A salvage pathway for protein synthesis: tmRNA and trans-translation.

* Abstract Transfer-messenger RNA (tmRNA, or SsrA), found in all eubacteria, has both transfer and messenger RNA activity. Relieving ribosome stalling by a process called trans-translation, tmRN[A.sup.a1a] enters the ribosome and adds its aminoacylated alanine to the nascent polypeptide. The original mRNA is released and tmRNA becomes the template for translation of a 10-amino-acid tag that signals for proteolytic degradation. Although essential in a few bacterial species, tmRNA is nonessential in Escherichia coli and many other bacteria. Proteins known to be associated with tmRNA include SmpB, ribosomal protein S1, RNase R, and phosphoribosyl pyrophosphate. SmpB, having no other known function, is essential for tmRNA activity, trans-translation operates within ribosomes stalled both at the end of truncated mRNAs and at rare codons and some natural termination sites. Both the release of stalled ribosomes and the subsequent degradation of tagged proteins are important consequences of trans-translation.

Key Words SsrA, 10Sa RNA, translation, ribosomes, RNA structure

CONTENTS

INTRODUCTION ESTABLISHMENT OF THE TRANS-TRANSLATION MODEL A MORE INCLUSIVE ROLE FOR TRANS-TRANSLATION THE BIOLOGICAL ROLES OF TRANS-TRANSLATION FACTORS REQUIRED FOR tmRNA ACTION STRUCTURE AND FUNCTION OF tmRNA FORMATION OF MATURE tmRNA EXPLOITING THE TAG SEQUENCE UNANSWERED QUESTIONS

INTRODUCTION

When originally proposed, the operon model visualized the repressor as an RNA molecule (40). However, that idea fell out of favor when it became clear first from genetic studies and then later from biochemical studies that the lac and [lambda] repressors are proteins. Beyond its various roles in translation (e.g., mRNA, tRNA, and components of ribosomes) and a few notable exceptions, the role of RNA as a contributor to cellular activities was viewed for many years as extremely limited. However, over the past several years there has been what can best be called a renaissance in the study of the contribution of RNAs to many facets of cellular physiology. Studies in both prokaryotic and eukaryotic cells have identified a number of small RNAs that play a variety of roles in maintaining cellular homeostasis (27, 93, 98). The activities of these small RNAs vary from regulating gene expression at the levels of either transcription or translation to blocking viral growth. Many of these activities rely on the ability of the RNA to hybridize to other RNAs. In fact, the story has come full circle since RNAs have been reported to be the functional repressors of some phages (57).

One of the most unusual of the small RNAs is tmRNA (transfer-messenger RNA). Also known as 10Sa RNA and SsrA, this RNA is unique in having both tRNA and mRNA activities. Originally identified in Escherichia coli by David Apirion as a small RNA of unknown function (12, 52, 71), the mechanism of action of tmRNA was determined by Robert Sauer and coworkers (46) and is discussed in detail below.

Although tmRNA is apparently universally conserved among Eubacteria (48, 101), E. coli mutants that do not express tmRNA are viable (70). A number of groups have studied how the absence of tmRNA affects the physiology of E. coli and other bacteria (as discussed below). One notable phenotypic characteristic of E. coli mutants lacking tmRNA activity, discovered early on, is the failure of these mutants to support growth of hybrid phages constructed between coliphage [lambda] and Salmonella phage P22, [lambda]immP22. This phenotype, Sip (selective inhibition of P22), was first described for a mutation called sipB391 (89) that only later was found to be an allele of ssrA. Kirby et al. (47) showed that sipB391 resulted from the excision of the cryptic phage CP4-57, whose attachment site is at the 3' end of the ssrA gene (77). The Sip phenotype holds true for E. coli with complete knockouts of ssrA, highlighting the utility of [lambda]immP22 for determining the SsrA phenotype in E. coli. Phage P22 itself can be similarly employed in studying the SsrA phenotype in Salmonella (44). Adding to the utility of the phage assay for tmRNA activity is the observation that growth of variants of [lambda]immP22 with mutations in the C1 gene is supported by E. coli lacking tmRNA activity (78, 89). Thus, the pair of [lambda]immP22 phages having either wild-type or mutated C1 provides a simple and specific test to determine whether an E. coli is expressing active tmRNA.

With this brief introduction, we begin the review with the hope that we will enlighten the reader on the nature and mechanism of action of tmRNA. We point out, however, that there have been a number of reviews of tmRNA over the past few years (25, 42, 63, 106), and the reader is referred to those reviews for additional information on the subject. The reader is also directed to websites dedicated to tmRNA: http://www.indiana.edu/~tmrna/(101) and http://psyche.uthct.edu/dbs/tmRDB/rna/tmrna.html (48).

ESTABLISHMENT OF THE TRANS-TRANSLATION MODEL

tmRNA was first identified in 1978 as an RNA species having an electrophoretic mobility of 10S (52); however, the function of tmRNA remained unknown until 1996, when the trans-translation model for tmRNA was proposed by Keiler et al. (46). Several studies prior to this landmark paper paved the way for the work to follow by describing potential tRNA- and mRNA-like activities for tmRNA. The first indications that tmRNA may be involved in translation were the observations that the 5' and 3' ends of tmRNA sequences from E. coli, Mycoplasma capicolum, and Bacillus subtilis resembled those of alanyl-tRNAs (50, 96). Experiments using purified alanyl-tRNA synthetase demonstrated that tmRNA could be charged with alanine in vitro, albeit at a much lower level than the charging observed with alanyl-tRNA. Ushida et al. (96) also showed association of tmRNA with 70S ribosomes in B. subtilis but not with 50S or 30S subunits. This finding was later extended to include E. coli (49, 92), with the additional observation that charging with alanine is required for the interaction of tmRNA with the 70S ribosome (92).

Although tmRNA has a short open reading frame encoded within its sequence, the prevailing thought, until 1995, was that tmRNA was not translated. However, Tu et al. (94) found that when murine interleukin-6 (IL-6) is overexpressed in E. coli, a collection of truncated forms of IL-6 bearing the same C-terminal extension are produced along with the full length IL-6. The amino acid sequence of this C-terminal extension, AANDENYALAA, with the exception of the N-terminal alanine, coincides with the C-terminal portion of the tmRNA open reading frame. A strain of E. coli having a disruption in the ssrA gene did not produce C-terminal extensions when IL-6 was overexpressed, suggesting that the C-terminal extensions were translated from tmRNA.

The apparent contradiction raised by the observations that tmRNA has both tRNA-like and mRNA-like qualities was resolved with the publication of the peptide tagging model for tmRNA function by Keller et al. (46), now known as the trans-translation model (5). The authors recognized that the C-terminal amino acid sequence of the peptide encoded by tmRNA resembles the amino acid sequence recognized by a periplasmic protease, Tsp, and therefore might be involved in targeting proteins for proteolysis. In confirmation of this hypothesis, the protein expressed from a cloned fragment in which DNA encoding the N-terminal domain of the [lambda] repressor was fused to DNA encoding the tmRNA peptide identified by Tu et al. (94) exhibited a greatly reduced half life; similar experiments using the periplasmic cytochrome [b.sub.562] protein fused to the tmRNA peptide yielded a similar result. The final, key experiments described by Keller et al. used a construct from which an mRNA that lacked an in-frame translation termination codon would be transcribed. Thus, it was predicted that the translating ribosomes would reach the 3' end of the mRNA, become stalled, and render the nascent peptide a target for tmRNA tagging. Confirming this prediction, the authors found tmRNA-dependent addition of the 11-amino-acid sequence identified by Tu et al. to the C termini of proteins translated from this mRNA lacking a stop codon, and again observed a greatly reduced half life for proteins having the tmRNA-encoded peptide sequence.

The trans-translation model may be broken into four steps (Figure 1). First, a translating ribosome becomes stalled, and alanyl-tmRNA enters this stalled ribosome much like a tRNA, in complex with EF-Tu and guanosine triphosphate (GTP) (85). Second, the alanine on tmRNA is transferred to the nascent peptide, followed by translocation of tmRNA from the ribosomal A site to the P site. Third, the original mRNA is released from the ribosome, and tmRNA becomes the template for translation. The 10-amino-acid tag sequence encoded by tmRNA is translated, followed by termination at a stop codon. Fourth, the translational ternary complex dissociates, releasing the tagged peptide. The tagged peptide is then degraded by proteolysis.

[FIGURE 1 OMITTED]

Work from several laboratories supports the basic mechanism of the transtranslation model. As mentioned above, tmRNA associates only with 70S ribosomes, and not with 30S or 50S subunits or polysomes, consistent with a role for tmRNA in active translation and with dissociation of the ribosome containing tmRNA from the original mRNA (49, 92, 96). Further, charging of tmRNA is required for association with ribosomes, as mutations affecting the nucleotides in tmRNA required for recognition by the alanyl-tRNA synthetase abolish the association of tmRNA with ribosomes (92). Experiments using an in vitro $30 extract system showed that translation of polyuridine produced polyphenylalanine followed by the tmRNA-encoded tag when tmRNA was added to the reaction (36). Again, charging of tmRNA with alanine was required for addition of the tmRNA-encoded tag, and translation of the tmRNA-encoded tag was not observed in the absence of polyuridine. These results confirm that tmRNA is the translational template …