In recent years, we have observed an increase in reports of infections caused by the pathogen Shewanella (Ng et al., 2022). Shewanella spp. are Gram-negative rods, widely disseminated in aquatic niches (mainly marine environments), sediments and soil, that can cause infections in humans. Most infections are attributed to a few Shewanella species that may cause skin and soft tissue infections, bacteriemia, hepatobiliary infections, otitis media, etc. (Sharma and Kalawat, 2010; Yousfi et al., 2017; Ng et al., 2022). Although most infection-related species were identified as Shewanella algae and Shewanella putrefaciens, in the last years the employment of molecular based-methods improved the identification of other pathogenic species, such as S. xiamenensis (Parmeciano Di Noto et al., 2016; Thorell et al., 2019; Huang et al., 2022). Furthermore, some Shewanella species can also cause infections in aquatic animals (e.g., S. algae) or coexist as symbionts or epibionts (e.g., Shewanella pealeana and Shewanella woodyi) (Hau and Gralnick, 2007; Janda and Abbott, 2014); while other species have shown to be important dissimilatory metal-reducing bacteria (Zhong et al., 2018). In order to thrive in each environment Shewanella spp. encode a plethora of genes that may contribute to their diversification and niche adaptation.

Previous studies with few isolates showed that this genus is capable of acquiring a wide variety of mobile elements, such as plasmids, prophages, group II (GII) introns, integrons, and integrative and conjugative elements (ICEs) (Pembroke and Piterina, 2006; Larouche and Roy, 2009; Quiroga and Centrón, 2009; Ramírez et al., 2010; Carattoli, 2013; Parmeciano Di Noto et al., 2016; Fang et al., 2018; Parmeciano Di Noto et al., 2019; Thorell et al., 2019; Huang et al., 2022). Recently, a comprehensive analysis of integron integrases reported that multiple horizontal genetic transfer (HGT) events were implicated in the emergence and spread of novel mobile integrons (Nuñez et al., 2022). Genome analyses have also led to the identification of a few resistance plasmids in Shewanella spp. isolates that participate in antimicrobial resistance (AMR) dissemination (Carattoli, 2013; Zago et al., 2020).

In the last decade Shewanella spp. have become more relevant due to its role as reservoir of AMR genes bla_OXA–48 and qnrA, from which were most likely transferred to Enterobacteriaceae and other frequent human pathogens (Dabos et al., 2018; Tacão et al., 2018; Araújo et al., 2021). In addition, there has been an increase in reports on multidrug resistant (MDR) Shewanella spp. isolated from clinical samples (Ramírez et al., 2010; Di Noto et al., 2016; Almuzara et al., 2017). Few studies are limited and focused on the analysis of the accessory genome of Shewanella spp., particularly on the mobilome, resistome, and virulome of S. algae isolates (Deng et al., 2019; Wang et al., 2020; Huang et al., 2022), which encode beneficial genetic traits that contribute to its survival. Recently, Zhong et al. (2018) suggested that Shewanella spp. can acquire different mechanisms by HGT, in order to adapt to diverse environments.

In addition, the mobilome of S. algae and Shewanella baltica were further analyzed revealing the presence of strain-specific mobile genetic elements (MGEs) (Uhrynowski et al., 2019; Zago et al., 2020). Although these reports provided helpful information on the accessory genome of this genus, no studies have addressed an integrated analysis. Here, we performed an in-depth comparative study of the accessory genome focusing on the mobilome, resistome, and virulome. We found that infections caused by Shewanella spp. are not restricted to these two species, but virulence factors are distributed in two main lineages. Taken together, our results suggest that the genus Shewanella has the capability to evolve as a threat if correct identification and surveillance is not taken into consideration.

Shewanella genus has a versatile and diverse accessory genome that contributes to its adaptation and survival to different niches. This genus showed a wide variety of genes related to the mobilome, the virulome, and the resistome scattered heterogeneously throughout all lineages and habitats. In-depth analysis of these genetic traits in a large set of genomes from a wide variety of species proceeding from different niches revealed a few associations that may provide new insights into the pathogenicity and evolution of Shewanella.

Shewanella mobilome consists of various MGEs widely distributed along all lineages. The comparative analysis of lineages with more representation in each cluster, such as S. algae, S. xiamenensis, and S. baltica, reflected the MGE diversity within species and the plastic nature of the Shewanella genome. A previous work suggested that the rate at which HGT happens is up to two orders of magnitude higher than per-gene point mutations (Puigbò et al., 2014), which indicates that the gene content of microbes can vary in relatively short timescales. MGEs found in this genus were quite diverse, reflecting their continuous evolution, contributing with the adaptation and survival of the host to different environments. Among the MGEs detected, ISs showed the broadest diversity with the highest variety and copy number per element within the CG2 group, S. putrefaciens, and S. baltica (Supplementary Figure 2). It is well known that ISs can collaborate with the reshaping and rearranging of bacterial genomes, but it is less clear which factors govern their dissemination throughout a genome and how it impacts on the host physiology and evolution (Vandecraen et al., 2017). It is likely that the target site availability is a major contributor; however, interference with other ISs or host factors may be limiting features specific to each lineage (Siguier et al., 2014). A similar pattern was observed for GII introns, which have been previously found in some Shewanella spp. genomes (Quiroga and Centrón, 2009) although their distribution was not related to the source or the lineage. Incidence of ISs and GII introns was lower in CG1 and Cluster A, suggesting that a common ancestor to lineages encompassing species from Cluster B up to Cluster C may have acquired some of these MGEs (such as, ISSod19, ISSpu8, ISSod20, ISSba4, and ISSba5) and maintained them thereafter (Supplementary Table 10).

On the other hand, prophages were heterogeneously distributed in all lineages regardless of their source, where a single genome may contain up to six different prophages. We did not observe their conservation within specific lineages, which may be explained by the independent evolution of each genome leading to gain/loss of these MGEs. Conversely, ICEs from the SXT/R391 family showed a limited dissemination among Shewanella spp. These ICEs have been frequently found in the genus Vibrio, in some Enterobacterales (Proteus, Providencia, and Klebsiella) and in a few Shewanella spp. isolates, which reveals a narrow spectrum of hosts. Their modest distribution may be the result of several factors that affect the mating process, such as availability of surface receptors, environmental conditions that perturb mating-pair stabilization, the presence of exclusion mechanisms or defense systems in the recipient cell (Neil et al., 2021).

Furthermore, our work provides insightful data regarding plasmids circulating in Shewanella spp. We were able to identify replicases from broad host dissemination incompatibility groups (IncA/C, IncP, and IncX) frequently found in clinical isolates (Sekizuka et al., 2011; Carattoli et al., 2014; Jousset et al., 2018), as well as unique replicases in this genus, reflecting an unexpected diversity. In addition, several isolates from different lineages harbored more than one plasmid. The ability of these elements found in isolates from various sources (clinical or environmental) to replicate and coexist in different Shewanella spp. unfolds new scenarios involving HGT that could lead to a serious problem in the near future. Accordingly, recent reports showed that Shewanella spp. can have ARG-bearing plasmids, which share the same backbone as those found in common bacterial pathogens, such as Enterobacterales, from which they most likely were acquired (Sekizuka et al., 2011; Carattoli, 2013; Jousset et al., 2018).

Increase of ARGs in Shewanella spp. indicates that this genus is evolving steadily toward extensively drug resistance (XDR) phenotypes (Ohama et al., 2021). In this regard, several MDR and XDR isolates from our dataset were recovered from clinical samples or hospital environments. Notwithstanding, strain Shewanella sp. ALD9 had an MDR phenotype and it was isolated from a marine environment with limited human impact. The fact that this MDR isolate can be recovered from a niche where antibiotic pressure is probably slim, reflects that the ARG exchange can occur in their natural habitats and allows us to forecast its impact on the emergence of XDR Shewanella spp. in clinical settings that will lead to infections difficult to treat.

Evolution toward XDR in Gram-negative bacteria is commonly related to mobile integrons (Centrón and Roy, 2002; Hall, 2012). The link between class 1 integrons and ecosystems with high anthropic impact has been previously reported, as well as their contribution to the spread of ARGs (Chamosa et al., 2017). These integrons were also found in Shewanella spp. isolated from a bovine fecal sample (Barlow et al., 2004). Similarly, S. baltica CW2 carrying a class 1 integron was isolated from the gut of a lake trout (Castillo et al., 2018). The symbiotic relationship between some Shewanella species with marine animals provides an alternative environment where conditions may promote the acquisition of ARGs leading to an increase of the resistance level in non-human niches by single HGT events that may involve plasmids circulating in clinical isolates harboring mobile integrons.

Regarding the ARGs detected in this genus, we found the non-mobile colistin resistance gene mcr in a few Shewanella species. Zhang et al. (2019b) proposed that these genes may have emerged from bacteria found in aquatic niches. Since a few genes were found in only five Shewanella spp. isolates, it is unlikely that this bacterium is its native host. Furthermore, we observed quinolone and carbapenem resistance genes that were ubiquitous within specific pathogenic lineages. The association of bla_{OXA–48–like} and bla_OXA–729 (former bla_OXA–55) genes with S. xiamenensis and S. algae, respectively, has been previously reported (Tacão et al., 2018; Ohama et al., 2021). Here we propose a possible third link between bla_OXA–900 and the S. putrefaciens lineage. Interestingly all these species belonged to the respective pathogenic lineages CG2 and CG1, which suggest that either they may have been acquired before speciation or that these genes were transferred by MGEs capable of inserting at specific loci in the chromosome. Further studies focused on the genetic surroundings of each of these genes employing a larger number of genomes may provide new insights into their ubiquitous nature. On the other hand, we found qnrA variants in Shewanella spp. from CG1, which is consistent with previous reports (Araújo et al., 2021; Huang et al., 2022). The fact that some Shewanella spp. encoding ubiquitous ARGs exposes a troublesome scenario where an isolate may be able not only to acquire MGEs carrying several ARGs, but also to become resistant to a broader spectrum of antibiotics by introducing point mutation in key genes. In this regard, enrichment analysis also showed the overrepresentation of GO terms related to ARGs ampC, emrE, folM, folX, nagA, nfnB, and rfbC.

Although Shewanella spp. can acquire ARGs-bearing MGEs (Parmeciano Di Noto et al., 2016; Jousset et al., 2018), S. algae, the most frequently reported species that causes infections in humans, showed the least diversity and frequency of plasmids and very few ARGs, which is consistent with previous reports (Huang et al., 2022). Conversely, most MDR and XDR isolates were found in the monophyletic group CG2. This cluster along with lineages S. xiamenensis, S. baltica, and S. putrefaciens had a higher rate of plasmids and ISs in their genomes, showing remarkable plasticity. The unique behavior of each lineage may be due to intrinsic features of the host that prevent plasmid invasion or maintenance, such as entry exclusion, restriction/modification and CRISPR-Cas systems (Hall, 2012). The formers are capable of hindering the invasion of plasmids and other MGEs, such as phages or ICEs, into a new host (Koonin and Makarova, 2009). Nevertheless, we found that isolates carrying these systems can also harbor different MGEs. Although a negative correlation among these elements has been previously reported for other bacteria; the opposite association has also been observed (Gophna et al., 2015; O’Meara and Nunney, 2019), which suggests that there are other factors that may govern this interaction. Accordingly, the sole presence of a CRISPR-Cas system in a genome does not denote activity. The type I-E from Escherichia coli is known for being tightly regulated and it is possible to turn it on under controlled laboratory conditions (Pul et al., 2010). Activity of CRISPR-Cas systems in Shewanella spp. may also be inhibited during DNA invasion by MGEs, and particular environmental conditions might be necessary to activate them. In the appropriate conditions, these systems may be fully functional and capable of acquiring immunity against new MGEs, evidenced by the presence of arrays with >40 CRISPRs. On the other hand, CRISPR-Cas systems may have been acquired by HGT events, and therefore its acquisition may have occurred after the invasion of an MGE. Last, it is known that MGEs may encode proteins, known as anti-CRISPRs, that overcome CRISPR-Cas machinery and evade host immunity (Stanley and Maxwell, 2018). Of note, Westra and Levin (2020) analyzed the interaction between phages and plasmids and proposed that it is dependent not only on MGE abilities but also on the strong complexity of each ecosystem. Further studies are necessary to unravel the interplay between these defense systems and MGE dissemination in a unique bacterium such as Shewanella.

A major question that requires further analysis is the identification of virulence genes that may be involved in Shewanella spp. pathogenesis. Our work led to the identification of various virulence genes in many species using two different approaches, comparative sequence analysis and GSEA. The implementation of GSEA allowed us to identify additional functional GO terms enriched in lineages comprising clinical samples that may encode certain molecular functions or biological processes contributing to their virulence. Among the virulence determinants detected or overrepresented in CG genomes, we found several genes related to adherence and colonization of different hosts (wecE and ycdX), toxicity (e.g., toxA, cysC, plcN, and cylR), secretion (T2SS, T6SS, and T3SS), swarming and swimming motility (e.g., acf, che, fla, fle, fli, flg, ycdX, and yeaJ), iron metabolism (e.g., acs, pch, ccm, sit, ftnA, hmuR, and tuf), among others. Interestingly, the T6SSs were found in all S. algae strains while gene vgrG1 was enriched in CG1. Linares et al. (2016) suggested that these systems probably participate during Shewanella’s infection process. In order to confirm whether they are key pathogenicity factors in S. algae, further studies should be conducted. In addition, we found several other virulence genes encoded in this species, which were consistent with previous results (Wang et al., 2020; Zago et al., 2020). We also identified other potential candidate genes overrepresented in CG genomes, such as speCEG (spermidine synthesis), potE (putrescine transport) (Fang et al., 2017; Guerra et al., 2018), cobS (cobalamin biosynthesis) or genes related to oxidative stress response, specifically the peroxidase encoding gen katG. KatG has been previously reported as an important factor that may generate a stronger response during oxidative stress and increase survival during infection (Jiang et al., 2014). Many of these predicted virulence factors may have as main purpose to facilitate long-term interactions with other organisms or to collaborate with Shewanella’s survival and adaptation to different habitats (i.e., in aquatic niches or within a host). In this regard, metagenomic studies have shown the relevant abundance of members of the Shewanella genus in fish gut and invertebrates, which may vary under different conditions (Parris et al., 2016; Tepaamorndech et al., 2020; Jia et al., 2022; Johny et al., 2022). Moreover, these bacteria were also detected in the gastric and colorectal mucosal microbiomes of healthy and sick patients (Angelakis et al., 2016; Yang et al., 2022), showing an unexpected incidence. The true role of these factors during Shewanella’s colonization or infection processes in different hosts is still unknown.

The in-depth study of GSEA results also allowed us to identify an overrepresentation of certain paralogs genes in isolates from the CGs, suggesting that they may be involved in Shewanella spp. pathogenicity, i.e., katG, cysQ, hemB, kefB, and modA genes. katG paralogs are unevenly distributed between CGs and NCGs. Our results showed that an early duplication followed by a sequence divergence process may have occurred in the environment in both CGs. These genomes tend to maintain both paralogs, which might result in better fitted bacteria capable of thriving in clinical settings or during host infection/colonization. On the other hand, variants for genes cysQ, hemB, kefB, and modA showed lower identity values and they were mostly present in isolates from the CG. Each of these genes has the potential to contribute to Shewanella spp. virulence at different levels during a colonization or infection process, either adapting the environment to its own benefit or to increase its chances of survival against the host immune system. The differential retention of paralogs and its effects on the virulence of Shewanella spp. is something that deserves further investigation.

Shewanella spp. may cause infectious diseases in humans and in aquatic animals as well as to thrive as symbiont or in environmental niches. In the context of the One Health strategy, the exchange and evolution of genetic elements, such as MGEs or other GIs, seem to be contributing to the emergence of more virulent and resistant species and will affect their treatment with first line antibiotics in the near future. Our data allow us to evidence the ongoing transformation of the genus Shewanella into a worrisome pathogen with profound consequences for human and animal health. In this regard, we advocate that a more active role in the correct identification of this pathogen may reduce its potential hazard, thus improvements on the molecular identification of known and potential pathogenic lineages described in this study are necessary. In addition, since some virulent lineages, such as S. algae and S. xiamenensis, have ubiquitous resistance genes in their genomes, we propose their use as biomarkers.

Based on our results, we can attest that the evolution toward pathogenicity in this genus is a complex process that probably occurs idiosyncratically in each lineage. This process may involve long-term evolution of multiple genes, i.e., point mutations and duplications, in parallel with the continuous genetic exchange among bacteria. Further studies on Shewanella spp. virulence will provide an insight into the colonization and/or infection processes.

The authors confirm all supporting data, code, and protocols have been provided within the article or through Supplementary material and they can be found in the following links https://github.com/LBC-Iriarte/Shewanella_genomics and https://figshare.com/authors/QuirogaLab_IMPaM_UBA-CONICET_/14600396. All genome sequences (144) are listed in Supplementary Table 1, and GFF, FAA, and FNA files of the genomes are available at https://figshare.com/authors/QuirogaLab_IMPaM_UBA-CONICET_/14600396.

AI and CQ: conceptualization, resources, and supervision. GC, GT, TA, AI, and CQ: data curation and visualization. GC, GT, TA, GP, MR, DC, AI, and CQ: formal analysis and writing—review and editing. CQ: funding acquisition. GC, GT, AI, and CQ: investigation, methodology, software, and writing—original draft. All authors contributed to the article and approved the submitted version.