Promiscuous restriction is a cellular defense strategy that confers fitness advantage to bacteria

Abstract

Most bacterial genomes harbor restriction-modification systems, encoding a REase and its cognate MTase. On attack by a foreign DNA, the REase recognizes it as nonself and subjects it to restriction. Should REases be highly specific for targeting the invading foreign DNA? It is often considered to be the case. However, when bacteria harboring a promiscuous or high-fidelity variant of the REase were challenged with bacteriophages, fitness was maximal under conditions of catalytic promiscuity. We also delineate possible mechanisms by which the REase recognizes the chromosome as self at the noncanonical sites, thereby preventing lethal dsDNA breaks. This study provides a fundamental understanding of how bacteria exploit an existing defense system to gain fitness advantage during a host-parasite coevolutionary "arms race."

Fig. 1.

Promiscuous activity of KpnI REase confers better protection. (A) Oligonucleotides (10 nM, 5′ end-labeled) containing a canonical sequence (-GGTACC-) or one of the most preferred noncanonical sequences (-GaTACC-) were incubated with Mg²⁺ and different dilutions of cell-free extracts of K. pneumoniae strain OK8 for 30 min at 37 °C and analyzed on an 8 M urea – 12% (wt/vol) polyacrylamide gel. The enzyme-mediated DNA cleavage of the 20-mer substrate generates a 12-mer end-labeled product. Lane C is substrate DNA with no enzyme. (B) Rates of DNA cleavage by WT and HF enzymes were assayed in the presence of canonical and noncanonical substrates. Reactions contained 1 nM (for canonical DNA cleavage) or 15 nM (for noncanonical DNA cleavage) of the enzyme and 100 nM of substrate in 10 mM Tris⋅HCl (pH 7.4). The reactions were initiated by the addition of 2 mM Mg²⁺ and incubation at 37 °C for different time intervals. The plot depicts the product formed vs. time. Data are presented as mean ± SEM. (C) Experimental design of the phage titration assay. Plaque-forming units (PFU) with cells harboring WT and HF were compared. A similar PFU count on both of the strains would indicate a comparable restriction by the WT and HF variants. Alternatively, a lower PFU count on WT-harboring cells compared with the HF variant would indicate greater protection against phages. (D) P1vir phage restriction by R.KpnI and its variants. The plaque-forming units of P1vir phage on cells harboring the WT, the HF variant, or the catalysis-deficient mutant (H149A) relative to cells containing empty vector are shown. (Left) Values from two independent experiments conducted in quadruplicate are plotted. (Right) Representative titer values are shown.

Fig. 2.

Promiscuous activity counters the antirestriction strategies. (A) Experimental design of phage titration assay with modified phage. Phage P1* (methylated at GGTACC sites in the genome) was prepared with the KpnI MTase overexpression strain as described in Materials and Methods. The HF variant should not restrict the modified phage, whereas the cells harboring a promiscuous WT REase would still decrease the EOP. (B) Plaque-forming units of phage P1* on cells harboring the WT, HF, or catalysis-deficient mutant (H149A) relative to cells harboring empty vector are shown. The measurements from two separate experiments conducted in quadruplicate are plotted.

Fig. 3.

In vivo analysis of KpnI R-M system-containing cells. (A) Organization of KpnI R-M system. (B) Western blot analysis with R.KpnI polyclonal antibodies. In lane 1, 0.2 μg of purified R.KpnI was used as a marker (M). In lanes 2–4, increasing amounts (250, 500, and 750 μg) of total protein from K. pneumoniae strain OK8 cell lysate (KpOK8) were used. In lanes 5–7, cell lysates (250 μg) prepared from three individual clones of E. coli cells expressing both MTase and REase from their respective endogenous promoters (KpnI R-M) are shown. (C) Phage P1* titer on E. coli cells harboring KpnI MTase alone (KpnI M) or both KpnI MTase and REase (KpnI R-M). The measurements from two separate experiments conducted in quadruplicate are shown.

Fig. 4.

Catalytic versatility confers fitness advantage. Growth profiles of E. coli BL26 cells harboring vector, the WT, or the HF variant in the presence of P1vir (A) or P1* (P1vir phage methylated at KpnI sites) (B). (C) Growth curve analysis of the cells in the absence of phage. Growth was monitored at OD₆₀₀. Growth curves shown are representative of three independent experiments.

Fig. 5.

Suppression of promiscuous activity of R.KpnI by polyamines and NAPs. Plasmid DNA cleavage activity of R.KpnI was assayed in the presence of different concentrations of polyamines, namely, spermidine and spermine or NAPs (i.e., HNS, Fis, HU). Protamine, a small arginine-rich protein was used as a control. Reactions contained 30 nM R.KpnI, and titration was carried out by preincubation of the supercoiled pUC18 DNA (14 nM) in buffer containing 10 mM Tris⋅HCl (pH 7.4) or 2 mM Mg²⁺ on ice with spermidine (0.5, 1, 2.5, 5, and 10 mM), spermine (10, 25, 50, 100, and 500 μM), or protamine (0.5, 1, 2.5, 5, and 25 μM) (A) or NAPs (25, 100, and 250 nM) (B). The reactions were initiated by addition of the enzyme and incubation at 37 °C for 1 h. Lane C is substrate DNA with no enzyme. Lane 0 shows the DNA cleavage reaction in the absence of polyamines or NAPs. Sp and P indicate specific and promiscuous DNA cleavage products, respectively. SC, L, and OC indicate the positions of the supercoiled, linear, and open circular forms of the plasmid, respectively.

Fig. 6.

Effect of canonical recognition sequence on the promiscuous activity of R.KpnI. Plasmids (14 nM) with (pUC18) or without (pUCΔK or pBR322) the KpnI recognition sequence were incubated with the enzyme (30 nM) on ice for 10 min. The reactions were then initiated by adding 2 mM Mg²⁺ to the plasmid enzyme mixture. The reactions were stopped at the indicated time points by adding a stop dye containing 10 mM EDTA. (A) pUC18 DNA containing a single KpnI recognition sequence. (B) pUCΔK DNA (pUC18 plasmid lacking the KpnI site). (C) Linearized pUCΔK DNA. (D) pBR322 DNA. (E) Linearized pBR322 DNA. P indicates promiscuous DNA cleavage products. SC, L, and OC indicate the positions of the supercoiled, linear, and open circular forms of the plasmid, respectively.

Fig. 7.

Coevolutionary arms race between the phages and their hosts results in the utilization of promiscuous activity as a defense strategy. Phage as the winner indicates successful infection of the host. Bacteria as the winner indicate efficient restriction of phage DNA. (A) In the absence of an R-M system, the phage emerges as the winner because of its ability to infect the bacteria. (B) Host adapts by acquiring an R-M system that can now restrict the invading DNA elements. (C) This, in turn, leads to the development of antirestriction strategies in the phages by (i) acquisition of DNA modification systems (e.g., methylation, glucosylation) or (ii) avoiding palindromic DNA sequences. (D) However, a promiscuous REase would target even those phages that are equipped with antirestriction mechanisms, thus conferring a survival advantage. The promiscuous cleavage characteristics may be acquired by the site-specific REase or retained during the evolution. Irrespective of the directionality, possessing promiscuous activity is advantageous to the bacteria in better restricting the invading genome elements.

Fig. P1.

Coevolutionary “arms race” between the phages and their hosts results in the utilization of promiscuous activity as a defense strategy. Phage as the “winner” indicates successful infection of the host. Bacterium as the “winner” indicates efficient restriction of phage DNA. (A) In the absence of the restriction–modification (R-M) system, the phage can infect the bacterium. (B) Host adapts by acquiring an R-M system that can now restrict the invading DNA elements. (C) This, in turn leads, to development of various antirestriction strategies in the phages. (D) However, a promiscuous REase would target even those phages with antirestriction mechanisms, thus conferring a survival advantage. The bidirectional arrow indicates that promiscuous cleavage characteristics may be acquired by a site-specific REase or retained during the evolution. Promiscuous activity is advantageous to the bacterium in better restricting invading genome elements.