Skip to main content
NCBI home page
Nucleic Acids Research logoLink to Nucleic Acids Research
. 2024 Oct 16;52(21):12976–12986. doi: 10.1093/nar/gkae896

Various plasmid strategies limit the effect of bacterial restriction–modification systems against conjugation

Tatiana Dimitriu 1,2,✉, Mark D Szczelkun 3, Edze R Westra 4
PMCID: PMC11602122  PMID: 39413206

Abstract

In bacteria, genes conferring antibiotic resistance are mostly carried on conjugative plasmids, mobile genetic elements that spread horizontally between bacterial hosts. Bacteria carry defence systems that defend them against genetic parasites, but how effective these are against plasmid conjugation is poorly understood. Here, we study to what extent restriction–modification (RM) systems—by far the most prevalent bacterial defence systems—act as a barrier against plasmids. Using 10 different RM systems and 13 natural plasmids conferring antibiotic resistance in Escherichia coli, we uncovered variation in defence efficiency ranging from none to 105-fold protection. Further analysis revealed genetic features of plasmids that explain the observed variation in defence levels. First, the number of RM recognition sites present on the plasmids generally correlates with defence levels, with higher numbers of sites being associated with stronger defence. Second, some plasmids encode methylases that protect against restriction activity. Finally, we show that a high number of plasmids in our collection encode anti-restriction genes that provide protection against several types of RM systems. Overall, our results show that it is common for plasmids to encode anti-RM strategies, and that, as a consequence, RM systems form only a weak barrier for plasmid transfer by conjugation.

Graphical Abstract

Graphical Abstract.

Graphical Abstract

Open in a new tab

Introduction

Plasmids are extra-chromosomal DNA molecules that replicate independently from the bacterial chromosome; conjugative plasmids can transmit horizontally between bacterial cells as they encode a conjugation machinery allowing transfer between adjacent cells (1). Conjugative plasmids are central to bacterial ecology and evolution because they frequently confer adaptive traits to bacteria, such as metal resistance, antimicrobial resistance (AMR) or virulence traits (2). In particular, conjugative plasmids play a major role in the spread of AMR genes among commensal and pathogenic bacterial species (3). In Enterobacteriaceae, a few major conjugative plasmid families are responsible for AMR including resistance to clinically relevant antibiotics (4,5). It is thus key to understand what shapes conjugative plasmid transmission among bacteria.

Plasmid population dynamics is primarily controlled by two parameters: the effect of a plasmid on its bacterial host’s fitness (which impacts plasmid vertical transmission) and its rate of horizontal transmission between cells by conjugation (6). There is now an extensive literature on plasmid fitness effects and their molecular determinants across plasmids and hosts (7), whereas variation in conjugation rates is less well understood, despite spanning many orders of magnitude (8). Conjugation will impact not only a plasmid’s ability to persist within a population (9), but also plasmid dissemination at larger scales, driving epidemics within species (10) and AMR gene dissemination across species (11). Conjugation rates themselves are determined by multiple factors: expression of the plasmid’s conjugative machinery, regulated by plasmid-encoded genes as well as the donor regulatory network (1,12), but also successful mating pair formation (MPF) and mating pair stabilization, which can depend on the expression of specific receptors on the recipient cell’s surface (13), and finally successful establishment of plasmids in the recipient cell. For this last step to be successful, plasmids must overcome any defence systems present in the recipient.

Bacterial defence systems defend bacteria against molecular parasites, and their activity is mostly studied against bacterial viruses, or phages (14). However, the most prevalent defence systems, restriction–modification (RM) systems and CRISPR–Cas systems, both act by recognizing specific DNA sequences, and as such can target plasmids as well as phages (15). At evolutionary timescales, these defence systems have been shown to impact horizontal gene transfer (16,17), and specifically the distribution of AMR mobile elements in bacterial pathogens (18–20). RM systems are by far the most common defences in bacteria, with on average more than two RM systems per bacterial genome (14). These innate immune systems use DNA modification to distinguish self from non-self. Specifically, they encode a restriction endonuclease, which cleaves DNA following the recognition of an unmodified recognition site, and a methyltransferase, which protects cellular DNA by methylating the recognition sites. Incoming DNA with unmodified recognition sites is recognized as non-self, and restricted.

Restriction endonucleases can cleave any double-stranded (ds) DNA molecules carrying unmodified recognition sequences. Recognition sites are commonly present in several copies on any DNA sequences longer than a few kilobases, which include conjugative plasmids. Indeed, restriction was shown in the early days of molecular biology to have an effect against plasmids, and specifically against DNA entry into cells by conjugation (21). Yet, restriction efficiency (RE) is often found to be low against conjugation (22), especially when compared to other gene transfer mechanisms (23). For instance, when first investigated, EcoKI restriction was observed to restrict phage λ by >105-fold, but conjugation by only 100-fold (24). Various causes have been suggested for this relative resistance to restriction, including the fact that conjugating plasmids enter the recipient as single-stranded DNA (ssDNA) before being replicated (25), avoidance of recognition sites in some plasmid sequences (26), and carriage of anti-restriction and protective functions by plasmids (27). For a few model plasmids, recognition site avoidance and carriage of specific genes have been shown experimentally to provide protection against restriction in the context of conjugation (28–30). However, the experimental data are scarce, limiting our ability to infer general causal relationships between the presence of RM systems and plasmid conjugation rates, and consequently our understanding of this relationship is lagging that of RM–phage interactions (31,32).

Here, we provide a more systematic analysis of how efficient different classes of RM systems are against AMR plasmid conjugation. To study plasmids representative of the known diversity of plasmid types, with a focus on AMR plasmids in bacterial pathogens, we assembled a collection of 13 AMR plasmids, which include well-studied model plasmids and plasmids isolated from clinical environments (9,33,34), and which belong to nine different plasmid families playing a key role in AMR transmission (5,35). These plasmids vary in several key aspects including size, replicon and conjugative transfer type (Supplementary Table S1). We measure conjugation of these plasmids towards a collection of recipient strains carrying 10 different restriction systems all native to Escherichia coli (36). These systems belong to RM Types I, II and III, the three types of classical RM systems differing in molecular structure and mechanistic properties (37). Type I RM are multiprotein complexes, with cleavage requiring the interaction of two endonuclease subunits, and happening at a distance from the recognition site. Type II RM cleave within or close to their recognition site. Finally, type III RMs are multiprotein complexes with an asymmetric recognition site, and cleavage requires two recognition sites on the same DNA in inverted repeat. We use four Type I systems and four Type II systems, with several subtypes of each, and two Type III systems, mirroring the relative abundance of these different types across prokaryotic genomes (38). Finally, we explore potential causes for the observed variation in RE.

Materials and methods

Strains, plasmids and growth conditions

Bacterial strains were grown at 37°C in lysogeny broth (LB), with agitation at 180 rpm.

The 13 wild-type conjugative plasmids used in this study are listed in Supplementary Table S1 with relevant characteristics: they vary in size, replicon, mobilization (MOB) and MPF types, as well as AMR gene content and antibiotics used for selection (Supplementary Table S1).

For selection of plasmid-carrying strains, antibiotics were used at the following concentrations: trimethoprim 25 mg/l, streptomycin 100 mg/l, tetracycline 10 mg/l or ampicillin 100 mg/l. Individual gene deletion mutants of these conjugative plasmids were also used and described in the next section.

The 10 E. coli RM systems we used (36) are listed in Supplementary Table S2 with their recognition sequence: they vary in mechanistic type (four Type I systems, four Type II systems and two type III systems). Type I systems are encoded on the chromosome, Type II and Type III systems are encoded on plasmids maintained with ampicillin (100 mg/l) and chloramphenicol (25 mg/l), respectively.

Conjugation assays used E. coli MG1655 ΔdapA::ErmR ΔhsdS::KnR or MG1655 ΔdapA::ErmR ΔhsdS::frt Δdcm::frt (abbreviated as RM0 Δdap). ΔdapA strains cultures were supplemented with diaminopimelic acid (DAP) 300 μM. To test the effect of EcoRV RM system presence in donors on restriction, we generated spontaneous nalidixic acid (Nal) resistant (NalR) mutants of MG1655 ΔhsdS::KnR and MG1655 ΔhsdS::KnR pEcoRV to use as recipients.

Cloning

Escherichia coli MG1655 ΔdapA::ErmR and MG1655 ΔdapA::ErmR ΔhsdS::KnR were generated by P1 transduction of the ΔdapA::ErmR locus from MFDpir (39) to MG1655 and MG1655 ΔhsdS::KnR (40), using selection with erythromycin 400 mg/l + DAP 300 μM + citrate 5 mM. To obtain RM0 Δdap, the KnR kanamycin resistance cassette was first excised from MG1655 ΔhsdS::KnR using pCP20 (41), then the dcm gene was inactivated by insertion of cat from pKD3 plasmid (41), and the cat cassette was excised using pCP20. Finally, the ΔdapA::ErmR locus was transduced from MFDpir as described earlier.

Plasmid gene knockouts were performed using λred recombination (41): pKD3’s chloramphenicol resistance cassette was amplified using primers including a tail with 50-bp homologies to regions adjacent to the target genes, using Q5 High-Fidelity DNA Polymerase (New England Biolabs). Diagnostic polymerase chain reactions (PCRs) were performed using Thermo Dream Taq™ PCR. Recombineering and diagnostic primers are listed in Supplementary Table S3.

Type II and Type III RM system plasmids were transformed into MG1655 ΔhsdS::KnR (40); conjugative plasmids were moved to the donor strains by conjugation using selection with erythromycin 300 mg/l + DAP 300 μM.

Conjugation assays and measure of restriction efficiency

Plasmid-carrying donors and recipient strains were first grown overnight in glass vials, without antibiotic selection for conjugative plasmids. Conjugation assays were run in a total volume of 1 ml of LB broth supplemented with DAP 120 μM in 24-well plates: for each conjugation, 50 μl of donor culture and 50 μl of recipient culture were added to pre-warmed medium and incubated at 37°C with shaking at 180 rpm. Each donor × recipient combination was run in four or five replicates depending on the experiment, and control assays with donor only and recipient only were run in each experiment. After 3–4 h, the conjugation mixes were serially diluted in M9 salts and droplets of appropriate dilutions were plated on LB-agar + antibiotic + DAP, LB-agar and LB-agar + antibiotic to estimate densities of donors, recipients and transconjugants respectively. As R6K plasmid had low transfer efficiency, for this plasmid the whole conjugation mix was also plated without serial dilution in order to improve the detection threshold. The antibiotics used for selection of plasmid-bearing cells are shown in Supplementary Table S1 for each plasmid. For pCF12, pOXA-48 and pCT plasmids, restriction by Type II RM systems was not measured as these plasmids carry only β-lactamase genes as selective markers, and the Type II RM systems are located on plasmids that also confer resistance to β-lactams. To test the effect of EcoRV RM system presence in donors on restriction, we identified NalR recipients and transconjugants by plating on LB-agar with 30 μg/ml Nal.

Transfer efficiency was estimated as γ (ml/cell/h) = T/DRt, where T, D and R, respectively, indicate the cell density of transconjugants, donors and recipients (cells/ml), and t is the incubation time (h). When no transconjugant colony was present, a threshold transfer efficiency was calculated by assuming that 0.5 transconjugant colony was detected.

RE for a given plasmid and RM system was calculated as RE = γRM+/γRM- where γRM+ is the transfer efficiency towards the recipient strain carrying the RM system and γRM+ is the transfer efficiency from the same donor towards MG1655 ΔhsdS::KnR. To calculate RE, replicates were paired arbitrarily by replicate number.

Sequencing and bioinformatics analyses

Two conjugative plasmids for which a sequence was not available publicly at the time of study, pCU1 and RIP113, were sequenced via MicrobesNG (http://www.microbesng.com) using hybrid whole-genome sequencing, and deposited in GenBank, with accession numbers PQ358084 and PQ358085, respectively. Plasmids' MOB type, MPF type and replicon were determined using COPLA (42).

The number of recognition sites on conjugative plasmids for each RM system was determined using Geneious Prime® 2023.2.1. In order to look for putative anti-restriction functions, we re-annotated all sequences using Bakta (43), and then used both Bakta output and the original plasmid annotations.

Statistical analyses

All statistical analyses were performed with R version 4.3.2 (44). For data processing and plotting, we also used packages reshape (45) and dplyr (46), and packages cowplot (47), ggpmisc (48) and broom (49) to plot statistical parameters.

For each experiment, statistical parameters are reported in the figure legends or within the results section. Transfer efficiency and RE were always log-transformed before statistical analysis. To analyse the effect of recognition site number across the whole dataset of wild-type plasmids, a value of 0.5 was added to the data for two plasmid × RM combinations (EcoKI sites for RIP113, EcoVIII sites for pOXA-48) for which the number of recognition sites was zero.

To analyse the effect of plasmid mutants on restriction, we first asked whether plasmid mutants influenced RE for specific RM systems with a linear model, RE ∼ mutant identity × RM system; in the case of pKAZ3 anti-restriction mutants, the experiment was performed twice and ‘experiment’ was added as a factor (RE ∼ experiment + mutant identity × RM system). When the interaction term between mutant and RM system was significant (P< 0.05), each mutant RE was compared to wild-type RE with an unpaired t-test, and we applied a Bonferroni correction to correct for multiple testing (10 × n tests, where 10 is the number of RM systems and n is the number of mutants tested).

Results

Restriction efficiency across E. coli conjugative plasmids and RM systems

To evaluate how effective restriction is against conjugative transmission, we first measured transfer efficiency for 13 wild-type plasmids in liquid culture, towards recipients carrying no or one of 10 E. coli RM systems. Transfer efficiency in the absence of RM systems varied between ∼10−11 and ∼10−16 ml/cell/h depending on the plasmids (Supplementary Figure S1). We then calculated RE by comparing transfer efficiency for the same plasmid between recipients that do or do not carry a RM system. Using one of the plasmid–RM pairs with the strongest RE, pRK100 and the EcoRV system, we confirmed that restriction in the recipient is dependent on the absence of the same RM system in the donor (Supplementary Figure S2).

RE values varied by five orders of magnitude depending on the plasmid and RM system (Figure 1). Most combinations had relatively low RE, with a median effect of 14-fold and average effect of 20.7-fold (Figure 1A). Overall, RE varied according to RM system identity and plasmid identity, but also their combination (Figure 1B), with all these factors having a significant impact: in a linear model with log10(RE) ∼ RM system × plasmid, RM system identity (F9,449= 211.9, P < 2e−16), plasmid identity (F12,449= 57.1, P< 2e−16) and their interaction (F96,449= 14.8, P < e−16) all had significant effects on RE. Type II EcoRII did not restrict plasmid conjugation at all, whereas Type II EcoRV had the strongest RE, with on average 1000-fold restriction across plasmids. When grouping RM systems by types, we found that Type II RM systems were significantly more efficient than Types I or III [log10(RE) ∼ RM type × plasmid, interaction effect F21,531= 2.45, P= 4 × 10−4; Tukey HSD test on RM type effect, Type II versus Type I 6.8-fold effect, P= 0.000, Type III versus Type II, 3.5-fold effect, P= 0.00003]. Conjugative plasmids also varied in their average sensitivity to restriction: RP4 plasmid was restricted <3-fold on average, whereas pRK100 was restricted 290-fold on average. Yet, the strong effect of the interaction term, with some plasmids such as pB10 displaying very variable sensitivity to restriction, suggests that specific factors determine a plasmid’s sensitivity to a given RM system.

Figure 1.

Figure 1.

Open in a new tab

RE of E. coli RM systems across wild-type AMR plasmids. In panel (A), a histogram of the distribution of RE across plasmid × RM combinations is shown. The solid and dashed lines indicate, respectively, median and mean RE. In panel (B), the left graph shows average RE per plasmid × RM system combination (geometric mean of n ≥ 4 replicates) as a colour scale. Grey squares indicate non-measured combinations (see the ‘Materials and methods’ section). On the right, each plasmid’s transfer rate in the absence of restriction is shown as a colour scale.

Effect of the number of RM recognition sites on RM efficiency

As RM systems are characterized by their sequence specificity, cleaving DNA after recognizing specific unmethylated recognition sites (Supplementary Table S2), we explored the effect of RM recognition site number for each plasmid on RE (Figure 2). Recognition site number varies across RM systems and plasmids (Supplementary Figure S3). Overall, it is lower for Type I systems (median 12 per plasmid) and Type II systems (median 13), and highest for Type III systems (median 108).

Figure 2.

Figure 2.

Open in a new tab

Effect of RM recognition site number on RE across wild-type conjugative plasmids. Log-transformed RE is shown as a function of recognition site number for each RM system. Dots and vertical lines indicate, respectively, the average RE per plasmid and the SEM (n ≥ 4 replicates), with the colour indicating plasmid identity. The grey line represents the model fit on log10(RE) ∼ log10(nsites) for each system, and shaded areas show 95% confidence intervals, with P-values indicated. RM systems are ordered per mechanistic type (top row: type I; middle row: type II; bottom row: type III).

We asked whether the number of recognition sites has a statistically significant effect on RE for each RM system (RM). We included potential plasmid effects not due to their number of recognition sites (nsites) for a particular enzyme, using the linear model: log10(RE) ∼ RM × log10(nsites) + RM × plasmid. We found that both RM and nsites had a highly significant effect (respectively, F9,449 = 71, P< 2 × 10−16 and F1,449 = 103, P< 2 × 10−16), and the interaction between RM and nsites was also significant (F9,449 = 8.99, P< 2 × 10−16). In addition, plasmid identity also had a significant effect after accounting for recognition site number (F12,449 = 11.3, P< 2 × 10−16), as well as plasmid interaction with RM type (F86,449 = 13.4, P< 2 × 10−16).

Thus, an increased number of recognition sites overall increases restriction, but this varies depending on the specific RM system considered (Figure 2). When considering RM systems individually, nsites has a significant effect for three of four Type II systems (EcoRI, EcoRV and EcoVIII), with the remaining Type II system, EcoRII, having little RE against any plasmids, independently of nsites. nsites effect was also significant for both Type III RM systems (EcoP1 and EcoP15), but with relatively low RE overall despite plasmids carrying more restriction sites than for Type II systems. Finally, nsites had a significant effect for only one (EcoR124I) of the four Type I systems tested, with EcoAI, EcoBI and EcoKI being unaffected by nsites despite strong variation in plasmid susceptibility to restriction (in particular, EcoBI RE varied by four orders of magnitude, but independently of nsites).

This absence of dependence of some RM systems on nsites, in particular Type I RM systems, and the significant effect of plasmid identity on RE both suggest that additional plasmid traits must impact RE. Next, we searched for putative anti-restriction functions carried by conjugative plasmids.

Plasmid carriage of putative anti-restriction genes

The genomes of the 13 conjugative plasmids contain a variety of genes that annotations suggest anti-restriction activity (Table 1). First, many plasmids carry annotated DNA methylase genes: RIP113 plasmid carries the full EcoRII RM system, and all large plasmids (> 90kb) carry at least one orphan methyltransferase (i.e. not associated to a restriction enzyme), some of which annotated specifically as dam (DNA adenine methylase) or dcm (DNA cytosine methylase). The largest plasmid, pKAZ3, carries as many as five methyltransferase genes. Second, other genes have annotations related to known anti-restriction functions. These include ard genes (alleviation of restriction of DNA), ardA, ardB and ardC, which have been shown to block restriction by various mechanisms (50), as well as klcA genes, which are homologues of ardB, and the vcrx089–093 operon, which has recently been shown to provide anti-restriction and anti-CRISPR functions in Vibrio (29).

Table 1.

Summary of putative anti-restriction functions identified on plasmids

Plasmid Size (bp) Methylation Anti-restriction
pKAZ3 147 957 dam*2,dcm*3 vcrx089–093
pRK100 142 357 methylase,dam klcA; undefined
pESBL-283 110 137 dam ardA*2
R1 97 566 methylase,dam klcA
pCT 93 629 dam ardA,klcA
pB10 64 508 klcA
pOXA-48 61 881 klcA; undefined
RP4 60 096 klcA
pCF12 43 704
RIP113 42 851 EcoRII dcm ardA,ardB
pCU1 41 131 ardA,klcA
R6K 39 872
R388 33 913 ardC
Open in a new tab

Plasmids are ordered from top to bottom by decreasing plasmid size. ‘Undefined’ indicates genes for which ‘anti-restriction’ was the only information provided. Details are given in the text.

In addition to these plasmid functions, we noted that the chromosome of E. coli K-12 also carries methyltransferase and anti-restriction functions: this includes the well-known dam methyltransferase, yhdJ methyltransferase [which appears to have little effect without overexpression (51)], two copies of klcA in the cryptic prophages CP4-57 and CP4-6, and dcm. In particular, Dcm in E. coli K-12 strains is related to EcoRII methyltransferase, recognizes the same sequence and provides protection against cell killing linked to the loss of EcoRII RM system (52). This suggests that Dcm activity might lead to the absence of EcoRII restriction in our assays. We next analysed how a subset of these candidate genes affect plasmid transmission in the face of restriction.

Effect of methyltransferases on restriction efficiency

We first asked whether inactivating Dcm encoded on the E. coli K-12 chromosome allows the EcoRII RM system to be active against plasmids, by measuring plasmid conjugation for all wild-type plasmids, with and without Dcm in the donor and EcoRII RM system in the recipient. For most plasmids, the only combination displaying reduced conjugation was the combination of a dcm– donor with a EcoRII+ recipient (Supplementary Figure S4), as expected if EcoRII restricts sites unprotected by Dcm methylation. However, interestingly RIP113 plasmid displayed lower conjugation efficiency from the dcm– donor even towards the EcoRII– recipient, and EcoRII had no additional effect on transfer efficiency (Supplementary Figure S4). Lower conjugation efficiency could be due to gene expression effects of altered methylation in the donor. Subsequent analysis of the RIP113 sequence revealed that it carries its own version of the EcoRII RM system (Table 1), explaining why it is protected from EcoRII restriction. Across plasmids, when plasmids were conjugated from the dcm– donor, the number of EcoRII recognition sites became a significant factor explaining EcoRII RE (Figure 3A, log10(RE) ∼ log10(nsites), F1,38 = 10.54, P= 0.0024), similarly to other Type II RM systems (Figure 2). As before, this was not the case when using the dcm+donor (Figure 3A, F1,38 = 2.08, P= 0.158). Still, when focusing on individual plasmids, we can see that three plasmids did not become susceptible to EcoRII, explaining why the effect across plasmids is relatively weak compared to other Type II RM systems. Among these three plasmids (pCU1, RIP113 and pKAZ3), RIP113 carries EcoRII RM system, and pKAZ3 encodes three dcm genes, which might also contribute to methylation of EcoRII recognition sites.

Figure 3.

Figure 3.

Open in a new tab

Effect of chromosome and plasmid methylases on RE. In panel (A), log-transformed EcoRII RE is shown as a function of recognition site number on the conjugative plasmids, with the donor strain being dcm+ (top) or dcm– (bottom). Dots and lines indicate, respectively, the average RE per plasmid and the SEM (n = 4 replicates), with the colour indicating plasmid identity. The grey lines represent the model fit on log10(RE) ∼ log10(nsites), and shaded areas show 95% confidence intervals, with P-values indicated. In panel (B), log-transformed RE for 10 E. coli RM systems is shown from a dcm– donor strain for wild-type pKAZ3, pKAZ3 Δdam and pKAZ3 Δdcm. The centre value of the boxplots shows the median, boxes show the first and third quartiles, and whiskers represent 1.5 times the interquartile range. Individual replicates are shown as dots (n ≥ 4). Asterisks show RM treatments for which the mutant plasmid is significantly different from the wild-type plasmid after Bonferroni correction (0.005 < *P < 0.05; ***P < 0.0005).

We next tested whether plasmid-encoded methyltransferases can also provide protection from restriction in our dataset, choosing pKAZ3 as an example as it contains both predicted adenine- and cytosine-methyltransferase genes (Table 1). We obtained knockout mutants of one dam and one dcm gene, and compared RE of each mutant with wild-type pKAZ3 (Figure 3B). Surprisingly, pKAZ3 Δdam was overall slightly less restricted than wild-type pKAZ3 across RM systems (2.1-fold increase, F1,89 = 11.0, P= 0.0013). However, there was no specific effect on a particular RM system (interaction F9,80 = 0.455, P= 0.9; all individual effects P> 0.05). On the other hand, for pKAZ3 Δdcm there was a significant interaction between mutant status and RM system (F9,136 = 22.5, P< 2 × 10−16); by far the strongest effect was on restriction by EcoRII (t14 = 18.5, P= 3.18 × 10−11 after Bonferroni correction), pKAZ3 Δdcm being restricted ∼ 2700-fold more than wild-type pKAZ3. In addition, dcm had a small but significant protective effect against EcoBI and EcoRV (Figure 3B, respectively, t14 = 3.38, P= 0.044 and t14 = 3.44, P= 0.04, after Bonferroni correction). Thus, we found that the dam gene we tested has no protective effect against restriction here, but dcm had a large protective effect against EcoRII, and a smaller effect against two other RM systems.

Effect of other anti-restriction functions on restriction efficiency

Finally, we tested for an effect of other plasmid-encoded putative anti-restriction functions on restriction with our panel of RM systems. We focused again on pKAZ3, which carries an anti-defence operon that has been relatively less studied in comparison to better-known ard genes. The vcrx089–vcrx093 operon was shown recently to have anti-restriction and anti-CRISPR activity in Vibrio cholerae in another IncC plasmid (29). Genes vcrx089 and vcrx090 had anti-restriction activity against V. cholerae’s Type I system, and genes vcrx091–093 form a recombination system related to phage λ’s Red system, which repairs double-strand breaks due to CRISPR–Cas targeting (29). We hypothesized that these genes might also provide anti-restriction activity in E. coli, and tested this hypothesis by constructing two pKAZ3 variants, one with both vcrx089 and vcrx090 deleted, and one with a deletion of vcrx093[shown to be essential for anti-CRISPR activity in (29)].

We compared RE for wild-type, Δvcrx089–90 and Δvxrx093 pKAZ3 variants (Figure 4). The knockouts affected restriction overall, in a way that was dependent on the RM system tested (interaction effect, F18,204 = 5.88, P= 2.71 × 10−11). Each of the deletions led to increased sensitivity to specific RM systems, with variable amplitude. The Δvcrx089–90 deletion strongly increased RE by the related EcoBI and EcoKI systems (424-fold increase, t14 = 9.01, P= 6.6 × 10−6 and 906-fold increase, t14 = 10.9, P= 1.29 × 10−6, respectively) and had a smaller effect on EcoRV and EcoVIII RE (9.6-fold increase, t14 = 6.41, P= 3.24 × 10−4 and 12.2-fold increase, t14 = 11, P= 5.64 × 10−7, respectively). The Δvxrx093 deletion also had a large effect on EcoBI RE (426-fold increase, t14 = 6.76, P= 1.82 × 10−4), and smaller but significant effects on EcoAI, EcoRII and EcoRV restriction (respectively, 13.1-fold increase, t14 = 3.74, P= 0.044; 5.67-fold increase, t14 = 3.85, P= 0.035; and 22.1-fold increase, t14 = 7.13, P= 10−4). Thus, both vcrx089–90 and vxrx093 confer protection against restriction by various Type I and Type II RM systems.

Figure 4.

Figure 4.

Open in a new tab

Effect of plasmid anti-restriction functions on RE. Log-transformed RE for 10 E. coli RM systems is shown from a dcm– donor strain for wild-type pKAZ3, pKAZ3 Δvcrx089–90 and pKAZ3 Δvxrx093. The centre value of the boxplots shows the median, boxes show the first and third quartiles, and whiskers represent 1.5 times the interquartile range. Individual replicates are shown as dots (n ≥ 4). Asterisks show RM treatments for which the mutant plasmid is significantly different from the wild-type plasmid after Bonferroni correction (0.005 < *P < 0.05; ***P < 0.0005).

Finally, we tested other genes with putative anti-restriction or repair activity (ardB and umuCD on RIP113, and klcA and klcB on pB10 plasmid); however, we did not detect any significant effect of these gene deletions (P > 0.05 for all, Supplementary Figure S5).

Discussion

In this work, we tested the activity of naturally occurring RM systems against a collection of conjugative plasmids that can transmit horizontally within E. coli. We found a large variation in RE, which depends on the specific plasmid–RM pair; and uncovered some drivers of this variation.

First, RE correlates with the number of RM recognition sites across conjugative plasmids. For a similar number of recognition sites, restriction was most efficient for Type II RM systems, possibly because they mostly require only one site for cleavage, whereas Type I and III systems require at least two. Indeed, the only Type II RM in our experiments that requires two sites for cleavage, EcoRII, achieved the same RE than other Type II systems, but with roughly 10 times more recognition sites on the target conjugative plasmids. When analysed separately for each RM system, the correlation is significant mostly for Type II RM systems. This is in agreement with data obtained using phage λ mutants and EcoRI Type II system, in which the probability of phage escape from restriction decreased with each additional restriction site, and effects were multiplicative (53). This pattern is consistent with a simple mechanism of escape from restriction due to a low probability of stochastic methylation at each recognition site, and independent between sites. This also suggests that for Type II systems, additional plasmid-specific factors like anti-restriction functions have little role overall.

By contrast, for several Type I RM systems, we observe no significant correlation between RE and recognition site number, suggesting that other factors drive the observed patterns. Indeed, we find many putative anti-restriction functions encoded on plasmids, particularly on the larger plasmids, in accordance with bioinformatic studies (26). We then show experimentally that some of these genes provide protection against specific RM systems. First, both host- and plasmid-encoded methyltransferases provide some protection. The effect of donor host methyltransferases on conjugation is well known as a factor that can ameliorate the transmission of engineered conjugative vectors across species (54,55). Here, we show that the chromosomal cytosine methylase Dcm protects all conjugative plasmids tested from restriction by EcoRII. This is likely to be more than an artefact from using a laboratory strain, as Dcm is widespread in E. coli natural isolates (56). Yet, some plasmids are still resistant to EcoRII when conjugating from a dcm mutant, and we show that in the case of pKAZ3, this is due to the presence of a plasmid-encoded cytosine methyltransferase. In contrast, we find no effect of inactivating a pKAZ3 adenine methyltransferase, but that plasmid contains two dam genes, likely providing functional redundancy. Adenine methylation was shown to provide resistance to restriction in another model conjugative plasmid (28). Still, methyltransferases could also have roles unrelated to restriction, including regulatory roles. For RIP113 plasmid, we observed a 10-fold decrease in conjugation even in the absence of restriction, when using a dcm mutant donor, suggesting that methylation contributes to the regulation of transfer, as shown for other conjugative plasmids (57).

In addition to methyltransferases, plasmids also carry other genes with putative anti-restriction (‘ard’) functions. Known ard genes are predominantly active against Type I restriction, possibly because Type II systems are too diverse in terms of molecular structure and mechanism for a generic anti-restriction mechanism to evolve (25). Even within Type I systems, they are often active against only a subset of RM systems (58–60). In our assays, we detected no effects of klcA in pB10 or ardB in RIP113. klcA and ardB are related genes, and their activity in vivo has been shown to depend on their expression levels (61), explaining why klcA was previously observed to have no effect on restriction in RP4 plasmid (62). Instead, we find specific anti-restriction effect of other genes on pKAZ3 plasmid, for which an anti-defence role was demonstrated more recently in another IncC plasmid, pVCR94 (29). We find that vcr089–vcr090 genes are active against some Type I as well as Type II RM systems, but quantitatively much more effective against Type I systems. We then demonstrate an anti-restriction role for gene vcr093, extending results from (29), which found a role against CRISPR–Cas interference. The anti-restriction mechanism is very likely the same as for its anti-CRISPR effect, namely by double-strand DNA break repair (29). Interestingly, this vcrx089–093 anti-defence operon is not annotated as such by the annotation tools we used, and we only noticed it by using PaperBLAST (63). This strongly suggests that other anti-restriction functions on these plasmids still await detection.

Overall, our data show that anti-restriction functions significantly contribute to the relative resistance of plasmids against RM systems, especially Type I systems, with 100- to 1000-fold increase in susceptibility to some Type I systems in anti-restriction gene knockouts. Alternatively, for some other plasmid–RM pairs, low plasmid sensitivity to restriction arises from carrying a low number of restriction sites, especially for Type II RM systems (Figure 2). Interestingly, for several of the plasmids analysed here, the number of restriction sites is lower than expected by chance (Supplementary Figure S3): for instance, RP4, R388 and RIP113 would be expected to have seven, five and nine EcoRI restriction sites, respectively, based on their size and GC content, but they only have one. This is consistent with bioinformatic studies showing a pattern of Type II RM target avoidance in mobile plasmids, particularly in smaller plasmids (26). In small plasmids, the total number of recognition sites is lower, meaning that a mutation removing a single target site can have a large effect on restriction. Indeed, experiments on RP4 plasmid showed that adding three EcoRI sites led to a 104-fold increase in restriction (30).

Finally, some general trends in our data could be explained by core features of conjugation. To contrast our results to measures of restriction against phages (the other main type of MGEs infecting bacteria), we compared our results to data from the BASEL collection of E. coli phages (32), which included the measure of RE by six RM systems, four of which we tested against plasmids. We can discern two interesting patterns (Supplementary Figure S6). First, many phages are totally resistant to restriction, despite containing a large number of restriction sites. This is likely due to hypermodification of their genomes, as most of these phages belong to the Tevenvirinae family, which display cytosine hypermodification (a strategy that to our current knowledge is not available to plasmids). Second, plasmids have low sensitivity to restriction by Type III systems compared to many phages. This pattern is apparent across plasmids, suggesting it is not caused by plasmid-specific anti-restriction functions, but might instead be linked to conjugation itself. Indeed, experiments in the naturally transformable species Neisseria gonorrhoeae showed a 104-fold decrease in RE for plasmid conjugation compared to transformation. This effect was attributed to restriction endonucleases acting on dsDNA and not ssDNA (64). Type III restriction requires the recognition of two unmodified recognition sites in inverse orientation (65,66). Moreover, the endonuclease and methyltransferase sub-units are part of the same complex, with the Mod subunit providing specificity. Type III EcoPI was shown to methylate ssDNA in vitro (67). Thus, the dynamics of restriction and methylation on incoming plasmids might make conjugative entry intrinsically less susceptible to Type III restriction, similarly to filamentous, ssDNA phages that are also relatively resistant to Type III restriction (68). More generally, several factors point to a key role for the early dynamics of host defences and plasmid anti-defences and establishment functions in determining whether conjugation is successful. Protective functions are expressed early during conjugation (69), and anti-defence genes are enriched in the leading region of plasmids (the first region entering the recipient during conjugation), suggesting that the speed of early events is critical (70).

Overall, our results show that restriction is on average low, and depends on the number of recognition sites on the plasmid (in particular for Type II restriction), and on the complement of anti-restriction functions carried by the plasmid (in particular for Type I restriction). The collection of plasmids we tested includes most types of plasmids carrying AMR genes in Enterobacteriaceae, in particular plasmids of groups IncF, IncI and IncN (5,35), suggesting that these patterns might hold true for most clinically relevant AMR plasmids. In terms of RM systems, we only tested a limited subset of the large diversity in existing RM subtypes. Still, Type I and Type III systems are very similar in functional mechanism within Type, as most variation between systems is variation in site recognition by the specificity determinant (hsdS for Type I, mod for Type III) (71,72), suggesting that our results might be representative. For Type II systems, there is much more variation in mechanism; thus, it might be difficult to establish general rules and other subtypes will need to be tested to establish generality.

Our results suggest that the overall effect of RM systems on plasmids in natural populations will be relatively low, due to multiple plasmid strategies mitigating restriction. Still, even a relatively poor defence system should have relevant effects on plasmid and bacterial eco-evolutionary dynamics. For a given plasmid, even a 2-fold restriction barrier would lead to a large competitive disadvantage in competition with another plasmid unaffected by restriction. On the host side, RM system distribution will determine which bacteria preferentially receive a plasmid in mixed populations. Restriction against a costly plasmid should provide a fitness benefit, while restriction against a beneficial plasmid should be counter-selected (73). Moreover, RM distribution might also impact selective pressures acting on the donor: investment in conjugative transmission is costly for the donor bacteria, so is not directly selected for. However, because clonemates share the same RM systems, RM will bias transfer preferentially to bacterial clonemates (40). In turn, this can allow increased investment by donors to transmit beneficial plasmids, via kin selection (74), because the benefits of transfer will be borne preferentially by the donors’ kin. Overall, these direct and indirect effects of RM-plasmid interactions will combine with the (likely stronger) effects of defence against phage infection (31) to impact RM distribution across bacteria. Finally, bacteria in nature commonly carry multiple defence systems, including on average more than two RM systems (14). The next step will thus be to study not only clinically relevant plasmids, as we have done here, but also clinically relevant and genetically more complex bacterial hosts (3), to understand patterns of AMR plasmid spread in clinically relevant bacterial communities (75).

Supplementary Material

gkae896_Supplemental_Files
gkae896_supplemental_files.zip (582.9KB, zip)

Acknowledgements

We thank Calin Guet for sharing RM system carrying bacterial strains, and Allison Lopatkin and Jeronimo Rodriguez-Beltran for sharing conjugative plasmids.

Contributor Information

Tatiana Dimitriu, Biomedical Sciences Research Complex, School of Biology, University of St Andrews, North Haugh, St Andrews KY16 9ST, UK; Environment and Sustainability Institute, Biosciences, University of Exeter, Penryn Campus, Penryn TR10 9FE, UK.

Mark D Szczelkun, DNA–Protein Interactions Unit, School of Biochemistry, University of Bristol, University Walk, Bristol BS8 1TD, UK.

Edze R Westra, Environment and Sustainability Institute, Biosciences, University of Exeter, Penryn Campus, Penryn TR10 9FE, UK.

Data availability

The data underlying this article are available in the article and in its online supplementary material. Sequences for plasmids pCU1 and RIP113 are available in GenBank with accession numbers PQ358084 and PQ358085, respectively.

Supplementary data

Supplementary Data are available at NAR Online.

Funding

Biotechnology and Biological Sciences Research Council [BB/X003051/1 to E.R.W.]; European Research Council [ERC-2017-ADG-788405 to M.D.S.]; Royal Society [URF\R1\231740 to T.D.]. Funding for open access charge: University of St Andrews.

Conflict of interest statement. None declared.

References

  • 1. Frost L.S., Koraimann G.. Regulation of bacterial conjugation: balancing opportunity with adversity. Future Microbiol. 2010; 5:1057–1071. [DOI] [PubMed] [Google Scholar]
  • 2. Frost L.S., Leplae R., Summers A.O., Toussaint A.. Mobile genetic elements: the agents of open source evolution. Nat. Rev. Microbiol. 2005; 3:722–732. [DOI] [PubMed] [Google Scholar]
  • 3. MacLean R.C., San Millán A.. The evolution of antibiotic resistance. Science. 2019; 365:1082–1083. [DOI] [PubMed] [Google Scholar]
  • 4. Partridge S.R., Kwong S.M., Firth N., Jensen S.O.. Mobile genetic elements associated with antimicrobial resistance. Clin. Microbiol. Rev. 2018; 31:e00088-17. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5. Carattoli A. Resistance plasmid families in Enterobacteriaceae. Antimicrob. Agents Chemother. 2009; 53:2227–2238. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6. Stewart F.M., Levin B.R.. The population biology of bacterial plasmids: a priori conditions for the existence of conjugationally transmitted factors. Genetics. 1977; 87:209–228. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7. San Millán A., MacLean R.C.. Fitness costs of plasmids: a limit to plasmid transmission. Microbiol. Spectr. 2017; 5:0016. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8. Sheppard R.J., Beddis A.E., Barraclough T.G.. The role of hosts, plasmids and environment in determining plasmid transfer rates: a meta-analysis. Plasmid. 2020; 108:102489. [DOI] [PubMed] [Google Scholar]
  • 9. Lopatkin A.J., Meredith H.R., Srimani J.K., Pfeiffer C., Durrett R., You L.. Persistence and reversal of plasmid-mediated antibiotic resistance. Nat. Commun. 2017; 8:1689. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10. Baker K.S., Dallman T.J., Field N., Childs T., Mitchell H., Day M., Weill F.-X., Lefèvre S., Tourdjman M., Hughes G.et al.. Horizontal antimicrobial resistance transfer drives epidemics of multiple Shigella species. Nat. Commun. 2018; 9:1462. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11. Sheppard A.E., Stoesser N., Wilson D.J., Sebra R., Kasarskis A., Anson L.W., Giess A., Pankhurst L.J., Vaughan A., Grim C.J.et al.. Nested Russian doll-like genetic mobility drives rapid dissemination of the carbapenem resistance gene blaKPC. Antimicrob. Agents Chemother. 2016; 60:3767–3778. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12. Dimitriu T. Evolution of horizontal transmission in antimicrobial resistance plasmids. Microbiology. 2022; 168:001214. [DOI] [PubMed] [Google Scholar]
  • 13. Frankel G., David S., Low W.W., Seddon C., Wong J.L.C., Beis K.. Plasmids pick a bacterial partner before committing to conjugation. Nucleic Acids Res. 2023; 51:8925–8933. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14. Georjon H., Bernheim A.. The highly diverse antiphage defence systems of bacteria. Nat. Rev. Microbiol. 2023; 21:686–700. [DOI] [PubMed] [Google Scholar]
  • 15. Dimitriu T., Szczelkun M.D., Westra E.R.. Evolutionary ecology and interplay of prokaryotic innate and adaptive immune systems. Curr. Biol. 2020; 30:R1189–R1202. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16. Oliveira P.H., Touchon M., Rocha E.P.C.. Regulation of genetic flux between bacteria by restriction–modification systems. Proc. Natl Acad. Sci. U.S.A. 2016; 113:5658–5663. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17. Wheatley R.M., MacLean R.C.. CRISPR–Cas systems restrict horizontal gene transfer in Pseudomonas aeruginosa. ISME J. 2020; 15:1420–1433. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18. Botelho J. Defense systems are pervasive across chromosomally integrated mobile genetic elements and are inversely correlated to virulence and antimicrobial resistance. Nucleic Acids Res. 2023; 51:4385–4397. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19. Pursey E., Dimitriu T., Paganelli F.L., Westra E.R., van Houte S.. CRISPR–Cas is associated with fewer antibiotic resistance genes in bacterial pathogens. Philos. Trans. R. Soc. B Biol. Sci. 2021; 377:20200464. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20. Li D., Li P., Peng M., Zhao X., Jiang X., Wang D., Yuan Y., Guo Q., Wang M., Xu X.et al.. Transmission barrier of the bla KPC plasmid mediated by type I restriction–modification systems in Escherichia coli. J. Antimicrob. Chemother. 2022; 77:952–956. [DOI] [PubMed] [Google Scholar]
  • 21. Arber W., Morse M.L.. Host specificity of DNA produced by Escherichia coli. VI. Effects on bacterial conjugation. Genetics. 1965; 51:137. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22. Roer L., Aarestrup F.M., Hasman H.. The EcoKI type I restriction–modification system in Escherichia coli affects but is not an absolute barrier for conjugation. J. Bacteriol. 2015; 197:337–342. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23. Read T.D., Thomas A.T., Wilkins B.M.. Evasion of type I and type II DNA restriction systems by Incl1 plasmid Collb-P9 during transfer by bacterial conjugation. Mol. Microbiol. 1992; 6:1933–1941. [DOI] [PubMed] [Google Scholar]
  • 24. Boyer H. Genetic control of restriction and modification in Escherichia coli. J. Bacteriol. 1964; 88:1652–1660. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25. Wilkins B.M. Plasmid promiscuity: meeting the challenge of DNA immigration control. Environ. Microbiol. 2002; 4:495–500. [DOI] [PubMed] [Google Scholar]
  • 26. Shaw L.P., Rocha E.P.C., MacLean R.C.. Restriction–modification systems have shaped the evolution and distribution of plasmids across bacteria. Nucleic Acids Res. 2023; 51:6806–6818. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27. Chilley P.M., Wilkins B.M.. Distribution of the ardA family of antirestriction genes on conjugative plasmids. Microbiology. 1995; 141:2157–2164. [DOI] [PubMed] [Google Scholar]
  • 28. Fomenkov A., Sun Z., Murray I.A., Ruse C., McClung C., Yamaichi Y., Raleigh E.A., Roberts R.J.. Plasmid replication-associated single-strand-specific methyltransferases. Nucleic Acids Res. 2020; 48:12858–12873. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29. Roy D., Huguet K.T., Grenier F., Burrus V.. IncC conjugative plasmids and SXT/R391 elements repair double-strand breaks caused by CRISPR–Cas during conjugation. Nucleic Acids Res. 2020; 48:gkaa518. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30. Guiney D.G. Promiscuous transfer of drug resistance in Gram-negative bacteria. J. Infect. Dis. 1984; 149:320–329. [DOI] [PubMed] [Google Scholar]
  • 31. Korona R., Levin B.R.. Phage-mediated selection and the evolution and maintenance of restriction–modification. Evolution. 1993; 47:556–575. [DOI] [PubMed] [Google Scholar]
  • 32. Maffei E., Shaidullina A., Burkolter M., Heyer Y., Estermann F., Druelle V., Sauer P., Willi L., Michaelis S., Hilbi H.et al.. Systematic exploration of Escherichia coli phage–host interactions with the BASEL phage collection. PLoS Biol. 2021; 19:e3001424. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33. Herencias C., Rodríguez-Beltrán J., León-Sampedro R., Alonso-del Valle A., Palkovičová J., Cantón R., San Millán Á.. Collateral sensitivity associated with antibiotic resistance plasmids. eLife. 2021; 10:e65130. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34. Alonso-del Valle A., León-Sampedro R., Rodríguez-Beltrán J., DelaFuente J., Hernández-García M., Ruiz-Garbajosa P., Cantón R., Peña-Miller R., San Millán A.. Variability of plasmid fitness effects contributes to plasmid persistence in bacterial communities. Nat. Commun. 2021; 12:2653. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35. Rozwandowicz M., Brouwer M.S.M., Fischer J., Wagenaar J.A., Gonzalez-Zorn B., Guerra B., Mevius D.J., Hordijk J.. Plasmids carrying antimicrobial resistance genes in Enterobacteriaceae. J. Antimicrob. Chemother. 2018; 73:1121–1137. [DOI] [PubMed] [Google Scholar]
  • 36. Pleška M., Lang M., Refardt D., Levin B.R., Guet C.C.. Phage–host population dynamics promotes prophage acquisition in bacteria with innate immunity. Nat. Ecol. Evol. 2018; 2:359–366. [DOI] [PubMed] [Google Scholar]
  • 37. Roberts R.J. A nomenclature for restriction enzymes, DNA methyltransferases, homing endonucleases and their genes. Nucleic Acids Res. 2003; 31:1805–1812. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38. Oliveira P.H., Touchon M., Rocha E.P.C.. The interplay of restriction–modification systems with mobile genetic elements and their prokaryotic hosts. Nucleic Acids Res. 2014; 42:10618–10631. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39. Ferrieres L., Hemery G., Nham T., Guerout A.-M., Mazel D., Beloin C., Ghigo J.-M.. Silent mischief: bacteriophage mu insertions contaminate products of Escherichia coli random mutagenesis performed using suicidal transposon delivery plasmids mobilized by broad-host-range RP4 conjugative machinery. J. Bacteriol. 2010; 192:6418–6427. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40. Dimitriu T., Marchant L., Buckling A., Raymond B.. Bacteria from natural populations transfer plasmids mostly towards their kin. Proc. R. Soc. B Biol. Sci. 2019; 286:20191110. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41. Datsenko K.A., Wanner B.L.. One-step inactivation of chromosomal genes in Escherichia coli K-12 using PCR products. P. Natl Acad. Sci. U.S.A. 2000; 97:6640–6645. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42. Redondo-Salvo S., Bartomeus-Peñalver R., Vielva L., Tagg K.A., Webb H.E., Fernández-López R., de la Cruz F.. COPLA, a taxonomic classifier of plasmids. BMC Bioinformatics. 2021; 22:390. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43. Schwengers O., Jelonek L., Dieckmann M.A., Beyvers S., Blom J., Goesmann A.. Bakta: rapid and standardized annotation of bacterial genomes via alignment-free sequence identification: find out more about Bakta, the motivation, challenges and applications, here. Microb. Genom. 2021; 7:000685. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44. R Core Team R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing. 2017; Vienna, Austria. [Google Scholar]
  • 45. Wickham H. Reshaping data with the reshape package. J. Stat. Softw. 2007; 21:1–20. [Google Scholar]
  • 46. Wickham H., François R., Henry L., Müller K., Vaughan D.. dplyr: a grammar of data manipulation. 2023; (9 October 2024, date last accessed)https://dplyr.tidyverse.org.
  • 47. Wilke C.O. cowplot: streamlined plot theme and plot annotations for ‘ggplot2’. 2017; R package version 0.8.0 www.rdocumentation.org/packages/cowplot/versions/1.1.3.
  • 48. Aphalo P.J. ggpmisc: miscellaneous extensions to ‘ggplot2’. 2023; (9 October 2024, date last accessed)https://cran.r-project.org/web/packages/ggpmisc/index.html.
  • 49. Robinson D., Hayes A., Couch S.. broom: convert statistical objects into tidy tibbles. 2023; (9 October 2024, date last accessed)https://broom.tidymodels.org.
  • 50. Tock M.R., Dryden D.T.. The biology of restriction and anti-restriction. Curr. Opin. Microbiol. 2005; 8:466–472. [DOI] [PubMed] [Google Scholar]
  • 51. Broadbent S.E., Balbontin R., Casadesus J., Marinus M.G., van der Woude M.. YhdJ, a nonessential CcrM-like DNA methyltransferase of Escherichia coli and Salmonella enterica. J. Bacteriol. 2007; 189:4325–4327. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 52. Takahashi N., Naito Y., Handa N., Kobayashi I.. A DNA methyltransferase can protect the genome from postdisturbance attack by a restriction–modification gene complex. J. Bacteriol. 2002; 184:6100–6108. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 53. Pleška M., Guet C.C.. Effects of mutations in phage restriction sites during escape from restriction–modification. Biol. Lett. 2017; 13:20170646. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 54. Butler C.A., Gotschlich E.C.. High-frequency mobilization of broad-host-range plasmids into Neisseria gonorrhoeae requires methylation in the donor. J. Bacteriol. 1991; 173:5793–5799. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 55. Elhai J., Vepritskiy A., Muro-Pastor A.M., Flores E., Wolk C.P.. Reduction of conjugal transfer efficiency by three restriction activities of Anabaena sp. strain PCC 7120. J. Bacteriol. 1997; 179:1998–2005. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 56. Militello K.T., Simon R.D., Qureshi M., Maines R., Horne M.L., Hennick S.M., Jayakar S.K., Pounder S.. Conservation of dcm-mediated cytosine DNA methylation in Escherichia coli. FEMS Microbiol. Lett. 2012; 328:78–85. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 57. Camacho E.M., Serna A., Casadesús J.. Regulation of conjugal transfer by lrp and Dam methylation in plasmid R100. Int. Microbiol. 2010; 8:279–285. [PubMed] [Google Scholar]
  • 58. Belogurov A.A., Delver E.P., Rodzevich O.V.. Plasmid pKM101 encodes two nonhomologous antirestriction proteins (ArdA and ArdB) whose expression is controlled by homologous regulatory sequences. J. Bacteriol. 1993; 175:4843–4850. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 59. McMahon S.A., Roberts G.A., Johnson K.A., Cooper L.P., Liu H., White J.H., Carter L.G., Sanghvi B., Oke M., Walkinshaw M.D.et al.. Extensive DNA mimicry by the ArdA anti-restriction protein and its role in the spread of antibiotic resistance. Nucleic Acids Res. 2009; 37:4887–4897. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 60. Serfiotis-Mitsa D., Herbert A.P., Roberts G.A., Soares D.C., White J.H., Blakely G.W., Uhrín D., Dryden D.T.F.. The structure of the KlcA and ArdB proteins reveals a novel fold and antirestriction activity against type I DNA restriction systems in vivo but not in vitro. Nucleic Acids Res. 2010; 38:1723–1737. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 61. Goryanin I.I., Kudryavtseva A.A., Balabanov V.P., Biryukova V.S., Manukhov I.V., Zavilgelsky G.B.. Antirestriction activities of KlcA (RP4) and ArdB (R64) proteins. FEMS Microbiol. Lett. 2018; 365:fny227. [DOI] [PubMed] [Google Scholar]
  • 62. Larsen M.H., Figurski D.H.. Structure, expression, and regulation of the kilC operon of promiscuous IncPot plasmids. J. Bacteriol. 1994; 176:5022–5032. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 63. Price M.N., Arkin A.P.. PaperBLAST: text mining papers for information about homologs. mSystems. 2017; 2:e00039-17. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 64. Stein D.C., Gregoire S., Piekarowicz A.. Restriction of plasmid DNA during transformation but not conjugation in Neisseria gonorrhoeae. Infect. Immun. 1988; 56:112–116. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 65. Meisel A., Bickle T.A., Kruger D.H., Schroeder C.. Type III restriction enzymes need two inversely oriented recognition sites for DNA cleavage. Nature. 1992; 355:467–469. [DOI] [PubMed] [Google Scholar]
  • 66. Van Aelst K., Tóth J., Ramanathan S.P., Schwarz F.W., Seidel R., Szczelkun M.D.. Type III restriction enzymes cleave DNA by long-range interaction between sites in both head-to-head and tail-to-tail inverted repeat. Proc. Natl Acad. Sci. U.S.A. 2010; 107:9123–9128. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 67. Sistla S., Krishnamurthy V., Rao D.N.. Single-stranded DNA binding and methylation by EcoP1I DNA methyltransferase. Biochem. Biophys. Res. Commun. 2004; 314:159–165. [DOI] [PubMed] [Google Scholar]
  • 68. Krüger D.H., Bickle T.A.. Bacteriophage survival: multiple mechanisms for avoiding the deoxyribonucleic acid restriction systems of their hosts. Microbiol. Rev. 1983; 47:345–360. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 69. Althorpe N.J., Chilley P.M., Thomas A.T., Brammar W.J., Wilkins B.M.. Transient transcriptional activation of the Incl1 plasmid anti-restriction gene (ardA) and SOS inhibition gene (psiB) early in conjugating recipient bacteria. Mol. Microbiol. 1999; 31:133–142. [DOI] [PubMed] [Google Scholar]
  • 70. Samuel B., Burstein D.. A diverse repertoire of anti-defense systems is encoded in the leading region of plasmids. 2023; bioRxiv doi:16 February 2023, preprint: not peer reviewed 10.1101/2023.02.15.528439. [DOI] [PMC free article] [PubMed]
  • 71. De Ste Croix M., Vacca I., Kwun M.J., Ralph J.D., Bentley S.D., Haigh R., Croucher N.J., Oggioni M.R.. Phase-variable methylation and epigenetic regulation by type I restriction–modification systems. FEMS Microbiol. Rev. 2017; 41:S3–S15. [DOI] [PubMed] [Google Scholar]
  • 72. Seib K.L., Srikhanta Y.N., Atack J.M., Jennings M.P.. Epigenetic regulation of virulence and immunoevasion by phase-variable restriction–modification systems in bacterial pathogens. Annu. Rev. Microbiol. 2020; 74:655–671. [DOI] [PubMed] [Google Scholar]
  • 73. Gandon S., Vale P.F.. The evolution of resistance against good and bad infections. J. Evol. Biol. 2014; 27:303–312. [DOI] [PubMed] [Google Scholar]
  • 74. Dimitriu T., Misevic D., Lotton C., Brown S.P., Lindner A.B., Taddei F.. Indirect fitness benefits enable the spread of host genes promoting costly transfer of beneficial plasmids. PLoS Biol. 2016; 14:e1002478. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 75. Alonso-del Valle A., Toribio-Celestino L., Quirant A., Pi C.T., DelaFuente J., Canton R., Rocha E.P.C., Ubeda C., Peña-Miller R., San Millan A.. Antimicrobial resistance level and conjugation permissiveness shape plasmid distribution in clinical enterobacteria. Proc. Natl Acad. Sci. U.S.A. 2023; 120:e2314135120. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

gkae896_Supplemental_Files
gkae896_supplemental_files.zip (582.9KB, zip)

Data Availability Statement

The data underlying this article are available in the article and in its online supplementary material. Sequences for plasmids pCU1 and RIP113 are available in GenBank with accession numbers PQ358084 and PQ358085, respectively.

Articles from Nucleic Acids Research are provided here courtesy of Oxford University Press

RESOURCES