Diverse anti-defence systems are encoded in the leading region of plasmids

Abstract

Plasmids are major drivers of gene mobilization by means of horizontal gene transfer and play a key role in spreading antimicrobial resistance among pathogens^1,2. Despite various bacterial defence mechanisms such as CRISPR-Cas, restriction-modification systems and SOS-response genes that prevent the invasion of mobile genetic elements³, plasmids robustly transfer within bacterial populations through conjugation^4,5. Here we show that the leading region of plasmids, the first to enter recipient cells, is a hotspot for an extensive repertoire of anti-defence systems, encoding anti-CRISPR, anti-restriction, anti-SOS and other counter-defence proteins. We further identified in the leading region a prevalence of promoters known to allow expression from single-stranded DNA⁶, potentially facilitating rapid protection against bacterial immunity during the early stages of plasmid establishment. We demonstrated experimentally the importance of anti-defence gene localization in the leading region for efficient conjugation. These results indicate that focusing on the leading region of plasmids could lead to the discovery of diverse anti-defence genes. Combined, our findings show a new facet of plasmid dissemination and provide theoretical foundations for developing efficient conjugative delivery systems for natural microbial communities.

Fig. 1. Enrichment of anti-defence genes in the leading region.

a, Anti-defence gene frequency in 21,907 plasmids and potential conjugative elements (MOBs F, P1, Q, V, H, B). The x axis shows ORF indices relative to the oriT, with 0 representing the first ORF in the leading region. Only positions represented in at least 500 sequences are shown (additional positions in Extended Data Fig. 2c). The y axis indicates the average frequency of anti-defence genes (combining SOS inhibition, anti-restriction and anti-CRISPR genes) over a five-ORF window. b, Breakdown of anti-defence gene frequency by functional category. c, The 100 largest gene families significantly enriched in the leading regions (one-sided Fisher’s exact test, α = 0.001 after FDR correction for multiple testing), categorized into six groups: (1) anti-defence genes, which are anti-CRISPRs, anti-restriction genes and SOS inhibitors; (2) DNA methyltransferases (MTases); (3) toxin–antitoxin genes; (4) SSBs; (5) other, which are annotated genes with no known association to anti-defence and (6) uncharacterized genes. The y axis shows the family size, whereas the x axis shows the families ranked on the basis of their size. Note that some of the most prevalent families are not enriched specifically in the leading region and are thus omitted from this analysis. Diamonds indicate gene families encoded opposite to the T-strand, which cannot be transcribed from the leading ssDNA. Putative annotations are indicated with striped bars. Structural comparison between a DNA-methyltransferase from Escherichia coli (gene family 8, NCBI accession CP029982.1) and a putative DNA-methyltransferase from E. coli (gene family 14, NCBI accession MCJK01000027.1) are presented above the respective families. The inset focuses on the 20 largest gene families (Supplementary Table 1). Source Data

Fig. 2. Representative anti-defence islands and ssDNA promoters.

a–d, Representative anti-defence islands from leading regions of conjugative elements in: (a) Salmonella enterica (NCBI assembly accession AAEPNF010000010.1), (b) Serratia marcescens (NCBI accession CP047692.1), (c) an insect metagenome (NCBI accession OFEI01000013.1) and (d) Streptococcus pneumoniae (NCBI accession CPMX01000004.1). The oriT location is marked in red on the left. Genes are colour-coded by functional category: anti-defence (red), MTase (peach), toxin–antitoxin genes (orange), SSB (yellow), mobility (transfer genes, blue), other (gene without known association to anti-defence, teal), uncharacterized genes enriched in the leading regions (grey), other uncharacterized genes (white). Asterisks (*) next to gene annotations indicate a potential anti-defence function. Frpo promoters are indicated by arrows: a solid arrow for promoters with significant similarity to known Frpo sequences, a dashed arrow for Frpo candidates and dashed with an asterisk (*) for low-certainty candidates. Further islands are presented in Extended Data Fig. 5. e, Predicted secondary structure of the Frpo in S. marcescens plasmid from b. f, Putative Frpo candidate in the conjugative element from c. Regions corresponding to the −10, −35 and UP elements, as well as their complementary regions, are coloured and highlighted above the structure, along with the canonical sequences of these elements. Uppercase letters indicate nucleotides conforming to canonical sequences of the −35 and −10 elements (5′-TTGACA-3′ and 5′-TATAAT-3′, respectively; full sequences in Extended Data Fig. 6b,c).

Fig. 3. The effect of anti-CRISPR in the leading region on conjugation efficiency.

a, Schematic representation of the donor and recipient cells during the conjugation experiments. The F plasmid’s T-strand is transferred into the recipient, starting with its leading region. In the recipient, a separate plasmid expresses Cas9 targeting the F plasmid. b, Representative example of transconjugant cell growth for each F plasmid variant. Droplet rows represent serial 1:5 dilutions. c, Conjugation efficiency as a function of the positioning and orientation of the anti-CRISPR acrIIA4 on the F plasmids. Bars indicate the mean conjugation efficiency of each F plasmid variant relative to the control, which is an F plasmid with no anti-CRISPR gene transferred to a recipient with a non-targeting gRNA. Red bars represent recipients with a targeting gRNA, whereas grey bars represent recipients with a non-targeting gRNA. Conjugation efficiency is calculated as the transconjugant frequency (T/(R + T)) per conjugation, divided by the transconjugant frequency of the control (T, transconjugants and R, recipient cells). Data are presented as mean values ± s.e.m. from n = 3 biologically independent experiments. Individual data points from each experiment are overlaid on the corresponding bars. Source Data

Fig. 4. Proposed model of the plasmid protection by diverse defence evading systems encoded in the leading region.

The figure illustrates how anti-defence genes with ssDNA promoters in the leading region can protect plasmids during the very early stages of transfer to a recipient cell. As the bacterial immune response activates defence systems against foreign DNA, various plasmid-encoded genes counteract these defences: anti-CRISPRs can inhibit CRISPR–Cas systems; SOS inhibitors (such as PsiB) can repress the cell’s SOS response by preventing RecA protein activation, thus inhibiting the cleavage of LexA, an SOS-response transcriptional repressor. Single-stranded binding (SSB) proteins, involved in the SOS-response inhibition, may protect transferred ssDNA from host nucleases. Anti-restriction proteins may prevent DNA cleavage by restriction–modification systems. MTases, methylating the transferred DNA can impede recognition by the host restriction–modification (RM) systems; and toxin–antitoxin genes potentially act against competitive MGEs or host defence systems (Supplementary Discussion). The top-right panel shows a schematic genetic organization of a conjugative plasmid. The four main functional gene groups are colour-coded: propagation (blue), adaptation (purple), replication (green) and the anti-defence genes (red) within the establishment module (orange).

Extended Data Fig. 1. Workflow overview.

a, All assembled genomes and metagenomes available in NCBI’s and EBI’s databases were analysed. In the first phase, we considered only sequences explicitly annotated as plasmids. In the second phase, we included all sequences that contained a detectable relaxosome component gene in proximity to a known oriT sequence. We termed the combined set of the two phases “potential conjugative elements”. Redundant elements of this set were omitted based on sequence similarity. The non-redundant sequences were then classified according to their MOB types. b, We mapped to the potential conjugative elements the leading and lagging regions, known anti-defence genes (anti-restriction, anti-CRISPR, and anti-SOS), and transfer-related genes. We focused on the genes enriched in the leading region and characterized them further. These gene families were classified based on sequence and structural similarity, into the following groups: anti-defence, putative anti-defence, DNA-methyltransferases, SSB genes, toxin-antitoxin genes, uncharacterized genes, and other functional families.

Extended Data Fig. 2. Frequency of anti-defence genes relative to the origin of transfer (oriT).

a, Anti-defence gene frequency in 2,259 non-redundant sequences annotated as plasmids. The x-axis shows ORF indices relative to the oriT, with 0 representing the first ORF in the leading region. Positions that were represented in at least 150 sequences annotated as plasmids are plotted. The y-axis indicates the average frequency of anti-defence genes (combining SOS inhibition, anti-restriction, and anti-CRISPR genes) over a five ORF window. b, Analysis of 22,897 out of 26,327 non-redundant potential conjugative elements that could be reliably mapped to a MOB type. The x-axis shows ORF indices relative to the oriT, with 0 representing the first ORF in the leading region. The y-axis indicates the average frequency of anti-defence genes (combining SOS inhibition, anti-restriction, and anti-CRISPR genes) over a five-ORF window, with frequencies for each MOB type colour-coded and stacked. Only positions represented in at least 1,000 sequences are shown. c, Anti-defence gene frequency within the 21,907 potential conjugative elements retrieved from genomic and metagenomic databases (MOB types F, P1, Q, V, H, B). Only positions represented in at least 50 sequences are shown. Source Data

Extended Data Fig. 3. Phylogenetic distribution of the analysed conjugative elements.

The phylogenetic distribution of the 13,738 non-redundant plasmids and potential conjugative elements. This set excluded 8,169 elements originating from metagenomes and sequences that could not be reliably mapped to the tree. The bacterial tree of life was acquired from iTOL, with bars colour-coded according to phyla, representing the conjugative element count on a log₁₀ scale.

Extended Data Fig. 4. Structural comparison of known anti-CRISPRs and anti-CRISPR candidates identified based on their location and structural similarity.

a, AcrIIA8 anti-CRISPR (NCBI accession VDB32352.1) compared to a putative anti-CRISPR found in a conjugative element from a human gut metagenome (Mgnify analysis accession ERZ1741958, NODE_63). b, Anti-CRISPR AcrIIA8 (NCBI Protein accession VDB32352.1) compared to a putative anti-CRISPR found in a conjugative element of Staphylococcus epidermidis (NCBI accession VYVG01000002.1). c, Anti-CRISPR AcrIIA1 (NCBI accession WP_003722518.1) compared to a putative anti-CRISPR found in a conjugative element from a human gut metagenome (NCBI accession BABC01000244.1). d, Anti-CRISPR AcrVA5 (NCBI accession WP_046699157.1) compared to a putative anti-CRISPR found in a conjugative element of Salmonella enterica (NCBI accession AAEVVI010000002.1).

Extended Data Fig. 5. Additional examples of anti-defence islands from various bacterial hosts.

a–d, Islands from leading regions of conjugative elements in: (a) Streptomyces sp. DJ (NCBI accession PKSK01000906.1), (b) Enterococcus durans (NCBI accession VMRQ01000005.1), (c) Shigella sonnei (NCBI accession CM012291.1) and (d) Klebsiella variicola (NCBI accession CP008701.1). The oriT location is marked in red on the left. Genes are coloured-coded by functional category: anti-defence (red), DNA-methyltransferase (MTase, peach), toxin-antitoxin genes (orange), ssDNA-binding protein (SSB, yellow), mobility (transfer genes, blue), other (gene without known association to anti-defence, teal). Frpo promoters are indicated by an arrow. Asterisks (*) indicate unannotated genes with a putative anti-defence function. e, Position distribution of anti-defence genes at each ORF position relative to umuD homologues (set as position 0) in the leading region. f, Similar analysis for umuC homologues. In cases of multiple umuD/umuC genes in the same leading region, the homologue closest to the oriT was used as reference.

Extended Data Fig. 6. Sequences and predicted secondary structures of Frpo promoters.

a, Candidate Frpo* sequences found in S. enterica conjugative elements (Fig. 2a), showing limited similarity to known Frpo. Promoter elements (−10, −35, and UP) are indicated by red boxes. b, Frpo identified upstream of an SSB protein in S. marcescens plasmid (see Fig. 2b). c, Candidate Frpo’ detected in a conjugative element from an insect gut metagenome (Fig. 2c). Sequences in b,c exhibit high sequence similarity to known Frpo sequences.