SmpB, a unique RNA‐binding protein essential for the peptide‐tagging activity of SsrA (tmRNA)

A. Wali Karzai, Miriam M. Susskind, Robert T. Sauer

Author Affiliations

  1. A. Wali Karzai1,
  2. Miriam M. Susskind2 and
  3. Robert T. Sauer*,1
  1. 1 Department of Biology, Massachusetts Institute of Technology, Cambridge, MA, 02139, USA
  2. 2 Department of Biological Sciences, University of Southern California, Los Angeles, CA, 90089‐1340, USA
  1. *Corresponding author. E-mail: bobsauer{at}
View Full Text


In bacteria, SsrA RNA recognizes ribosomes stalled on defective messages and acts as a tRNA and mRNA to mediate the addition of a short peptide tag to the C‐terminus of the partially synthesized nascent polypeptide chain. The SsrA‐tagged protein is then degraded by C‐terminal‐specific proteases. SmpB, a unique RNA‐binding protein that is conserved throughout the bacterial kingdom, is shown here to be an essential component of the SsrA quality‐control system. Deletion of the smpB gene in Escherichia coli results in the same phenotypes observed in ssrA‐defective cells, including a variety of phage development defects and the failure to tag proteins translated from defective mRNAs. Purified SmpB binds specifically and with high affinity to SsrA RNA and is required for stable association of SsrA with ribosomes in vivo. Formation of an SmpB–SsrA complex appears to be critical in mediating SsrA activity after aminoacylation with alanine but prior to the transpeptidation reaction that couples this alanine to the nascent chain. SsrA RNA is present at wild‐type levels in the smpB mutant arguing against a model of SsrA action that involves direct competition for transcription factors.


SsrA RNA, also known as 10Sa RNA or tmRNA, is encoded in 12 of the 13 completely sequenced bacterial genomes as well as in many additional bacteria whose genome sequences are incomplete (Ushida et al., 1994; Williams and Bartel, 1996, 1998). The inclusion of this RNA of ∼360 nucleotides in even the limited genetic repertoire of Mycoplasma genitalium (482 genes) as well as in Gram‐negative and Gram‐positive species suggests an important biological role. Nevertheless, mutation of the ssrA gene in Escherichia coli results in a viable strain. SsrA‐defective strains have a variety of subtle phenotypes including temperature‐sensitive growth, reduced motility, inability to support growth of λimmP22 hybrid phage, induction of Alp protease activity and enhanced activity of several repressor proteins (Kirby et al., 1994; Komine et al., 1994; Retallack et al., 1994; Retallack and Friedman, 1995).

The SsrA molecule has features of both a tRNA and an mRNA, including the ability to be charged with alanine, bind 70S ribosomes and to encode sequences found at the C‐termini of some intracellular proteins (Komine et al., 1994; Ushida et al., 1994; Tu et al., 1995; Felden et al., 1996, 1997; Keiler et al., 1996; Muto et al., 1996, 1998; Tadaki et al., 1996; Williams and Bartel, 1996, 1998). The dual role of SsrA as a tmRNA was first suggested in a model in which alanine‐charged SsrA RNA binds in the A site of ribosomes stalled on damaged mRNA, donates the charged alanine to the nascent polypeptide chain and then replaces the damaged message as a surrogate mRNA, directing translation of a short peptide tag encoded by a sequence internal to SsrA (Keiler et al., 1996). The final translation product of this process carries an 11‐residue C‐terminal peptide tag (AANDENYALAA) that makes the tagged protein a substrate for proteolysis by C‐terminal specific proteases (Tu et al., 1995; Keiler et al., 1996; Gottesman et al., 1998). This system prevents the accumulation of ribosomes stalled at the end of damaged mRNAs that lack a stop codon and also provides a general quality‐control system that degrades incomplete protein fragments which otherwise might have inappropriate cellular activities.

The current understanding of SsrA function is based largely on the tRNA‐like properties of SsrA RNA and on the structures of tagged protein products. It is not known whether additional cellular factors are required for SsrA function or how the diverse phenotypes of SsrA‐defective cells are related to the activities of SsrA. In Salmonella typhimurium, insertion mutations in smpB (small protein B), a gene immediately upstream of ssrA, prevent plating of bacteriophage P22 but not c1 mutants of P22 (M.M.Susskind and K.Hughes, unpublished data). This phenotype is the same as that originally reported by Retallack et al. (1994) for plating of λimmP22 hybrid phage on SsrA‐defective E.coli. Disruption of the smpB gene was previously reported to decrease the virulence of S.typhimurium in mice and to reduce bacterial survival within macrophages (Bäumler et al., 1994). Here, we demonstrate that deletion of the smpB gene in E.coli prevents SsrA‐mediated peptide tagging and mimics other phenotypes of SsrA‐deficient strains. We also show that SmpB is an RNA‐binding protein that binds specifically to SsrA and that SmpB is required for stable association of SsrA with ribosomes in the cell.

SsrA RNA is required for phage λimmP22 development and has been proposed to prevent phage growth by competing with DNA sites for binding of transcription factors (Retallack and Friedman, 1995). We find, however, that SsrA RNA is processed and present at normal levels in the smpB mutant and yet this strain still fails to plate λimmP22. This result indicates that SsrA cannot affect phage development by a simple competition mechanism.


SmpB mutants have the same phenotypes as SsrA mutants

To address the possible functions of the SmpB protein in SsrA‐mediated protein tagging, we constructed an SmpB‐defective strain by deleting >80% of the smpB gene in the W3110 strain of E.coli (Figure 1A). This in‐frame deletion (ΔsmpB‐1) does not alter the normal initiation and termination signals for smpB transcription and translation. The parental strain (W3110) and the otherwise isogenic ΔsmpB‐1 and ssrA::kan strains displayed no differences in growth rate or colony morphology at temperatures ranging from 16 to 42°C, on rich or minimal media. At temperatures >42°C, the ΔsmpB‐1 and ssrA::kan strains had slightly increased doubling times (∼10%) in comparison with the parental strain. Temperature‐sensitive growth of W3110 ssrA::kan has been reported previously (Komine et al., 1994).

Figure 1.

(A) The smpB and ssrA genes are arranged tandemly in E.coli and S.typhimurium, although these genes are not adjacent in many other bacterial genomes. ΔsmpB‐1 is an in‐frame deletion of the E.coli SmpB coding sequence. (B) Northern blot probed with 32P‐labeled DNA oligonucleotides complementary to SsrA RNA (top) or 16S ribosomal RNA (bottom) in strains W3110, W3110 ΔsmpB‐1, W3110 ssrA::cat and W3110 ΔsmpB‐1/pSsrA.

We first tested whether SmpB is required for the synthesis of SsrA RNA. Northern blot hybridization experiments (Figure 1B) show that SsrA RNA is present at wild‐type levels in ΔsmpB‐1 cells and is overproduced in ΔsmpB‐1 cells transformed with a plasmid bearing the ssrA gene. Moreover, in this and other experiments using gels capable of ±1 nucleotide resolution, SsrA RNA from ΔsmpB‐1 cells had the same electrophoretic mobility as SsrA from wild‐type cells, indicating that SmpB is not required for the 3′ and 5′ processing events that convert precursor SsrA RNA into its mature, active form (Subbarao and Apirion, 1989; Komine et al., 1994; Li et al., 1998).

Some of the most distinctive phenotypes of SsrA‐defective strains involve phage development. For example, certain λimmP22 hybrid phage fail to plate efficiently on E.coli strains lacking functional SsrA (Retallack et al., 1994) and, as we now show, SmpB‐defective strains are similarly compromised. The plating efficiency of λimmP22 dis c2‐5 on ΔsmpB‐1 cells was reduced by four orders of magnitude to a level comparable with that observed with ssrA::kan cells (Figure 2). This defect in λimmP22 plating on ΔsmpB‐1 cells could be complemented by a low copy‐number plasmid expressing the smpB gene but not by a plasmid that overexpresses SsrA RNA (Figure 2). Thus, λimmP22 fails to plate efficiently on the ΔsmpB‐1 strain because these cells lack the SmpB protein.

Figure 2.

Plating efficiency of phage λimmP22 dis c2‐5 on wild‐type (W3110), SmpB‐defective (W3310 ΔsmpB‐1) and SsrA‐defective (W3310 ssrA::kan) strains with or without plasmids expressing SmpB (pSmpB) or SsrA (pSsrA).

Phage Mu carrying a temperature‐sensitive variant of Mu c repressor (c‐ts) cannot be induced to grow lytically in lysogenic strains lacking SsrA RNA (T.Baker, personal communication). Figure 3 shows that a Mu c‐ts lysogen of the wild‐type strain lysed ∼1 h after a temperature shift to 42°C, whereas c‐ts lysogens of the ΔsmpB‐1 and ssrA::kan strains failed to lyse and bacterial growth continued at normal rates. This defect in phage Mu induction in ΔsmpB‐1 and ssrA::kan strains was complemented by low copy‐number plasmids expressing SmpB and SsrA, respectively. Thus, the ΔsmpB‐1 and ssrA::kan strains both fail to support induction of the lytic development of Mu c‐ts lysogens.

Figure 3.

Time course of induction of phage Mu c‐ts62 pAp1 in lysogens of wild‐type (W3110), SmpB‐defective (W3310 ΔsmpB‐1) and SsrA‐defective (W3310 ssrA::kan) strains following a 30–42°C temperature shift to inactivate the Mu c‐ts62 repressor. The decrease in absorbance at 600 nm is caused by phage‐induced cell lysis.

SmpB is required for SsrA‐mediated peptide tagging

SsrA RNA is essential for the C‐terminal addition of degradation tags to polypeptides synthesized from mRNAs lacking a stop codon (Keiler et al., 1996). To investigate whether SmpB protein is required for this process, we used a gene (λ‐M2‐H6‐trpAt) that encodes the N‐terminal domain of λ repressor with M2 epitope and His6 sequences and a trpAt transcriptional terminator (Keiler et al., 1996). The mRNA transcribed from this gene has no in‐frame stop codons, and, in wild‐type cells, the λ‐M2‐H6‐trpAt protein is tagged and degraded. Western blots showed that the λ‐M2‐H6‐trpAt protein was present at equivalent levels in both ΔsmpB‐1 and ssrA::kan cells (Figure 4A, lanes 1 and 2) but did not accumulate in wild‐type cells or in a ΔsmpB‐1 strain in which SmpB was provided from a plasmid (Figure 4, lanes 3 and 4). These differences in steady‐state protein levels are expected if the λ‐M2‐H6‐trpAt protein is tagged and degraded in cells containing both SmpB and SsrA but is stable in cells lacking either SmpB or SsrA because the ssrA‐encoded degradation tag is not added.

Figure 4.

λ‐M2‐H6‐trpAt protein, expressed from an mRNA with no stop codon, accumulated in cells lacking SmpB or SsrA (lanes 1 and 2) but not in cells containing functional SmpB and SsrA (lanes 3 and 4). λ‐M2‐H6‐trpAt protein was detected by immunoblotting with anti‐FLAG M2 antibodies or anti‐DD‐tag antibodies. Comparison of the upper and lower panels in lanes 5–7 indicates that the accumulation of tagged λ‐M2‐H6‐trpAt protein in cells expressing SsrADD required SmpB. Plasmid pPW500 encodes λ‐M2‐H6‐trpAt; SsrADD encodes the protease‐resistant AANDENYALDD tag.

To monitor directly for SsrA‐mediated tagging, we characterized the tagging of the λ‐M2‐H6‐trpAt protein in a strain containing the SsrADD RNA variant, which mediates the addition of a peptide tag (AANDENYALDD) that does not result in degradation (E.Roche and R.Sauer, unpublished). Proteins modified with this mutant tag can be detected using antibodies to the DD‐peptide tag. Each of the strains in lanes 5–7 of Figure 4 contains the λ‐M2‐H6‐trpAt gene and SsrADD RNA but only the strain lacking SmpB failed to accumulate the DD‐tagged λ‐M2‐H6‐trpAt protein (Figure 4, lane 5, lower panel). Untagged λ‐M2‐H6‐trpAt protein was, however, expressed in this smpB strain (Figure 4, lane 5, upper panel) indicating that the defect is in SsrADD‐mediated tagging of λ‐M2‐H6‐trpAt rather than in its synthesis. Overall, these results provide strong evidence that the SmpB protein is required for efficient SsrA‐mediated tagging of proteins synthesized from mRNAs that lack in‐frame termination codons.

The λ‐M2‐H6‐trpAt proteins were purified using Ni–NTA chromatography and analyzed by mass spectrometry (data not shown). The proteins from strains lacking functional SmpB or SsrA had masses within 1 Da of that predicted for untagged λ‐M2‐H6‐trpAt protein lacking the N‐terminal formyl methionine (residues 2–117). The proteins purified from strains containing both SmpB and SsrADD had masses expected for residues 2–117 of λ‐M2‐H6‐trpAt plus the AANDENYALDD tag. In strains lacking SmpB, no addition of alanine or any part of the SsrA‐degradation tag to the λ‐M2‐H6‐trpAt protein was detected, although the possibility that a small amount of protein is tagged and then degraded cannot be absolutely excluded. We note, however, that similar quantities of untagged λ‐M2‐H6‐trpAt protein were expressed in the ΔsmpB‐1 and ssrA::kan strains (Figure 4, lanes 1 and 2). Hence, if tagging does occur in the ΔsmpB‐1 strain, then it must be inefficient.

SmpB protein is required for SsrA RNA association with ribosomes

In lysates prepared from wild‐type cells, SsrA RNA co‐fractionates with 70S ribosomes (Komine et al., 1996; Tadaki et al., 1996). To test whether SmpB is required for stable association of SsrA with ribosomes, we fractionated cell lysates prepared from wild‐type and ΔsmpB‐1 cells and monitored the localization of SsrA RNA by Northern blot hybridization. As expected from previous work (Komine et al., 1996; Tadaki et al., 1996), SsrA RNA in wild‐type cell lysates co‐sedimented with 70S ribosomes near the bottom of the final sucrose gradient (Figure 5). By contrast, in ΔsmpB‐1 cells SsrA was found in the top few fractions of the sucrose gradient and was not associated with ribosomes. Furthermore, a greater fraction of the total SsrA RNA was recovered in the S100 fraction in ΔsmpB‐1 cells than in wild‐type cells. Taken together, these results suggest that SmpB protein directly or indirectly facilitates the stable association of SsrA RNA with 70S ribosomes.

Figure 5.

SsrA RNA co‐sediments with 70S ribosomes in lysates of wild‐type but not ΔsmpB‐1 cells. (A) W3110 and W3110 ΔsmpB‐1 cell lysates were fractionated using the centrifugation and washing steps shown (Komine et al., 1996). (B) The presence of SsrA RNA in each fraction was assayed by Northern blot hybridization using 32P‐labeled oligonucleotides complementary to SsrA RNA. The lower band appears to be an SsrA degradation product. The position of 70S ribosomes in the gradient was determined by probing with oligonucleotides complementary to 16S RNA.

SmpB protein binds SsrA RNA

To test directly whether SmpB protein interacts with SsrA RNA, the protein was overexpressed, purified and characterized in vitro. Sequential Edman degradation of the purified protein gave the N‐terminal sequence TKKKAHK as expected for bona fide SmpB after removal of the initiator formyl methionine. Purified SmpB behaved as a stably folded protein, displaying cooperative denaturation and a far‐UV circular dichroism (CD) spectrum indicative of a predominantly β‐sheet protein (data not shown).

The interaction of SmpB protein with SsrA RNA was investigated using a gel‐mobility shift assay. As shown in Figure 6A, addition of increasing concentrations of SmpB resulted in formation of an SmpB–SsrA RNA complex. The observed binding is saturable with half‐maximal binding at a free SmpB concentration of ∼20 nM. To probe the specificity of SmpB–SsrA binding, unlabeled SsrA RNA and total yeast tRNA were used to compete for binding of 200 nM SmpB to 100 pM 32P‐labeled SsrA (Figure 6C). Approximately 400‐fold higher molar concentrations of yeast tRNA than SsrA RNA were required to achieve the same degree of competition. These results demonstrate that SmpB is an RNA‐binding protein with a high degree of affinity and specificity for SsrA RNA.

Figure 6.

(A) Gel‐mobility shift assay of the binding of purified SmpB to 100 pM of 32P‐labeled SsrA RNA produced by transcription in vitro. (B) Quantified binding data from (A). (C) Competition for binding of SmpB to 100 pM 32P‐labeled SsrA by unlabeled SsrA RNA or total yeast tRNA.

SmpB does not affect charging of SsrA

SsrA can be charged with alanine by alanyl‐tRNA synthetase in vitro (Komine et al., 1994; Ushida et al., 1994). As shown in Figure 7, SsrA RNA produced by transcription in vitro was charged with alanine by purified alanyl‐tRNA synthetase at the same rate in the presence and absence of 2 μM SmpB, a concentration that results in complete binding in the gel‐mobility shift assay. Thus, SmpB protein is neither required for nor significantly affects SsrA charging. Furthermore, SsrA RNA purified from wild‐type and ΔsmpB‐1 cells could be charged with 3H‐alanine by alanyl‐tRNA synthetase with similar kinetics (data not shown), providing additional evidence that SsrA RNA is properly processed in SmpB‐defective cells.

Figure 7.

Kinetics of charging of SsrA RNA (18 pmol) produced by transcription in vitro with [3H]alanine by purified alanyl‐tRNA synthetase in the presence and absence of SmpB.


The E.coli SmpB protein contains 160 amino acids and is reasonably basic (36 Lys, His and Arg residues; 18 Asp and Glu residues). Homology searches using PSI‐BLAST (Altschul et al., 1997) revealed related sequences with pairwise identities from 32 to 78% in all 13 bacterial species with known genome sequences. Interestingly, Rickettsia prowazeki contains an SmpB homolog but no identified SsrA homolog (Williams, 1999), suggesting either that SmpB may also have SsrA‐independent roles or that SsrA orthologs are very difficult to identify in some species. Sequence searches failed to detect significant homologies between the SmpB family and any other bacterial proteins or with eukaryotic or archaeal proteins. Thus, SmpB appears to play a sufficiently important biological role to be universally present in bacteria and may constitute a unique protein family.

Escherichia coli strains lacking the SmpB protein have the same phenotypes as strains lacking SsrA RNA, suggesting that SmpB is essential for SsrA function. Both SmpB‐defective and SsrA‐defective strains fail to add any part of the AANDENYALAA degradation tag to proteins synthesized from damaged mRNA, and do not support growth of λimmP22 hybrid phage or induction of Mu c‐ts phage. SmpB‐defective strains contain normal levels of SsrA RNA and are not complemented by SsrA‐overproducing plasmids, indicating that the absence of SsrA‐mediated function in these strains is not caused by a deficiency of SsrA RNA. Finally, SmpB is not required for the processing of SsrA RNA to its mature and active form.

SmpB protein binds to SsrA RNA in vitro with an affinity (∼20 nM) that would ensure substantial binding at the concentrations of SsrA that have been measured in vivo (∼10 μM; Altuvia et al., 1997). Because our binding studies were performed with SsrA RNA transcribed in vitro, the base modifications that occur in vivo (Felden et al., 1998) cannot be essential for formation of the SmpB–SsrA complex. Retallack and Friedman (1995) have reported that SsrA RNA binds to a number of bacterial repressor and activator proteins in low‐salt buffer (50 mM KCl), raising the question of whether the SmpB–SsrA interaction that we observe is specific. In competition experiments, however, SmpB binds to SsrA RNA ∼400‐fold more tightly than to bulk tRNA, indicating that this interaction is specific with respect to the identity of the bound RNA. In addition, the binding assays shown in Figure 6 were performed in buffer containing 200 mM KCl, an ionic strength comparable with or higher than physiological. Finally, the fact that SmpB‐defective cells are also defective in SsrA function makes it exceedingly likely that the SmpB–SsrA interaction observed in vitro is biologically significant.

The ability of SsrA to function as both a tRNA and an mRNA has been proposed to require charging with alanine, recognition of the A site of ribosomes in which mRNA translation has stalled, participation in the transpeptidation reaction that adds the charged alanine to the growing nascent chain, movement to the P site and finally the ability to release the stalled mRNA and to serve as a message that directs translation of the last 10 residues of the degradation tag (Keiler et al., 1996). Regardless of the model, the results presented here indicate that SmpB is not required for charging of SsrA but is required to permit SsrA RNA to associate stably with 70S ribosomes. Moreover, because no tagging or addition of alanine to the λ‐M2‐His6‐trpAt protein was detected in cells lacking SmpB, it is possible that SmpB is required for some tRNA‐like function of SsrA following charging but prior to the transpeptidation reaction.

A ternary complex of elongation factor Tu (EFTu), GTP and aminoacylated tRNA protects the labile ester linkage between normal tRNAs and their charged amino acids and mediates binding to the A site of the ribosome (for review, see Sprinzl, 1994). Although SmpB is not homologous to EFTu, it might serve one or both of these EFTu‐like functions for SsrA. However, EFTu–GTP binding to tRNA mainly involves sugar–phosphate backbone contacts in the acceptor and TΨC stems (Nissen et al., 1995) and because SsrA appears to contain these tRNA‐like elements of secondary structure (Komine et al., 1994; Felden et al., 1996; Williams and Bartel, 1996), it is plausible that EFTu itself could serve these functions for SsrA. Another possible role for SmpB would be to serve as a protein mimic of the anticodon helix and loop of a normal tRNA, in a manner analogous to domain 4 of elongation factor G (Ævarsson et al., 1994; Czworkowski et al., 1994; Nissen et al., 1995). SsrA is not believed to have an anticodon helix or to require codon–anticodon interactions for ribosome binding. If SmpB acts as an anticodon–helix mimic, it might provide additional contacts required for high‐affinity ribosome binding, regulate the GTPase activity of a ribosome‐bound EFTu–GTP–SsrA complex or mediate interactions with ribosomal elements required for translocation.

The regulatory activities of proteins such as P22 C1, Lac repressor, LexA repressor and λ repressor are increased in SsrA‐defective cells (Retallack et al., 1993, 1994; Retallack and Friedman, 1995). This observation explains many of the phenotypes of SsrA‐defective cells. For example, hyperactivity of the Mu c‐ts repressor in SsrA‐defective cells could explain why phage Mu induction is not observed in SsrA‐defective cells. Enhanced activity of the P22 C1 activator could lead to excess transcription from the paQ promoter in λimmP22 (Hoopes and McClure, 1985; Ho and Rosenberg, 1985; Stephenson, 1985; Ho et al., 1992), resulting in interference with lytic phage development and explaining the inability of this phage to plate on SsrA‐defective strains. λimmP22 phage that are c1 or that have an altered‐specificity C1, which would not be expected to activate paQ, do plate on SsrA‐defective strains (Strauch et al., 1986; Retallack et al., 1993, 1994). Purified SsrA RNA binds P22 C1, Lac repressor, LexA repressor and λ repressor in vitro, and it has been proposed that SsrA RNA also binds these regulatory proteins in vivo, reducing their free concentrations and thereby diminishing their biological activities (Retallack and Friedman, 1995). Our results, however, cast doubt on this binding model because wild‐type SsrA RNA is present at normal levels in ΔsmpB‐1 strains and at higher than normal levels in ΔsmpB‐1 pSsrA strains, and yet both of these strains have the same phage development phenotypes as SsrA‐defective strains. It is not known how the SmpB–SsrA system does function to increase the activities of proteins like P22 C1 and Mu c‐ts repressor.

A number of questions remain to be answered. Are macromolecules in addition to SmpB required for SsrA function? How does SmpB binding support SsrA‐mediated peptide tagging? Does SmpB have functions independent of SsrA? The identification of SmpB as an essential participant in SsrA function represents an important step forward that should help to guide future studies of detailed mechanism and biological function in the SmpB–SsrA peptide‐tagging system.

Materials and methods

Strains, plasmids and phage

Escherichia coli strains were grown in Luria–Bertani (LB) medium (10 g tryptone, 5 g yeast extract and 5 g NaCl/l) or on LB–agar plates containing 1–1.5% agar/l. Ampicillin (100 μg/ml) and/or tetracycline (30 μg/ml) were added to plates or media as needed. Plasmid pKW1 and its derivatives, pKW11 (pSsrA) and pKW23 (pSsrADD), are tetracycline‐resistant plasmids with pACYC replicons (K.Williams and E.Roche, unpublished). Plasmid pPW500 (ampicillin resistance; colE1 replicon) directs transcription of an mRNA with no in‐frame stop codons that encodes the N‐terminal domain of λ repressor with an M2 FLAG epitope and His6 tag (Keiler et al., 1996). To generate plasmid pSmpB, the promoter and coding sequences of the smpB gene were amplified by PCR to add NcoI and EcoRI sites at the 5′ and 3′ ends, respectively. This DNA was digested with NcoI and EcoRI and cloned into NcoI–EcoRI‐digested pKW1. The method described by Link et al. (1997) was used to generate a precise deletion of the smpB gene, first in a derivative of plasmid pKO3 and then in the chromosome of E.coli strain W3110. The structure of the chromosomal deletion (ΔsmpB‐1) was verified by PCR sequencing. Strain W3110 ssrA::kan was a gift of Hachiro Inokuchi; in this strain, a portion of the promoter and approximately two‐thirds of the ssrA gene have been replaced with a kanamycin‐resistance gene (Komine et al., 1994). Strain W3110 ssrA::cat was constructed by phage P1 transduction from strain X90 ssrA::cat (Keiler et al., 1996). The plating efficiency of phage λimmP22 dis c2‐5 (a gift from David Friedman) on the wild‐type strain W3110 and its otherwise isogenic derivatives W3110 ΔsmpB‐1 and W3110 ssrA::kan was determined as described previously (Retallack et al., 1994). Phage Mu c‐ts62 pAp1 was a gift of Tania Baker; Mu c‐ts62 pAp1 lysogens of strain W3110 were selected by ampicillin resistance.

Overexpression and purification of SmpB protein

The coding sequence of the smpB gene of E.coli was amplified by PCR to add a NdeI site at the 5′ end and a BamHI site at the 3′ end. This DNA fragment was cloned into the NdeI–BamHI digested pET9a expression vector (Novagen) to generate the SmpB‐overexpressing plasmid pSmpB9a. High levels of SmpB protein were expressed from pSmpB9a but were largely present in inclusion bodies. As a result, we purified SmpB under denaturing conditions and refolded the purified protein.

Escherichia coli strain BL21(DE3)/pLysS/pSmpB9a was grown in 6 l of LB broth at 37°C to an A600 of 0.6, induced with 1 mM IPTG and grown for an additional 3 h. The cells were harvested by centrifugation, resuspended in 150 ml of buffer A [50 mM HEPES pH 7.5, 5 mM dithiothreitol (DTT), 1 mM EDTA and 8 M urea] and lysed by sonication. The cell lysate was centrifuged at 15 000 r.p.m. in a Sorvall SS‐34 rotor to remove cellular debris, and the supernatant was loaded on a Source™ 15S column (Pharmacia Biotech). The column was washed with buffer A and eluted with a linear gradient from 0 to 600 mM KCl in buffer A. SmpB protein eluted from the column between 210 and 250 mM KCl. Fractions containing SmpB protein were combined and loaded directly onto a C4 reverse‐phase HPLC column (Vydak™) equilibrated in 30% acetonitrile and 0.1% trifluoroacetic acid. A linear gradient from 36 to 46% acetonitrile in 0.1% trifluoroacetic acid was applied and SmpB‐containing fractions were combined and lyophilized. The protein was resuspended in buffer B (50 mM potassium phosphate pH 6.0, 200 mM KCl and 5 mM β‐mercaptoethanol) containing 3 M guanidinium hydrochloride and dialyzed extensively against buffer B without denaturant. This method produces SmpB protein that is >95% pure with a yield of 1–2 mg/l of cell culture. Purified SmpB is most soluble below pH 7, in the presence of reducing agents, and with NaCl or KCl at concentrations of 100 mM or more.

The N‐terminal sequence of SmpB was determined by sequential Edman degradation by the MIT Biopolymers Facility. The CD spectrum of 10 μM SmpB was determined at 25°C using an AVIV 60DS spectrapolarimeter in buffer containing 50 mM potassium phosphate pH 6.0, 250 mM KCl, 5% glycerol and 1 mM DTT. Denaturation of SmpB protein by GuHCl was monitored by changes in CD ellipticity at 228 nm under the same temperature and buffer conditions but in the presence of increasing concentrations of denaturant.

Northern and Western blots

Cellular RNA was purified from log‐phase cultures using the RNAeasy Total RNA Purification Kit (Qiagen), electrophoresed on 1.5% agarose and 20% formaldehyde gels, and transferred in 20× SSC to Nytron membranes (Schleicher and Schuell) using a Turboblot apparatus. Prehybridization and hybridization with appropriate 32P‐labeled DNA oligonucleotides complementary to SsrA RNA (5′‐TAA AGC GTA GTT TTC GTC GTT TGC‐3′) or 16S ribosomal RNA (5′‐CCG TCC GCC ACT CGT CAG CAA‐3′) were carried out according to Sambrook et al. (1989). To assay for correct processing of SsrA RNA, cellular RNA was electrophoresed on 4% polyacrylamide TBE gels in 8 M urea and detected by Northern blotting using the antisense SsrA probe (Li et al., 1998).

For Western blot analysis of λ‐M2‐H6‐trpAt protein expression, cells were grown to mid‐log phase in LB broth, induced with 1 mM IPTG, grown for varying times and lysed in SDS sample buffer. Total cellular proteins were electrophoresed on 15% acrylamide Tris–tricine gels and transferred to PVDF membranes. These blots were probed either with anti‐FLAG M2 antibodies (Eastman Kodak) or with antibodies that recognize the mutant AANDENYALDD tag (a gift of Eric Roche).

λ‐M2‐H6‐trpAt purification and characterization

Cells were lysed in 50 mM Tris–HCl pH 8, 10 mM imidazole, 6 M GuHCl and λ‐M2‐H6‐trpAt protein was purified by Ni‐NTA chromatography (Qiagen) and dialyzed into 50 mM Tris–HCl pH 8, 50 mM KCl. Matrix‐assisted laser desorption ionization time‐of‐flight (MALDI‐TOF) mass spectrometry was performed using a Voyager‐DE STR Biospectrometry Workstation (PerSeptive Biosystems).

Gel‐mobility shift assay

A form of SsrA corresponding to the mature, processed RNA was transcribed in vitro using T7 polymerase in the presence of [α‐32P]UTP. Binding reaction mixtures (20 μl) contained 50 mM MES pH 6.5, 200 mM KCl, 5% glycerol, 5 mM β‐mercaptoethanol, 0.01% NP‐40, 0.1 mg/ml bovine serum albumin (BSA), labeled SsrA RNA (100 pM) and different amounts of SmpB protein. Mixtures were incubated at room temperature for 30 min and loaded onto a 5% acrylamide gel in TBE buffer (50 mM Tris pH 8.3, 25 mM borate and 2 mM EDTA). The gel was electrophoresed at 200 V, dried on gel‐drying paper and exposed to BIOMAX MR film (Eastman Kodak) for 8–12 h at −80°C. Competition experiments were performed in the same manner except for inclusion of unlabeled SsrA or yeast tRNA (Sigma) in the reaction mixtures.

Fractionation and charging

Cell fractionation and ribosome association experiments were performed by the method of Komine et al. (1996). Mature SsrA RNA produced by transcription in vitro was charged with [3H]alanine as described (Komine et al., 1994) using purified alanyl‐tRNA synthetase (Ribas de Pouplana and Schimmel, 1997). His6‐tagged alanyl‐tRNA synthetase was purified by Ni‐NTA chromatography from E.coli strain TG1/pQE‐875/pREP4 (a gift of Lluis Ribas de Pouplana and Paul Schimmel). SsrA RNA from wild‐type and ΔsmpB cells was purified as described (Komine et al., 1994).


We thank Dereth Philips, George Church, Kelly Williams, David Friedman, Hachiro Inokuchi, Lluis Ribas de Pouplana and Paul Schimmel for providing materials, and Eric Roche, Ramon Tabtiang, Kelly Williams, David Bartel, Kelly Hughes and Tania Baker for helpful discussions. We are especially grateful to Eric Roche who developed and allowed us to use the SsrADD system for detecting tagging. Supported in part by an NIH NRSA postdoctoral grant to A.W.K. and by NIH grant AI‐16892.


View Abstract