Introduction

A large body of theoretical and experimental work has shown that dynamics of microbiomes are shaped by the network of interbacterial interactions[1–4]. These cooperative and competitive interactions are often achieved via the secretion of cross-feeding metabolites[5], antimicrobial peptides[6] and bacterially produced small molecules[7, 8] and are crucial to ecological properties including stability and ability to respond to external perturbations[9, 10]. Among the competitive interactions, bacteriocin production is proposed to be a prominent mediator of microbiome dynamics[11] and, specifically, several reports including ours have shown that a bacteriocin subclass, Class IIb microcins, mediates Enterobacteriaceae dynamics in vivo[12–14].

Class IIb microcins are ribosomally synthesized bacteriocins between 5 kDa to 10 kDa in size with activity against closely related strains or species[13, 15–18]. Unlike all other microcins, they carry a serine-rich C-terminal motif for a posttranslational modification with a siderophore, here an enterobactin or an enterobactin derivative, before they are secreted into the extracellular space[19]. Siderophores are iron-chelating molecules commonly employed by various bacteria to scavenge free iron to compete with other bacteria, particularly in resource-scarce environments such as the gastrointestinal tract[16, 17, 20] and are often associated with increased pathogenicity or virulence[21–23]. The iron chelating moiety of these posttranslationally modified antimicrobial peptides is recognized by high-affinity receptors and functions as a Trojan Horse key to susceptible bacteria as it triggers import into the periplasmic space, where the peptide inhibits the molecular target of susceptible bacteria[20, 24–26]. Because of these features, delivery of class IIb microcins by wildtype and engineered probiotics has been recently proposed as a strategy to combat drug-resistant enteric bacteria[12, 13, 17, 27], which is in line with a growing body of work from the past decade that explores siderophore conjugation, including with enterobactin, to specifically deliver antibiotics and other small molecules to drug-resistant Gram-negative pathogens[28–30].

To date only five class IIb microcins have been described and only four have been characterized in terms of their antimicrobial activity. Specifically, the class IIb microcins MccE492 and MccG492 (uncharacterized) are solely present in Klebsiella pneumoniae (Kp), whereas MccH47 is specific for Escherichia coli (Ec)[16]. Additionally, truncated versions of mciA (MccI47) and mcmM (MccM) are present in Kp RYC492, whereas they are intact in the Ec CA46 genome[16]. Interestingly, while the genes encoding for microcin posttranslational modifications are highly conserved between Ec and Kp, suggesting a conserved pathway for microcin maturation, the toxin and corresponding immunity genes are significantly more variable (Fig S1). We hypothesized that class IIb microcin production extends beyond these specific compounds and organisms and identified a total of twelve novel class IIb microcins in seven additional Enterobacteriaceae species. Utilizing heterologous expression of these compounds in our E. coli system optimized for enterobactin conjugation, we show potent antimicrobial activity by the encoded toxins against a library of bacteria, including Gram-negative ESKAPE and plant pathogens. This demonstrates that class IIb microcin genes are more prevalent in the microbial world than previously recognized and that synthetic hybrid microcins can be a viable tool to target clinically relevant drug-resistant pathogens.

Materials and Methods

Bioinformatic class IIb microcin identification

We developed a pipeline that by leveraging BLAST[31] enabled us to mine publicly available genome databases for novel, previously undescribed class IIb microcins. We included mchCDEF and mcmL for Ec as well as mceCDGHIJ for Kp for posttranslational modification and export, expecting more reliable hits for longer and functionally conserved proteins in close proximity to class IIb microcin and immunity genes. Thus, we first ran tblastn[31] against RefSeq[32], to screen for all genes related to biosynthesis pathways, known microcin genes, as well as immunity gene sequences exhibiting homology to the microcin gene clusters found in Ec CA46 and Kp RYC492. Homology to the microcin gene clusters were guided by BLAST parameters sseqid (Genome ID), pident (percentage of identical positions) along with sstart (start of alignment in Genome) and send (end of alignment position in Genome). Resulting hits were concatenated by genome ID and assessed for their proximity to one another in the genome. These gene clusters should, at best, contain all the known genes required for toxin maturation, including mchCDEF and mcmL[16, 17]. In addition to genomic hits to the known microcins, small ORFs of 50 to 150 amino acids in size close to the biosynthesis genes were screened and annotated manually using the criteria described below as well as their domains were predicted using SMART[33]. The ORFs were meticulously examined and assessed against established class IIb microcin criteria known from Ec H47, Ec I47, Ec M, Kp E492, and Kp G492: (i) a serine-rich C-terminus culminating in a final serine, (ii) the presence of fewer than two cysteine residues, (iii) a signal peptide within the initial 15 amino acids ending with GG or GA, and (iv) close proximity (≤200 bp) to an ORF featuring a predicted transmembrane domain, typically encoding an immunity peptide. The identified genes were included in the pipeline’s input to expand the scope of gene detection. We repeated this process iteratively through the pipeline until no additional genes were added to the output. Subsequently, blastp was used to assess microcin similarity shown in Table 1 and Table S1.

Table 1:

Phylogenetic analyses

For the native full length coding sequence of the microcin and immunity genes a codon-based sequence alignment was generated using the MUSCLE algorithm[34]. For phylogeny of all microcins, the nucleotide sequences without the respective signal peptides were codon-aligned. Subsequently, we determined the best fit substitution models for maximum likelihood phylogenetic analyses, resulting in the General Time Reversible model with discrete gamma distribution (GTR+G) and the Hasegawa-Kishino-Yano model with discrete gamma distribution (HKY+G), respectively. A bootstrap test with 1000 replicates for maximum likelihood and random seed was conducted for all trees. Alignment, model testing, and tree building was performed in MEGA11[35].

antiSMASH analyses

To test if similar results of class IIb microcin identification could be obtained with automated bioinformatic tools, we ran antiSMASH 7.0[36], a widely used tool for microbial genome mining and biosynthetic gene cluster detection. As input, we utilized the seven genomes from the newly identified class IIb microcins: (i) Bg CP014137, (ii) Gq CP014136, (iii) Ko CP033844, (iv) Ps CP034363, (v) Ro CP008886, (vi) Se CP030220, (vii) Sf CP033055. As a positive control for the well-established microcins Kp E492 and Kp G492 as well as Ec H47 and Ec M we used the accession numbers CP127839 (Kp RYC492) and CP148105 (Ec Nissle 1917), respectively. Notably, using the “loose” setting, in none of the cases a class IIb microcin biosynthesis gene cluster was detected, nor were any microcin genes identified. This was the case for both, the novel microcins and the original, well-annotated, microcins.

Plasmids and heterologous class IIb microcin expression

ORFs of identified microcin and immunity genes were codon optimized for frequent Ec codon usage without creating repetitive sequences and synthesized by Integrated DNA Technologies (Coralville, IA) with 18 bp of native 5’ upstream sequence and 20 bp of native 3’ downstream sequence, respectively. Using Gibson Assembly[37], the genes were cloned into our previously established Ec class IIb microcin expression system that results in mature class IIb microcins posttranslationally modified with an MGE[13, 17]. Briefly, the antimicrobial and the immunity genes are co-expressed under the control of an arabinose-inducible pBad/araC promoter in a high copy plasmid with a pUC-derived origin of replication. All assemblies were verified using whole plasmid sequencing. DNA files for all used plasmids can be found as supplementary material.

Static inhibition assays

Cultures of strains with confirmed plasmid assemblies were spread in LB agar plates containing 100 µg/ml ampicillin. In addition to a pUC19 control without microcin expression, single colonies for each microcin were picked with a sterile pipet tip and all placed into the same solid LB agar plate containing 100 µg/ml ampicillin for plasmid retention, 0.2 mM 2,2-dipyridyl to create iron-limited conditions during the growth phase, and 0.4% L-arabinose for induction of gene expression. Plates were incubated at 37°C for up to 72 h, before they were overlaid with the target bacterial isolates. Note that testing all microcin-expressing stains on the same plate allowed us to confidently assess differential inhibitory activity between all 17 tested microcins. For the overlay, the microcin-producing bacteria in the stabs were inactivated using chloroform vapors and ten minutes under ultraviolet light. Then, target bacteria were diluted 1:2000 from overnight culture in LB media containing 100 µg/ml ampicillin and 0.2 mM 2,2-dipyridyl. Ec and S. flexneri strains were diluted 1:200 to acquire dense bacterial lawns. Finally, 0.5 ml of molten agar was added to the liquid media and the resulting soft agar was spread on the plate with the inactivated bacteria and incubated for 16 h at 37°C. The pUC19 control strain was unable to create any zone of inhibition against any of the tested target bacteria.

Relative MIC dilution factors

For enrichment of microcin Se G492, an MBP-microcin fusion protein was expressed from pHMT-SeG492 in E. coli BL21 cells as previously described[13, 17]. Harvested cells were resuspended in column buffer (200 mM NaCl, 20 mM Tris-HCl, pH 7.5), lysed by sonication, and passed through a high flow amylose resin (New England Biolabs, Ipswich, MA) as recommended by the manufacturer. The protein was eluted with 10 mM maltose, cleaved with Tobacco etch virus (TEV) protease, and further processed as previously reported[13, 17]. The relative MIC assays were conducted using sterile 96-well round bottom microplates. The plates were prepared as follows: the first row contained 20 µl of 2x LB with 0.4 mM 2,2’-dipyridyl and 20 µl of Se G492 containing solution in amylose resin elution buffer (200 mM NaCl, 20 mM Tris-HCl, 10 mM maltose, pH 7.5). All other wells were filled with 20 µl of 1x LB, 0.2 mM 2,2’-dipyridyl, and 0.5× amylose resin elution buffer, and a two-fold serial dilution was performed across the plate. The target bacteria were grown overnight in LB at 200 rpm and 37°C and were added to a final dilution of 10⁻⁴ into the wells. The plates were then incubated in the dark at 37°C with gentle agitation. Relative MICs were determined as the lowest concentration at which no growth was observed after 24 hours. All reported values represent the median of at least three biological replicates.

Results

With the hypothesis that class IIb microcin production is a common trait among Enterobacteriaceae, we anticipated that the genes encoding for the antimicrobial and immunity would exhibit a high degree of dissimilarity to already known peptides as target specificity may result in accelerated adaptive coevolution[38]. Therefore, in addition to the known microcin and immunity genes from MccE492, MccG492, MccH47, MccI47, and MccM, in our informatic approach we included the genes that are necessary for mature class IIb microcin biosynthesis, extending our search to longer sequences for more reliable Basic Local Alignment Search Tool (BLAST) results[31]. Moreover, we hypothesized that the amino acid sequences of genes responsible for posttranslational modification and microcin export would be less prone to evolutionary changes, thereby maintaining the functional integrity of the gene cluster[16, 39]. We then assessed their proximity in the respective genome location, because microcin genes are typically flanked by genes essential for toxin maturation[16]. Further, we manually assessed and annotated small open reading frames (ORFs) upstream and downstream of the maturation genes, allowing us to also identify novel class IIb microcins without significant sequence similarity to the known antimicrobials, enabling the discovery of compounds with new molecular targets or modes of action (see Methods).

Our informatics-driven analysis identified twelve promising class IIb microcin candidates from seven gene clusters with high similarity to Ec CA46 and Kp RYC492 in seven species across the Enterobacteriaceae family (Fig 1A, Fig S2): (i) Brenneria goodwinii (Bg; 2; GenBank: CP014137), (ii) Gibbsiella quercinecans (Gq; 1; CP014136), (iii) Klebsiella oxytoca (Ko; 1; CP033844), (iv) Pantoea sp. (Ps; 1; CP034363), (v) Raoultella ornithinolytica (Ro; 4; CP008886), (vi) Salmonella enterica (Se; 2; CP030220), (vii) Serratia fonticola (Sf; 1; CP033055). Although it has traditionally been a defining characteristic of class IIb microcins that all required genes are encoded within the chromosome[12], the gene cluster we discovered for Se is situated on a 159 kbp plasmid. Phylogenetic sequence analysis of both the antimicrobial and immunity peptide genes revealed the presence of eight different clades represented in both trees, respectively (Fig 1B,C). Regarding the well-established class IIb microcins MccH47, MccI47, MccM, MccG492, and MccE492, we identified novel members for each group, supported by nucleotide sequence similarity, amino acid identity, the closest blastp match, and domain predictions (Table 1, Fig S3). It is important to note that application of established tools for secondary metabolite identification (e.g., antiSMASH 7.0)[36] to these genomes did not yield identification of any of the old or novel microcins providing support of the relevance of our approach. In order to then ensure that these novel microcins are unique and not part of any other microcin class, we performed phylogenetic analysis for all known microcin genes from the classes I, IIa, and IIb and show distinct clustering for all newly described sequences (Fig S4). In light of this discovery, we propose a new nomenclature for class IIb microcins that includes the species initials in which they were identified (e.g., Ec, Kp), the closest relative already characterized class IIb microcin (G492, E492, H47, I47 or M), as well as the identifiers ‘A’ for antimicrobial or ‘I’ for immunity gene.

Fig 1:

Based on this, the novel G492 relative found in Salmonella enterica will be called Se G492 with the antimicrobial peptide identified as Se G492A and the immunity peptide identified as Se G492I. It is worth highlighting that in the case of the G492 group, all its members have the immunity gene located downstream of the antimicrobial gene, whereas for the other clades, this arrangement is reversed. In addition to uncovering eight novel variants of the five previously characterized microcins, we have identified four additional microcins through manual curation of ORFs in proximity to the microcin maturation genes. These novel microcins, which we name microcin W (MccW), microcin X (MccX), and microcin Z (MccZ), seem to belong to three entirely new clades based on nucleotide similarity (Fig 1 B,C). The two members of the microcin X group, found in B. goodwinii (Bg X) and R. ornithinolytica (Ro X), only show significant similarity between one another, but not to any of the other antimicrobial or immunity peptides. This holds true for the nucleotide similarity (Fig 1 B,C) as well as amino acid identity and the closest blastp hits (Fig. 1D, Table 1, Table S1). Similarly, MccW from Gibbsiella quercinecans (Gq W) does not show any sequence similarity to either the known or novel antimicrobial or immunity peptides in terms of nucleotide similarity, amino acid identity, the respective blastp hits, or phylogenetic localization (Fig 1,Table 1, Table S1). Lastly, MccZ from R. ornithinolytica (Ro Z) shows insignificant amino acid similarity with Ec MA (mcmA) for the antimicrobial, whereas the immunity peptide does not have any match among the known or the novel microcins (Fig 1 C, Table 1, Table S1). Crucially, the identification of MccX, and MccZ within the same gene clusters as representatives of the E492 (Bg E492), H47 (Ro H47), and I47 (Ro I47) groups strongly implies that they are functional components of a microcin gene cluster.

To test the newly identified microcins for antimicrobial activity we used our previously established Ec overexpression system[13, 17]. All antimicrobial and immunity peptides were codon optimized, synthesized, and cloned into an inducible high copy vector (see Methods). Thus, we extracted the novel microcins out of their native genomic context of siderophore biosynthesis and transferred them into a heterologous expression background optimized for microcin-monoglycosylated enterobactin (MGE) linkage[13, 17]. This allowed us to create hybrid compounds that could be efficiently tested for antimicrobial activity in an E. coli background. Through static plate inhibition assays involving live-producing cells[13, 17, 27], we successfully validated robust antimicrobial activity of eleven out of the twelve newly discovered microcins (Fig 2A). Notably, antimicrobial activity was only observed in iron-depleted media (Fig 2B,C). The hybrid microcins exhibit a range of specificities, with some inhibiting targets narrowly (e.g., Ps G492AI), while others exert a broader effect against multiple bacteria (e.g., Se G492AI). Moreover, our study also provides the first evidence of inhibitory activity by Kp G492, a microcin whose existence and function have only been proposed in the scientific literature based on genetic sequence[16].

Fig 2:

To this date class IIb microcins have been only shown to be very selective and only active against different species within the Enterobacteriaceae family[13, 16, 17, 27]. While the activity for the novel microcins varies, we here report, for the first time, antimicrobial activity outside of the Enterobacteriaceae family utilizing hybrid antimicrobial peptides. We demonstrate that microcins Ps G492 and Se G492 have activity against Gram-negative multidrug-resistant ESKAPE pathogens with both being capable of inhibiting Acinetobacter baumannii (BAA 1790), and with microcin Se G492 alone also showing activity against Pseudomonas aeruginosa (PA14) (Fig 2). Specifically, compared to K. pneumoniae (BAA 1705) Se G492 is 256-times more effective against A. baumannii (BAA 1790), 128-times more effective against E. coli (BAA 196), and 8-times more effective against P. aeruginosa (PA14) (Fig. 2B).

Discussion

With a comprehensive analysis of publicly available bacterial genomes, we unraveled twelve previously undiscovered class IIb microcins. Among these findings, we identified three novel microcin clades, specifically MccW, MccX, and MccZ. Through heterologous expression, we showed antimicrobial activity for all but one novel microcins and are the first to demonstrate activity for the known class IIb microcin Kp G492. Hence, this research demonstrates that class IIb microcin genes exhibit a higher prevalence in Enterobacteriaceae genomes than previously reported. As a result, their impact on ecological community dynamics in natural environments, including the growth of Pseudomonadales species, might be broader than previously thought. For antimicrobial activity testing, microcin and immunity genes were overexpressed recombinantly in our Ec-derived expression system optimized for microcin-MGE production[13, 17]. The common process of posttranslational modification with the siderophore consolidated the import mechanism of the hybrid microcins towards enterobactin, the most characteristic siderophore of the Enterobacteriaceae family. This allowed us to test the target-specific antimicrobial activity of the microcins irrespective of siderophore production in the native genomic background. However, it is important to note that class IIb microcin activity is dependent on active import through siderophore receptors and consequently some of these microcins might display different activity spectrums when tested in their native genomic background of siderophore biosynthesis. Furthermore, static plate inhibition assays exhibit lower sensitivity compared to purification approaches with quantitative minimum inhibitory concentration (MIC) assays. Thus, the activity spectrums of the hybrid microcins could encompass a wider range than what has been described in this study when tested as a purified product. However, historically the microcin literature proves that ideal approaches for purification and MIC testing can vary between the antimicrobials[13, 16, 17, 25].

We were able to expand the origins of class IIb microcins from the enteric bacteria Ec and Kp to other members of the Enterobacteriaceae family, including well-known phytopathogens[40–42]. Specifically, B. goodwinii and G. quercinecans are associated with Acute Oak Decline (AOD) and are frequently isolated together[43] and the two strains containing microcin genes were isolated within the same research project. Notably, these bacteria grow synergistically[44], while upregulating iron transporters during co-culture[45], hinting at class IIb microcin-related competition. We were able to show activity of the overexpressed hybrid microcins against human-derived enteric isolates, however, their native spectrum might have evolved to target more frequently encountered strains from the genus Brenneria or Gibbsiella. Further, we demonstrated activity of several class IIb microcins against the three tree pathogen genera Brenneria, Gibbsiella, as well as Rahnella[44, 46–49]. Thus, treatment with potent microcins, purified or produced in live bacteria, could present a viable option to target bacteria-caused plant diseases.

In health care settings the burden by Gram-negative ESKAPE pathogens and multidrug-resistant Enterobacteriaceae weighs heavily on modern medicine and novel antimicrobials are needed to develop new treatment options[50, 51]. In addition to enteric pathogens and pathobionts, bacteria outside of the Enterobacteriaceae family have also been shown to scavenge for and to import enterobactin, including P. aeruginosa and A. baumannii [52, 53]. Therefore, different siderophore conjugates could be a viable option to target these pathogens as well or to finetune the desired target range[28–30]. Antimicrobial peptides and particularly microcins are promising candidates for selective eradication of enteric pathogens and have been demonstrated to potently reduce pathogen colonization in vivo, when produced by a live probiotic[12, 13]. Here we present the most comprehensive library of class IIb microcins created so far, that is suited for heterologous expression and in vivo application for the development of novel live biotherapeutic products against drug-resistant enteric bacteria and Gram-negative ESKAPE pathogens.

In this study, we challenge the prevailing notion that class IIb microcin production is limited to Ec and Kp. Through comprehensive genomic analysis of publicly available bacterial genomes, coupled with heterologous overexpression, we unveiled a set undiscovered class IIb microcins across Enterobacteriaceae species. Our findings not only expand the known repertoire of class IIb microcins but also hold significant implications for synthetic hybrid compounds. We demonstrate that these newly identified class IIb microcins exert remarkable inhibitory effects on ESKAPE pathogen species when expressed in a system for enterobactin-derived conjugation. This discovery underscores their potential as agents against a broader spectrum of pathogens, including those affecting humans and plants, thus opening new avenues for antimicrobial research and applications.

Acknowledgements

The author would like to thank the administrative staff in the Microbiology Department at the University of Massachusetts Chan Medical School, especially Annette Bohigian, Amy Parker, Dhruti Desai, Marie Berardi, Richard Fish, and Tracey Rae, for their support.

Competing Interests

The authors of this manuscript have the following competing interests: V.B. receives support from a sponsored research agreement from Vedanta Biosciences, Inc. The authors B.M.M. and S.K.B have declared that no competing interests exist.

Data Availability Statement

The genomes are accessible with the following GenBank numbers: Brenneria goodwinii (CP014137), Gibbsiella quercinecans (CP014136), Klebsiella oxytoca (CP033844), Pantoea sp. (CP034363), Raoultella ornithinolytica (CP008886), Salmonella enterica (CP030220), Serratia fonticola (CP033055). All information is included in the manuscript or supporting files. All plasmid sequences as well as annotation files to produce Figs 1A, S1, S2, and S3 are available as supplementary material.

Funding Disclosure

This work was supported by the CDMRP PRMP W81XWH2020013 to V.B., by the NIH NIA 1R01AG075283-01A1 to V.B., and by the Deutsche Forschungsgemeinschaft (DFG) project 457837076 to B.M.M. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Author Contributions

B.M.M, S.K.B, and V.B conceptualized the study. S.K.B. performed bioinformatic data mining. B.M.M and V.B designed the experiments. B.M.M performed plasmid design, verification, and in vitro activity testing. B.M.M, S.K.B, and V.B wrote the manuscript.