tmRNA Is Essential in Shigella flexneri

Nonstop mRNAs pose a challenge for bacteria, because translation cannot terminate efficiently without a stop codon. The trans-translation pathway resolves nonstop translation complexes by removing the nonstop mRNA, the incomplete protein, and the stalled ribosome. P1 co-transduction experiments demonstrated that tmRNA, a key component of the trans-translation pathway, is essential for viability in Shigella flexneri. tmRNA was previously shown to be dispensable in the closely related species Escherichia coli, because E. coli contains a backup system for trans-translation mediated by the alternative release factor ArfA. Genome sequence analysis showed that S. flexneri does not have a gene encoding ArfA. E. coli ArfA could suppress the requirement for tmRNA in S. flexneri, indicating that tmRNA is essential in S. flexneri because there is no functional backup system. These data suggest that resolution of nonstop translation complexes is required for most bacteria.

Introduction mRNAs that lack a stop codon can originate from many events, including premature transcription termination, physical or chemical damage to a complete mRNA, or nucleolytic activity. Translation of a nonstop mRNA is problematic, because termination requires a stop codon. Release factors specifically recognize a stop codon in the ribosomal A site and promote hydrolysis of the peptidyl-tRNA, releasing the newly-synthesized protein and the ribosome [1,2]. Eukaryotes have mRNA proofreading mechanisms to limit translation initiation on nonstop mRNAs [3]. However, bacteria lack most mRNA proofreading mechanisms and ribosomes frequently translate to the end of a nonstop mRNA [4], generating a nonstop translation complex composed of a truncated mRNA, an incomplete nascent polypeptide, and a ribosome that cannot elongate or terminate translation by the canonical reactions. These nonstop translation complexes are resolved by trans-translation, a reaction mediated by tmRNA and a small protein, SmpB [5]. tmRNA contains a tRNAlike acceptor stem and a reading frame encoding a short peptide. SmpB binds tmRNA with high affinity [6]. During transtranslation, tmRNA-SmpB recognizes the nonstop translation complex and promotes translation of the tmRNA-encoded peptide onto the end of the nascent polypeptide [7,8]. This reaction releases the ribosome at a stop codon within tmRNA. The tmRNA-encoded peptide is recognized by several proteases, so the incomplete protein is rapidly degraded [9,10,11,12]. trans-Translation also stimulates degradation of the nonstop mRNA, so all components of the nonstop translation complex are efficiently removed [13,14].
trans-Translation occurs with high frequency in bacteria, and is found throughout the bacterial kingdom. Estimates from E. coli suggest that 2-4% of translation reactions end in trans-translation [4]. Genes encoding tmRNA (ssrA) and SmpB (smpB) have been identified in all sequenced bacterial genomes, indicating that transtranslation confers a selective advantage in all environments that can support bacterial life [15].
The abundance and ubiquity of trans-translation suggest that it is very important, and in some species ssrA and smpB are essential [16,17,18,19]. However, mutants of E. coli K12 lacking transtranslation activity are viable and have only mild growth defects in typical culture conditions [20,21]. A screen for E. coli genes that cannot be deleted in DssrA cells identified arfA, which encodes an alternative release factor [22]. ArfA binds nonstop translation complexes and recruits RF-2 to hydrolyze the peptidyl-tRNA, releasing the nascent polypeptide and ribosome [23,24]. ArfA is a backup system for trans-translation, because it is only produced when trans-translation activity is limiting [25,26]. In E. coli, arfA mRNA contains an RNase III cleavage site 59 of the stop codon, so expression of arfA will result in a nonstop complex [25]. When trans-translation is functional, ArfA will be tagged and degraded. However, if trans-translation is limiting, the truncated but active ArfA will be released [25]. arfA genes have been found in the genome sequences of many bacteria, including most enteric gamma-proteobacteria [22,26].
In this paper we show that ssrA is essential in S. flexneri, a human pathogen that causes acute dysentery. S. flexneri is closely related to E. coli [27]. In fact, the Shigella and Escherichia genera are phylogenetically indistinguishable [27]. S. flexneri lacks arfA, but when E. coli arfA is expressed in S. flexneri, ssrA can be deleted. These results suggest that trans-translation is essential in S. flexneri because it is the only available mechanism to resolve nonstop translation complexes.

Results and Discussion
ssrA is essential in S. flexneri Efforts to replace ssrA in the chromosome of S. flexneri 2a 2457T with a kanamycin-resistance gene using Red-mediated recombination were not successful in wild-type cells. However, when a second copy of ssrA was provided on a plasmid (pSsrA), kanamycin-resistant colonies were recovered. Diagnostic PCR reactions confirmed that ssrA was deleted in kanamycin-resistant cells containing pSsrA (Fig. 1). These results suggested that ssrA is essential in S. flexneri.
The requirement for ssrA was confirmed using a co-transduction experiment. A marker linked to the chromosomal ssrA locus was introduced into S. flexneri by transducing zfg-2003::Tn10 from a donor E. coli strain. The Tn10 insertion in zfg-2003 confers tetracycline resistance, and is located ,0.25 minutes from ssrA. S. flexneri zfg-2003::Tn10 was then transformed with pSsrA and the chromosomal copy of ssrA was replaced with a kanamycinresistance gene to produce S. flexneri zfg-2003::Tn10 ssrA::kan pSsrA. P1 lysates were prepared from this strain and used to measure cotransduction of ssrA::kan and zfg-2003::Tn10 into S. flexneri strains. The co-transduction frequencies were measured by selecting for tetracycline-resistant transductants and screening these transductants for kanamycin resistance. Based on the map distance between zfg-2003 and ssrA, the tetracycline-resistance gene and kanamycin-resistance gene should be co-transduced with a frequency of ,70% if ssrA were not essential. When S. flexneri pSsrA was used as a recipient, the co-transduction frequency was 7264%, close to the theoretical value. However, when wild-type S. flexneri with no additional copy of ssrA was used as a recipient, no kanamycin-resistant colonies were recovered from 550 tetracycline-resistant transductants. If ssrA were not essential, the probability of obtaining no co-transductants in these experiments would be (0.3) 550 , or ,10 2288 . These results show that unlike E. coli K12, S. flexneri requires ssrA for viability.
S. flexneri strains do not have arfA E. coli can survive without trans-translation activity because arfA is expressed in the absence of trans-translation and ArfA activity can resolve nonstop translation complexes [22,23,24,25]. Deletions of ssrA and arfA in E. coli are synthetically lethal [22]. Searches of genome sequences of Shigella species using BLAST [28] revealed that arfA homologs are present in S. boydii, S. sonnei, and S. dysenteriae, but not in S. flexneri. In Escherichia and Shigella species that have arfA, the gene is encoded between mscL and zntR, 39 of trkA. This chromosomal locus in S. flexneri strains contains an insertion element 39 of trkA, suggesting that arfA has been deleted by genetic rearrangement (Fig. 2A).
E. coli ArfA can suppress the lethal phenotype of ssrA deletion in S. flexneri Given the close phylogenetic relationship between E. coli and S. flexneri, it was surprising that the phenotypes caused by deleting ssrA were so different. However, the absence of arfA in S. flexneri suggested that the difference might be due to the absence of a backup mechanism for trans-translation in S. flexneri. To determine if E. coli ArfA could suppress the requirement for ssrA in S. flexneri, a plasmid encoding His6-ArfA from the ASKA collection (pCA24N-His 6 -ArfA) was transformed into S. flexneri and these cells were used as the recipient in a co-transduction experiment with P1 lysates from S. flexneri zfg-2003::Tn10 ssrA::kan pSsrA. ssrA::kan and zfg-2003::Tn10 were co-transduced into S. flexneri pCA24N-His 6 -ArfA with a frequency of 6968%, indicating that ssrA is not essential in cells with pCA24N-His 6 -ArfA. Diagnostic PCR reactions confirmed that ssrA was deleted in the kanamycinresistant cells (Fig. 2B). These results indicated that ArfA can suppress the requirement for ssrA in S. flexneri.
ssrA could be deleted in S. flexneri pCA24N-His 6 -ArfA cells even when IPTG was not added to induce arfA expression. Western blotting revealed that amount of ArfA in uninduced S. flexneri ssrA::kan pCA24N-His 6 -ArfA was 15-20% the amount in cells that had been induced (Fig. 3A). When S. flexneri ssrA::kan pCA24N-His 6 -ArfA cells were inoculated into fresh medium containing IPTG the lag phase of growth was shorter than when no IPTG was added, but the doubling time during logarithmic growth was similar   (Fig. 3B). These data indicate that the amount of ArfA produced in uninduced cultures is sufficient for viability of S. flexneri in the absence of trans-translation, but higher levels of ArfA are required for optimal growth in culture.
The results described here suggest that all species of the Escherichia/Shigella lineage require a mechanism to resolve nonstop translation complexes. For most species in this group ssrA is not essential because ArfA acts as a backup system, but because S. flexneri does not have arfA, ssrA is essential. In other proteobacteria, such as Caulobacter crescentus, ssrA is not essential [29], but there is no ArfA homolog. Perhaps nonstop translation complexes are not as severe a challenge in these species. Alternatively, these species may have a distinct mechanism for releasing nonstop translation complexes.
Like E. coli, S. flexneri has a gene encoding ArfB (YaeJ), a second alternative release factor. Purified ArfB can release nonstop translation complexes in vitro, and multicopy expression of arfB in E. coli can suppress the synthetic lethality of ssrA and arfA deletions [30,31]. However, endogenous arfB does not support deletion of ssrA in S. flexneri or simultaneous deletion of ssrA and arfA in E. coli [30], suggesting that it is not expressed under culture conditions even when nonstop translation complexes accumulate to lethal levels.

Bacterial strains and plasmids
All strains were grown at 37uC in lysogeny broth supplemented with 30 mg/ml kanamycin, 20 mg/ml chloramphenicol, or 12 mg/ ml tetracycline as appropriate ( Table 1). Transformation of plasmids into S. flexneri was performed by electroporation [32].
The sequences of ssrA genes from E.coli and S. flexneri are identical. Plasmid pSsrA was made by amplifying ssrA from E. coli K-12 MG1655 using primers ssrAU_BamHI and ssrAL_HindIII, digesting the product with BamHI and HindIII, and ligating the resulting DNA into pJS14 cut with the same enzymes. Redmediated recombination was performed using the Wanner method [33]. S. flexneri cells containing pKD20 and pSsrA were transformed with a PCR product made using primers Shi_ssrA_del-F and Shi_ssrA_del-R with plasmid pKD4 as the template. E. coli strain BD1467 was used as the donor strain to transduce zfg-2003::Tn10 mutation into S.flexneri. Plasmid pCA24N-His 6 -ArfA was a gift from the Ades lab, and the sequence of arfA on the plasmid was verified prior to use.

PCR to verify gene replacement
Replacement of ssrA in S. flexneri ssrA::kan pSsrA was verified by colony PCR using primers ssrA_KO_check-F and ssrA_KO_check-R, which flank the ssrA gene. As a control, colony PCR using the same primers was also performed on wild-type S. flexneri. To verify replacement of ssrA in S. flexneri ssrA::kan pCA24N-His 6 -ArfA, genomic DNA was prepared from the deletion strain [36], and used as template for PCR amplification using primers ssrA_KO_check-F and ssrA_KO_check-R. As a control, genomic DNA was prepared from wild-type S. flexneri and used as a template for PCR amplification using the same primers. The expected product size for wild-type was 681 bp, and for ssrA::kan the expected product size was 1724 bp.

ArfA expression and growth
Expression of ArfA in S. flexneri ssrA::kan pCA24N-His 6 -ArfA was examined under three different conditions. Saturated cultures of S. flexneri ssrA::kan pCA24N-His 6 -ArfA grown with or without IPTG were diluted 1:100 into growth medium with 1 mM final concentration of IPTG, or cells were grown without any exposure to the inducing agent. As a negative control, wild-type S. flexneri without plasmid was tested. Cultures were grown to OD 600 = 0.4 at 37uC, and cells were harvested by centrifugation and analyzed by Western blotting.
Growth curves were obtained by diluting saturated cultures of S. flexneri ssrA::kan pCA24N-His 6 -ArfA 1:100 in growth medium with or without 1 mM IPTG at 37uC with constant shaking, and sampling cultures every 20 min to measure OD 600 . Points between 80 min and 160 min were fit to the single exponential function OD 600 = c(e bt ), where t is time, and the value for b was used as the growth rate.

Western blotting
Cell pellets were lysed by boiling in SDS sample buffer (63 mM Tris-HCl (pH 6.8), 2% SDS, 10% glycerol, 0.1% 2-mercaptoethanol, 0.0005% bromophenol blue). The samples were resolved on a 15% SDS polyacrylamide gel, blotted to PVDF membrane, and probed with 1:5000 dilution anti-PentaHis antibody (Qiagen) [37]. Goat anti-mouse antibody (GE Healthcare) was added at 1:5000 dilution for 1 h at room temperature prior to addition of ECF reagent and imaging with a Typhoon 9410 (GE Healthcare). The relative amounts of ArfA protein were determined by quantifying the bands using InageQuant software (GE Healthcare).