The SsrA-SmpB Ribosome Rescue System Is Important for Growth of Bacillus subtilis at Low and High Temperatures

Bacterial strains and genetic methods.The B. subtilis strains used in this study are listed in Table 1. Standard recombinant DNA methods were according to reference 49, and transformation of B. subtilis PB2 and derivatives was according to reference 14. Plasmids used for these constructions are shown in Table 2. To locate promoter activity, fragments were PCR amplified and cloned into transcriptional fusion vector pDG268, which integrates in single copy at the amyE locus (2). To make in-frame smpBΔ1, removing triplets 4 to 153, a four-primer fragment (24) was cloned into the pCP115 vector (6) to create pJH13; smpBΔ1 was then substituted for the wild-type allele via a two-step replacement procedure (51). To complement the ssrA::cat insertion in trans, a fragment was PCR amplified and cloned into the pDG1730 vector (19), placing P₃, P_HS, and ssrA in single copy at amyE. DNA sequences were confirmed for all of the constructs.

Reverse transcription (RT)-PCR and rapid amplification of cDNA ends (RACE)-PCR experiments.Shake cultures were grown at 37°C to mid-exponential phase in buffered Luria broth (BLB) medium lacking salt (5) and then diluted 1:25 into fresh BLB medium. At mid-exponential phase, some cultures were subjected to a 10-min ethanol stress (5%, vol/vol) while others were allowed to grow untreated before harvesting and RNA extraction. Total RNA was isolated with the RNeasy kit (QIAGEN Inc., Valencia, CA) according to the manufacturer's instructions, except that the lysozyme concentration was increased to 15 mg/ml and a homogenization step (passage through a 20-gauge needle) was added.

For RT-PCR, RNA was extracted from wild-type (PB2) cells and the RT reaction was done according to reference 49. One microgram of total RNA and 25 pmol of gene-specific primer 5′-GATTACTTAAGCGTCTACG-3′ were incubated at 65°C for 10 min, and then Transcriptor reverse transcriptase (Roche Applied Science, Mannheim, Germany) was added and incubation was continued at 50°C for 90 min. The resulting cDNA was amplified in a PCR with Taq polymerase (Promega, Madison, WI). As a negative control, the RT reaction was performed without the addition of reverse transcriptase. To detect a yvaK-to-ssrA complement, the forward primer FP1 (5′-CGGATGTAAGGATGCTGGGA-3′) was used with the reverse primer RP (5′-GGTTTCACTCATCTTCTTGCC-3′); to detect a secG-to-ssrA complement, the FP2 forward primer (5′-GGTTATCGTCAGCATTGC-3′) was used with the same reverse primer. PCR products were analyzed by agarose gel electrophoresis.

For RACE-PCR (16), RNA was extracted from wild-type (PB2) or sigB mutant (PB153) cells. RACE-PCRs were done as previously described (47), with a primer complementary to the poly(A) tail (5′-GACCACGCGTATCGATGTCGACT₁₆V-3′ [where V = A, C, or G]) and two nested, gene-specific primers, (i) YVAK1 (outer, 5′-GTATGTACAAGTTCTTCAGGC-3′) and YVAK2 (inner, 5′-GACGCCATGTCCTTCATATTGAG-3′); (ii) SMPB1 (5′-GTTATAGCGGTTTCCCTG-3′) and SMPB2 (5′-GGGCTGACGTGCATATTGTGG-3′); or (iii) SSRA1 (5′-TCTTCTTACGTTCTCAGA-3′) and SSRA2 (5′-CGCAAGCGTAGCCTACTTGGA-3′). PCR products were separated on agarose gels, purified and sequenced.

β-Galactosidase accumulation assays.Shake cultures were grown at 37°C in BLB medium and diluted 1:25 into fresh medium as described for the RNA experiments above. At mid-exponential phase, cultures were subjected to ethanol stress (5%, vol/vol). Samples were collected at the indicated times and assayed according to Miller (39), by using sodium dodecyl sulfate and chloroform to permeabilize the cells. Protein levels were determined with the Bio-Rad protein assay reagent (Bio-Rad Laboratories, Hercules, CA). Activity was defined as ΔA ₄₂₀ × 1,000 per minute per milligram of protein.

Growth experiments.High-temperature growth experiments were done essentially as described by Holtmann et al. (26), and low-temperature experiments were done essentially as described by Brigulla et al. (7). For high-temperature experiments, wild-type and mutant cells were precultured at 37°C in shake flasks containing either BLB medium or Spizizen's minimal medium with 0.5% glucose (SMM; reference 1). At A ₅₇₈ = 0.5, exponentially growing cells were diluted 1:5 into fresh BLB medium or SMM and transferred to a 52°C water bath shaker. For low-temperature experiments, cells were similarly precultured in SMM, diluted 1:5 into fresh SMM, and transferred to a 16°C air shaker. Growth was monitored with a Klett-Summerson colorimeter with a number 66 (red) filter; units were plotted on a linear scale versus time, in accordance with references 7 and 26.

The secG-ssrA gene cluster can be transcribed as a single polycistronic message.As shown in Fig. 1A, the secG-ssrA region encodes only a single rho-independent terminator sequence that lies downstream from ssrA, leading us to hypothesize that the five-gene cluster could be transcribed as a single message. We used RT-PCR to look for such a message in RNA isolated from unstressed, exponential-phase B. subtilis cells. Two different forward primers were designed, one to anneal to the secG message and the other to anneal to the yvaK message. In conjunction with a reverse primer designed to anneal to the ssrA message, the secG primer was expected to amplify a product of 4,508 bp and the yvaK primer was expected to amplify a product of 4,087 bp.

• Open in new tab
• Download powerpoint

FIG. 1.

The five genes in the secG-ssrA cluster can be transcribed as a polycistronic message. (A) Map of the B. subtilis secG-ssrA region showing promoter activities located in this work (P₁, P_B, P₂, and P₃), the heat shock promoter described by Muto et al. (P_HS; reference 41), and the proposed rho-independent terminator sequence following ssrA. A full-length transcript of the region is drawn beneath the kilobase scale, together with the predicted sizes of the RT-PCR products resulting from its amplification with primers FP1, FP2, and RP. (B) RT-PCR full-length transcript detection in total RNA isolated from wild-type B. subtilis. Lanes +RT and −RT indicate with and without reverse transcriptase, lane M contains size markers, and the predicted lengths of the two RT-PCR products are on the right.

As shown in Fig. 1B, products of the predicted size were found in the reaction mixtures with reverse transcriptase. These products were missing from the reaction mixtures lacking transcriptase, ruling out the possibility that they arose from amplification of contaminating DNA. We conclude that at least some full-length message is transcribed from the secG-ssrA cluster and that these genes constitute an operon. However, further analysis revealed the existence of multiple promoters within this region, some stress inducible. These promoters fell into three sets: (i) two in the secG-yvaK region, (ii) one in the rnr-smpB interval, and (iii) two in the smpB-ssrA interval.

The secG-yvaK region.Expression profiling experiments previously identified yvaK as a candidate element of the σ^B regulon (22, 45, 47); its product is a carboxylesterase of unknown function (23). While a putative σ^B-dependent promoter for yvaK could be located by inspection, it has not been characterized experimentally. With RNA isolated from cells that had been ethanol stressed to increase σ^B activity, RACE-PCR experiments found several potential 5′ ends for the yvaK message. We could obtain sequences from the two most prominent PCR products (Fig. 2A and B), locating one 5′ end to a site 24 nucleotides (nt) upstream from the ATG initiation triplet of yvaK (Fig. 2E) and the other to a site 31 nt upstream from the ATG of secG (Fig. 2D). The yvaK 5′ end was preceded by sequences previously suggested to define the σ^B-dependent promoter, with GtTTTt at −35 and tGGaAa at −10 (22, 45); this 5′ end was not apparent when RNA extracted from ethanol-stressed cells of a sigB null mutant was used (Fig. 2B). We provisionally call this promoter P_B. By contrast, the secG 5′ end was preceded by sequences resembling a σ^A promoter, with gTGACA at −35 and TAaAAT at −10. This signal was diminished but still detectable in the sigB null mutant, while more rapidly migrating minor bands increased in intensity, suggesting that they were decay products of the secG message. We provisionally call this promoter P₁. The 5′ end associated with P1 was found with a primer within yvaK, indicating that the secG message extends into the yvaK reading frame.

• Open in new tab
• Download powerpoint

FIG. 2.

One σ^A-like and one σ^B-dependent promoter in the secG-yvaK region. (A) Diagram showing promoter location, cDNA from the RACE-PCR experiments, and fragments fused to a lacZ reporter for in vivo assay. (B) RACE-PCR products of RNA extracted from ethanol-stressed cells of wild-type PB2 (lane 1) or sigB null mutant PB153 (lane 2), with negative controls in which the poly(dA) tailing step was omitted for first-strand synthesis products from wild-type (lane 3) or sigB mutant (lane 4) RNA. Size markers are on the left, and the positions of the 596- and 233-bp products (determined by direct DNA sequencing) are indicated on the right. The 5′ ends of these products are located on the DNA sequences in panels D and E. (C) Promoter activity arising from the fragments shown in panel A was measured indirectly by β-galactosidase (β-Gal) accumulation from lacZ transcriptional fusions. Early-exponential-phase cells were stressed by the addition of 5% ethanol at time zero. Longer fragment: filled circles, PB978 (wild type); open circles, PB981 (sigBΔ3). Shorter fragment: filled triangles, PB918 (wild type); open triangles, PB919 (sigBΔ3). (D) yvaM-secG intercistronic region, with the translation start sites in bold (cat for divergently transcribed yvaM and atg for secG). The 5′ end of the 596-bp product from panel B is marked by a filled inverted triangle (P₁), with potential −35 and −10 sequences for σ^A in bold and underlined. (E) secG-yvaK intercistronic region, with the secG termination site (tag) and the yvaK start site (atg) in bold. The 5′ end of the 233-bp product from panel B is marked by a filled inverted triangle (P_B), with potential −35 and −10 sequences for σ^B in bold and underlined.

To test if the secG-yvaK region contained functional promoter activities, we made two transcriptional fusions to a lacZ reporter gene (Fig. 2A). One carried a fragment extending from the divergently transcribed yvaM gene preceding secG to a site within yvaK; this would be expected to contain both the P₁ and P_B activities suggested by RACE-PCR. The other carried a shorter fragment that removed 387 bp from the 5′ end of the first; this would be expected to contain only P_B activity. These fusions were placed in single copy at the amyE chromosomal locus. As shown in Fig. 2C, in the wild type the longer fusion manifested promoter activity in unstressed cells, and this activity increased following ethanol stress. In the sigB null mutant, this same fusion had slightly less activity in unstressed cells and only a modest increase following stress. By contrast, in the wild type the shorter fusion had very low activity in unstressed cells and a significant increase following ethanol stress, as is the case with well-characterized σ^B-dependent fusions. Moreover, this stress-induced increase in activity was abolished in the sigB null mutant. The fusion data are consistent with the RACE-PCR assays and led us to conclude that at least two promoters lie in the secG-yvaK region, i.e., a σ^B-independent P₁ activity preceding secG and a σ^B-dependent P_B activity preceding yvaK. On the basis of the β-galactosidase accumulation assays, P_B appears to be relatively weak, consistent with its sequence divergence from more typical σ^B-dependent promoters.

The rnr-smpB interval.A similar analysis located promoter activity in the rnr-smpB interval. RACE-PCR produced a single abundant product of about 290 bp that did not appear to differ quantitatively among the three sources of RNA—wild-type cells with or without ethanol stress and sigB mutant cells with ethanol stress (Fig. 3A and B). Sequencing of the PCR product located its 5′ end to a site 47 nt upstream from the ATG initiation triplet of smpB (Fig. 3D). This 5′ end was preceded by sequences resembling an extended σ^A promoter (21), here provisionally called P₂, with TTGtag at −35 and TGgTAaAAT at −10. We made two lacZ fusions to determine whether the region containing these sequences had promoter activity (Fig. 3A). The first carried a fragment extending from within the 748th triplet of the upstream rnr gene to a site within smpB, and the second removed 198 bp from the 5′ end of this fragment, deleting the proposed −35, −10, and +1 sequences of P₂. The longer fragment indeed contained promoter activity, and this increased less than twofold in ethanol-stressed cells (Fig. 3C). By contrast, the fusion bearing the shorter fragment had no detectable activity. From the RACE-PCR and fusion results, we conclude that the rnr-smpB interval contains at least one promoter, P₂, which is modestly induced by ethanol stress.

• Open in new tab
• Download powerpoint

FIG. 3.

One σ^A-like promoter in the rnr-smpB interval. (A) Diagram of interval labeled as in the Fig. 2 legend. (B) RACE-PCR products of RNAs extracted from ethanol-stressed cells of wild-type PB2 (lane 1) or sigB null mutant PB153 (lane 2) and from unstressed wild-type cells (lane 3). Size markers are on the left, and the 292-bp product is labeled on the right; the 5′ end of this product is located on the sequence in panel D. (C) Promoter activity from the fragments shown in panel A measured by β-galactosidase (β-Gal) accumulation as described in the Fig. 2 legend. Longer fragment: filled circles, PB970. Shorter fragment: filled triangles, PB972. (D) rnr-smpB intercistronic region, with the rnr termination site (tag) and the smpB start site (atg) in bold. The 5′ end of the 292-bp product from panel B (determined by DNA sequencing) is marked by a filled inverted triangle (P₂), with potential −35 and −10 sequences for an extended σ^A promoter in bold and underlined.

The smpB-ssrA interval.Because of the rapid ssrA transcript processing to yield mature 10S tmRNA, fusion analysis was key in locating two potential promoter activities in the smpB-ssrA interval. RACE-PCR produced three products of 241, 295, and 370 bp from RNAs isolated from ethanol-stressed wild-type and sigB mutant cells (Fig. 4A and B, lanes 1 and 2). Of these, the 370-bp product appeared to be reduced or absent when RNA from unstressed wild-type cells was used (Fig. 4B, lane 3). DNA sequencing was able to locate the 5′ end for the 241-bp principal product, which was the 5′ end of the mature 10S tmRNA (Fig. 4D). However, no sequence data could be obtained for the 295- and 370-bp minor products, so we estimated the locations of their 5′ ends on the basis of gel mobility. The 295-bp product appeared to correspond to a transcript from the ssrA heat shock promoter mentioned by Muto and colleagues (41); its estimated 5′ end lay downstream from sequences resembling an extended σ^A promoter, here called P_HS, with TTGAaA at −35 and TGtTATAAT at −10 (Fig. 4D). The 370-bp product had an estimated 5′ end that lay downstream from another set of sequences resembling an extended σ^A promoter, here provisionally called P₃, with TTGAtt at −35 and TGcTATAcT at −10 (Fig. 4D).

• Open in new tab
• Download powerpoint

FIG. 4.

Two σ^A-like promoter activities in the smpB-ssrA interval. (A) Diagram labeled as in the Fig. 2 legend. (B) RACE-PCR products of RNAs extracted from ethanol-stressed cells of wild-type PB2 (lane 1) or sigB null mutant PB153 (lane 2) and from unstressed wild-type cells (lane 3). Size markers are on the left, and the 241-, 295-, and 370-bp products are labeled on the right. (C) Promoter activity from the fragments shown in panel A, measured as described in the Fig. 2 legend. Longest fragment: filled circles, PB979. Middle fragment: filled triangles, PB982. Shortest fragment: filled squares, PB973. β-Gal, β-galactosidase. (D) smpB-ssrA intercistronic region, with the smpB termination site (taa) and the 5′ end (g) of the mature tmRNA in bold. The estimated 5′ ends of the 370- and 295-bp products from panel B are marked by open inverted triangles (P₃ and P_HS, respectively), with potential −35 and −10 sequences for extended σ^A promoters in bold and underlined. The 5′ end of the 241-bp product from panel B (determined by DNA sequencing) is indicated by the filled inverted triangle (ssrA) and represents the processed (or mature) 10S tmRNA.

We made three lacZ fusions to determine whether the sequences containing P₃ and P_HS were important for promoter activity (Fig. 4A). The first carried a fragment extending from within the 86th triplet of the upstream smpB gene to a site within ssrA; the second removed 269 bp from the 5′ end of this fragment, deleting the proposed −35, −10, and +1 sequences for P₃; and the third removed an additional 110 bp, deleting the proposed −35, −10, and +1 sequences for P_HS. The longest fragment had strong promoter activity, and this increased about threefold in ethanol-stressed cells (Fig. 4C). With the second fragment, this activity significantly decreased, consistent with the loss of P₃, and ethanol induction was no longer apparent. With the third fragment, there was no detectable activity, consistent with the loss of P_HS. From the fusion results shown in Fig. 4C, P₃ is ethanol inducible, a conclusion in accord with the signals present in the RACE-PCR experiment shown in Fig. 4B (compare the strengths of the 370-bp signals in lanes 1 and 3). From the sum of the RACE-PCR and fusion results, we conclude that the smpB-ssrA interval contains at least two promoter activities, P₃ and P_HS.

Loss of smpB or ssrA function has a significant effect on growth at both high and low temperatures.Muto et al. (41) previously showed that ssrA function was required for robust growth of B. subtilis at high temperature and under conditions of ethanol or cadmium stress. Moreover, additional genes in the B. subtilis cluster are known to have stress- or tmRNA-related functions in other bacteria. For example, RNase R is a processive 3′-to-5′ exonuclease that can degrade RNA with significant secondary structure (11, 44); it has been implicated in quality control of rRNA, as well as degradation of certain aberrant mRNAs recognized by the trans-translation system (12, 48). This latter activity is consistent with the copurification from E. coli extracts of a complex containing tmRNA, SmpB, and substoichiometric amounts of RNase R (32). A role for E. coli RNase R in adapting to environmental transitions is suggested by its elevated level following cold stress, starvation, or entry into stationary phase (9, 10).

In order to uncover their possible roles in stress resistance, we constructed null alleles in the four genes downstream from the P_B promoter in the B. subtilis cluster and tested their effects on growth at high and low temperatures. It proved difficult to make strains bearing in-frame deletions within yvaK or rnr, so these were built with the pMUTIN4 plasmid, which disrupts the gene of interest while ensuring continued transcription of downstream genes (53). We were able to construct a strain bearing a deletion (smpBΔ1) that removed most of the smpB coding region; this in-frame deletion did not affect the initiation or termination signals for smpB transcription and translation. We also moved the ssrA::cat insertion of Muto et al. (41) into our genetic background so that all of the constructed strains were isogenic (Table 1).

In agreement with the results of Muto et al. (41), we found that the ssrA::cat insertion had a major impact on growth when cells were shifted from 37 to 52°C and that this phenotype was complemented in trans by a fragment containing P₃, P_HS, and ssrA (Fig. 5A and B). We also established that the smpBΔ1 null allele had the same high-temperature phenotype, as expected if SmpB is required for ssrA function (33, 34, 58). Our significant new findings are that (i) ssrA::cat had an equally striking impact on growth when cells were shifted from 37 to 16°C; (ii) this low-temperature phenotype could be fully complemented by a fragment containing P₃, P_HS, and ssrA; and (iii) the smpBΔ1 allele had the same phenotype (Fig. 5C). smpBΔ1 is an in-frame deletion that should not affect the expression of downstream genes, and the complementation test indicated that loss of ssrA function alone was sufficient to cause the low-temperature phenotype. Together, these results argue that it was the absence of a functional ribosome rescue system and not the production of a tmRNA fragment from the interrupted ssrA::cat allele that had the harmful effect. We conclude that smpB and ssrA are each required for robust growth at both low and high temperatures and that the promoters on the P₃-P_HS-ssrA fragment provide sufficient ssrA expression to support growth under both conditions.

• Open in new tab
• Download powerpoint

FIG. 5.

Growth of an smpB or ssrA mutant is impaired at both high and low temperatures. For high-temperature experiments, cells were grown at 37°C in shake flasks containing either BLB medium or SMM. At A ₅₇₈ = 0.5, cells were diluted 1:5 into fresh BLB medium (A) or SMM (B) and transferred to a 52°C water bath shaker. For low-temperature experiments, cells were grown in SMM, diluted 1:5 into fresh SMM (C), and transferred to a 16°C air shaker. All strains were Trp⁺ prototrophs to avoid possible transport effects. Filled circles, PB1017 (wild type); open circles, PB1018 (sigBΔ1), which served as a negative control in the low-temperature experiments (28); filled triangles, PB1019 (smpBΔ1); open triangles, PB1020 (ssrA::cat); filled squares, PB1021 (ssrA::cat with P₃, P_HS, and ssrA in trans).

In contrast, the yvaK or rnr null allele had no substantial effect on growth at either temperature extreme (data not shown). These rnr results are in agreement with the earlier study of Höper et al. (28), who noted that loss of rnr function had no significant effect on growth or survival in heat or cold.