RNase J1 and J2 Are Host-Encoded Factors for Plasmid Replication

Abstract

Plasmids need to ensure their transmission to both daughter-cells when their host divides, but should at the same time avoid overtaxing their hosts by directing excessive host-resources toward production of plasmid factors. Naturally occurring plasmids have therefore evolved regulatory mechanisms to restrict their copy-number in response to the volume of the cytoplasm. In many plasmid families, copy-number control is mediated by a small plasmid-specified RNA, which is continuously produced and rapidly degraded, to ensure that its concentration is proportional to the current plasmid copy-number. We show here that pSA564 from the RepA_N-family is regulated by a small antisense RNA (RNA1), which, when over-expressed in trans, blocks plasmid replication and cures the bacterial host. The 5' untranslated region (5'UTR) of the plasmid replication initiation gene (repA) potentially forms two mutually exclusive secondary structures, ON and OFF, where the latter both sequesters the repA ribosome binding site and acts as a rho-independent transcriptional terminator. Duplex formation between RNA1 and the 5'UTR shifts the equilibrium to favor the putative OFF-structure, enabling a single small RNA to down-regulate repA expression at both transcriptional and translational levels. We further examine which sequence elements on the antisense RNA and on its 5'UTR target are needed for this regulation. Finally, we identify the host-encoded exoribonucleases RNase J1 and J2 as the enzymes responsible for rapidly degrading the replication-inhibiting section of RNA1. This region accumulates and blocks RepA expression in the absence of either RNase J1 or J2, which are therefore essential host factors for pSA564 replication in Staphylococcus aureus.

Keywords: RNase J; Staphylococcus aureus; antisense RNA; essential host factors; plasmid replication control.

FIGURE 1

The 5′UTR of the repA mRNA can form putative regulatory structures that are modified by hybridization with the antisense RNA1. Fat lines indicate RNA, with key nucleotides shown. GG^UP, CC^MID, and GG^DW are indicated on the repA UTR RNA (in green) and the CC^RNA1 complementary to GG^UP are indicated on RNA1. The poly-U stretches that are part of the putative rho-independent transcriptional termination structures are underlined. The Ribosome Binding Site (RBS) has been highlight with a red box. (A) Overview of the pSA564 genetic locus where the repA regulatory elements are found. The repA and RNA1 promoters are shown as black and blue arrows, respectively. RBS indicates the GAGG RBS. The repA start codon is in italics and the first nucleotide in the start codon is at position +198 of the transcript. (B) The proposed OFF-structure of the repA UTR, where CC^MID and GG^DW base-pairs in the UTR-SLIII(TT) stem-loop to form a rho-independent transcriptional terminator, and at the same time sequester the RBS. This structure is favored by pairing of the mRNA with the complementary RNA as shown in (E). (C) The proposed ON-structure of the repA UTR, where GG^UP and CC^MID base-pairs in the UTR-SLII(ON) stem-loop, which prevents the formation of UTR-SLIII(TT) (B). (D) The putative secondary structure of RNA1, where RNA1-SLII doubles as kissing loop and a transcription terminator. (E) RNA1 in a duplex along its full length with the repA UTR. UTR-SLI cannot form, the UTR-SLII stem is much shorter (only 14 base-pairs), but formation of UTR-SLIII(TT) is possible.

FIGURE 2

An incompatibility factor is encoded within 269 bp of pSA564. (A) The region upstream of the repA gene contains two divergent genes, rac and rep1, where the latter has been split in two (and presumably inactivated) by a frameshift mutation to form rep1_N-truncation and rep1_C-truncation. The 2,216 bp region cloned into pRacUTR (light gray), includes the putative rac promoter (question mark) and the repA promoter. The 269 bp region cloned into pUTR269 (dark gray) includes the repA start codon and the RNA1 promoter, but not the repA promoter. (B) pSA564 is lost upon acquisition of either pRacUTR or pUTR269. Colonies from the transformation were picked, sequentially diluted 10-fold and spotted on Mueller-Hinton plates containing either no antibiotic, chloramphenicol, penicillin G, or both antibiotics. (C) Northern blotting shows a small antisense RNA, transcribed from pSA564 as well as from pRacUTR and pUTR269. The minor band observed below the main RNA1 signal is presumably a fragment of RNA1, since it is absent from the S. aureus strain RN4220 which carries no plasmid. Probe R1 was used to detect RNA1, and a probe against 5S rRNA was used as control.

FIGURE 3

Mutations that alters RNA1-repA mRNA interactions. (A) The CC^MIDGG mutation (red nucleotides in insert) prevents the formation of the putative OFF-structure by shortening the duplex of the stem and liberating some of the nucleotides of the ribosome binding site. (B) The CC^MIDGG mutation (red nucleotides in insert) removes two base-pairs from the foot of the 46 base-pair UTR-SLII(ON) stem, in addition to preventing the putative OFF-structure from forming (A). The GG^UPCC mutation (red nucleotides in insert) also weakens the UTR-SLII stem, which presumably shifts the equilibrium toward formation of UTR-SLIII(TT). The GG^UPCC,CC^MIDGG mutation (red nucleotides in insert) allows the full length of the UTR-SLII(ON) stem to form, and at the same time weakens/prevents the UTR-SLIII(TT). (C) The P_RNA1* mutations in the –10 region of the RNA1 promoter (red nucleotides) do not disrupt base-pairing within the UTR-SLII (wild-type in blue). See also Supplementary Figure 2 for full nucleotide sequence. (D) The kiss* mutations (in red) weaken the base-base interactions between RNA1-SLII and UTR-SLI. See also Supplementary Figure 2 for full nucleotide sequence.

FIGURE 4

Incompatibility between pVG1 derivatives and pSA564. (A) Schematic linear map of the pVG1 construct. The orange section is from the vector backbone, which includes a ColE1 origin of replication for E. coli (but not for S. aureus), an ampicillin resistance cassette for E. coli and a chloramphenicol resistance cassette for S. aureus. The green section was cloned from pSA564, and includes (from left to right) a truncated rep1_N gene, the rac gene, the RNA1 gene and the repA_N gene. The location of the rac promoter is unknown. Further details about the pSA564 insert can be found in Supplementary Figure 1. (B) The pVG1 plasmid and its derivatives were transformed into PR01, plated on MHC agar plates and incubated over night at 37°C. Colonies were picked directly from the transformation MHC agar plate and resuspended in MH medium. These suspensions were used for serial dilutions that were then spotted on MH, MHC, MHP, and MHCP agar-plates and incubated over night at 37°C. Growth on MHC indicates the presence of pVG1 or its derivatives (or the pEB01 control), growth on MHP indicates the presence of pSA564, and growth on MHCP indicates the presence of both pSA565 and a pVG1 plasmid. Details of the mutations in pVG1 can be found in Table 2.

FIGURE 5

RNA1 mutations affect plasmid copy-number. (A) Schematic linear map of the pVG1 construct. The orange section is from the vector backbone, which includes a ColE1 origin of replication for E. coli (but not for S. aureus), an ampicillin resistance cassette for E. coli and a chloramphenicol resistance cassette for S. aureus. The green section was cloned from pSA564, and includes (from left to right) a truncated rep1_N gene, the rac gene, the RNA1 gene, and the repA_N gene. The location of the rac promoter is unknown. Further details about the pSA564 insert can be found in Supplementary Figure 1. (B) The copy number of pVG1 and its mutant derivatives was estimated by plating on medium containing increasing concentrations of chloramphenicol. Growth at higher concentrations indicate higher chloramphenicol resistance gene dosage (i.e., higher plasmid copy number). The experiment was carried out in the PR02 (a RN4220 derivative) host strain, and plasmid replication is therefore whole dependent on cis-produced RepA. Details of the mutations in pVG1 can be found in Table 2.

FIGURE 6

Mutating the RNA1 promoter permits the detection of RepA by Western blotting.(A) Schematic linear map of the pVG9 construct. The orange section is from the vector backbone, which includes a pT181 replication cassette (with replication genes and origin) for replication in S. aureus, a ColE1 origin of replication for E. coli, an ampicillin resistance cassette for E. coli and a chloramphenicol resistance cassette for S. aureus. The green section was cloned from pSA564, and includes (from left to right) a truncated rep1_N gene, the rac gene, the RNA1 gene, and the repA_N gene. The latter is translationally fused to a streptavidine-FLAG tag (pink). The location of the rac promoter is unknown. Further details about the pSA564 insert and the backbone cloning vector can be found in Supplementary Figure 1. (B) RepA is undetectable under normal circumstances (pVG9) but becomes easily detectable when the RNA1 promoter is mutated (pVG9[P_RNA1^∗]). An additional mutation of the RepA start codon abolishes detection (pVG9[P_RNA1^∗, Met1Pro]), thus confirming the specificity of the antibody. Mutations that prevent formation of the putative OFF-structure (pVG9[GG^UPCC,CC^MIDGG] and pVG9[CC^MIDGG]) also permits the detection of RepA. In the construct “P_repA^∗” the -10 box of the repA promoter was mutated from TAATAT to TAATGG, in “Met1Pro” the AUG start codon of RepA was mutated to CCG. In the pVG9 constructs the RepA was C-terminally fused to a streptavidine-FLAG tag, and was detected using anti-FLAG antibodies. Antibodies against the CshA protein were used as loading control.

FIGURE 7

pSA564 is lost in RNase J1 and J2 mutants. SA564-derived strains, mutated for different components previously identified to be part of the S. aureus RNA degradation machinery, were spotted on MH and MHP. The presence of pSA564 was determined by growth on MHP. The strains mutated for the two genes, rnjA and rnjB, encoding the 5′–3′ ribo-exonucleases RNase J1 and RNase J2 were no longer able to form colonies on MHP plates.

FIGURE 8

RNA1 is essential for RNase J-deficient plasmid-loss. (A) Both pVG1 and pVG1[P_RNA1*] are able to establish themselves in RN4220 (an S. aureus strain which does not harbor any plamids), although the colonies are visibly smaller for pVG1[P_RNA1*]. pVG1[P_RNA1*] is able to establish itself in the ΔrnjA strain to form colonies in 24 h. The RN4220 and the ΔrnjA colonies are from two different MHC-plates. (B) The colony counts from MHC-plates (pVG1 and pVG1[P_RNA1*] transformants) at ∼24 h were normalized with counts from MHT-plates (pCN36 transformants, counted at ∼42 h). RN4220 exhibits similar colony, colony ratios for pVG1 and pVG1[P_RNA1*] (blue and green bars, respectively), whereas only pVG1[P_RNA1*] replicates rapidly in ΔrnjA to form colonies within 24 h. Note that ΔrnjA transformed with pVG1 will eventually form colonies when the plates are left long enough in the incubator.

FIGURE 9

RNA1 degradation and abundance. (A) Overview of where probe R1 and R2 hybridizes on RNA1. (B) Level of RNA1 is similar for pSA564 and pRacUTR in PR01, but the overall intensity of the combined RNA1 bands are much higher in ΔrnjA. (C,D) Rifampicin was used to block transcription and RNA was extracted at different time points to follow the rate of degradation by Northern blotting. The intensity of longest RNA1 species detected for each time-point was used to calculate the RNA1 half-life. However, RNA1 fragments accumulate in the ΔrnjA mutant, and these shorter RNA1 fragments cannot be readily quantified individually, so we here present the half-life of the combined signal of the R1 probe (i.e., all bands pooled). (E,F) A second probe (R2) against RNA1, this time targeting the 5′-end was used to re-probe the same membrane as in (C,D), respectively. (G) Probe against 5S rRNA to normalize the membrane shown in (C,E). (H) Probe against 5S rRNA to normalize the membrane shown in (D,F).