In situ activation and heterologous production of a cryptic lantibiotic from a plant-ant derived Saccharopolyspora species

Most clinical antibiotics are derived from actinomycete natural products (NPs) discovered at least 60 years ago. Repeated rediscovery of known compounds led the pharmaceutical industry to largely discard microbial NPs as a source of new chemical diversity but advances in genome sequencing revealed that these organisms have the potential to make many more NPs than previously thought. Approaches to unlock NP biosynthesis by genetic manipulation of the strain, by the application of chemical genetics, or by microbial co-cultivation have resulted in the identification of new antibacterial compounds. Concomitantly, intensive exploration of coevolved ecological niches, such as insect-microbe defensive symbioses, has revealed these to be a rich source of chemical novelty. Here we report the novel lanthipeptide antibiotic kyamicin generated through the activation of a cryptic biosynthetic gene cluster identified by genome mining Saccharopolyspora species found in the obligate domatia-dwelling ant Tetraponera penzigi of the ant plant Vachellia drepanolobium. Heterologous production and purification of kyamicin allowed its structural characterisation and bioactivity determination. Our activation strategy was also successful for the expression of lantibiotics from other genera, paving the way for a synthetic heterologous expression platform for the discovery of lanthipeptides that are not detected under laboratory conditions or that are new to nature. Importance The discovery of novel antibiotics to tackle the growing threat of antimicrobial resistance is impeded by difficulties in accessing the full biosynthetic potential of microorganisms. The development of new tools to unlock the biosynthesis of cryptic bacterial natural products will greatly increase the repertoire of natural product scaffolds. Here we report an activation strategy that can be rapidly applied to activate the biosynthesis of cryptic lanthipeptide biosynthetic gene clusters. This allowed the discovery of a new lanthipeptide antibiotic directly from the native host and via heterologous expression.

have the potential to make many more NPs than previously thought. Approaches to ecological niches, such as insect-microbe defensive symbioses, has revealed these 27 to be a rich source of chemical novelty. Here we report the novel lanthipeptide 28 antibiotic kyamicin generated through the activation of a cryptic biosynthetic gene cluster identified by genome mining Saccharopolyspora species found in the obligate 30 domatia-dwelling ant Tetraponera penzigi of the ant plant Vachellia drepanolobium. 31 Heterologous production and purification of kyamicin allowed its structural 32 characterisation and bioactivity determination. Our activation strategy was also 33 successful for the expression of lantibiotics from other genera, paving the way for a  Here we describe activation of the cryptic Saccharopolyspora lanthipeptide BGCs and 112 the characterization of their product, a new class II lantibiotic that we called kyamicin. 113 We also exemplify a heterologous expression platform for lanthipeptide production that 114 may be particularly useful for strains that are refractory to genetic manipulation. The 115 methodologies reported should be applicable for the activation of cryptic BGCs from a 116 wide range of actinomycetes.

119
Origin, characteristics and genome sequencing of Saccharopolyspora strains. 120 The Saccharopolyspora strains were isolated from ants taken from the domatia of T.    Amongst these was a BGC for a cinnamycin-like lanthipeptide. The BGC architecture 141 was conserved across all three genomes, including an identical pro-peptide sequence 142 encoded by the precursor peptide gene, suggesting they all encode the same molecule which we named kyamicin (Fig. 1B). The sequence and annotations for 144 these three BGCs have been deposited at GenBank under the accession numbers 145 MK251551 (KY3), MK251552 (KY7) and MK251553 (KY21).

146
Through comparison to the cinnamycin BGC (24), and cinnamycin biosynthesis (32), 147 we assigned roles to each of the genes in the kyamicin (kya) BGC (Table 1). The kya 148 BGC is more compact than that for cinnamycin, and the genes missing from the 149 kyamicin BGC are dispensable for cinnamycin production (38). The cinorf11 gene is 150 not required for cinnamycin production but a homologue is present in the kyamicin 151 cluster. While cinorf11 lacks a plausible stop codon and its reading frame extends 570 152 bp into the cinR1 gene, its homologue, kyaorf11, has a stop codon and does not run 153 into the kyaR1 gene suggesting it may encode a functional protein.

154
To detect production of kyamicin we grew all three strains on a range of 13 liquid media 155 (Table S2) and extracted after four, five, six and seven days of growth, using 156 (individually) methanol and ethyl acetate. Analysis of the extracts using UPLC/MS 157 failed to identify the anticipated product (the methods were validated using authentic 158 duramycin). This was consistent with parallel bioassays which failed to show any 159 antibacterial activity for the extracts against Bacillus subtilis EC1524, which is sensitive 160 to cinnamycin (24). Similarly, no activity was observed in overlay bioassays.

161
Activation of the kyamicin BGC. Cinnamycin production and self-immunity 162 ultimately rely on two gene products (38). The transcription of the biosynthetic genes 163 is driven by CinR1, a SARP (Streptomyces Antibiotic Regulatory Protein, which usually 164 act as pathway specific transcription activators), and self-immunity is conferred by a 165 methyl transferase (Cinorf10) that modifies PE in the membrane to prevent binding of 166 cinnamycin. We reasoned that transcription of the homologues of these two genes 167 (kyaR1 and kyaL, respectively), driven by a constitutive promoter, would circumvent 168 the natural regulatory mechanism and initiate production of kyamicin. To achieve this, 169 we made a synthetic construct, pEVK1, containing kyaR1-kyaL (in that order), with a  (Table 2).     Consequently, the duramycin BGC was reconfigured in pOJKKH, which contains all 251 the biosynthetic genes, but lacks immunity and regulatory genes, and has a SARP 252 binding site upstream of durN that is similar to that upstream of kyaN (    Saccharopolyspora strains which led to activation of the BGC and production of 285 kyamicin. Since we were unable to isolate enough kyamicin from these strains for 286 further study, a heterologous production platform was developed using S. coelicolor 287 M1152 which allowed us to confirm the structure of kyamicin and assess its 288 antibacterial activity.

289
Having demonstrated the utility of a constitutively expressed SARP/self-immunity 290 cassette for driving expression of the otherwise silent kya BGC we utilised this 291 knowledge to activate duramycin production in a heterologous host.

292
Contemporaneous with our experiments, the duramycin BGC was also identified by 293 genome sequencing of S. cinnamoneus ATCC 12686 (33). This analysis described 294 the same genomic region containing durN to durH and surrounding genes (Table 1) 295 but failed to reveal putative regulatory and immunity genes. Co-expression of durA,  Table S1 in the supplemental material.    (Fig. S2A) followed by colony hybridization to give pDWCC2 and pDWCC3, respectively.

374
Analysis of the sequence of these plasmids identified 15 genes (shown in Fig. S7)