Plasmidome of Listeria spp.—The repA-Family Business

Bacteria of the genus Listeria (phylum Firmicutes) include both human and animal pathogens, as well as saprophytic strains. A common component of Listeria spp. genomes are plasmids, i.e., extrachromosomal replicons that contribute to gene flux in bacteria. This study provides an in-depth insight into the structure, diversity and evolution of plasmids occurring in Listeria strains inhabiting various environments under different anthropogenic pressures. Apart from the components of the conserved plasmid backbone (providing replication, stable maintenance and conjugational transfer functions), these replicons contain numerous adaptive genes possibly involved in: (i) resistance to antibiotics, heavy metals, metalloids and sanitizers, and (ii) responses to heat, oxidative, acid and high salinity stressors. Their genomes are also enriched by numerous transposable elements, which have influenced the plasmid architecture. The plasmidome of Listeria is dominated by a group of related replicons encoding the RepA replication initiation protein. Detailed comparative analyses provide valuable data on the level of conservation of these replicons and their role in shaping the structure of the Listeria pangenome, as well as their relationship to plasmids of other genera of Firmicutes, which demonstrates the range and direction of flow of genetic information in this important group of bacteria.


Introduction
The genus Listeria (Firmicutes) comprises 26 species, but studies have mainly focused on only one, L. monocytogenes, a pathogenic bacterium causing rare but severe food poisoning that can be fatal to unborn children and individuals with a compromised immune system [1]. Listeria spp. strains are ubiquitous in natural and agricultural environments. They are also commonly found in different food processing scenarios where they can tolerate a wide range of stress conditions, such as the presence of sanitizers, high salinity, acidity and non-optimal temperatures (e.g., [2][3][4][5]).
Many genes that may contribute to the ability of Listeria spp. to colonize various environments have been identified within mobile genetic elements, mainly plasmids. These genes can confer resistance to antibiotics [6,7], sanitizers [8,9], heavy metals [10][11][12] or elevated temperatures [13]. Several studies have identified plasmids that are commonly present and highly conserved among Listeria strains isolated from various sources and different geographic locations, which suggests a strong selection pressure for the maintenance of these replicons [12,[14][15][16][17][18][19].
Recently, a large-scale survey of plasmid sequences extracted from whole genome sequencing (WGS) data from almost 2000 L. monocytogenes strains was performed by Schmitz-Esser and colleagues [15]. This study indicated that plasmids are common in this species (54% of tested strains harbored such replicons), although their distribution varied greatly depending on the L. monocytogenes sequence type (ST) and lineage. WGS data are a valuable source of information on the distribution and diversity of bacterial plasmids, but they provide limited information on the structure, genetic load and evolution of these replicons due to the often incomplete sequences obtained.
Little is known about plasmids occurring in Listeria species other than L. monocytogenes, even though the conjugative transfer of replicons conferring heavy-metal and sanitizer resistance between nonpathogenic and pathogenic species of this genus has been documented [20][21][22]. Apart from the plasmid sequences obtained in our laboratory, 56 unique complete plasmid nucleotide sequences from L. monocytogenes were present in GenBank (NCBI) as of 6 December 2020, compared to only three from L. innocua and one from L. grayi.
In the present study, we obtained the complete nucleotide sequences of 50 plasmids isolated from strains representing six Listeria species, originating from natural and food processing environments. Combined with data available in GenBank and from our previous studies, we analyzed a set of 113 complete plasmid sequences from different Listeria species to provide an in-depth insight into their diversity, genetic load and potential contribution to the stress responses of these bacteria, as well as the impact of transposable elements (TEs) on the evolution of these replicons.

Identification and Classification of Plasmids Occurring in 782 Listeria spp. Strains
We gathered a large collection of Listeria spp. strains, isolated from various environments, that are subject to different anthropogenic pressures. In our previous studies [4,20,23,24] we analyzed the plasmid content of 566 strains. In the present study, the plasmid profiles of an additional 216 strains were determined. The entire collection of 782 strains represents seven Listeria species: L. monocytogenes (526 strains), L. innocua (105), L. seeligeri (71), L. welshimeri (59), L. ivanovii (16), L. marthii (4) and L. grayi (1) (Supplementary Table S1). The majority of the strains (627) originated from various food and food production-associated sources, while the others (155) came from environmental soil and water samples. The strains carried 374 plasmids in total: 18 small replicons (<10 kb; present in 2% of tested strains) and 356 larger replicons (>15 kb; harbored by 46% of strains). Only four strains maintained two plasmids each.
The Listeria spp. strains were classified into 59 groups based on their plasmid restriction fragment length polymorphism (RFLP) patterns ( Figure 1, Table S1). The highest numbers of plasmid RFLP profiles were found in strains of L. monocytogenes (35 groups) and L. innocua (13 groups), i.e., the two most numerous species in our strain collection. In three cases, plasmids with the same RFLP profile were observed in strains of different species, suggesting their horizontal transmission ( Figure 1). The majority of RFLP profiles were found only in single strains. However, the most common plasmid group was present in as many as 102 L. monocytogenes strains. Plasmids with identical RFLP profiles were found in strains originating from both food-associated and environmental sources in only three cases (Figure 1).

A dataset of Fully Sequenced Listeria spp. Plasmids
The obtained sequence data were combined with complete plasmid sequences available in the GenBank database (as of 6 December 2020): 56 of L. monocytogenes, three of L. innocua, 1 of L. grayi. A few other sequences (deposited as "complete plasmid sequences") were excluded from the dataset because they either appeared to be partial sequences or large fragments of chromosomal DNA (Table S2). As a result, a total of 113 unique complete plasmid sequences were further analyzed (their characteristics and GenBank accession numbers are presented in Table S2).
Interestingly, all of the replicons referred to as large plasmids (106 sequences, ranging in size from 15 kb to 153 kb) have related replication systems (REP) encoding a RepA replication initiation protein. These RepAs, containing the helical PriCT_1 domain (PFAM: PF08708) found in the C-termini of primases, show sequence similarity to the replication initiation protein of the well-studied theta-replicating plasmid pAMβ1 of Enterococcus faecalis, representing the Inc18 group of plasmids [12,27].

A Dataset of Fully Sequenced Listeria spp. Plasmids
The obtained sequence data were combined with complete plasmid sequences available in the GenBank database (as of 6 December 2020): 56 of L. monocytogenes, three of L. innocua, 1 of L. grayi. A few other sequences (deposited as "complete plasmid sequences") were excluded from the dataset because they either appeared to be partial sequences or large fragments of chromosomal DNA (Table S2). As a result, a total of 113 unique complete plasmid sequences were further analyzed (their characteristics and GenBank accession numbers are presented in Table S2).
Interestingly, all of the replicons referred to as large plasmids (106 sequences, ranging in size from 15 kb to 153 kb) have related replication systems (REP) encoding a RepA replication initiation protein. These RepAs, containing the helical PriCT_1 domain (PFAM: PF08708) found in the C-termini of primases, show sequence similarity to the replication initiation protein of the well-studied theta-replicating plasmid pAMβ1 of Enterococcus faecalis, representing the Inc18 group of plasmids [12,27].

Conserved Backbone and Structure of the repA-Family Plasmids
The plasmidome of Listeria spp. is dominated by a large group of replicons, designated in this study as repA-family plasmids. Comparative analysis revealed that the conserved backbone of the vast majority (103/106) of these replicons consists of three components, placed in a conserved order: (i) the repA gene of the REP module, (ii) two genes probably involved in active partitioning of plasmid molecules (PAR), and (iii) two or three genes encoding putative subunits of a Y-family DNA polymerase (POL). The plasmids also contain numerous transposable elments (TEs), which have influenced the structure of these replicons.

Conserved Backbone and Structure of the repA-Family Plasmids
The plasmidome of Listeria spp. is dominated by a large group of replicons, designated in this study as repA-family plasmids. Comparative analysis revealed that the conserved backbone of the vast majority (103/106) of these replicons consists of three components, placed in a conserved order: (i) the repA gene of the REP module, (ii) two genes probably involved in active partitioning of plasmid molecules (PAR), and (iii) two or three genes encoding putative subunits of a Y-family DNA polymerase (POL). The plasmids also contain numerous transposable elments (TEs), which have influenced the structure of these replicons.

Replication Systems
Most of the analyzed Listeria spp. plasmids can be categorized into two major RepA phylogenetic groups (G1 and G2), distinguished in previous studies [12,15,16] ( Figure S1). The majority of the plasmids occur in L. monocytogenes and L. innocua strains (Figure 2a). Two plasmids, pN1-011A (149 kb; NC_022045.1) and pCFSAN021445 (152 kb; NZ_CP022021.1), each carry two separate replication systems (G1 and G2), proving that they are composite molecules, probably generated by recombination events between two individual replicons. The RepA proteins of several plasmids sequenced in this study fall into three additional phylogenetic groups (named G4-G6). Plasmids of the group G4 were described in two recent studies [15,29]-they rarely occur, but encode a putative sanitizer resistance module. Plasmids classified within groups G1, G4 and G5 are on average smaller than those in the other groups. The G5 group is the most divergent, and two plasmids from this group, pLIS46 (MZ147617) and pLIS47 (MZ147618), were harbored by one L. ivanovii strain (Sr19), suggesting that they are compatible and can stably coexist in the same cell. The RepA proteins of several plasmids sequenced in this study fall into three additional phylogenetic groups (named G4-G6). Plasmids of the group G4 were described in two recent studies [15,29]-they rarely occur, but encode a putative sanitizer resistance module. Plasmids classified within groups G1, G4 and G5 are on average smaller than those in the other groups. The G5 group is the most divergent, and two plasmids from this group, pLIS46 (MZ147617) and pLIS47 (MZ147618), were harbored by one L. ivanovii strain (Sr19), suggesting that they are compatible and can stably coexist in the same cell.
To obtain a more comprehensive picture of the diversity and distribution of RepA proteins in the genus Listeria, a PSI-Blast search of the GenBank protein database (as of 24 April 2021) was performed. This analysis resulted in the identification of 418 unique complete RepA proteins from 13 Listeria species (phylogenetic tree presented in Figure 2b; a more detailed description of the dataset is presented in Figure S2). To estimate the prevalence of RepA proteins, the number of sequenced Listeria spp. strains encoding specific RepA proteins was retrieved from the NCBI database and shown in Figure S2. RepA proteins belonging to G1 and G2 groups were the most commonly found in the sequenced strains. However, it should be noted that most sequenced strains available in the GenBank database belong to L. monocytogenes species. Interestingly, over 90% of the Listeria strains possessing a RepA protein encoded one of the eight most common variants, with the dominant variant (WP_003728500.1-G1) found in over 6000 strains ( Figure S2), and in 48 of the completely sequenced plasmids. Plasmids encoding RepA proteins were found in different Listeria species, although some host preferences could be observed, e.g., RepA groups G1 and G2 were not common in L. seeligeri and L. ivanovii strains, while members of G6 were detected exclusively in the former (Figure 2b).
Interestingly, seven of the repA-family plasmids-pLIS2 (MW934262; G1), pLIS9 (MZ043156; G1), pLIS10 (MZ043158, G1), pLIS35 (MZ127844, G1), pLIS37 (MZ127846, G1), pCFSAN010068_01 (NZ_CP014251.1, G1) and pLI100 (NC_003383.1, G2)-carry additional (1 or 2) complete or partial REP modules, unrelated to the repA family, or PAR modules located in separate regions of the plasmid genome. Such intra-replicon gene transfer may be associated with transposition and recombination events, since the aforementioned genetic modules are often part of putative composite transposons, flanked by insertion sequences (ISs). All but one of the transferred REP modules (one of the additional REP present in pLI100) are related to each other and show sequence similarities to many Lactococcus spp. plasmids. The REP of pLI100 is similar to those of Listeria RCR plasmids pDB2011 and pIP823. Such natural bi-replicon plasmids may potentially have the properties of shuttle replicons and extend the host range.
There have been few studies on the host range of large Listeria plasmids, and these are usually limited to species within this genus or do not involve plasmid DNA sequencing, e.g., [21]. To test the host range of the analyzed repA-family plasmids, a series of mobilizable shuttle plasmids was constructed in Escherichia coli. These were composed of (i) vector pDKE2 (contains the R6K origin of replication that is functional exclusively in E. coli strains encoding R6K replication initiation protein Pi in trans) and (ii) the REP + PAR or REP + PAR + POL modules ( Figure S3) of five Listeria plasmids (pLIS36-MZ127845, pLIS1, pLIS26-MZ090006, pLIS6, pLIS50-MZ151535) representing the G1, G2, G4 and G5 RepA phylogenetic groups, respectively ( Figure S1).
The constructed plasmids were capable of replication in all tested Listeria species (L. monocytogenes, L. seeligeri, L. ivanovii, L. innocua, L. grayi) and B. subtilis. With Staphylococcus aureus, numerous transconjugants were obtained, but autonomous forms of the introduced plasmids were observed only for the G2 and G4 REP systems. None of the tested plasmids was able to replicate in E. coli DH5α (Gammaproteobacteria) or Paracoccus pantotrophus KL100 (Alphaproteobacteria). These observations indicate that the host range of these repA-family plasmids may be limited to gram-positive bacteria.

Stable Maintenance Systems
The observed repA genes are associated and adjacent with putative parAB operons encoding partitioning systems (PAR) that promote stable maintenance of bacterial plasmids. The PAR is responsible for active segregation of newly replicated plasmid copies into daughter cells during cell division, although the putative PAR systems present in Listeria plasmids were not yet functionally tested.
The PAR systems of Listeria plasmids (classified to group Ib) are related, and can be roughly divided into 11 phylogenetic groups ( Figure S4). They mostly occur together with specific RepA types (e.g., G1, G4), but in some cases (e.g., G2, G6) similar repA genes are associated with PAR systems representing different phylogenetic groups and vice versa, i.e., related PAR systems can be associated with distinct RepA proteins. For example, plasmids pLIS4 and pLIS5 encode highly similar ParA and ParB proteins (95% and 87% amino acid (aa) identity, respectively), while their RepA proteins are more divergent (41% aa identity, groups G6 and G5) ( Figure 3). This reflects both the modular structure of Listeria spp. plasmids and recombinational gene exchange occurring between different repA-family replicons. As many as 54 plasmids encode an additional putative ParA-family protein (100% aa sequence identity), located close to the conserved plasmid backbone (PAR; Figure 3; G1 and G2 RepA groups). PAR systems were not found in any of the small Listeria RCR plasmids. associated with PAR systems representing different phylogenetic groups and vice versa, i.e., related PAR systems can be associated with distinct RepA proteins. For example, plasmids pLIS4 and pLIS5 encode highly similar ParA and ParB proteins (95% and 87% amino acid (aa) identity, respectively), while their RepA proteins are more divergent (41% aa identity, groups G6 and G5) ( Figure 3). This reflects both the modular structure of Listeria spp. plasmids and recombinational gene exchange occurring between different repA-family replicons. As many as 54 plasmids encode an additional putative ParA-family protein (100% aa sequence identity), located close to the conserved plasmid backbone (PAR; Figure 3; G1 and G2 RepA groups). PAR systems were not found in any of the small Listeria RCR plasmids. Many repA-family plasmids (but none of the small Listeria RCR plasmids) contain additional stabilization modules, which act by a mechanism involving post-segregational killing of plasmid-less cells. Such systems are usually two-gene operons, encoding a stable toxic protein and a labile antitoxin that neutralizes the toxin's activity [30].
The most common of these stabilization modules in Listeria plasmids are type II toxinantitoxin (TA) systems, representing two TA families: (i) mazEF/pemIK-occurring in 49 replicons (47 of these systems are highly similar) and (ii) relBE/parDE-two such TA loci are present in a total of eight plasmids. In addition, putative toxins of type III TA systems were Many repA-family plasmids (but none of the small Listeria RCR plasmids) contain additional stabilization modules, which act by a mechanism involving post-segregational killing of plasmid-less cells. Such systems are usually two-gene operons, encoding a stable toxic protein and a labile antitoxin that neutralizes the toxin's activity [30].
The most common of these stabilization modules in Listeria plasmids are type II toxinantitoxin (TA) systems, representing two TA families: (i) mazEF/pemIK-occurring in 49 replicons (47 of these systems are highly similar) and (ii) relBE/parDE-two such TA loci are present in a total of eight plasmids. In addition, putative toxins of type III TA systems were identified in 24 plasmids (Table S2). Plasmids belonging to the G1 RepA group mostly encode type II TA systems, although type III and other putative TA systems are also present. Type III systems are more common in plasmids from the G2 group (Table S2).
None of the TA modules has been experimentally analyzed so far, although a few studies have been performed on chromosomally encoded TAs of L. monocytogenes strains [31][32][33].

Y-Family DNA Polymerase Modules
Intriguingly, most of the analyzed repA plasmids contain 2 or 3 conserved genes located near the PAR modules (POL module) (Figure 3), which encode putative UvrX and YolD subunits of the Y-family DNA polymerase-possibly equivalents of UmuC and UmuD from gram-negative bacteria [34,35]. Y-family polymerases act as DNA repair systems, which are activated in response to DNA damage (e.g., caused by UV radiation) and specialize in translesion DNA synthesis, bypassing damaged bases that would otherwise block the replication forks [36]. The activity of these polymerases may stimulate spontaneous deletions during DNA repair [37], and could therefore contribute to variation and adaptation of both plasmid and host genes [12].
The genes encoding Y-family polymerases are components of the global SOS system, and their expression may be regulated by the transcriptional repressor LexA [34,35]. Sequences similar to L. monocytogenes SOS operator sequences (LexA-boxes) [34,38] were identified upstream of the POL modules in most of the analyzed plasmids (data not shown).
The larger UvrX subunits display a high level of aa sequence conservation (62-100% aa sequence identity) and form three main phylogenetic groups (POL1-POL3; Figures 3 and S5). The polymerase module of the POL2 group contains two putative yolD genes. One plasmid (pLIS51; MZ151536) carries an additional copy of these genes, located far from the REP module ( Figure 3).
In six of the repA plasmids, the transposition of TEs has caused disruption of the gene encoding one of the polymerase subunits ( Figure 3). In some of those cases (e.g., pLIS28, pLIS4) transposition caused shortening of the uvrX gene by only about 30 bp; it is possible that the proteins lacking some C-terminal aa are still functional. Interestingly, in five cases (pLIS4, pLIS5, pLIS28, pLIS48, pLIS54) the inserted element was a Tn554-like transposon that also harbors a homologous uvrX gene. Related UvrX proteins are also commonly encoded by the chromosomes of Listeria spp. ( Figure S5).

Conserved Structure of the repA-Family Plasmids
Comparative analysis revealed that sequence conservation among the repA-family plasmids is not limited to the core REP, PAR and POL modules. Many plasmids have similar structures and share numerous highly conserved DNA segments (Figure 4), as was observed in previous studies [12,[14][15][16][17]. The pan-genome of these plasmids consists of around 900 genes, which is a relatively small number considering the fact that these large plasmids carry about 70 genes on average (range: . The structure and gene content of the vast majority of Listeria repA plasmids is highly conserved, and several of the plasmids differ only by point mutations or the presence of small indels of <1 kb in size (Table S2). Sequence conservation between plasmids of different RepA lineages is shown in Figure 4 (a more detailed version, including plasmid names, is available in Figure S6).
In most cases, RepA clustering is associated with a similar gene content (Figure 4; also visualized by a protein-based similarity network of Listeria plasmids in Figure S7). Most of these plasmids contain numerous complete and partial TEs (including insertion sequences, mostly of the IS3, IS6 and IS21 families) and recombinase genes, with an average of around nine such elements and genes per plasmid. The highest number (24) was found in plasmid pCFSAN021445. In many cases, DNA regions of high nucleotide sequence similarity shared by different plasmids are bordered by TEs (Figures 4 and S6), suggesting that these elements may play an important role in the evolution of Listeria spp. plasmids by mediating diverse recombinational rearrangements. suggesting that these elements may play an important role in the evolution of Listeria spp. plasmids by mediating diverse recombinational rearrangements.

Conjugative Transfer Modules of Listeria spp. Plasmids
DNA transfer enables bacteria to acquire new traits. To our knowledge, conjugative transfer of only three of the sequenced Listeria plasmids has been experimentally confirmed: pLIS1 [20], pLM1686 (NZ_MK134858.1) [39,40] and a small mobilizable plasmid pIP823 [6]. In addition, it was demonstrated that a plasmid vector encoding a relaxase (an enzyme necessary for the initiation of plasmid conjugal transfer) from pLIS3 could be mobilized between E. coli strains in the presence of the helper conjugational system of plasmid RP4 [24]. Several approaches to transfer other plasmids in biparental mating experiments failed, which suggests that the tested replicons are either not self-transmissible or require specific conditions for efficient transfer [7,13,20,23,41] (indicated in Table S2).
Based on in silico predictions, it seems probable that 24 of the analyzed Listeria repA plasmids contain complete transfer modules (TRA) encoding a MobL-family relaxase [42],
In addition, it was demonstrated that a plasmid vector encoding a relaxase (an enzyme necessary for the initiation of plasmid conjugal transfer) from pLIS3 could be mobilized between E. coli strains in the presence of the helper conjugational system of plasmid RP4 [24]. Several approaches to transfer other plasmids in biparental mating experiments failed, which suggests that the tested replicons are either not self-transmissible or require specific conditions for efficient transfer [7,13,20,23,41] (indicated in Table S2).
Based on in silico predictions, it seems probable that 24 of the analyzed Listeria repA plasmids contain complete transfer modules (TRA) encoding a MobL-family relaxase [42], cell wall hydrolase, type IV secretion system (T4SS) proteins, a putative toprim domain-containing protein, proteins similar to pilus assembly complex components, surface adhesion proteins, plus many other hypothetical proteins of unknown function ( Figures S8 and S9). The predicted TRA modules of plasmids pLIS1 and pLM1686, which were successfully conjugatively transferred between Listeria strains, are approx. 25-30 kb-long and are present in 11 and nine plasmids, respectively.
Interestingly, over half of the repA-family plasmids from group G2 and all of the G4 group contain complete predicted TRA modules, while plasmids from the G1, G3 and G5 groups seem to carry only partial modules. However, any plasmid carrying a relaxase gene can still potentially be mobilized for conjugative transfer when provided with a functional TRA module in trans, for example by another plasmid present in the same cell, or by an integrative conjugative element (ICE) integrated within the chromosome. Most of the small Listeria plasmids have the potential to be mobilized for transfer, except two replicons, pAUSMDU00000235 and pLIS55, that do not encode predicted relaxases.

Adaptive Genes of Listeria spp. Plasmids
Listeria spp. strains can be subjected to diverse environmental stressors; during food production processing, they are often exposed to antimicrobials, high salinity, acidity and non-optimal temperatures. The gene content of the Listeria spp. plasmids was analyzed in order to identify genes possibly involved in the response to stress factors, which may contribute to the adaptation of Listeria spp. to diverse and changeable environments. Many of the identified modules are shared among plasmids from various Listeria species, as well as with bacteria of other genera, suggesting their horizontal transmission. Many of these modules occur within putative composite and non-composite transposons. A summary of this analysis is presented in Figure 5, and the distribution of the identified genes in individual plasmids is presented in Table S2.
Interestingly, even though plasmid pLIS48 (MZ147619) contains a putative novel cadAC module, the strains carrying pLIS48 or pLIS48-like plasmids were not resistant to cadmium-so either the genes are non-functional, or they confer resistance to other factors [23].
Most of the resistance determinants were harbored by non-composite transposons Tn5422 (cadA1C1), Tn6870 (cadA6C6) or Tn7130 (putative cadAC cassette in pLIS48), or a composite transposon Tn6869 (pLIS4), which also encodes a set of proteins involved in carbohydrate transport and metabolism [10,23]. The cadA2C2 cassette is adjacent to serine recombinase and transposase genes, suggesting that it could also be an integral part of a mobile element. Arsenic resistance in Listeria spp. is usually determined by the presence of chromosomally inserted genomic island LGI2 (arsR1D2R2A2B1B2 and upstream arsA1D1 cassette) or a Tn554-like transposon (arsCBADR) [44,46]. These loci were linked with plasmids in only a few cases: (i) a Tn554-like transposon was found within five plasmids, inserted near the PAR modules (Figure 3), (ii) a partial sequence (arsADR) was identified in two plasmids, and (iii) an arsR1D2R2A2B1B2 operon (flanked by TEs), 99% identical to the locus within island LGI2, was detected in one plasmid (pLI100) ( Table S2). Two plasmids (pLIS24 and pLIS36) carry another 3-gene cluster, arsRBC, highly similar (99% nt identity) to operons occurring in many S. aureus plasmids and chromosomes, which was shown to confer arsenic resistance in S. aureus and B. subtilis [49]. The strains harboring pLIS24 and pLIS36 were capable of growth in the presence of (meta)arsenite at high concentration (500 μg/mL), which suggests the functionality of this plasmid resistance locus in Listeria. Interestingly, most (8/10) plasmids harboring the aforementioned arsenic resistance determinants were isolated from species other than L. monocytogenes ( Figure 5).
Numerous Listeria spp. plasmids (44/113) harbor genes that are likely to be involved in copper detoxification. A copper transporter system gene ctpA was identified in a plasmid of L. monocytogenes DRDC8 (U15554.3), and is part of a 10-gene cluster involved in copper homeostasis and virulence [50][51][52]. This cluster is present in nine complete L. monocytogenes plasmids, of which eight are highly similar replicons (Table S2).
A complete putative copper-resistance cluster consisting of genes for a CopB-family copper-translocating P-type ATPase, multi-copper oxidase Mco and a YdhK-family protein of unknown function, is present in two plasmids, pLIS28 and pLIS54 (Table S2). Two Numbers in boxes represent the number of plasmids harboring specific genes. Abbreviations are as follows: QAC-quaternary ammonium compound, Cd-cadmium, Cu-copper, As-arsenic, LGI2-like-the arsenic resistance operon similar to that in genomic island LGI2, HM1-HM3heavy metal ATPases, Gly/Bet-glycine/betaine ABC transporter, Fe/OXY-proteins putatively involved in iron export or the oxidative stress response, CPA-cation/proton antiporter, MFS-major facilitator superfamily transporters, MATEmultidrug and toxic compound extrusion family pumps, ARG-antibiotic resistance genes. Frame-shifted genes were included when present.
Arsenic resistance in Listeria spp. is usually determined by the presence of chromosomally inserted genomic island LGI2 (arsR1D2R2A2B1B2 and upstream arsA1D1 cassette) or a Tn554-like transposon (arsCBADR) [44,46]. These loci were linked with plasmids in only a few cases: (i) a Tn554-like transposon was found within five plasmids, inserted near the PAR modules (Figure 3), (ii) a partial sequence (arsADR) was identified in two plasmids, and (iii) an arsR1D2R2A2B1B2 operon (flanked by TEs), 99% identical to the locus within island LGI2, was detected in one plasmid (pLI100) ( Table S2). Two plasmids (pLIS24 and pLIS36) carry another 3-gene cluster, arsRBC, highly similar (99% nt identity) to operons occurring in many S. aureus plasmids and chromosomes, which was shown to confer arsenic resistance in S. aureus and B. subtilis [49]. The strains harboring pLIS24 and pLIS36 were capable of growth in the presence of (meta)arsenite at high concentration (500 µg/mL), which suggests the functionality of this plasmid resistance locus in Listeria. Interestingly, most (8/10) plasmids harboring the aforementioned arsenic resistance determinants were isolated from species other than L. monocytogenes ( Figure 5).
Numerous Listeria spp. plasmids (44/113) harbor genes that are likely to be involved in copper detoxification. A copper transporter system gene ctpA was identified in a plasmid of L. monocytogenes DRDC8 (U15554.3), and is part of a 10-gene cluster involved in copper homeostasis and virulence [50][51][52]. This cluster is present in nine complete L. monocytogenes plasmids, of which eight are highly similar replicons (Table S2).
A complete putative copper-resistance cluster consisting of genes for a CopB-family copper-translocating P-type ATPase, multi-copper oxidase Mco and a YdhK-family protein of unknown function, is present in two plasmids, pLIS28 and pLIS54 (Table S2). Two variants of the CopB are encoded by these plasmids, differing by a 17 aa duplication. Genes encoding the shorter or longer variants of CopB are present in 32 plasmids, but the Mco gene is truncated by a transposon insertion and subsequent recombination events. The copB-mco operon is highly similar (99% nt identity) to those found in several S. aureus chromosomes and plasmids, including plasmid SAP078A (NC_019009).
The copB-mco operon of plasmid SAP078A confers copper hypertolerance in S. aureus strains and enhances bacterial survival inside macrophages [53]. Whether this operon performs a similar function in L. monocytogenes strains remains unknown. Interestingly, expression of the mco gene of some Listeria plasmids was shown to be upregulated during growth under salt and acid stress conditions [16,54]. In plasmids pLIS28 and pLIS54, the operon is located between two serine recombinase genes (this resistance locus is also adjacent to recombinase genes in most S. aureus sequences, including SAP078A).
Plasmid pLIS41 (MZ147614) encodes another putative copper detoxification cluster-a copper translocating P-type ATPase and a CopY-family repressor. This cluster shares 99-98% nt sequence identity with several Lactobacillus spp. and Lacticaseibacillus spp. plasmid and chromosome sequences. Ten other plasmids harbor the copY gene and a partial ATPase gene.
The analyzed plasmids also carry genes for three other putative heavy metal translocating P-type ATPase types, but the substrates of these proteins remain unknown. One of these proteins is encoded by 44 plasmids, of which 14 also carry genes encoding a second type of P-type ATPase (Table S2). The third type is encoded by only two plasmids-pLGUG1 (NC_014496) and pNH1 (NZ_MH277333) [7,12]. In pLGUG1 the ATPase gene forms part of a putative IS1216-driven (IS6 family) composite transposon, flanked by directly repeated sequences (DRs)-scars of a transposition event. The genes encoding ATPases of types two and three display high similarity (98-99% nt identity) to sequences present in many Enterococcus spp. strains.

Sanitizer Resistance
Resistance to quaternary ammonium compounds (QACs), which are commonly used as sanitizers in food-processing environments, is often observed in Listeria spp. strains [20,21,55,56]. Several different QAC efflux pumps have been identified in Listeria spp., two of which are commonly associated with plasmids-bcrABC [9] and emrC [8].
The bcrABC cassette is present in 20 plasmids (Table S2). In many cases, this cassette is associated with cadA2C2 and a triphenylmethane reductase gene tmr (present in some Listeria plasmids), and may be part of a putative composite transposon, although no flanking DRs were identified [9]. In plasmids pLIS33 and pCFSAN022990 [57], the cassette is part of another putative IS1216-driven composite transposon, flanked by DRs. The emrC gene was only detected in three nearly identical small plasmids, which differ by point mutations (Table S2).
Additionally, plasmids pLIS12 and pLIS26 encode a transcriptional regulator and a SMR protein, closely related (93% aa identity) to the qacH QAC resistance efflux pumps of Tn6188 [58]. Strains harboring this gene cassette were resistant to benzalkonium chloride (BC), but further experiments are required to confirm its specific function. Interestingly, the cassette is bordered by 37-bp-long inverted repeats (with 32/37 bp identity to those of the Tn3-family transposon Tn7129 present in 8 Listeria plasmids) and flanked by 5-bp-long DRs. This is a structure typical for non-autonomous mobile insertion cassettes (MIC). Closely related MICs (99% nt identity) are also present in plasmids from Enterococcus faecalis (LR962259.1) and Clostridioides difficile (LR595854.1) strains.

Antibiotic Resistance
Surprisingly, very few Listeria spp. plasmids have been connected with antibiotic resistance [7,22,[59][60][61]. Two small plasmids, pIP823 and pDB2011, were shown to confer resistance to trimethoprim (drfD) [6,22,60]. The latter replicon additionally determines resistance to spectinomycin (spc), and carries a probably non-functional erythromycin resistance gene (erm(A)) [22]. These two broad-host-range plasmids show little similarity to other Listeria plasmids, and were probably recently acquired from Staphylococcus spp., as has also been suggested in the case of two multidrug resistance plasmids pIP811 and pWDB100 (37 kb and 39 kb; undetermined sequences) [59,61]. Sequences similar to plasmids pIP823 and pDB2011 were identified in only a few Listeria strains (WGS data; GenBank), indicating that they are probably a recent acquisition by this genus.
Plasmid pNH1 contains multiple copies of IS1216, which can result in the generation of various composite transposons, one of which was shown to be functional (Tn6659) [7]. Plasmid pNH1 also carries a trimethoprim resistance gene (drfG) introduced by a putative TE (3 kb), which is also present in numerous Enterococcus spp. and Staphylococcus spp. plasmid and chromosome sequences (99-100% nt identity). This element was previously identified as a component of transposon Tn6198 [62].
Among the plasmids sequenced in this study only pLIS35 harbors two putative aminoglycoside phosphotransferase family proteins. However, none of the seven L. innocua and L. welshimeri strains harboring pLIS35 and pLIS35-like plasmids were resistant to any of the nine aminoglycosides when tested in antimicrobial susceptibility assays.

Heat Stress
An ATP-dependent protease ClpL from plasmid pLM58 was shown to be involved in increased heat resistance of L. monocytogenes strains [13]. The ClpL gene is present in more than half (58/113) of the analyzed Listeria plasmids. Interestingly, this gene was also found to be upregulated in response to low pH [16,54]. Genes encoding two other shorter proteins belonging to the ClpA/ClpB family, are present in eight plasmids (Table S2); they may also be potentially involved in a heat-shock response, but this has not been experimentally tested.

Oxidative, Acid and High Salinity Stresses
Listeria spp. strains are often exposed to high salinity and low pH conditions in food production facilities and food products. Genes putatively involved in osmotic, acid and oxidative stress responses were identified in Listeria plasmids based on similarity to homologs from other bacteria [12], but their functions have not been experimentally verified. Comparisons of wild-type and plasmid-cured strains indicate that some Listeria plasmids could play a role in growth and survival under salt, acid and oxidative stress conditions [63], but the underlying mechanisms remain unknown.
A region encoding a heavy metal resistance ATPase, a glycine betaine ABC transport system substrate-binding protein and an oxidoreductase (NADH peroxidase; npr) is present in 44 Listeria plasmids (Table S2). The latter two genes may play a role in responses to high osmolarity and oxidative stress, and were shown to be upregulated under high salt conditions [16,[64][65][66][67]. This DNA segment is flanked by TEs, and a similar region (99% nt identity) is present in an E. faecalis plasmid (LR961976.1). In three plasmids (pLIS2, pLIS9, pLIS37), this region forms part of a large ISLmo13-driven composite transposon (31 kb), bordered by DRs. Moreover, several E. faecalis, Weissella spp. and Aerococcus spp. strains harbor homologous genes encoding glycine betaine permease and oxidoreductase (97-99% nt identity).
Another DNA region, encoding a cation/proton antiporter (potential role in maintaining pH), an enolase and CrcB-family fluoride efflux transporters, is present in 11 plasmids, while two plasmids carry only the enolase and antiporter genes (Table S2). A region containing a putative iron export ABC transporter subunit and a protein with a ferritin-like iron-binding domain is present in 14 plasmids (Table S2). It has been suggested that these genes are involved in the oxidative stress response [63]. A highly similar region (99% nt similarity) is present in many Enterococcus spp. plasmids and chromosome sequences.
Finally, three plasmids (pLIS54, pLI100, pNH1) from three different Listeria species carry a six-gene cluster related to the Kdp-type potassium transport system adjacent to a serine recombinase gene. Homologous systems, present in many bacterial genomes, including L. monocytogenes chromosomes, are involved in osmotic and pH adaptation [68,69].

Other Genes of Potential Adaptive Value
Eleven Listeria plasmids encode a triphenylmethane reductase (tmr), which was shown to be involved in toxic dye (e.g., crystal violet) detoxification [70]. In all cases, the tmr gene is located close to the bcrABC cassette. Triphenylmethane dyes have industrial applications, and may be released into the natural environment. They are also illegally used in fish farming to prevent infections in fish [71]. All strains for which complete plasmid sequences were obtained were tested for crystal violet detoxification, and only those with plasmids encoding the tmr gene (pLIS1, pLIS8, pLIS16, pLIS29; Table S2) were capable of growth at high concentrations of crystal violet. However, strains harboring pLIS35 and pLIS35-like plasmids, which also carry the tmr gene, were, for unknown reasons, incapable of crystal violet detoxification.
Other genes potentially involved in the efflux of toxic compounds include two distinct MATE (multidrug and toxic compound extrusion) family efflux pumps. One is present in plGUG1 and the other in nine closely related plasmids (Table S2) where the MATE transporter is adjacent to the ctpA gene, a copper resistance determinant.
In addition, plasmids pLIS4 and pLIS5 carry a gene encoding a MFS (major facilitator superfamily) transporter and eight nearby genes putatively involved in carbohydrate transport and metabolism. Plasmid pLIS34 encodes a bi-functional aspartate transaminase/aspartate 4-decarboxylase and aspartate-alanine antiporter that are highly similar (99% aa identity) to proteins found in multiple Enterococcus spp. strains, and form part of a putative IS1062-driven composite transposon flanked by DRs.
Finally, the small plasmid pAUSMDU00000235 (2.7 kb) encodes a putative bacteriocin and bacteriocin immunity protein. A plasmid (2.9 kb) producing a functional bacteriocin was previously reported in a L. innocua strain [72]; the bacteriocin and immunity protein region was sequenced (AF330821), and it shares high nt similarity with the related region of pAUSMDU00000235.

Putative Virulence Factors
Most Listeria plasmids do not seem to enhance virulence. In fact, a survey of plasmids in the NCBI database suggested that extrachromosomal replicons are markedly less prevalent in clinical L. monocytogenes strains than in food-associated or environmental strains [15]. The first potential virulence plasmid described in L. monocytogenes was pLMIV-the only known Listeria plasmid that encodes four internalin-like proteins. However, the invasion efficacy in human epithelial cells of the pLMIV-cured strain did not differ from the parental strain [41]. Studies on the small pLMST6-like plasmids suggest that they may contribute to increased virulence of L. monocytogenes strains, although the mechanism remains unknown [73].

Listeria Plasmids Pangenome Estimation
Global comparative analyses of the Listeria spp. plasmid pangenome, in relation to the whole pangenome of Listeria spp. as well as to the plasmidome of bacteria of other genera of the type Firmicutes, were performed. The main goal was to collect information on: (i) the rate of DNA transfer between Listeria spp. plasmids and chromosomes, (ii) the role of plasmids in shaping the pangenome of these bacteria and (iii) the prevalence of related plasmids in other Firmicutes.
A total of 893 gene clusters were identified in the complete plasmid sequences of Listeria spp. analyzed using the bacterial pangenome analysis pipeline Panaroo (Table S3). Of these, 411 clusters were found only in plasmids sequenced in this study. Apart from transposases, subunits of the Y-family polymerase are among the most prevalent gene clusters found in the analyzed Listeria plasmids.
A total of 14,765 gene clusters were identified when the available complete sequences of both chromosomes (244 sequences) and plasmids of Listeria spp. were analyzed (Table S4). Only 119 of these gene clusters were found in both the analyzed chromosomes and the plasmids. However, it should be noted that the clusters created by Panaroo include both complete and partial genes, and that while the sequence identity is set to 95%, the Panaroo algorithm allows for clusters below this threshold based on additional context support. Interestingly, the Y-family polymerase gene cluster is also the most commonly found cluster shared between chromosomes and plasmids, being present in over 97% of the analyzed chromosomes and in around 85% of the plasmids.
Almost a quarter of the common gene clusters (present both in chromosomes and plasmids) are part of putative TEs (including genes carried by composite transposons) in some of the plasmids. Another quarter are parts of a prophage region present in one plasmid, pCFSAN008100b, and multiple chromosomes. Around 8% are part of the arsenic operon shared between plasmid pLI100 (where it is flanked by two ISs) and the chromosomal LGI2 genomic islands. Many of the remaining shared genes are located in plasmids in close proximity to genes encoding serine recombinases.
Panaroo identified a total of 73,928 gene clusters in the 5533 complete plasmid sequences from the phylum Firmicutes available in GenBank as of 6 December 2020 (Table S5). Only 178 of these clusters are shared between plasmids of Listeria and other Firmicutes-commonly with Enterococcus spp. and Lactococcus spp. plasmids. Around half of these plasmidborne clusters are connected with TEs, while around a quarter are putatively involved in adaptation to stress conditions. A protein-based similarity network (less stringent identity thresholds than the Panaroo analyses) was constructed to visualize similarities between Listeria spp. plasmids and those of other Firmicutes genera. Most of the repA-family plasmids of Listeria form a tight cluster of highly related replicons (Figure 6), which correlates with the conserved structure of these replicons (Figure 4). However, the location of the Listeria plasmids within the network indicates that they share many genes with plasmids of other Firmicutes, which proves their common evolutionary history. In particular, most of them locate between Lactococcus and Enterococcus plasmids. Of these, the Enterococcus ones seem to act as the main bridging hub between Staphylococcus and Bacillus plasmids, as well as the mosaic cluster at the top of the network. In fact, this cluster is composed mostly of Leuconostoc and Lactiplantibacillus plasmids but the majority of the grey nodes in that part of the network correspond to plasmids carried by other representatives of closely related genera of the family Lactobacillaceae. Each node (n = 5533) represents a complete plasmid sequence with the size of the node corresponding to the size of the plasmid. Colors indicate the host genus of the most prevalent groups (>2% of the presented plasmids belonging to this genus). Plasmids originating from the remaining hosts are represented in grey. Edges reflect similarity between connected nodes.

Bacterial Strains and Growth Conditions
The 782 analyzed Listeria spp. strains, originating from food sources, food production plants and environmental soil and water samples, are listed in Supplementary Materials Table S1. We have previously described 566 of these strains with regard to their resistance profiles and the presence of plasmids: 287 strains from fish, fish products and food-producing factories [24], 155 strains from natural, agricultural and soil samples [23] and 124 strains from other foods and food-processing environments [4,20]. Apart from 12 USA strains, indicated in Table S1, all of the analyzed strains were isolated in Poland between 2001-2017.
Unless otherwise specified, Listeria spp., B. subtilis and S. aureus strains were grown in brain-heart infusion (BHI) broth (BioMaxima, Lublin, Poland) or on BHI agar plates (1.5% w/v agar) at 37 °C. E. coli, and P. pantotrophus strains were grown in lysogeny broth (LB, BioMaxima) or LB agar plates (LA). When necessary, the media were supplemented with kanamycin (BioMaxima)-50 μg/mL, streptomycin (BioMaxima)-100 μg/mL, erythromycin (Sigma-Aldrich, St. Louis, MO, USA)-2 μg/mL for Listeria spp., B. subtilis, S. aureus and 250 μg/mL for E. coli, and diaminopimelic acid-60 μg/mL (DAP) (Sigma-Aldrich). For long-term storage, the isolates were suspended in BHI containing 20% glycerol and frozen at −80 °C. Each node (n = 5533) represents a complete plasmid sequence with the size of the node corresponding to the size of the plasmid. Colors indicate the host genus of the most prevalent groups (>2% of the presented plasmids belonging to this genus). Plasmids originating from the remaining hosts are represented in grey. Edges reflect similarity between connected nodes.

Bacterial Strains and Growth Conditions
The 782 analyzed Listeria spp. strains, originating from food sources, food production plants and environmental soil and water samples, are listed in Supplementary Materials Table S1. We have previously described 566 of these strains with regard to their resistance profiles and the presence of plasmids: 287 strains from fish, fish products and foodproducing factories [24], 155 strains from natural, agricultural and soil samples [23] and 124 strains from other foods and food-processing environments [4,20]. Apart from 12 USA strains, indicated in Table S1, all of the analyzed strains were isolated in Poland between 2001-2017.

Plasmid DNA Isolation and Restriction Fragment Length Polymorphism (RFLP) Profile Analysis
Plasmid DNA was isolated from Listeria spp. and B. subtilis strains, as described previously [23]. In the case of S. aureus strains, lysostaphin (A&A Biotechnology, Gdańsk, Poland) was used instead of lysozyme (Sigma-Aldrich). Plasmid DNA for sequencing was isolated using large-scale preparation methods [80], and sequence assembly was verified by RFLP analysis. Purified plasmid DNA was digested with restriction enzymes EcoRI, BamHI or NcoI, according to the manufacturer's protocol (Thermo Scientific, Waltham, MA, USA). The resulting DNA fragments were separated by electrophoresis in 0.8% agarose gels.  Table S2).

Heavy Metal, Metalloid, Benzalkonium Chloride and Crystal Violet Susceptibility
The susceptibility of selected Listeria spp. strains to cadmium chloride, sodium(meta)arsenite and BC was determined using an agar dilution method, as described previously [24]. Strains were classified as resistant if they produced confluent growth on plates containing ≥75 µg/mL cadmium chloride, ≥500 µg/mL sodium (meta)arsenite or ≥10 µg/mL BC. Positive and negative control strains were included in all experiments, with inhibitory concentrations that were determined in a previous study [24]. The susceptibility of Listeria spp. to crystal violet (CV) was determined using an agar dilution method, as described previously [70], with some modifications. Briefly, several colonies picked from BHI agar plates were suspended in saline solution to obtain a turbidity of 0.5 McFarland units (corresponding to approximately 10 8 CFU/mL). Spots of 5 µL of each bacterial suspension were applied to BHI plates supplemented with 15 µg/mL CV (Chempur, PiekaryŚląskie, Poland). The plates were then incubated at 37 • C for 48 h. Strains were classified as resistant to CV if they produced confluent growth.

Antimicrobial Susceptibility
The resistance of Listeria spp. strains harboring pLIS35 and pLIS35-like plasmids (Table S1) to 9 antimicrobial agents was tested using a disc diffusion method, performed according to European Committee on Antimicrobial Susceptibility Testing (EUCAST) guidelines. L. welshimeri ATCC 35987 and L. innocua PZH 5/04 were used as susceptible control strains. The following antimicrobial discs were used: spectinomycin (10 µg), spectinomycin (25 µg), spectinomycin (100 µg), streptomycin (10 µg), netilmicin (30 µg), tobramycin (10 µg), kanamycin (30 µg), amikacin (30 µg), gentamycin (10 µg) and neomycin (30 µg). Single colonies were selected after 24 h of incubation on BHI agar and suspended in sterile saline solution to obtain a turbidity of 0.5 McFarland units. Mueller-Hinton agar (Thermo Scientific), supplemented with 5% defibrinated horse blood, was inoculated by swabbing in three directions with a sterile cotton swab dipped in the suspension. Antimicrobial discs were then applied, and the plates incubated at 37 • C for 18 ± 2 h, after which inhibition zone diameters were measured. The inhibition zones of the tested strains were compared with those of the control strains.

Shuttle Vector Construction
E. coli-Listeria spp. shuttle vectors were constructed by cloning replication and stabilization systems from plasmids pLIS1, pLIS6, pLIS26, pLIS36 and pLIS50 into the multicloning site of mobilizable E. coli vector pDKE2 [23], containing kanamycin and erythromycin resistance cassettes as selection markers for gram-negative and gram-positive bacteria, respectively. Oligonucleotide primers and restriction enzymes used for cloning, as well as the cloned plasmid fragments, are shown in Supplementary Figure S3. The modules were amplified from purified plasmid DNA by PCR using Phusion High-Fidelity DNA Polymerase (Thermo Scientific), digested with restriction enzymes and ligated with linearized vector DNA using T4 DNA ligase (Thermo Scientific). The ligated DNAs were used to transform chemically competent E.coli DH5α pir with selection on LA plates containing kanamycin. Recombinant constructs isolated from transformants were verified by restriction enzyme analysis. Isolated plasmid DNA was then introduced into chemically competent DAP auxotrophic E. coli β2163 cells, with selection on LA plates containing kanamycin and DAP.

Assessing Host Range of Shuttle Vectors
Biparental mating between donor (E. coli β2163 strains harboring constructed shuttle vectors) and recipient strains (L. monocytogenes 10403S, L. innocua PZH 5/04, L. ivanovii PZH 7/04, L. seeligeri ATCC 25967, L. grayi ATCC 25401, L. welshimeri ATCC 35897, S. aureus ATCC 29213, B. subtilis 168 and P. pantotrophus KL100) was used to assess the host range of the constructed plasmids. Overnight cultures of the donor and recipient strains were mixed at a 1:1 ratio, and 100 µL of the mixtures were spread on BHI agar plates containing DAP. After 24 h of incubation at 37 • C, the bacteria were washed off the plates with LB medium, collected by centrifugation and re-suspended in 1 mL of LB. Suitable dilutions were plated on BHI medium supplemented with erythromycin (gram-positive recipients) and incubated at 37 • C for 48 h. In the case of P. pantotrophus KL100, kanamycin was used and the plates were incubated at 30 • C. The resulting colonies were streaked onto BHI medium containing erythromycin, as well as in the case of Listeria spp. transconjugants, also on ALOA medium (BioMaxima) with erythromycin. The presence of autonomous transferred plasmids in transconjugants was confirmed by plasmid DNA isolation. Transformation of chemically competent cells with purified plasmid DNA (selection on LA plates with kanamycin) was used to test if the constructed vectors were capable of replicating in E. coli DH5α (not containing the pir gene). No transconjugants were obtained for any of the recipient strains when the donor strain harbored the empty pDKE2 vector.

Bioinformatic Analyses
Automatic annotation of all the complete plasmid sequences obtained in this study (available in the GenBank database) was performed using RAST on the PATRIC platform [82,83], followed by manual refinement in Artemis [84], based on BLAST homology searches performed at the National Center for Biotechnology information (NCBI) website (https://www.ncbi.nlm.nih.gov (accessed on 24 April 2021)) and HHpred (https: //toolkit.tuebingen.mpg.de/tools/hhpred (accessed on 24 April 2021)) [85]. Comparative genomics of the plasmids was visualized using EasyFig [86]. Comparative analyses of the TRA systems was performed with Clinker [87]. Novel transposable elements identified in this study were deposited in the ISfinder database [88] and The Transposon Registry [89]. Phylogenetic trees of RepA, ParA, UvrX and MobL proteins were constructed in MEGAX [90], with parameters and accession numbers of the analyzed proteins available in Supplementary Figures S1, S4, S5 and S9, respectively. Visualization of the RepA phylogenetic tree was performed using iTOL [91].
All data used for comparative genomic analyses were downloaded from the NCBI GenBank or RefSeq databases on 6 December 2020. Listeria spp. and Firmicutes spp. plasmids were recovered from the nucleotide database using the following query terms: Listeria[ORGANISM] AND plasmid[filter] AND "complete sequence" and Firmicutes [ORGANISM] AND plasmid[filter] AND "complete sequence", respectively. Plasmids for which both Gen-Bank and RefSeq records were present were filtered, and only the latter was retained for further analysis. Moreover, additional filtering was applied after manual inspection of their content, i.e., based on descriptions including keywords such as: "partial", "ribosomal", "lncRNA", "cloning vector", "artificial sequence" (Tables S2 and S5). Listeria spp. genomes were recovered from the NCBI assembly database using Listeria[ORGANISM] as the query. In the case of genome assemblies for which GenBank and RefSeq records were available, only the latter was retained for further analysis.
Listeria spp. plasmids identified in this study were compared in three schemes: against other Listeria spp. plasmids, against other Listeria spp. plasmids and chromosomes, as well as against other Firmicutes spp. plasmids present in public repositories. The comparisons were conducted based on Panaroo v1.2.7 pangenome analysis considering the data specific for each scheme [92]. Each analysis was run with the following parameters: -threshold 0.95 -len_dif_percent 0.95 -clean-mode sensitive -remove-invalid-genes -merge_paralogs -min_trailing_support 1. The resulting gene groups were further used for COG classification analysis using rspblast v2.10.1 with an e-value of 1 × 10 −5 as the threshold and considering only the best blast hit (-max_target_seqs 1 -max_hsps 1) [93,94]. Protein-based network analysis of Firmicutes spp. plasmids employed an all-against-all blastp search of plasmidencoded proteins. The search was conducted with an e-value of 1 × 10 −10 and minimum 95% query coverage per HSP (QCOV) and 75% sequence identity (PIDENT) as thresholds. The resulting similarities were parsed with a Python script and the weight of each edge connecting two nodes was calculated as the sum of multiplications of QCOV and PIDENT between all proteins encoded by each plasmid divided by the total number of proteins encoded by a pair of plasmids. The network was laid out in Gephi using the ForceAtlas2 algorithm [95,96].

Conclusions
This study provides the first large-scale in-depth insight into the structure and genetic composition of plasmids occurring in bacteria of the genus Listeria. Plasmids are common in this genus, and they were present in 47% of the 782 tested strains. We analyzed the sequences of 113 plasmids, representing 69 structural variants. These replicons were identified in six Listeria species, which enabled us to draw more general conclusions concerning gene flux in this genus.
Intriguingly, we found very low diversity among Listeria spp. plasmids. All plasmids considered as theta-replicating, irrespective of their size (15-153 kb), carried closely related REP systems similar to that of E. faecalis plasmid pAMβ1 (28 kb) [97]. Most available data concerns L. monocytogenes strains isolated from food-related sources. However, our analyses included 256 Listeria strains belonging to species other than L. monocytogenes (originating from various sources), and all plasmids found within those strains also belonged to the repA family, suggesting a more general trend among Listeria species. Related plasmids are prevalent in Firmicutes, although the possibility that repA-family plasmids are in some way favored in Listeria spp cannot be excluded. It would be useful to know whether other types of theta-replicating plasmids occurring in gram-positive bacteria can be maintained in Listeria spp. We have found two reports of the transfer of natural conjugative plasmids from E. faecalis (pAMβ1) and Streptococcus agalactiae (pIP501) into Listeria spp. recipients [98,99], but in both cases the introduced plasmids encoded RepA proteins of the pAMβ1-type. Interestingly, many pAMβ1 family plasmids found in other Firmicutes, including pAMβ1 and pIP501, encode antibiotic resistance genes [27], which are very rare in Listeria species.
Another interesting finding concerns the conserved backbone of the repA plasmids, which, apart from the replication and partitioning systems, typically includes POL modules, encoding Y-family polymerases. The evolutionary conservation of these genes and their proximity to the REP and PAR modules, strongly suggests that they may play an important role (direct or indirect, e.g., as a component of a plasmid gene regulatory network) in the maintenance of the plasmids. In a few cases, the polymerase genes were disrupted by a transposon, although their absence could be complemented by chromosomally-encoded enzymes, since homologous genes are present in over 97% of sequenced Listeria chromosomes. Undoubtedly, understanding the role of the POL modules in the biology of Listeria plasmids is an important task for future studies.
The repA-family plasmids show a very high degree of structural conservation. Moreover, highly similar plasmids were identified in bacterial strains isolated from different countries in separate years, sources and even from different Listeria species. For example, the most prevalent plasmid RFLP group (no. 1- Figure 1; pLIS13), gathering replicons from 102 L. monocytogenes strains tested in this study (Table S1), is very closely related to pLM6179-like plasmids isolated from numerous L. monocytogenes strains previously identified in many different countries [18,19].
The chromosomes of Listeria spp. are characterized by a stable and conserved genetic structure [44,100], which suggests that the phenotypic variation of these bacteria may be largely determined by extrachromosomal replicons. Due to their low diversity, the studied plasmids have a limited impact on the size of the Listeria pangenome, but they contain a large reservoir of predicted adaptive genes, which may play a vital role in the response of their host strains to different environmental stressors. It should be emphasized that such genes may potentially impact other phenotypic traits, as was previously observed in the case of some cadmium resistance determinants (cadAC), which also influenced virulence and biofilm formation of L. monocytogenes strains [45,101].