Conjugative plasmid-encoded toxin-antitoxin system PrpT/PrpA directly controls plasmid copy number

Abstract

Toxin-antitoxin (TA) loci were initially identified on conjugative plasmids, and one function of plasmid-encoded TA systems is to stabilize plasmids or increase plasmid competition via postsegregational killing. Here, we discovered that the type II TA system, Pseudoalteromonas rubra plasmid toxin-antitoxin PrpT/PrpA, on a low-copy-number conjugative plasmid, directly controls plasmid replication. Toxin PrpT resembles ParE of plasmid RK2 while antitoxin PrpA (PF03693) shares no similarity with previously characterized antitoxins. Surprisingly, deleting this prpA-prpT operon from the plasmid does not result in plasmid segregational loss, but greatly increases plasmid copy number. Mechanistically, the antitoxin PrpA functions as a negative regulator of plasmid replication, by binding to the iterons in the plasmid origin that inhibits the binding of the replication initiator to the iterons. We also demonstrated that PrpA is produced at a higher level than PrpT to prevent the plasmid from overreplicating, while partial or complete degradation of labile PrpA derepresses plasmid replication. Importantly, the PrpT/PrpA TA system is conserved and is widespread on many conjugative plasmids. Altogether, we discovered a function of a plasmid-encoded TA system that provides new insights into the physiological significance of TA systems.

Keywords: ParE; origin of replication; plasmid copy number; plasmid replication; toxin–antitoxin.

Fig. 1.

ParE/PF03693 pairs are abundant on conjugative plasmids. (A) ParE toxins are associated with multiple antitoxin PFAM families. Each PFAM is represented by a specific color. (B) Distribution of ParE/PF03693 pairs in representative phyla and genera. (C) The average number of ParE/PF03693 pairs per genome. (D) Multiple sequence alignment constructed by ClustalW to compare the amino acid sequence identity of ParE associated antitoxins from PF03693 and PF09386 in the conjugative plasmids.

Fig. 2.

PrpT and PrpA constitute a TA pair. (A) Circular map of pMBL6842 and the position of the prpA-prpT operon. (B and C) Viability of cells overexpressing prpT, prpA, and prpA-prpT in P. rubra. IPTG, isopropyl β-ᴅ-thiogalactoside. (D) Morphologies of cells overexpressing prpT with 0.5 mM IPTG for 2 h over time (see Movie S1). The ghost and lysed cells are marked with blue and red arrows. (E) PrpT and PrpA form a complex in vitro. His-tagged PrpT and untagged PrpA were coproduced via pET28b-prpA-prpT-His (lane 3) and copurified with increasing concentration of imidazole (lanes 4–6). Lane 1: size marker; lane 2: NC (no IPTG). (F) The BACTH assay showed that PrpT interacts with PrpA. The data are from three independent cultures. SDs are shown, and statistical significance (NS, no significant; *P < 0.05; **P < 0.01; ***P < 0.0001) is indicated with asterisks in Figs. 2–4 and 6. Images shown in Figs. 2–6 are representative images.

Fig. 3.

PrpA and PrpT/PrpA bind to the prpA-prpT operon. (A) Schematic of the prpA-prpT operon. The Shine–Dalgano sequences of prpA and prpT are highlighted in red. (B) Comparison of the RBS activities is shown using the two lacZ reporter plasmids, pRBS^prpA or pRBS^prpT in A. (C) Western blot showing that the production of PrpA exceeded PrpT (n = 3); the results were obtained by using FLAG-tagged PrpA (10.5 kDa) or PrpT (12.3 kDa). RNAP was used as a control. (D) The promoter activity was measured by overexpressing PrpA or PrpT/PrpA using pRBS^prpA. (E) EMSA results showed that PrpA and PrpA/PrpT complex bound and shifted the promoter in a dose-dependent manner. (F) The binding site of PrpA is analyzed by the DNase I footprinting assay using two different concentrations of PrpA. The 30-bp binding site of PrpA covers the −35 and partial −10 regions of the promoter.

Fig. 4.

PrpA controls plasmid copy number. (A) PrpT/PrpA TA system does not control the segregational stability of plasmid pMBL6842. The wild-type and deletion mutants were cultured in 2216E medium without antibiotic for 448 generations. Quantification of plasmid copy number after cultivation for 6 h (in the exponential phase) and 24 h (in the late stationary phase) from the starting point OD₆₀₀ ∼0.01, by qPCR (n = 4) (B) and by PCR-free whole-genome sequencing (C). (D) Quantification of the plasmid copy number when expressing prpA and prpAT under their native promoter in the ΔprpAT strain after 24 h by qPCR (n = 4).

Fig. 5.

PrpA competes with RepB for binding to the pMBL6842 ori. (A) Nucleotide sequence of the pMBL6842 ori region. The AT-rich region is marked in bold letters. The 5′-GATC sites are indicated by boxes above the sequence. The iterons are underlined using red arrows. The binding sites of RepB and PrpA are highlighted with blue and green, respectively. The conserved binding motif of PrpA in site 1 and 2 is indicated by a box in the sequences. (B) EMSA results showing that RepB and antitoxin PrpA bind and shift the pMBL6842 ori. (C) DNase I footprinting assays used to determine the binding sites of RepB and PrpA. EMSA results showing that PrpA and RepB compete for binding to ori when the two proteins are added sequentially (D and E) or added at the same time (F).

Fig. 6.

PrpA is labile and degradation occurs at both termini. (A) The stability of N^FLAG-PrpA-C^His was determined by Tricine-SDS-PAGE (Upper), Western blot using anti-His antibody (Middle), and anti-FLAG antibody (Lower) after being treated with P. rubra cell lysates. Lane 1, size marker; lane 2, cell lysate; lane 3, purified N^FLAG-PrpA-C^His, lanes 4–9, N^FLAG-PrpA-C^His treated with P. rubra lysates over time. (B) Schematic of the modular organization of PrpA. (C) A BACTH assay was performed to assess interactions between PrpA proteins of varying lengths and PrpT.

Fig. 7.

Proposed mechanism underlying how PrpT/PrpA controls plasmid replication. When PrpA is stable, it binds to the iterons in the ori, interfering with the binding of RepB to the ori, thus preventing overreplication of the plasmid. During stress, PrpA is degraded, thus derepressing the inhibition of the binding of RepB to ori.