Cryptic Streptococcus mutans 5.6-kb plasmids encode a toxin-antitoxin system for plasmid stabilization

Abstract

Background: In all Streptococcus mutans strains, 5-13% carry a 5.6-kb plasmid. Despite its frequency, little is known about its mediated functions with most of the information coming from a single study focussing on plasmid pUA140.

Objective: Here, we describe the sequence and genetic organization of two S. mutans 5.6-kb plasmids, pDC09 and pNC101.

Results: Based on PicoGreen dsDNA quantification and Real-Time quantitative PCR (RTQ-PCR), the plasmid copy number was found to range between 10 and 74, depending on the strain tested. In contrast to literature, we identified six instead of five open reading frames (ORFs). While the putative gene products of ORF1 (as a Rep-protein) and ORF2 (as a Mob-protein) could be confirmed as being identical to those from pUA140, the functions of ORF3 (unknown) and ORF 4 (possibly AtpE homologue) could not be further revealed. However, the product of ORF5 showed a fairly high identity (38-50%) and structural similarity (58-74%) to RelE of Streptococcus pneumoniae, Streptococcus equi, and Streptococcus downei. In addition, we identified a functionally corresponding ORF6 encoding a protein with 61-68% identity (81-86% similarity) to the S. equi and S. downei antitoxin of the RelB family. RelE and RelB together form a plasmid-encoded toxin-antitoxin (TA) system, RelBE(plas). Despite its rather limited sequence similarity with chromosomal TA systems in S. mutans (RelBE(chro), MazEF, HicBA), we found similar tertiary structures applying I-Tasser protein prediction analysis.

Conclusion: Type II-toxins, as the plasmid-encoded RelE, are RNA endonucleases. Depending on their mRNA cleavage activity, they might 1) kill every plasmid-free progeny, thereby stabilizing plasmid transfer at the expense of the host and/or 2) help S. mutans enter a dormant state and survive unfavourable environmental conditions. Whilst a function in plasmid stabilization has been confirmed, a function in persistence under nutritional stress, tested here by inducing amino acid starvation, could not be demonstrated so far.

Keywords: HicBA; MazEF; RelBE; Streptococcus mutans; plasmid addiction system; regulator of translation; toxin–antitoxin cassette.

Fig. 1

Physical map and organization of plasmids pDC09 and pNC101.

Fig. 2

Amino acid sequence comparison of three chromosomally encoded TA systems (HicBA, MazEF, RelBE) between plasmid carrier strains DC09/NC101 and plasmid-free strains UA159/KK23. None of the alterations are typical for plasmid carriers or non-carriers or were found to be essential for the principal toxin or antitoxin function. In case of RelB, the plasmid-encoded version (RelB_plas of DC09 and NC101) has little similarity with RelB_chro (of UA159, DC09, NC101, all identical). In case of RelE, the plasmid-encoded version (RelE_plas of DC09 and NC101) has little similarity with the RelE_chro of UA159, DC09, and NC101. However, we found similar tertiary structures applying I–Tasser protein prediction analysis (Fig. 3).

Fig. 3

Protein prediction model based on I-Tasser, A) left: RelE_chro UA159, right: RelE_plas DC09/NC101; B) left: RelB_chro UA159, right: RelB_plas DC09/NC101. Source: http://zhanglab.ccmb.med.umich.edu/I-TASSER/ (12, 13)