tr r re e e e e et t tr r ro no ntnt tn n nt t te ep uc coco no nomotmtmtmtmtmtocococn n nm mrh4 h4 hc c c4 h h h h a 4erh a 4erh a a a a o 4erh erh erh e h4 h4 hcococt t4 h h h h a 4e h a 4 e G 2 2 2 7 7 7 2t2t2t7t7t7rt2 2 2 7 7 7 2rt2rtea ea ea ea ea ea e e e e e e ii i i i iaiaiaiaiaia cc c c c c cia a a a a a a a cic c c c c cccccccccccg g g g g g g g riririririririririririravavavavavavavavseiata a a a a a a a a a a a a a a a aetetetetetetetetet t t t t t t t tatatcen n n ne e e e e e e e e e epin n n n n n n n n n n n n n n ndididididididididididididididididididid S A A A A A A A A A A A A A A A A A A A ADIeb c d b c dl42 42 42 a b c 4 4 4 a c d b4 c4 d4 a b c a4 b4p 7 7 7 2 2 2 7 7 7 2 2 2 7 7 7 2 2 X1 X1 X1 X1 X1 X1 Y1 Y1 Y1 Y1 Y1 Y1 X X X X X X Y Ym 8 8 8 8 8 8 8 8a D D D D D D D D D D D D D D D D D D D D S A A A A A A A A A A A A A A A A A A A A 11 61 5 7 11 1 01 01 7 17 1 1 4 7 7 4 44 44 44 R R R R R E R E R E R E %6%2% % ..6 519 2.9 2.8 48 22 2 2 9 88 03 ,0,6,01 6 4,5 2 7 55647 ,, ,20,2 9 2 4 23 38 ,8,2 03 4,3,3 88 80 4 0 3 1 0 ,7 0 49,3,4,2 1 6 9 4 7 ,0,6,64 7,6 9 46 16 46 0 4 6 4 834 3 ,3,7,3,2 57 96 9 17 28 ,36,4,1,6 9 6 3 4 0 ,13,85,63,3 41 6 9 6 46 08 ,5,5 37, ,3 9 2 3 3 19428 ,, ,09,6 30 77 2 1 7 40 4 4 0 9 41624 ,, ,03,5 82 24 2 329 ,,2 907 5,0 8,2 1 7 7 tn tn tn t e e e ne mtmtmtmth ae a a a4 h e 7th e e 2rtr7rth 7rtag ag a a ava g v a g v avatatataetn enene iidind did A A A A c42 a7 c7 d7 Y8 Y8 Y8 Y8 D D D D A A A A %3 %7 % % % % % % % % % % % % % % % % % % % % % % 7 2 00 03 4 2 9 6 3 9 0 3 4 8 4 5 2 1 2 7 1 8 4 6 24 5 248165 9 7 0 3 1 1 7 8 1 8 1 5 5 9 6 6 5 .. . . . .0.2.6 5 9 4 5 5 1 8 5 7 0 9 8 6 0 51 1 0 2 1 2 3 2.1.2. . . . . . . . . . . . . .% 11 5 1 2 2 2 7 2 2 1 3 3 9 8 s e 00 05 05 00 05 05 05 05 00 00 0 0 0 0 0 0 0 0 0 0 0 0 0 0 s 5 0 1 0 8 8 0 8 7 4 57 5 0 0 0 5 5 5 5 5 5 5 0 0 a 4 4 0 1 3 0 1 7 9 36 3 92 8 8 5 92 64 9 0 2 2 b 2 7 9 9 7 0 6 34 2 61 66 79 3 5 8 5 3 9 9 6 3 0 5 9 8 4 9 8 1 7 3 9 9 5 57 00 0 8 9 0 7 7 0 8 7 d 6 3 9 2 1 4 3 6 1 0 1 6 8 5 9 02 4 8 6 1 8 0 5 2 e 4 7 1 2 7 2 2 0 6 2 8 8 5 2 8 5 45 1 2 5 6 3 4 2 p 1 1 1 1 1 1 7 7 3 9 5 1 p 4 2 1 1 3 1 1 1 3 3 a M s e 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 s 01 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 a 4 10 6 3 1 6 8 9 5 9 5 8 6 1 8 5 4 8 4 6 2 2 2 0 b 1 4 46 6 2 3 4 7 4 8 5 5 7 0 2 4 4 0 0 5 5 7 3 2 la0 t 5 6 1 7 7 7 5 1 3 8 0 1 0 7 4 3 6 0 8 0 2 4 6 5 4 3 1 9 9 0 1 0 1 5 4 7 2 6 9 9 9 0 7 2 6 8 2 2 6 3 4 6 4 1 6 8 9 2 8 4 0 1 6 8 4 9 5 7 8 1 7 1 2 9 o 6 1 0 5 8 7 40 2 8 9 4 0 6 7 7 9 7 2 3 4 1 6 1 4 T 3 5 4 5 3 4 4 64 6 7 0 3 8 8 0 3 6 1 5 8 5 2 9 6.3 4 4 5 3 4 4 5 4 4 3 3 3 5 3 3em % % % % % % % %o 3 7 0 0 4 2 9 6 %3 % % % % % % % % % % % % % % %n7e22 0 3 4 1 5 9 7 90 03 3 4 8 4 5 2 1 2 7 1 8 4 6g.4.5.2.8.6.0.2.6.5.91 .41 7 8 1 8 1 5 5 9 6 6 5 .5.5 1 8 5 7 0 9 8 6 0 5% 1 1 0 2 1 2 3 2 1 2 11 5 1.2.2.2.7.2.2.1.3.3.9.805P s d 09 1 30 0 7 7 1 7 4 8 5 95 62 60 85 77 71 17 95 3 9 1 4 4 a 4 84 8 28 74 10 22 78 55 23 33 9 9 9 8 2 6Oileo r02 17 99 75 92 89 77 22 43 60 93 9 7 3 0 15 00 10 79 87 02 43 57 10 71 45cd 9 41 3 42 41 42 42 12 21 42 69 7 7 5 1 1 8 4 8 3 5 0 3 6 9 4 1 52 71 03 07 22 41 51 72 83 07 26. eEp p ota M gnis d 28 20 29 62 2 2 6 8 0 8 0 6 2 2 6 0 8 6 8 2 4 4 4 0p a 2 8 2 5 4 7 9 5 9 7 1 1 5 0 5 9 8 1 0 1 0 4 6 4p0 2 5 7 61 4 8 7 2 1 7 1 5 8 2 0 8 6 2 9 9 1 2 0aer5 2 9 2 8 9 8 1 43 4 2 7 58 4 2 6 9 3 8 2 2 l 0 1 3 2 0 4 9 5 7 9 0 0 2 7 9 9 0 5 6 2 5 22 4 9maa3tt2 o 7 3 0 0 0 1 8 0 5 1 7 6 1 7 5 9 4 7 8 8 3 5 4 5 5 7 8 3 2 9 9 08 29 37 59 08 60 7 7 1 7 3 2 0 6 0 5 8 2 1 7 9 8 01 9 8 7 7 7 01 7 7adTqese- a e i a ea ea ea ea ea ea ea e e eAciccicicicicicicia cia cia cica ic c c c c c c c c c cg ag ag ag ag ag ag ag ag ag ag ag riririririririririririravavava a a a a a a a aNv v v v v v v v vRs eatatatatatatatatatatatatatatatata a a a a a a awie e e e e e e e e e e e e e e et t t t t t t tacrenin n n n n n n n n n n n n n nene e e e e e efp didididididididididididididididindin n n n n ndididididididoS A A A A A A A A A A A A A A A A A A A A A A A AnoDiItb c d b c d b c d a b creol4 p 2 42 42 a7 b7 c7 42 42 42 a7 c7 d7 42 42 42 a7 b7 c7 42 42 42 a7 c7 d7p X X X X X X Y Y Y Y Y Y X X X X X X Y Y Y Y Y Yom 1 1 1 1 1 1 1 1 1 1 1 1 8 8 8 8 8 8 8 8 8 8 8 8ra D D D D D D D D D D D D D D D D D D D D D D D D P S A A A A A A A A A A A A A A A A A A A A A A A A % %4 %4 %3 %9 %6 %0 %9 %8 %6 %8 %3 %7 %6 %0 %0 %6 %6 %6 %1 %1 %3 %0 %9 % 0 0 0 0 0 1 0 0 0 0 6 1 0 1 1 1 3 1 1 1 2 2 3 2 00 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 04 .. . . . . . . . . . . . . . . . . . . . . .00 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0.0 s e 46 72 98 52 93 42 62 25 04 86 40 77 47 42 43 29 44 81 53 84 22 45 8 8 s 9 4 9 9 8 3 4 3 9 6 9 2 9 1 9 5 1 4 0 0 8 5 9 a 6 3 1 9 3 6 2 3 7 0 1 5 1 7 3 3 4 7 7 4 3 4 0 b 2 9 9 3 0 2 1 3 0 7 6 6 0 3 6 0 4 1 5 1 9 5 2 1 1 9 4 2 4 3 3 2 3 8 7 2 4 3 8 4 6 3 4 5 7 5 9 d 1 7 9 2 2 e 1 1 1 p pa M s e 00 63 06 89 46 44 89 64 4 4 6 0 0 2 6 4 8 0 0 8 2 2 8 0 s 2 4 5 9 3 7 2 6 1 3 8 3 8 2 1 3 8 9 3 2 9 8 3 a 6 3 7 0 8 3 1 6 9 3 9 4 2 9 2 0 7 4 7 4 2 0 6 0 b 1 4 2 1 8 7 6 3 4 9 3 2 2 7 3 1 8 7 1 1 6 4 2 7 2 7 7 1 3 4 7 6 0 2 7 5 4 0 1 3 2 5 2 5 5 6 7.leat01 00 0 4 3 2 64 32 10 82 70 04 21 50 52 07 08 17 59 69 70 59 42 23m o 3 7 18 90 85 14 06 42 44 04 89 95 16 24 57 30 20 58 92 16 22 48 32 80o T 3 4 3 5 3 4 3 4 3 4 2 4 3 4 3 5 4 3 3 3 3 4 3 3neg% %4 % % % % % % % % % % % % % % % % % % % % 0 40 30 90 60 01 90 80 60 80 36 71 60 01 01 61 63 61 1 1 3 %0 %9 %2050 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 01 1 2 2 3 4P.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.0.00.0 0 0 0 0 00.0.0.0.0.0.0Oilsod 0 9 4 0 9 7 7 1 8 0 8 1 6 2ca 45 68 89 08 70 35 42 76 41 24 42 2 2 6 46 48 79 54 15 28 26 00 71 94.er2 3 1 8 4 8 6 6 4 7 7 3 0 7 2 0 8 3 1 2 0 6 1 8 3 51 4 8 7 61 82 21 7 8 51 91 52 52Ed ote p g pnaiMppas d 43 67 08 47 6 2 0 6 0 4 6 0 6 2 8 4 8 0 2 4 8 8 4 0m a 6t7 04 9 5 79 79 00 53 58 74 47 37 71 78 03 67 88 48 36 37 83 07 10 04aera l22 2 92 41 77 46 70 46 61 37 62 99 14 03 66 36 97 95 50 80 92 31 46 97 6 04 26 91 61 28 12 84 88 08 69 81 22 48 05 70 50 17 06 42 44 0 6 6d at6 9 7 0 7 8 7 8 6 8 5 9 7 8 7 0 8 7 7 4 1q6 7 6 9 6 6eo T 1 1s-A e e e e e e e e e eN a a a a a a a a a e e iciciciciciciciciaiaiaia a a a a a a a a a a aR cicicicicicicicc c c cic c c cga ga ga ga ga ga ga ga ga ga ga gader r r r r r r ririririr v v v v v v v v v v v vrseetiata a a a a a a a a a a a a a a a aletetetetetetetetetetetetetet t t tatatatatataticfenfpindindindindindindindindindin n n nenenenenenenenenenendididididididididididididididoS A A A A A A A A A A A A A A A A A A A A A A A AnoDiItb c d breol42 42 42 a7 b7 c7 4 c 2 4 d4 a c d b4 c4 d4 a b c a4 b4 c4 a c d p 2 2 7 7 7 2 2 2 7 7 7 2 2 2 7 7 7p X X X X X X Yom 1 1 1 1 1 1 1 Y1 Y1 Y1 Y1 Y1 X8 X8 X8 X8 X8 X8 Y8 Y8 Y8 Y8 Y8 Y8ra D D D D D D D D D D D D D D D D D D D D D D D D P S A A A A A A A A A A A A A A A A A A A A A A A A % %2 % % % % % % % % % % % % % % % % % % % % % % % 6 7 3 0 9 9 3 1 2 1 4 3 6 2 7 2 5 4 7 4 9 5 0 8 .6.2 1 8 2 5 1 8 2 7 6 7 2 0 3 6 2 9 9 8 7 4 00 0.0.1.0.1.1.1.0.1.5.2.0.1.1.1.3.1.0.0.1.1.4.4 s e 0 0 0 0 0 0 0 0 0 0 05 05 0 0 0 0 05 0 0 0 0 0 0 0 s 0 5 51 00 08 5 0 0 5 0 0 0 5 5 5 0 5 5 5 5 5 a 50 32 9 5 0 9 1 5 9 9 2 9 3 2 3 4 3 9 1 2 0 0 b 9 7 7 98 64 56 56 18 76 08 52 5 7 8 6 6 4 7 7 2 2 6 1 7 4 7 0 6 9 7 0 0 5 1 8 7 5 0 89 7 d 4 6 3 6 4 4 9 0 1 8 1 9 3 5 5 1 0 1 2 0 4 1 1 9 e 2 4 9 0 4 1 1 1 0 7 3 3 9 9 3 1 7 1 4 6 6 2 7 8 p 2 3 6 3 6 6 5 3 5 p 2 1 2 5 4 7 1 5 3 3 6 9 1 41 a M s e 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 0 0 0 s 1 1 6 3 1 6 8 9 5 9 5 8 6 1 8 5 4 8 4 6 2 02 0 0. a 4 0 4 6 2 3 4 7 4 8 5 5 7 0 2 4 4 0 0 5 5 2 0eb 1 4 6 7 8 7 4 3 6 0 8 0 2 4 6 5 4 3 1 9 7 3 2cnl0ea5 67 71 31 05 42 69 92 98 94 00 71 26 68 84 29 25 67 38 4 9 0 1 0 1 6 4 1 6rto 16 51 00 15 78 87 40 26 86 97 40 03 68 78 70 93 7 2 3 4 71 16 21 94efeT 3 5 4 5 3 4 4 4 3 4 4 5 3 4 4 5 64 14 53 83 53 25 93 63rA % %2 % % % % % % % % % % % % % 6 7632019892351128124736672 % % % % % % % % % % 2 7 2 5 4 7 4 9 5 0 8N.0.0.0.1.0.1.1.1.0.1.5.2.0.01.3 6 2 9 9 8 7 4 01. . . . . . . . .Rr1 3 1 0 0 1 1 4 4S s d 01 74 38 08 63 11 00 83 34 01 91 91 4 8 7 5 7 9 6 9 3 5 1 13 a 8 4 5 7 9 3 3 6 3 6 5 2 3 8 4 4 4 0 3 8 4 0 029 3 7 3 8 9 9 1 3 7 4 75 5 3 2 5 9 3 6 5 1 1 6 4 er4 9 8 1 8 2 3 2 0 5 3 6 1 0 3 0 2 5 1 9 2 9 90 4 6 1 2 6 2 2 0 6 1 6 8 8 9 7 2 1 2 8 2 2 4 3 75 d 1 1 1 1 1 4 72 5 11 8 41 43 0 6 7 3 8 4 9P e 1 1 1 3 2O pilpoacM.Es d 28 20 29 62 24 27 6 8 0 8 0 6 2 2 6 0 8 6 8 2 4 4 4 0 a 2 8 2 5 6 4 9 5 9 7 1 1 5 0 5 9 8 1 0 1 0 4 6 4ot er0 25 5 72 1 9 82 78 29 18 71 14 5 8 2 0 8 6 2 9 9 1 2 0 3 4 2 7 58 4 2 6 9 3 82 2 2gi l0 a 1 3 2 0 4 9 5 7 9 0 0 2 7 9 9 0 5 6 2 5 4 9n3 t 27 3 0 0 0 0 57 65 90 42 73 85 80 6 37 57 41 7 53 52 70 86 30 25 28 92p o 1 8 1 7 9 8 9 7 9 8 0 7 9 8 0 9 8 7 7 7 0 7 7pT 1 1 1 1amaeaa ea ea ea e e e e e e e et idcicicia cia cia cia cia cia cia cia cica i g ag ag ag a a a a a a a a cc c c c c c c c c c ca a a a ga ga g g g g g geriririririri i i i i ia a a a a ad r r r r r r v v v v v v v v v v v vrseetiata a a aletetet tatatatatatatatatatatatatatatatatatataticfene e e e e e e e e e e e e e e e e e e e efpin n n n n n n n n n n n n n n n n n n n n n ndidididididididididididididididididididididididoS A A A A A A A A A A A A A A A A A A A A A A A AnoDiItb c d breol42 42 42 a b c 4 c4 d4 a c d b4 c4 d4 a b c a4 b4 c4 a c d p 7 7 7 2 2 2 7 7 7 2 2 2 7 7 7 2 2 2 7 7 7p X X Xom 1 1 1 X1 X1 X1 Y1 Y1 Y1 Y1 Y1 Y1 X8 X8 X8 X8 X8 X8 Y8 Y8 Y8 Y8 Y8 Y8ra D D D D D D D D D D D D D D D D D D D D D D D D P S A A A A A A A A A A A A A A A A A A A A A A A A nin wohsn 8sa 9 91 06 76 98 48 21 7 3 2 7 0 2 0 e 41 91 5 93 1 4 4 79 53 96 83 79 93 31n m 2 3 4 4 1ae tm ntenemtamteareTrt / ln 1 o a 1 31 29 32 34 99 28 52 92 08 10 58 18 7 e 2 1 9 1 1 5 0 1 1 2rt1 n m oclo drt)ennao Cmoat ctad d 4 8 1 1 2 0 7 0 2 0 0 3 6 9uot7 6 Y 1 8 1 4 5 1 4 01 3 3 1 63 36 2 2 3 5 3 1 6 5 7 85 0 0 1 4 21 71 1nEu 8Do DcAecetac 4nec7 92 08 56 45 40 13 64 03 11 02 56 46 5ui7lY 2p8 1 61 4 93 1 5 72 74 5 5 l 4 8 64 01 31 1feDniR Ay.lbgna 7 8 8 5 1 9 9 4 6 4ii7 5sp Y 0 36 47 38 82 8 6 6 7 0 16 24 60 61sop8 2 2 4 5 81 83 33 5 93 6 11 1aDpAdlmsuc 6 2 2 0 8 3 4 9 5oda7 81 9 6 89 3 8 1 8 7 9 0 7c2 51 93 7 7 5 1 3terX 0 8 1 1 2ah0 Dt,5 A.PeOb 68 09 49 87 53 8 0 8 3 0 4 7 1 0.i(h 7tiX 1 6 7 81 11 38 8 81 32 81 2e8 1nDew gsAleanea 2g 7 6 65 91 11 65 63 05 99 95 48 34 06 25 4c2of X 2 1 1 1 C 8 1 1 1 3 7 1 1otT DG AgnH aa i1ps 1 1 1 2 1 1 1 1 1 1 1 1 2pg daaavem.rsA 0 5dad PeretOaloqeu Ns-gert n 7 7 7 7 7 7 4 A 2 42 42 4 4 4 4 4 NpuiT T T T T T 2 2 2 2 2 o T T T T T T T TRy pltedenamiretcT li ifin 1 1 1 1 1 1 1 1 1 1 1ffg .tis1.t0.t.t.t.t.t.1t.t.t.t1t o9 55 98 29 2 1 9t.0 5t.9 2 0.9.2rr1– 4ef9 3 1 3 8 5 94 53 07 81 93 17 24 98bo 4 1 14 22 31 41 15 91 14 62 22 31 04 52 4mssDtnIg| g| g| g| g| g| g| g| g| g| g| g| g 1 | g|unnu m e G G G G G G G G G G G G G Glaouln e A A A A A A A A A A A A A A V V V V V V V V V V V V V VtoCocG A A A A A A A A A A A A A ATa. nid naweroh 1s (s atnaaadtle t en 4 a 5 4 9 3 2 3 3 7 5 3 7 9 8 0 6 6 4 5 7 2 9 7 1 D e 4 6 7 2 73 4 6 0 3 6 7 35 6 6 0 2 4 6 5 3 7 3 1 5 5 2 4 28 68 1 2 4 51 0 4 3 19 2 3 4 44 9 4 1 7 8 7 2 7 8 76m nuamtotncnemlmtaatoe rf mteortt / fn a oerloitT orncton 1 5 3 2 6 7 7 5 8 0 9 6 9 2 6 3 4 9 1 2 2 8 2nitorufna 94 6 69 65 91 72 05 01 01 64 8 5 82 9 18 3 75 8 9 3 8 1 8cop n e 5 3 1 0 m 1downrolapln o krtataaitn n otn uC danth stiw d 11 22 72 61 80 40 05 88 28 64 39 18 26 17 65 77 51 87 9 8 0 6 0u b 7 o 9 3 7 7 9 8 6 4 2 2 5 0 1 4 3 2 0 7 4 6c usniY 8 7 7 7 4 3 2 3 1 3 4 3 74 4 3 8 9 0 5et aet1 5 2 1 6 Darcilo ofrA tplc 99 88 11 40 60 13 71 61 22 61 83 78 47 68 77 69 7 7 6 5 8 9 8pna7 4 1 7 4 5 7 9 4 3 8 3 6 2 6 2 6 3 4 1euciY 5 4 4 4 2 3 3 2 8 1 9 3 5 6 5 6 7 R.oc te1 1 1 2 3 2 3 1 6 h D gnciatA o pdlpy a 25 38 89 90 38 66 22 77 26 73 04 02 86 17 43 7 9 9 6 3 8 4 6puh7 9 2 6 0 7 0 6 9 6 5 4 9 2 8 0 2 6 3ao Y 7 6 6 7 4 3 1 4 7 0 4 2 3 0 4 4 1 8 9 0 ca1 2 3 4 4 4 2 7msd o dtD Aa ae er r rdalic 9 7 6 9 5 2 0 6 8 6 1 9 4 8 9 1 0 7 8 0 7 15 7 49 94 12 82 25 9 11 04 8 4 13 01 27 4 46 9 7 3 7 1 5 13 02 1 30emX 05vis1 1PiretDOhdso At-i05msb 9 3 4 0 5 6 6 2 3 0 4 7 4 8 1 7 7 7 6 1 0 7 14 6 07 83 91 42 54 41 7 03 6 4 82 7 38 1 9 2 8 4 8 64 13 5 1 6wPsiX 1 0 Oe1 1eneilngoeD g A c C.sEih a 5 4 8 6 7 2 5 8 3 5 2 3 8 9 7 0 5 4 8 3 3 7 35 5 32 08 71 03 45 7 31 76 21 7 62 8 88 4 8 6 6 5 8 27 7 9 2 4T GlaT.X 3 1 1 1 1 1Hitn de .tD A eaiet th xco gi etas 1 1 2 1 1 1 2 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 1ci plrhed.ah hgita erAcihnidehsid0 5taw)l ruo4esP O gf f souco Ner espnnaid t 7 7 7 7 7 7 7 7 7 7 7uege tn7 7 7 7 7 7 7 7 7 7 7 7 T T T T T T T T T T T T T T T T T T T T T T T m oio yltyl lnnn osip eaorci e tn dmoniTfih n Tcgsa1 1 1 1 1 1 1 1 1 1 1i.s1r1usm –ru eirt.te5.t6.t8.t4.t2.t8.t0.t t1t1t1t1 1t1 1t1t1t1t1t1t0.8.5t. .9.9.8t. .8t. .0.4.8.7.4.5 t 29 45 15 41 95 77 99 63 32 22 55 64 12 51 31 00 18 9 6 6 0 2 9of4v ca25 31 04 45 12 92 14 34 64 22 97 1 9 4 5 1 3 99 54 81 12 24 60ssg DtnnibIg g g g g g g g g g g 4g 2g 2g 8g 3g 2g 2g 5g 4g 3g 1g 5gn m p o e | C | C | C | C | C | C | C | C | C | C | C | C | C | C | C | C | C | C | | | | |uu p etn eIRIRIRIRIRIRIRIRIRIRIRI I I I I I ICICICICICIol aoCocmrR R R R R R R R R R R R p G A A A A A A A A A A A A A A A A A A A A A A A 07 61 53 13 97 76 97 87 23 72 80 84 86 56 3 3 1 6 7 3 9 9 8 9 4 2 2 9 471 68 62 66 78 62 78 03 31 3 3 1 0 6 76 85 19 48 65 55 65 92 68 38 56 84 69 31 68 1 7 21 52 2 2 2 01 5 3 1 1 3 1 41 05 82 19 93 7 7 9 0 4 4 5 1 8 4 6 2 25 4 1 8 5 0 4 3 8 9 4 6 4 9 5 2 9 4 4 3 41 4 8 5 2 5 5 4 6 6 6 3 1 3 0 7 7 5 4 1 2 8 1 6 2 4 1 1 4 1 41 5 1 1 4 1 3 1 3 81 .)emoctu12 33 36 60 20 73 71 75 98 62 86 70 92 93 12 19 68 37 19 78 87 79 27 43 80 60 5 0 4o8 8 8 1 4 0 2 9 6 7 9 6 6 2 5 5 9 3 0 0 0 E1 7 2 5 7 3 8 2 0 7 1 5 2 5 4 5 7 4 7 9 9 9 1 1 42 1 2 2 01 5 3 1 1 3 31Decn08 8 2 4 4 1 4 3 2 3 7 2 6 3 4 4 3 5 9 2 0 9eu8 4 1 6 1 9 8 3 2 6 0 8 7 7 2 6 9 6 3 1 5 6 8 5 8 3 4 3 4 9 02 22 12 43 20 02 43l0 72 49 63 32 77 16 96 52 13 7 82 9 43 6 26 6 43 22 9 0 5 0 4 8fn 1 1 1 2 1 2 3 1 1 81iylb80 56 92 25 92 49 59 41 71 29 81 42 02 39 51 50 46 13 26 1 8 2 3 0 4 6 0 6 3iss6 3 2 1 2 1 6 6 9 4 8 4 8 0 1 6 3 0 8 9 5 o 1 8 2 6 8 2 8 2 0 9 0 5 7 3 7 7 4 1 4 9 8 8 5 2 9 0 8 1 6 8 32 2 2 2 8 5 2 1 1 3 1 11pdluo5 5 5 1 0 8 6 4 8 5 7 5 5 5 3 8c41 64 9 17 74 91 44 7 1 6 7 2 0 0 6 3 7 9 5 7 2t1 5 72 94 01 61 93 6 3 01 8 25 1 2 1 5 11 11 34 61 73 41 1 03 091aht,.e.1 5 7 9 0 9 4 0 7 9i(81 7 8 8 6 4 9 3 8 9 9 0 4 6 4 4 0 8 2 6 0 4 01 87 75 31 41 5e2 7 72 45 91 31 75 5 5 12 6 37 4 6 7 3 4 1 6 6 2 1 8 1 5 1 1 5 2 4 1 4 22neglac3 3 2 6 2 2 4 7 2 7 9 2 3 0 9 9 5 2 5 6 8 0 4 0 0 5 o2 4 7 1 5 8 6 4 0 0 4 8 9 3 3 6 5 4 5 7 9f1 3 7 4 41 5 2 4 1 3 1 1 3 6 1 8 4 4 4 1 4 6 1 4 1 4 1 2 1 2 41 otgnip 1 1 1 2 1 1 1 1 1 2 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 1 1 1 1pam sdaerqes- A 42 42 42 42 42 42 42 42 42 42 42 42 42 42 42 42 42 42 42 42 42 42 42 42 42 42 42 42 42 N T T T T T T T T T T T T T T T T T T T T T T T T T T T T TRderetlif1t1t1t1t1t1t1t1t1t1t1t1t1t1t1t1 1 1 1 1 1 1 1 1 1f.5. . . . . .1t1t1 1t t t t t t t t or4 5 6 8 4 2 9.4.8.0.0.8.5t. .9.9.8t. .8t.t. . . . . . . . . e4 29 45 15 41 95 8 3 7 9 6 3 2 5 6 1 5 3 0 1 2 09 46 69 86 70 42 30 59 b9 2 3 0 4 1 26 21 79 91 33 26 22 59 41 29 14 15 01 83 9 9 5 0 8 1 2 2 6 m2g 5| g 1 | g 4 | g 5 2 4 2 2 4 4 4 2 7 4 2 2 8 3 2 99 92 45 81 14 23 41 13 05 u | g| g| g| g| g| g| g| g| g| g| g| g| g| g| g| g g g g g g g g g g nCICRICRIC C C C C C C C C C C C C C C C| C | C | C | C | C | C | C | C | C | ClRIRIRIRIRIRIRIRIRIRIRIRIRIRIRIRIRIRIRIRIRIRIRIRIRIRIRatA A A A A A A A A A A A A A A A A A A A A A A A A A A A A o a T 9 6 0 0 9 9 8)8)479,38002100.% % % 3 7007.0.5.%40 %8 .d3, , , ,3,3,06 1 1 02 38 5 9 3 24:19.7F:F 9 9, .9 6 1et7 8 ,4 ,],](2( ne8,8 1 % % 9 46.0.7 2 3 3m,3g142,:43 6arD:6 2f,D%,2,F4%8;.7.5 1de:5S:ta. [S ci1.%[%lp507.89.u 193d 3:9C:Cdn1a5e006 0 7 1 9680 5 2 9.7 0 % % % ) 6)t0e0_,4,1,0,1,7 2 3735,3, .41.4.% 953 % .9,1.lpA81 1 1 1 8 1 9:1F:F 9 99 8 7 C 8m1 3 834 0o,86,,],](74( c G,6 8 % % 7 1 48.96.7 4 2 7,1Dag81, a:88,7;D:D 1yv1 14p.,%, oA9% .5 c;1 0.1 1el .:1 g 5S: 7[Sni%[%s307.4d427.n93 :9aC:Ce0t0e9l_pA moC G,e bc,ea) m e gSmn ;eic3oto Fi0 ) p 3 8in F p ptec7rceg Galipr.st=9so mm m A:p’isnre(= nac ’rt osrfmocsns se n eetn( tto p drf ,oin n o ao ire Careg‘yrotct s :sstg‘ fa zsdscn nioseecfokao nfo n uta drearetdocara pra p e er trniamd ptcke nb nasle bsdslO O e e n c Onam aid xa%aC C p p p p o er CSBu e e atm o C utS S o U U a art ronN M M M T G N T B B M MnIC UeBaGbevitan 3 6r6 1etl s1 1asg 1 nie 5 84 n g gglieke | |nsd n o ag ClICIe R Ru mfrkiA A,ol- 0; ;befr0 41 55.ngeSetigDr95 1 1 8 18F8 Co 3 4 4 1ir_pg g g 8s| | |n1nCICICInAra R R R ArtA A Aw g onis hBhce n e ekr1tae mie5 go l-n tr1m8 ht r2in hc Beto p4tiecoE t Por sp ng|w ajyc Ca) _yrtyGred(1la P 0 MilCiA ht -'Y54 a / Rm a / BmV 3 C P n Ra aAegfn Afroett ,se B nr- e -rsulrcete r gtre e dt+rS otoia o nprop niPps.esnliNse)netR n)ca amvilea orN-ar6 tatj rdtyla(tnrt pa CiiCsn a SeS ga-'B mSn a aKu D5 AafAesh c / F k n Mil / nPrdyae rat1.1.1.1.13.g v 79 5 0 2 5avitne2 8 1 4 1.at0 0 1 1 0 u m . 0 0 1 o 0 0 0 0 0Aep elN00 00 0 0 0 0 00htg pp n o 10 10 1 1 0 0 10fnioluibS sJ J J R J O OsnsO O O e N N N N Neimeic n c J J J J J A A A A ApsA C C C C Coerwcesoaeh79 58 01 32 51inscegsigi2 t 0 0 0 0 10 41 01 _ 0_ 0 0 0cirgseno 3.n0 3 _3 _3 _3Ainkecsn g1 0 0 C 0 00 00 00 ICICICICIuag2rR R R R R A A A A olfniA Agolsek tna1 1 1 1 1..3.4.9. .aolao 7 2 1r nf3 1 2 8 6 deN1 1 7 4pesien 00 00 10 10 10nto0 0 0 0 0snht is1eon ff0 10 10 10 1 ose J J J J 0JedQglosc Q Q Q Q u c E E E E E A O O O O Ogn Btati3 k 9.)s7 41 92 28 16n5 nao30 10 11 71 41l12isg 8it0 0 0 0 0f rf se 1nCICICICICIoA(r rAo R R R R R c A A no yp A A Ai lxetb 8 3 9 8 7alna9 6 1 6 5 8 3 0 8imes itS 2 D 5 5 9 1 1mrsnea erC 3g 1g 4g 5g 5gt1e_ 8 |e2ffi1 C| C | C | C | C rIRI I I ID 02dR R R R ( A A A A A Atnereffi-dn y olnts s s s :nae y s eto Ncodtbe n n dno d ,if ede etsinisnnr taln e seutale n r oioitg g i o o cu ercita hd rtciteirai s taststa p syl e n rearroln ev-o e h a e g hrcog oco lva ysis toehr ly hhtotto pic hp pytu mn itefoty yef ta a swenom 0 oepniyntinipnie d ma e a 1 s 3 n ecr tileg vit r lla i n t o sv etS gn 53ualcp si ibom it lcaliagaitialar(p yehfnmioty dd er(efmrailise e cp iy s nerns ytte wssiliivita- ciniomiv c nte n * s a u e a g o e birc h sgo hsolhsntpefonn e e c o nitn 0 a oi 7.efee icv sea nrahtsys d a n hp netenciag nini olayi r7 vmret18 t nhe hs petealsuys h eofir5 s gosn go s go er srei pDnIC 1 orai tshto ho ota so dtah hc ntaitniedavedra y eyis sisvb frah p hn d d d ep o pnrp d d d e arnerytde)4o vfetegveh g ehiviru).1 al teal taeel dtanel eta olsttieoh tietctch(7.0,sie e etsoo e pele y s h o nta nefur 3.yln R R R n R d R p H o P onI ts0(aoi )ntaissla ) )6.stn svre g(16 c o e oht, n0itp.g(ocistid enmsn sn e ooehtfistse) gsta c s o li6p ep o h hllat-l)f(6. .2 gcietao lhtsoca his n fi ts oi llaa 0 ,0pl tlcnicaootuq(oldey mrr lla ae o p aat sgretye) yti9.oatntilivit%5 )1.7 hp-moc otytitytnytivi ytiiib vic iyr tpefn 9( e(3 ,2noin.0olgsc Csoi litbtitclei tc an ei tc o f bitefoi(i cs near ta) ealactep r encip nituu gen ec nlo stot le )5.hnirrod(00 htM C-apnrrs e u e s t s g u e v s g e ota nc,0(iMtnnto s htohoc h peg ss ) deiIsoth a s p otn h apninio ehg no ntalc 6.i oia - ) 4 ac2,t m n tio n frfa inininio i nhtata p-trt (1.d o n ito n i o n i oitairirs n oi 0(ait t tp r airairair- a tvavh n ) s cicici ai4.s stml erumava a a o cv v v htetetd e e ) 3 b 13 i cic c c ne ne ne M(1 ,8e.tsedl tetietietiete go g g2(to ne ne ne ne nelyolyoly) n R e)bgo g lyo g lyo g lyo g h h h lyolyp-p p) n- -a 6.9.7 o n n(0 , nB(ybh h h h h o o P P P P P N N N0(op.ndmoen t e n g o o oita - P N P P - N cit ac htiN taihrid ara P v slga iievnietnoasit ts e a - - dmrfciotreao HirP P N P N N lea etlv dlts no iero A t M geLzd n elen )ol le o ) y m d pa) ) ) ) o(b(c(d(e()f) ) (g(h(A(h o prfMmo A c B

Claims

CLAIMS 1. A method comprising providing a microbe-resistant species of rotifer in culture, exposing the rotifer to a microbe, and isolating a compound produced by the rotifer following exposure to the microbe.

2. The method according to claim 1, wherein the compound is a secondary metabolite.

3. The method according to claim 1 or claim 2, wherein the microbe-resistant species of rotifer is a fungus-resistant species, and wherein the microbe to which the rotifer is exposed is a fungus.

4. The method according to claim 3, wherein the fungus is a conidium, preferably wherein the conidium is R. globospora, R. denticulospora, R. sp. ‘ARSEF 10274’, H. spirosporum, H. sp. ‘ARSEF 14199’, or T. subcuticularis.

5. The method according to any one of claims 1 to 4, wherein the rotifer is lysed prior to the isolation of the compound, and wherein the compound is isolated from the rotifer lysate.

6. The method according to any one of claims 1 to 5, wherein the genome of the rotifer comprises genetic material acquired via horizontal gene transfer (HGT) from non-metazoan taxa, preferably wherein said genetic material has increased expression following the exposure of the rotifer to the microbe.

7. The method according to any one of claims 1 to 6, wherein the microbe-resistant species of rotifer is a species of bdelloid rotifer, and preferably wherein the bdelloid rotifer species is Adineta ricciae, Adineta vaga, Adineta steineri or Habrotrocha ligula.

8. The method according to any one of claims 1 to 7, wherein the method further comprises assaying anti-microbial activity of the isolated compound.

9. A method comprising assaying anti-microbial activity of a compound isolated according to the method of any one of any one of claims 1 to 7.

10. A polynucleotide comprising one or more genes which are comprised within the genome of a microbe-resistant rotifer and have been acquired by the rotifer via HGT from non- metazoan taxa, preferably wherein said one or more genes have increased expression upon exposure of the rotifer to a microbe, preferably wherein the microbe is a fungus.

11. The polynucleotide according to claim 10, wherein the one or more genes encode one or more non-ribosomal peptide synthetases, polyketide synthetases, or hybrid non-ribosomal peptide and polyketide synthetases.

12. A vector comprising the polynucleotide defined according to claim 11.

13. A host cell comprising the polynucleotide according to claim 10 or claim 11, or the vector according to claim 12.

14. A compound produced and isolated by the method according to any one of claims 1 to 7 or produced by the host cell according to claim 13.

15. A method of treating a microbial infection, wherein the method comprises administering the compound according to claim 14 to a subject.

16. The method according to claim 15, wherein the microbial infection is a fungal infection.

17. A method of controlling microbial disease in a plant, the method comprising contacting the plant with a compound according to claim 14.

18. The method according to claim 17, wherein the microbial disease is a fungal disease.

19. The method according to claim 17 or claim 18, wherein a microbe is in contact with the plant, and the controlling comprises killing the microbe or inhibiting the growth of the microbe, preferably wherein the microbe is a fungus.

Images