Amber (UAG) suppressors affected in UGA/UAA-specific polypeptide release factor 2 of bacteria: genetic prediction of initial binding to ribosome preceding stop codon recognition Kuniyasu Yoshimura, Koichi Ito and Yoshikazu Nakamura* Department of Tumor Biology, The Institute of Medical Science, The University of Tokyo, Minato-ku, Tokyo 108-8639, Japan
Abstract Background: Prokaryotic translational release factors, RF1 and RF2, catalyse protein release at UAG/UAA and UGA/UAA stop codons, respectively. Mutations in RF1 and RF2 are known to cause non-sense suppression for UAG (amber) and UGA (opal) codons, respectively, and they do not exert a reciprocal (‘cross’) suppression phenotype. We aimed to isolate RF mutants of such cross-suppression activity, which we designated ‘Csu’ phenotype in this paper. Results: Using a lacZ (UAG) reporter, we selected amber suppressor alleles occurring in the plasmidbearing RF2 gene of Salmonella typhimurium. Of nine such RF2 csu alleles, five were mis-sense mutations and four were non-sense mutations. The former mis-sense mutants retained the RF2 activity and catalysed UGA termination both in vivo and in vitro.
Introduction The translational suppression of non-sense codons is a most effective way of examining the structure, function, and interactions of any translational macromolecules, as long as, and to the extent that, such molecules are involved in the specificity or accuracy of translation. Most of the translational suppressors that have been studied over the last three decades or so have been tRNAs that acquired an anticodon cognate to the termination codon; however other translational macromolecules are now coming under scrutiny. Non-sense suppression can be Communicated by: Akira Ishihama *Correspondence: E-mail:
[email protected] q Blackwell Science Limited
RF2 C-terminal deletions equivalent to the nonsense alleles exerted amber suppression as well as opal suppression activity. Moreover, the equivalent RF1 segments also showed both the suppression phenotypes. Conclusions: All the csu mutations were mapped at the C-terminal half of RF2 and are strikingly coincident with the highly conservative amino acids, suggesting that they affect the conserved function of bacterial RFs. We propose here that there should be an ‘initial binding’ step of RFs to the ribosome, preceding stop codon recognition (‘initial binding’ hypothesis) and that the N-terminal RF domain(s), that are truncated or affected by the csu mutations, are responsible for this step and interfere with the proper functioning of cognate release factors on the ribosome.
caused by mutations that affect the ribosome (reviewed by Murgola 1995) on the translation factors (Hughes et al. 1987; reviewed by Hinnebusch & Liebman 1991). Irrespective of whether tRNA, ribosome or translation factors become a non-sense-suppressor, they must compete with the normal termination mechanism. Translation termination requires two classes of polypeptide release factors (RFs): a class-I factor, codon-specific RFs (RF1 and RF2 in prokaryotes; eRF1 in eukaryotes), and a class-II factor, nonspecific RFs (RF3 in prokaryotes; eRF3 in eukaryotes) that bind guanine nucleotides and stimulate class-I RF activity (Capecchi & Klein 1969; Caskey et al. 1969; Goldstein & Caskey 1970; reviewed by Tate & Brown 1992; Nakamura et al. 1996; Buckingham et al. 1997). Genes to Cells (1999) 4, 253–266
253
K Yoshimura et al.
Although the mechanism of stop codon recognition has long been unknown, in spite of its considerable interest due to protein-, instead of tRNA-, catalysed genetic decoding; we have shed light on this problem by proposing a ‘release factor tRNA mimicry’ hypothesis (Ito et al. 1996; reviewed by Nakamura & Ito 1998). The RF-tRNA mimicry hypothesis was borne out of the discovery of the conservation of seven domains, A through G, among class-I RFs, two of which, domains D and E, also share an homology with the C-terminal portion, domain IV, of elongation factor EF-G. A threedimensional structural study of Thermus thermophilus EF-G revealed that domains III–V appear to mimic the shapes of the acceptor stem, the anticodon helix and the T stem of tRNA, respectively (Ævarsson et al. 1994; Czworkowski et al. 1994; Nissen et al. 1995). Therefore, it appears that domains D/E of RF constitute a ‘tRNA-mimicry’domain that is necessary for RF binding to the ribosomal A site (Ito et al. 1996). The model of RF-tRNA mimicry predicts a peptide anticodon mimic in RF to read the stop codon. There are now several lines of evidence for this prediction, showing that domains C and D are involved in the anticodon activity (reviewed by Nakamura & Ito 1998). In contrast to the highly conservative sequences of domains C–F, N-terminal domains A and B of prokaryotic RFs are much less conserved (Ito et al. 1996; as shown later in Fig. 6B). These probably reflect the nature of these domain functions: the C-terminal conservation may be due to its common functions of tRNA mimicry and codon decoding. On the other hand, the less conservative feature of N-terminal domains might mean that these domains interact with translational components whose sequences or structures are diverse among different organisms. Nevertheless, in contrast to domains C–E, the other domain functions have been poorly investigated. Tate and colleagues have argued for two functional regions in RFs using three chimeric constructs of RF1 and RF2; one, a hydrolysis domain and the other, a domain for codon recognition and ribosome binding (Moffat et al. 1993). They have also described that the chymotrypsin cleavage of RF2 at position 244 (in domain E) stabilizes the codon-specific RF–ribosome interaction and abolishes peptidyl-tRNA hydrolysis activity (Moffat & Tate 1994), claiming that position 244 may directly or indirectly interact with the peptidyltransferase centre. This is not inconsistent with our findings that the substitution of Thr for Ala or Ser conserved at position 246 of bacterial RF2 s affects the polypeptide release activity (Uno et al. 1996). It has, however, not been examined which part of RF domain(s) is involved in the interplay of the peptidyltransferase centre or in the binding to the ribosome. 254
Genes to Cells (1999) 4, 253–266
Mutations in the bacterial genes for RF1 (prfA) and RF2 (prfB) are known to cause non-sense suppression for UAG (amber) and UGA (opal) codons, respectively (reviewed by Ryde´n & Isaksson 1984; Kawakami et al. 1988; Elliott & Wang 1991; Nakamura et al. 1995). Up to now, no prfA or prfB mutants that show reciprocal (‘cross’) suppression phenotype have been described, i.e. a UGA suppressor affected in RF1, or a UAG suppressor affected in RF2. These stringent correlations between the suppressible codon and the species of altered RFs are thought to represent the selectivity of the stop codon by bacterial RF1 and RF2. Unlike RF1 and RF2, mutations in the RF3 gene (prfC) cause a suppression of all three stop codons (Kawazu et al. 1995; Matsumura et al. 1996; Yanofsky et al. 1996; Crawford et al. 1999). RF3 is thought to accelerate the dissociation of RF1 and RF2 from the ribosome in a GTP-dependent manner (Freistroffer et al. 1997; Pavlov et al. 1997). Therefore, the omnipotent suppression activity of RF3 mutants reflects its codon-nonspecific feature of class-II RF. In spite of this ‘textbook’ view on the specificity of non-sense suppression by prfA and prfB mutations, we aimed to isolate a novel prfB mutant harbouring the amber (UAG) suppressor activity, which we designated ‘Csu’ mutants for ‘cross suppression’ phenotype. A csu mutant, once isolated, should provide a novel clue to understanding of how RFs recognize the stop codon and terminate translation. One likely scenario is to create a potential antagonist against RF1 that could interfere with UAG codon recognition. Even if the interference is direct or indirect, it should shed light on the nature of the RF anticodon moiety or the mechanism of the RF–ribosome interaction. In this paper, we argue that most of the csu alleles interfere with proper functioning on the ribosome of the cognate RF, suggesting that the first step in the sequence of interactions between the ribosome and RF is the codon-independent formation of an initial complex and that this step is facilitated by the N-terminal domain(s) of release factors RF1 and RF2. This argument is comparable with the recent findings of the initial binding step to the ribosome of the elongation factor EF-Tu X GTP X aminoacyl-tRNA complex preceding codon recognition (see below; Rodnina et al. 1996).
Results Rationale of selection of csu mutants In this study, we aimed to isolate an RF2 mutant that q Blackwell Science Limited
RF-ribosome initial binding hypothesis
exerts, unlike the previously described alleles, amber (UAG) suppressor activity. We referred to this allele as csu for ‘cross suppression’. The rationale of selection of csu mutants was based on the assumption that, given that RF1 and RF2 share essential motifs for the selective recognition of each stop codon, the codon specificity might be swapped or appended by mutations, and some of the mutants whose specificity changes are incomplete may interfere with the cognate RF activity and show the Csu phenotype. Alternatively, given that there is a codon-independent step common to RF1 and RF2 preceding stop codon recognition during the sequential interaction with the ribosome of RF, a mutation interfering with each step could give rise to a Csu phenotype. We could easily discriminate these two types, since the former phenotype may be codondependent, while the latter phenotype may be codonindependent. Irrespective of which type is borne, most of the csu mutants seem to be dominant, because the phenotype is based on the interference with cognate RF action. Taking these into consideration, we employed mutagenesis of the plasmid-bearing prfB gene for selection of a dominant amber suppressor. Because the Escherichia coli RF2 gene is toxic when cloned in a multicopy plasmid, we used a Salmonella typhimurium RF2 gene which is sufficient for E. coli growth and does not show a toxic phenotype (Uno et al. 1996; Ito et al. 1998).
Isolation of dominant csu alleles in plasmidbearing RF2 The Salmonella prfB gene (Kawakami & Nakamura 1990) was fused to an isopropyl-1-thio-b-D-galactoside (IPTG)-inducible lac promoter in plasmid pTWV229. The plasmid DNA, referred to as pTWV-RF2, was
mutagenized in vitro with hydroxylamine (allowing both C→T and G→A substitutions), and transformed into the lacZ (UAG) strain (DEV1). Ampicillin-resistant (ApR) Lacþ (red or pink) colonies were selected on lactose–MacConkey plates at 37 8C. Plasmid DNAs were recovered from these Lacþ colonies, retransformed into the same parental strain, and those that gave a reproducible phenotype (i.e. red or pink colony formation on MacConkey plates) were thought to harbour csu alleles. Nine such plasmids, csu-1 through csu-9, were further characterized. As shown in Fig. 1, lacZ (UAG) cells transformed with pTWV-RF2 derivatives possessing csu alleles formed red or pink colonies on MacConkey plates (samples 4–12), while those transformed with the parental plasmid, vector alone or that carrying RF1 and RF2 formed white colonies (samples 1–3). The colour intensity varied depending on the csu alleles. This Csu phenotype was reproducible upon recloning the mutant prfB DNAs into a constitutive pBR322 plasmid, showing that this phenotype is not specific to a lac promoter expression system used in the parental plasmid (data not shown). Evidence for in vivo read-through of UAG was also demonstrated using an assay based on a gene which codes for three identical engineered antibody binding B-domains of protein A from Staphylococcus aureus (Bjo¨rnsson & Isaksson 1993; Mottagui-Tabar et al. 1994). The sequence with the UAG codon was inserted into a linker between the segments coding for the second and third identical IgG binding domains (pAB96 for UAG read-through assay). The predicted influence of csu mutants on read-through of the UAG codon in growing bacteria was shown by the appearance of the three-domain (3A0 ) protein product, a readthrough product, in addition to the two-domain (2A0 ) protein, a terminated product. As shown in Table 1, the
Figure 1 Lacþ phenotype of DEV1 lacZ (UAG) strain suppressed by RF2 csu mutations. DEV1 transformants bearing pTWV-Csu plasmids were grown on lactose-MacConkey plates supplemented with ampicillin (50 mg/mL) and 2% lactose at 37 8C. Plasmids: sample 1, pTWV229; sample 2, pTWV-RF2; sample 3, pTWVRF1; sample 4, pTWV-Csu1; sample 5, pTWV-Csu2; sample 6, pTWV-Csu3; sample 7, pTWV-Csu4; sample 8, pTWV-Csu5; sample 9, pTWV-Csu6; sample 10, pTWV-Csu7; sample 11, pTWV-Csu8; sample 12, pTWV-Csu9. q Blackwell Science Limited
Genes to Cells (1999) 4, 253–266
255
K Yoshimura et al. Table 1 Amino acid substitutions and in vivo activity of the csu mutants RT value (x 10¹2) Mutation (pTWV229) (prfAþ) (prfBþ) csu-1, csu-3 csu-2 csu-4, csu-6 csu-5 csu-7 csu-8 csu-9
Amino acid position
322 332 252 325 203 262 264
Substitution
UAG
UGA
Complementation of prfB::CmR
Q(CAG)→UAG R(CGC)→C(UGC) Q(CAG)→UAG S(UCC)→F(UUC) R(CGU)→C(UGU) K(AAA)→I(AUA) T(ACC)→I(AUC)
3.7 6 0.4 0.4 6 0.1 3.6 6 0.1 4.6 6 0.3 6.7 6 0.9 6.1 6 0.5 10.2 6 0.8 6.6 6 0.9 6.5 6 0.6 6.3 6 0.7
10.6 6 0.9 10.6 6 2.2 2.2 6 1.1 11.7 6 1.2 13.5 6 0.8 10.5 6 0.1 3.3 6 1.1 5.9 6 1.0 4.6 6 1.5 1.5 6 1.1
¹ ¹ þ þ/¹ ¹ þ/¹ þ ¹ ¹ þ
Influence on UAG and UGA read-through of expression of RF2 csu products was monitored using the 3A0 reporter assay as described in Experimental procedures. Read-through (RT) values, i.e. molar amounts of 3A0 domain protein (translation readthrough) relative to 2A0 domain protein (translation termination), were measured. Experiments were performed independently at least four times, and the values are expressed with standard deviations. Complementation of the RF2-knockout (prfB::CmR) strain by csu alleles was examined by P1 transduction of prfB::CmR into W3110 harbouring pTWV-Csu plasmids: þ means appearance of CmR transductants at normal frequency (normal size colony); þ/¹ means appearance of CmR transductants at one-tenth frequency compared with þ (smaller size colony); ¹ means no appearance of CmR transductants (at the frequency less than 10¹2 compared with þ).
expression of csu RF2 derivatives stimulated UAG read-through 2–3-fold compared with wild-type RF2, consistent with the selection strategy using lacZ (UAG).
Residual RF2 activity of the csu derivatives in vivo Using the 3A0 reporter plasmid pAB101, which encodes UGA instead of UAG, the effects of the expression of csu products on UGA read-through was monitored (Table 1). The expression of wild-type RF2 reduced UGA read-through frequency fivefold compared with the vector control. Under these conditions, most of the csu mutants, except for csu-9 and probably csu-5, exhibited the reduced or small capacity of stimulating UGA termination. The data suggested that csu-1, csu-3, csu-4, and csu-6 products may not possess any residual RF2 activity, while csu-7 and csu-8 products may have some. The csu-2 allele seemed to be exceptional, since it enhanced UGA read-through slightly but significantly. This suggests that csu-2 acts distinctly differently from the others on UGA termination. One plausible scenario would be that it exerts the suppression of not only UAG but also UGA (see below). However, it cannot be excluded at present that these different UGA read-through levels are due, at least in 256
Genes to Cells (1999) 4, 253–266
part, to different cellular levels of csu products in E. coli. Because the 3A0 reporter assay suggested that the residual activity of RF2 remained in some of the csu products, we asked whether they can complement the prfB null allele of E. coli or not. The RF2-geneknockout (prfB::CmR) strain (RM786) is viable only when wild-type RF2 protein is provided from plasmids or phages (Ito et al. 1998). The P1 phage transduction experiments revealed that the prfB::CmR allele can be transduced at the normal frequency into E. coli cells (W3110) containing wild-type, csu-5 and csu-9 derivatives of pTWV-RF2, and less frequently but significantly into those containing csu-1, csu-3, csu-4 and csu-6 derivatives, and failed to be transduced at any detectable level into those containing the csu-2, csu-7 and csu-8 derivatives (Table 1). The viability of the transductants that were formed was IPTG-dependent (data not shown). It was noteworthy that the two mutants, csu-5 and csu-9, that were capable of perfectly complementing the RF2 knockout mutation were those showing the strongest residual activity in the 3A0 reporter assay for UGA readthrough (see Table 1). Less frequent or no complementation with other csu alleles will be discussed later.
Sequence and product analyses of csu RF2 alleles DNA sequencing revealed one mutation in each csu q Blackwell Science Limited
RF-ribosome initial binding hypothesis
mutant within the prfB gene which caused a mis-sense or a non-sense substitution at different amino acid positions (Table 1). Surprisingly, csu-1 and csu-3 substitute the UAG stop codon for Gln 322, and csu-4 and csu-6 substitute the UAG stop codon for Gln 252. The other csu alleles cause single amino acid substitutions for Arg 332 (csu-2), Ser 325 (csu-5), Arg 203 (csu7), Lys 282 (csu-8) and Tyr 264 (csu-9). All these substitutions occurred at highly conservative residues in domains D–F (see below). Product sizes and cellular levels of csu RF2 derivatives expressed in each transformant were investigated by Western blot analyses using anti-RF2 antibody. To correlate, if necessary, the expression level to the suppression frequency, if any, of lacZ (UAG) used in the initial selection, the red transformants of DEV1 whcih appeared on lactose-MacConkey plates (see Fig. 1) were directly inoculated into LB broth containing 2% lactose, and mid-late phase cultures were used for the protein analyses. Because Salmonella RF2, from which the csu alleles were created, migrates slightly slower than E. coli RF2 (Fig. 2; compare lanes 1 and 2), a broad dense band or close doublet bands were detected by Western blotting (Fig. 2; compare lanes 3–5). Under these conditions, cells transformed with the csu-2, csu-5, csu-7, csu-8 and csu-9 plasmids gave rise to over-expressed RF2 protein patterns (lanes 7, 10, 12–14) similar to that of the parental plasmid transformant (lane 5), while those transformed with four amber mutants, csu-1, csu-3, csu-4 and csu-6, gave rise to smaller polypeptide bands in sizes of 31 and 40 kDa, instead of the full length product (lanes 6, 8, 9, 11). These fast migrating bands should represent RF2 amber fragments, and their polypeptide masses were consistent with the positions of amber alleles that were previously deduced by DNA sequencing.
In vitro polypeptide release activity of csu RF2 The above sequence and product analyses classified the csu alleles into mis-sense and non-sense (amber) substitutions. To explore the activity of the mis-sense csu proteins in vitro, wild-type as well as csu-2 (R332C), csu-5 (S325F), csu-8 (K282I) and csu-9 (T264I) gene products were marked with a C-terminal histidine tag (which does not affect the activity of wild-type RF2 in vivo or in vitro; Uno et al. 1996), and the tagged proteins were purified to homogeneity by affinity chromatography using Ni-NTA agarose. The ability to terminate protein synthesis was monitored by the rate of Nformylmethionine (fMet) release at the stop codons (Mikuni et al. 1994). As expected, the csu-9 protein, which was thought to retain the highest residual activity q Blackwell Science Limited
of RF2 through the in vivo analyses (see Table 1), catalysed fMet release at UGA and UAA with about 60% efficiency compared with wild-type RF2 (Fig. 3). Three other proteins, csu-2, csu-5 and csu-8, catalysed threefold less efficiently than csu-9 protein (Fig. 3). These results indicated that the mis-sense csu substitutions did not abolish the primary activity of RF2, although their specific activity was significantly lower than that of the wild-type RF2. This reduced activity could, at least in part, account for the failure of complementation of the RF2 knockout (prfB::CmR) allele by csu-2 and csu-8 proteins. However, since the csu-5 allele with similar in vitro release activity perfectly complemented prfB::CmR, there must be another reason(s) for the lack of complementation (see below).
RF2 C-terminal deletions showing amber suppressor activity The occurrence of Csu phenotype by amber mutations in plasmid-bearing prfB prompted us to ask whether amber fragments or read-through products (eventually suppressed by Csu activity) should be responsible for the Csu action. To answer this, we manipulated a series of C-terminal deletions of RF2, D1–D6, at positions of csu alleles or conserved residues: two of them, RF2D1 and RF2D3, were equivalent to the predicted amber fragments of csu-1 (csu-3) and csu-4 (csu-6), respectively (Fig. 4). The DEV1 lacZ (UAG) strain was transformed with pTWV229 derivatives carrying RF2D1 to D6, and plated on lactose-MacConkey plates. As shown in Fig. 5A, cells that were transformed with plasmids expressing RF2D1–D4 were red or pink (Lacþ; samples 3–6), whereas those with other deletion plasmids were white (Lac¹; samples 1, 2, 7 and 8). These results clearly demonstrate that the amber fragments, but not the suppressed products, of csu RF2 alleles exert Csu activity, suggesting that the N-terminal part of RF2 can interact with the ribosome. The data further demonstrated that at least the N-terminal 210 amino acid fragment, proximal to the csu-7 position, is sufficient for Csu action.
RF2 C-terminal deletions showing opal suppressor activity The same set of RF2D1-D6 plasmids were transformed into OM6 lacZ (UGA) cells. Surprisingly, two of them, RF2D1 and D2, enabled OM6 Lacþ (Fig. 5B, samples 3 and 4), whereas the other plasmids did not (samples 1, 2, 5–8). These results indicated that the Csu phenotype owing to the alteration of RF2 may not be restricted to Genes to Cells (1999) 4, 253–266
257
K Yoshimura et al.
Figure 2 Western immunoblot analysis of csu products of RF2. DEV1 transformant colonies shown in Fig. 1 were directly inoculated into LB media containing ampicillin (50 mg/mL) and 2% lactose, and grown to log phase at 37 8C. Equal amounts of cell lysates were analysed by SDS-PAGE and subjected to immunoblot staining with anti-RF2 antibody as described in Experimental procedures. Lanes: 1, E. coli RF1 with a histidine tag (50 ng; affinity purified protein using Ni-NTA Agarose); 2, Salmonella RF2 with a histidine tag (50 ng; affinity purified protein using Ni-NTA Agarose); 3, pTWV229 transformant (control); 4, pTWV-RF1 transformant (control); 5, pTWV-RF2 transformant; 6, pTWV-Csu1 transformant; 7, pTWV-Csu2 transformant; 8, pTWV-Csu3 transformant; 9, pTWV-Csu4 transformant; 10, pTWV-Csu5 transformant; 11, pTWV-Csu6 transformant; 12, pTWV-Csu7 transformant; 13, pTWV-Csu8 transformant; 14, pTWV-Csu9 transformant. Salmonella RF2 (noted as stRF2) migrates behind E. coli RF2 (noted as ecRF2). Two amber fragments are marked with their molecular masses.
its noncognate stop codon, UAG, but seems to be common to all stop codons, although the reason for not exerting opal (UGA) suppression by RF2D3 and D4 are not known (see Fig. 4).
RF1 C-terminal deletions showing Csu phenotype Because RF2 C-terminal deletions RF2D1 and D2 are capable of enhancing the read-through of not only UAG but also UGA, we planned to do reciprocal experiments using E. coli RF1. Both RF1 and RF2 share the same sequence at the region truncated by RF2D1 and D3, or the same amino acids hit by csu-1 (csu-3) and csu-4 (csu-6) amber mutations. Hence, we made pTWV-RF1 expression plasmids carrying RF1D1 and D3 equivalent to RF2D1 and D3, respectively. These plasmids were transformed into DEV1 lacZ (UAG) and OM6 lacZ (UGA) strains, and the transformants were grown on lactose-MacConkey plates (Fig. 5C). As expected, RF1D1 and D3 enabled both indicator strains to take up a red colour (samples 3 and 4), showing that the N-terminal D1 and D3 fragments of RF1 are able to interfere with the normal release factor action and to lead to the suppression of both UAG and UGA.
Discussion Csu phenotype caused by release factor mutations The RF2 csu mutations isolated in this study are in sharp 258
Genes to Cells (1999) 4, 253–266
contrast to the previously isolated prfA and prfB mutations for several reasons. First, suppression by csu is omnipotent rather than specific to cognate stop codons, while that by other suppressors is specific to the cognate codon. Second, csu was isolated as a dominant suppressor, while others are recessive suppressors. This probably means that suppression by csu is based on the interference with the other RF, while that by other suppressors is based on the sole inactivation of RF. Third, the suppression activity of csu is not always influenced by the residual RF2 activity, while that by other suppressors is influenced by the residual activity. Fourth, csu mutations hit highly conservative residues, while other classical suppressors are not restricted to the conserved residues. Of these characteristics, the crucial finding was that csu-driven suppression is omnipotent and can be reproduced using both RF1 and RF2 equally. Taking these findings into consideration, it is likely that these csu mutations affect a novel and important step of polypeptide termination catalysed by release factors, which must be common to both RF1 and RF2, probably preceding and/or independent of stop codon recognition.
Release factor csu alleles It is remarkable that csu mutations coincided with the highly conservative residues in the conserved domains D–F (Fig. 6), and both mis-sense and non-sense prfB mutations gave rise to the same Csu phenotype. They seem to possess specific alterations, since simple loss-offunction mutations in the conservative residues of RFs q Blackwell Science Limited
RF-ribosome initial binding hypothesis
Figure 3 Polypeptide release activity of wild-type and csu RF2 proteins at UGA and UAA codons. f [3H]Met release from the [f [3H]Met-tRNAf?AUG?ribosome] complex upon addition of histidine-tagged RF2 products and terminator triplets was determined as described in Experimental procedures. The values represent the radioactivity of net f [3H]Met release after subtracting the triplet-free control value. Reactions contained 20 mM UGA (upper panel) and UAA (lower panel), as well as equal molar amounts (50 pmol) of RF proteins. Experiments were performed independently at least three times, and the values are expressed with standard deviations.
do not always give rise to Csu activity (data not shown). This probably suggests that the altered RF2 proteins, as well as the truncated RF2 (and RF1) polypeptides, essentially possess the same defect by affecting the conserved function that is common to RF1 and RF2, but not the stop codon recognition function which is selective with RF1 and RF2. The fact that C-terminal deletions of RF1 and RF2 manifested the Csu phenotype could be interpreted as indicating that Csu activity is primarily caused by a loss-of-function of a certain activity. We assume that a noncognate RF, once bearing the csu alteration, binds to the ribosome in such a way as to lead to an interference of proper functioning of cognate RF, probably as a consequence of an altered q Blackwell Science Limited
recognition of the codon or altered interaction with the peptidyltransferase centre (‘hydrolysis domain’) by csu. There are some characteristic properties that are associated with some of the csu alleles which we have to address here. First, P1 transduction experiments demonstrated that amber csu alleles, csu-1 (csu-3) and csu-4 (csu-6), can support the viability of the RF2-knockout (prfB::CmR) strain, though weakly, in the absence of tRNA amber suppressor (see Table 1). This does not indicate that the amber fragments carry an RF activity sufficient for polypeptide termination and cell growth because the equivalent segments, RF2D1 and RF2D3, failed to complement the prfB::CmR allele (data not shown). Therefore, it is conceivable that the amber csu allele in prfB per se is suppressed by Csu action and the read-through RF2 product complements prfB::CmR. This represents a phenomenon of autogenous suppression that has been shown with an opal mutation in the Salmonella RF2 gene (Kawakami & Nakamura 1990) or depletion of eRF1 of Saccharomyces cerevisiae (Stansfield et al. 1996), although the mechanisms are distinct. An in vivo and in vitro functional assay also demonstrated that, of the three alleles, csu-2, csu-5 and csu-8, which possess similar f Met release activity in vitro (see Fig. 3), csu-5 only complemented prfB::CmR. This clearly argues that the reduced activity of RF2 per se could not account for the failure of complementation, but that there must be another reason for this. We assume that csu-2 and csu-8 may cause more severe damage in vivo which was not detectable in the simple in vitro f Met release assay.
N-terminal domain function of RF Of the seven domains of prokaryotic RF1 and RF2, the C-terminal tRNA mimicry part which includes domains D/E is highly conservative, whereas the Nterminal part which includes domains A/B is much less conservative. (see Fig. 6B). It is obvious that RF1 and RF2 are more than simple tRNA-mimicry proteins, since they can activate the hydrolysis of peptidyl-tRNA and can bind to the ribosome in the absence of an EFTu-like vehicle protein. These activities must be encoded in domains other than in the tRNA-mimicry part. The present findings strongly suggest that the release factor consists of at least two functional parts: an N-terminal ribosome-binding part, and a C-terminal codon-decoding and (probably) peptide-release (hydrolysis) part (see Fig. 7B). Genes to Cells (1999) 4, 253–266
259
K Yoshimura et al.
Figure 4 Truncated RF1 and RF2 polypeptides and their suppression activity on UAG and UGA alleles in lacZ. C-terminal deletions were generated by PCR using designed primers as described in Experimental procedures. Boxes indicate RF2 and RF1 segments cloned into expression plasmid pTWV229. UAG and UGA suppression activity was monitored by colony colours formed on lactoseMacConkey plates as shown in Figure 5, and was summarized based on at least five independent transformants: þ means suppression (‘red’); þ/¹ means weak suppression (‘pink’); and ¹ means no suppression (‘white’). Positions of csu alleles are indicated.
Figure 5 Suppression assay of lacZ (UAG) and lacZ (UGA) mutations by C-terminal deletion polypeptides of RF2 and RF1. (A and B) Colony colour of DEV1 lacZ (UAG) (A) and OM6 lacZ (UGA) (B) transformants on lactose-MacConkey plates supplemented with ampicillin (50 mg/mL) and 2% lactose at 37 8C. Plasmids: sample 1, pTWV-RF2; sample 2, pTWV229; sample 3, pTWV-BD1; sample 4, pTWV-BD2; sample 5, pTWV-BD3; sample 6, pTWV-BD4; sample 7, pTWV-BD5; sample 8, pTWV-BD6. (C) Colony colour of DEV1 lacZ (UAG) (upper) and OM6 lacZ (UGA) (lower) transformants on lactose-MacConkey plates supplemented with ampicillin (50 mg/mL) and 2% lactose at 37 8C. Plasmids: sample 1, pTWV229; sample 2, pTWV-RF2; sample 3, pTWV-AD3; sample 4, pTWVAD1.
260
Genes to Cells (1999) 4, 253–266
q Blackwell Science Limited
RF-ribosome initial binding hypothesis
Figure 6 Protein sequences conserved in prokaryotic RF1 and RF2, and positions of csu alleles. (A) Comparison of the amino acid sequences of RF domains D through F. Amino acid substitutions caused by csu mutations are indicated. RF2_SALTY, Salmonella RF2 (accession no. P28353); RF2_ECOLI, E. coli RF2 (accession no. P07012); RF2_HAEIN, Haemophilus influenzae RF2 (accession no. P43918), RF2_HELPY, Helicobacter pylori RF2 (accession no. P55999), RF2_CYANO, Synechocystis sp. RF2 (accession no. sll1865), RF1_SALTY, Salmonella RF1 (accession no. P13654); RF1_ECOLI, E. coli RF1 (accession no. P07011); RF1_HAEIN, Haemophilus influenzae RF1 (accession no. P43917), RF1_HELPY, Helicobacter pylori RF1 (accession no. P55998), RF1_CYANO, Synechocystis sp. RF1 (accession no. sll1110), RF1_YEAST, S. cerevisiae mitochondrial RF1 (accession no. P30775). The number of the amino acid position is counted from the N-terminal Met of the Salmonella RF2 sequence. (B) Average similarity plots of the 22 bacterial RF1 and RF2 sequences including those listed in (A) plus: Mycobacterium tuberculosis RF2 (accession no. O05782), Streptomyces coelicolor RF2 (accession no. Q53915), Bacillus subtilis RF1 (accession no. P45872), Mycoplasma pneumoniae RF1 (accession no. P75420), Mycoplasma genitalium RF1 (accession no. P47500), Mycoplasma capricolum RF1 (accession no. P71496), Coxiella burnetii RF1 (accession no. P47849), Thermus aquaticus RF1 (accession no. P96077), Mycobacterium leprae RF1 (accession no. P45833), Kluyveromyces lactis mRF1 (accession no. P41767), Schizosaccharomyces pombe mRF1 (accession no. Q09691). Domains D through F are thought to mimic tRNA (Ito et al. 1996). The average similarity score along the entire sequence is provided by the dashed line. Comparison scores (expressed in standard deviations) were calculated using the PILEUP and PLOTSIMILARITY program from the GCG program package (Devereux et al. 1984). The amino acid position refers to the coordinate of the similarity alignment, which is distinct from that shown in (A).
What is novel here is the proposal of a primary ribosomal binding domain in the N-terminal part of RFs. However we do not exclude the possibility that the other part also interacts with the ribosome. The less conservative nature of the N-terminal could be interpreted as indicating that it probably interacts with the less conservative structural component(s) of the ribosome, such as ribosomal protein(s) that are diverse in different organisms. Given that a hydrolysis domain is q Blackwell Science Limited
impaired by csu, the potential ribosome binding domain may lose its proper control, since it could be functionally coupled with the hydrolysis domain through the recognition of cognate or noncognate stop codons by the anticodon domain.
‘RF-ribosome initial binding’ hypothesis The kinetic mechanism of RF binding to the A site Genes to Cells (1999) 4, 253–266
261
K Yoshimura et al.
Figure 7 ‘RF-ribosome initial binding’ hypothesis. (A) The predicted sequence of interactions between the ribosome and RF. We propose initial binding to the ribosome of RF preceding stop codon recognition. (B) Schematic diagram of the predicted functional domain structure of RF bound to the ribosome. The N-terminal part (shown as N) encodes an initial binding domain, and the C-terminal (shown as C) encodes a codon-recognition and peptide-release (hydrolysis) domain.
of the ribosome that is associated with mRNA is not known. In the case of the elongation factor EF-Tu X GTP X aminoacyl-tRNA complex, the first step in the sequence of interactions between the ribosome and the ternary complex is the codonindependent formation of an initial complex (Rodnina et al. 1996). While the initial binding of the ternary complex to the ribosome is rapid and readily reversible, subsequent correct codon recognition stabilizes the complex. It induces the GTPase conformation of EFTu, instantaneously followed by GTP hydrolysis (Pape et al. 1998). The rapid and labile formation of the initial binding complex may fit the initial screening of ternary complexes by the ribosome. We could probably extend this analogy to RF, at least to some extent, since the initial screening needs to include RFs when the ribosome encounters the stop codon. We could propose here that the predicted Nterminal ribosome binding domain (see above) is involved in the initial binding of RF to the ribosome preceding stop codon recognition (‘RF-ribosome initial binding’ hypothesis; Fig. 7) although other possibilities, such as a simultaneous occurrence of these two steps, could not be entirely excluded at present. This model predicts that, while the initial 262
Genes to Cells (1999) 4, 253–266
binding to the ribosome of RF may be rapid and labile, a cognate RF codon recognition may stabilize the complex and induce the hydrolysis conformation of the peptidyltransferase centre, instantaneously followed by polypeptide-chain release. The csu mutant of RF2, even with or without the residual RF2 activity, may interfere with a proper functioning of both RF1 and RF2, probably by reducing their entry to the ribosome. Two mechanisms might be immediately possible: one, to assume that the initial binding complex is stabilized and that a rapid release of RF is blocked by csu; alternatively, the csu product might proceed to a step beyond a potential ejection period when noncognate and be trapped as a nonproductive form (the ‘entrapment’ model). We would favour the latter scenario because, as discussed above, the Csu activity seemed to be caused by a loss-of-function, which could not account for a potential increase in the affinity (seemingly a gain-offunction) to the ribosome of RF csu products. Given that the initial binding to the ribosome is true of RF, similar to the EF-Tu X GTP X aminoacyl-tRNA complex, one could speculate that the Csu activity might influence the sense codon decoding to some extent. This might be correlated with the observation that E. coli cells became sick upon overexpression of some csu alleles (our unpublished results). Nevertheless, the putative initial binding sites on the ribosome of both elongator and terminator components may not be largely or exclusively overlapping, since the Csu phenotype requires the entry of a misreading tRNA ternary complex. The present findings strongly suggest that the elongation and termination steps of protein synthesis probably share comparable mechanisms in the initial screening of the cognate decoding component by the ribosome. Further in vivo and in vitro analyses are needed in order to elucidate the mechanism of Csu action, as well as to clarify the predicted initial binding to the ribosome of release factors preceding stop codon recognition.
Experimental procedures Bacterial strains and media The E. coli K-12 strains used are listed in Table 2. The LB medium contained 1% bacto-tryptone, 0.5% yeast extract and 0.5% NaCl (Miller 1972) and was supplemented with ampicillin (50 mg/mL). The lactose-MacConkey plate contained 4% MacConkey agar base (Miller 1972) and was supplemented with lactose (2%) and ampicillin (50 mg/mL). RF2-knockout allele (prfB::CmR) described previously (Ito et al. 1998) was transduced into strain W3110 in the presence of the prfB plasmid by the P1 phage, giving rise to RM786. q Blackwell Science Limited
RF-ribosome initial binding hypothesis
Plasmids and manipulations
RF2 purification
The plasmids used are listed in Table 2. Plasmid pTWV-RF2 is an IPTG controllable prfB expression plasmid that encodes the Salmonella prfB gene at the BamHI-SacI sites of pTWV229. pTWV-Csu1 to -Csu9 are derivatives of pTWV-RF2, constructed as amber suppressors in this study. A set of C-terminal deletion segments of Salmonella prfB and E. coli prfA were constructed by polymerase chain reaction (PCR) using a pair of primers (i.e. primer no. 9 and one of the primers, nos 1–8; see Table 3), and were cloned into the BamHI-SacI sites of pTWV229, giving rise to pTWV-BD and pTWV-AD plasmids. For the overproduction of RF2 and its csu products, wild-type and prfB csu sequences in pTWV-RF2 and pTWV-Csu DNAs were amplified by PCR using a pair of primers, nos 10 and 11 (Table 3; see Uno et al. 1996). The amplified DNAs contained the entire coding sequence of RF2. They were cloned downstream of a T7 RNA polymerase promoter between the BglII and SacI sites of pET30-a-c(þ) (Novagen Inc., WI) according to the manufacturer’s instructions, giving rise to pET-RF2H6 and pETCsu2H6 to -Csu9H6 plasmids. For the read-through assay using the 3A0 reporter system, the wild-type and prfB csu segments in pTWV-RF2 and pTWV-Csu plasmids were re-cloned into the BamHI-SacI sites of plasmid pSTV29, giving rise to pSTV-RF2 and pSTV-Csu derivatives. pSTV-RF1 was constructed by cloning the E. coli prfA segment into the same sites of pSTV29. The structure of each plasmid construct was confirmed by DNA sequence analyses.
Plasmids pET-RF2H6, -Csu2H6, -Csu5H6, -Csu8H6 and -Csu9H6 were transferred to BL21 (DE3). BL21 (DE3) contains a lysogenic lambda phage derivative, DE3, carrying the gene for T7 RNA polymerase under the control of an inducible lacUV5 promoter. The over-expression of recombinant proteins was achieved by T7 RNA polymerase in BL21 (DE3) transformants in the presence of 0.5 mM IPTG for 2.5 h, and histidine-tagged RF proteins were purified to homogeneity from cell lysates by affinity chromatography using Ni-NTA Agarose (Qiagen).
Mutagenesis of RF2 and selection of crosssuppressors The pTWV-RF2 DNA was mutagenized by incubation with 0.4 M hydroxylamine at pH 6.0 for 20 h at 37 8C (Oshima et al. 1995; Ito et al. 1998). The plasmid was then precipitated with ethanol and rinsed several times with LB broth. The E. coli K12 strain DEV1 lacZ (UAG) was transformed with the mutagenized DNA and ApR colonies were selected on lactose-MacConkey plates containing 50 mg/mL of ampicillin at 37 8C. Plasmid DNAs were recovered from red or pink colonies, retransformed into the same parental strain, and those that gave a reproducible phenotype (i.e. red or pink colony formation on lactoseMacConkey plates) were further characterized.
Complementation of prfB knockout allele by csu mutations P1 phage was grown on an RM786 (prfB::CmR) strain carrying plasmid pSUIQT-RF2* (Ito et al. 1998), and was infected to W3110 strain transformed with pTWV-RF2 or pTWV-Csu plasmids. Chloramphenicol-resistant (CmR) colonies were selected on LB agar plates supplemented with chloramphenicol (15 mg/mL) and 1 mM IPTG. The appearance of CmR transductants at a normal frequency represents the complementation of prfB::CmR by the csu allele that was present in the host cell. q Blackwell Science Limited
In vitro termination assay In vitro fMet release assay with these wild-type and csu mutant RF2 proteins was carried out essentially as previously described (Mikuni et al. 1994). 500 pmol E. coli 70S ribosome, 25 nmol AUG initiator codon, and 250 pmol f[3H]Met-tRNAf (1 × 104 c.p.m.) were mixed in a solution (500 mL) containing 20 mM Tris?HCl (pH 7.5), 0.15 mM NH4Cl and 10 mM MgCl2. The mix was incubated at 30 8C for 30 min for the pre-formation of ribosomal initiation complexes, and its aliquot (5 mL) was mixed with histidine-tagged RF proteins, terminator triplets, and 5 mL 10 × reaction buffer [5 M Tris?HCl (pH 7.5), 0.75 M NH4Cl and 0.3 M MgCl2], and adjusted to 50 mL with distilled water. The reaction mix was incubated at 30 8C for 30 min, and terminated by the addition of 0.1 N HCl. f[3H]Met released from the ribosomal complex was extracted with ethylacetate, and its radioactivity was counted by liquid scintillation.
Analysis of protein products of the 3A0 gene Escherichia coli W3110 cells harbouring pSTV-RF2 csu derivatives (CmR) were transformed with the 3A0 reporter plasmids pAB96 (for UAG read-through assay; ApR) and pAB101 (for UGA readthrough assay; ApR) (for structures see Bjo¨rnsson & Isaksson 1993; Mottagui-Tabar et al. 1994). CmR ApR transformants were grown in LB media containing selective antibiotics and IPTG (1 mM), and exponentially growing cells were examined for the synthesis of 3A0 and 2A0 proteins as previously described (Uno et al. 1996). The read-through (RT) value is calculated by the following formula: RT ¼ 2½3A0 ÿ=f3½2A0 ÿ þ 2½3A0 ÿg where 3[2A0 ] and 2[3A0 ] represent the amounts of 2A0 and 3A0 products, respectively.
Western blots Protein was determined by means of a Bio-Rad Protein Assay (Bio-Rad Laboratories). Bulk proteins prepared from sonicated cells were solubilized in 120 mM Tris?HCl (pH 6.8), 4.6% SDS and 5% b-mercaptoethanol by boiling for 3 min, separated by SDS–polyacrylamide gel electrophoresis (SDS-PAGE) using 10% polyacrylamide gels, and electroblotted on to Hybond-ECL Genes to Cells (1999) 4, 253–266
263
K Yoshimura et al.
Strain or plasmid Escherichia coli K-12 DEV1 OM6 RM786 BL21 (DE3) W3110 Plasmid pUC118 pTWV229 pSTV29 pET30-a-c(þ) pAB96 pAB101 pTWV-RF2 pTWV-RF1 pTWV-Csu pTWV-BD pTWV-AD pSTV-RF2 pSTV-RF1 pSTV-Csu pET-RF2H6
pET-CsuH6
Relevant description
Source or reference
Hfr thi-1 relA spoT1 lacZ105(UAG) leu(UGA) lacZ659(UGA) trpA9605(UAG) his29(UAG) W3110 prfB::CmR Lysogenic for l DE3 carrying the gene for T7 RNA polymerase under the lacUV5 promoter Prototroph
Go¨ringer et al. (1991)
E. coli cloning vector, ApR E. coli cloning vector, ApR pACYC184 derivative expression vector; CmR Multi-cloning-site expression vector under the control of T7 RNA polymerase promoter, KmR 3 A0 reporter plasmid for UAG dependent termination; ApR 3 A0 reporter plasmid for UGA dependent termination; ApR S. typhimurium prfB coding segment cloned into multi cloning site in pTWV229, ApR Same as pTWV-RF2 but E. coli prfA substituted for S. typhimurium prfB, ApR Same as pTWV-RF2 but csu-1 through csu-9 alleles in S. typhimurium prfB, ApR Same as pTWV-RF2 but Cterminal deletions D1 through D6 in S. typhimurium prfB, ApR Same as pTWV-RF2 but Cterminal deletions D1 and D3 in E. coli prfA, ApR S. typhimurium prfB coding segment cloned into multicloning site in pSTV29, CmR Same as pSTV-RF2 but E. coli prfA substituted for S. typhimurium prfB, CmR Same as pSTV-RF2 but csu-1 through csu-9 alleles in S. typhimurium prfB, CmR Histidine tagged S. typhimurium prfB gene cloned downstream of a T7 RNA polymerase promoter in pET30-a-c(þ), KmR Same as pET-RF2H6 but csu-2 csu-5, csu-8 and csu-9 alleles in S. typhimurium prfB, KmR
membranes (Amersham). The blots were incubated with rat anti-RF2 serum (Uno et al. 1996) diluted 1:5000 in TBS buffer [20 mM Tris?HCl (pH 7.6), 137 mM NaCl] containing 0.1% Tween 20 (1 h, room temperature with shaking) after
264
Genes to Cells (1999) 4, 253–266
Table 2 Bacterial strains and plasmids
Mikuni et al. (1994) Ito et al. (1998) Novagen, Inc. Laboratory stock TaKaRa Co. TaKaRa Co. TaKaRa Co. Novagen, Inc. Mottagui-Tabar et al. (1994) Mottagui-Tabar et al. (1994) This work This work This work This work This work This work This work This work Ito et al. (1998)
This work
pretreatment with TBS buffer supplemented with 5% skimmed milk. After three washes, the blots were incubated with peroxidase-conjugated sheep secondary antibody and anti-rat immunoglobulin G (Amersham, 1:3000 dilution in blocking q Blackwell Science Limited
RF-ribosome initial binding hypothesis Table 3 List of PCR primers Primers*
50 -to-30 sequence
Target/purpose
# 1 (¹) # 2 (¹) # 3 (¹) # 4 (¹) # 5 (¹) # 6 (¹) # 7 (¹) # 8 (¹) # 9 (þ) #10 (¹)
GGGGAGCTCTTAGCTTCCCCAGCCGAT GGGGAGCTCTTACATCGCCTGTTTTTCAG GGGGAGCTCTTAACCGCCTGCGCCAGA GGGGAGCTCTTAGCCGGAGTCAAACGG GGGGAGCTCTTAGATAGTCGCGGACTTAATAC GGGGAGCTCTTAGCAATCGGCGCTATCG GGGGAGCTCTTAGCTGCGATCGCCACTC GGGGAGCTCTTAACCACCCGCCCCTGAC CAGGAAACAGCTATGAC GGGAGCTCTTAGTGGTGGTGGTGGTGGTG TGAATGCTCGCGGGTCTGGTGGAA GGGGATCCGCTCTTATCACCGCAT
RF2D1 RF2D2 RF2D3 RF2D4 RF2D5 RF2D6 RF1D1 RF1D3 Universal His-tag RF2
#11 (þ)
His-tag RF2
*(þ) means a sense sequence, and (¹) means an anti-sense sequence.
buffer) for 1 h. After washing, the membranes were developed by means of enhanced chemiluminescence using Hyperfilm-ECL (Amersham) according to the manufacturer’s instructions.
Other DNA procedures Single-or double-stranded DNAs were sequenced by means of dideoxynucleotide chain-termination (Sanger et al. 1977). PCR was carried out according to standard methods (Saiki et al. 1988), and other DNA manipulations were conducted as previously described (Sambrook et al. 1989).
Acknowledgements We thank K. Matsumura for participation in the early part of the construction of prfB::CmR strain; and L. Isaksson for the 3A0 reporter system. This work was supported in part by grants from The Ministry of Education, Science, Sports and Culture, Japan; the Human Frontier Science Program (awarded in 1993 and 1997); and the Basic Research for Innovation Biosciences Program of Bio-orientated Technology Research Advancement Institution (BRAIN).
References Ævarsson, A., Brazhnikov, E., Garber, M., et al. (1994) Threedimensional structure of the ribosomal translocase: elongation factor G from Thermus thermophilus. EMBO J. 13, 3669–3677. Bjo¨rnsson, A. & Isaksson, L.A. (1993) UGA codon context which spans three codons reversal by ms2i6A37 in tRNA, mutation in rpsD(S4) or streptomycin. J. Mol. Biol. 232, 1017–1029. Buckingham, R.H., Grentzmann, G. & Kisselev, L. (1997) Polypeptide chain release factors. Mol. Microbiol. 24, 449–456. Capecchi, M.R. & Klein, H.A. (1969) Characterization of three proteins involved in polypeptide chain termination. Cold Spring Harbor Symp. Quant. Biol. 34, 469–477. q Blackwell Science Limited
Caskey, T., Scolnick, E., Tompkins, R., Goldstein, J. & Milman, G. (1969) Peptide chain termination, codon, protein factor, and ribosomal requirements. Cold Spring Harbor Symp. Quant. Biol. 34, 479–488. Crawford, D.-J.G., Ito, K., Nakamura, Y. & Tate, W.P. (1999) Indirect regulation of termination efficiency at highly expressed genes and recoding sites by the factor recycling function of Escherichia coli release factor RF3. EMBO J. 18, 727–732. Czworkowski, J., Wang, J., Steitz, T.A. & Moore, P.B. (1994) The crystal structure of elongation factor G complexed with GDP, at 2.7 A˚ resolution. EMBO J. 13, 3661–3668. Devereux, J., Haeberili, P. & Smithies, O. (1984) A comprehensive set of sequence analysis programs for the Vax. Nucl. Acids Res. 12, 387–395. Elliott, T. & Wang, X. (1991) Salmonella typhimurium prfA mutants defective in release factor 1. J. Bacteriol. 173, 4144–4154. Freistroffer, D.V., Pavlov, M.Y., MacDougall, J., Buckingham, R.H. & Ehrenberg, M. (1997) Release factor RF3 in E. coli accelerates the dissociation of release factors RF1 and RF2 from the ribosome in a GTP-dependent manner. EMBO J. 16, 4126–4133. Goldstein, J.L. & Caskey, C.T. (1970) Peptide chain termination: effect of protein S on ribosomal binding of release factors. Proc. Natl. Acad. Sci. USA 67, 537–543. Go¨ringer, H.U., Hijazi, K.A., Murgola, E.J. & Dahlberg, A.E. (1991) Mutations in 16S rRNA that affect UGA (stop codon)directed translation termination. Proc. Natl. Acad. Sci. USA 88, 6603–6607. Hinnebusch, A.G. & Liebman, S.W. (1991) Protein synthesis and translational control in Saccharomyces cerevisiae. In: The Molecular and Cellular Biology of the Yeast Saccharomyces: Genome Dynamics, Protein Synthesis, and Energetics, Vol. 1, pp. 627–735. Cold Spring Harbor, New York: Cold Spring Harbor Laboratory Press. Hughes, D., Atkins, J.F. & Thompson, S. (1987) Mutants of elongation factor Tu promote ribosomal frameshifting and nonsense readthrough. EMBO J. 6, 4235–4239. Ito, K., Ebihara, K., Uno, M. & Nakamura, Y. (1996) Conserved motifs of prokaryotic and eukaryotic polypeptide release factors: tRNA-protein mimicry hypothesis. Proc. Natl. Acad. Sci. USA 93, 5443–5448. Genes to Cells (1999) 4, 253–266
265
K Yoshimura et al. Ito, K., Uno, M. & Nakamura, Y. (1998) Single amino acid substitution in prokaryote polypeptide release factor 2 permits it to terminate translation at all three stop codons. Proc. Natl. Acad. Sci. USA 95, 8165–8169. Kawakami, K., Inada, T. & Nakamura, Y. (1988a) Conditionally lethal and recessive UGA-suppressor mutations in the prfB gene encoding peptide chain release factor 2 of Escherichia coli. J. Bacteriol. 170, 5378–5381. Kawakami, K. & Nakamura, Y. (1990) Autogenous suppression of an opal mutation in the gene encoding peptide chain release factor 2. Proc. Natl. Acad. Sci. USA 87, 8432–8436. Kawazu, Y., Ito, K., Matsumura, K. & Nakamura, Y. (1995) Comparative characterization of release factor RF-3 genes of Escherichia coli, Salmonella typhimurium and Dichelobacter nodosus. J. Bacteriol. 177, 5547–5553. Matsumura, K., Ito, K., Kawazu, Y., Mikuni, O. & Nakamura, Y. (1996) Suppression of temperature-sensitive defects of polypeptide release factors RF-1 and RF-2 by mutations or by an excess of RF-3 in Escherichia coli. J. Mol. Biol. 258, 588–599. Mikuni, O., Ito, K., Moffat, J., et al. (1994) Identification of the prfC gene, which encodes peptide-chain-release factor 3 of Escherichia coli. Proc. Natl. Acad. Sci. USA 91, 5798–5802. Miller, J.H. (1972) Experiments in Molecular Genetics. Cold Spring Harbor, New York: Cold Spring Harbor Laboratory Press. Moffat, J.G., Donly, B.C., MacCaughan, K.K. & Tate, W.P. (1993) Functional domains in the Escherichia coli release factors: activities of hybrids between RF-1 and RF-2. Eur. J. Biochem. 231, 749–756. Moffat, J.G. & Tate, W.P. (1994) A single proteolytic cleavage in release factor 2 stabilizes ribosome binding and abolishes peptidyl-tRNA hydrolysis activity. J. Biol. Chem. 269, 18899–18903. Mottagui-Tabar, S., Bjo¨rnsson, A. & Isaksson, L.A. (1994) The second to last amino acid in the nascent peptide as a codon context determinant. EMBO J. 13, 249–257. Murgola, E.J. (1995) Translational suppression: when two wrongs do make a right. In: Trna: Structure, Biosynthesis, and Function (eds D. So¨ll & U.L. RajBandary), pp. 491–509. Washington DC: ASM Press. Nakamura, Y. & Ito, K. (1998) How protein reads the stop codon and terminates translation. Genes Cells 3, 263–278. Nakamura, Y., Ito, K. & Isaksson, L.A. (1996) Emerging understanding of translation termination. Cell 87, 147–150. Nakamura, Y., Ito, K., Matsumura, K., Kawazu, Y. & Ebihara, K. (1995) Regulation of translation termination: conserved structural motifs in bacterial and eukaryotic polypeptide release factors. Biochem. Cell Biol. 73, 1113–1122. Nissen, P., Kjeldgaard, M., Thirup, S., et al. (1995) Crystal
266
Genes to Cells (1999) 4, 253–266
structure of the ternary complex of Phe-tRNAPhe, EF-Tu, and a GTP analog. Science 270, 1464–1472. Oshima, T., Ito, K., Kabayama, H. & Nakamura, Y. (1995) Regulation of lrp gene expression by H-NS and Lrp proteins in Escherichia coli: dominant negative mutations in lrp. Mol. Gen. Genet. 247, 521–528. Pape, T., Wintermeyer, W. & Rodnina, M.V. (1998) Complete kinetic mechanism of elongation factor Tu-dependent binding of aminoacyl-tRNA to the A site of the E. coli ribosome. EMBO J. 17, 7490–7497. Pavlov, M.Y., Freistroffer, D.V., MacDougall, J., Buckingham, R.H. & Ehrenberg, M. (1997) Fast recycling of Escherichia coli ribosomes requires both ribosome recycling factor (RRF) and release factor RF3. EMBO J. 16, 4134–4141. Rodnina, M.V., Pape, T., Fricke, R., Kuhn, L. & Wintermeyer, W. (1996) Initial binding of the elongation factor TuGTPaminoacyl-tRNA complex preceding codon recognition on the ribosome. J. Biol. Chem. 271, 646–652. Ryde´n, S.M. & Isaksson, L.A. (1984) A temperature-sensitive mutant of Escherichia coli that shows enhanced misreading of UAG/A and increased efficiency for some tRNA nonsense suppressors. Mol. Gen. Genet. 193, 38–45. Saiki, R.K., Gelfand, D.H., Stoffel, S., et al. (1988) Primerdirected enzymatic amplification of DNA with a thermostable DNA polymerase. Science 239, 487–491. Sambrook, J., Fritsch, E.F. & Maniatis, T. (1989) Molecular Cloning: a Laboratory Manual, 2nd edn. Cold Spring Harbor, New York: Cold Spring Harbor Laboratory Press. Sanger, F., Nicklen, S. & Coulson, A.R. (1977) DNA sequencing with chain-terminating inhibitors. Proc. Natl. Acad. Sci. USA 74, 5463–5467. Stansfield, I., Eurwilaichitr, L., Akhamaloka & Tuite, M.F. (1996) Depletion in the levels of eRF1 causes a reduction in the efficiency of translation termination in yeast. Mol. Microbiol. 20, 1135–1143. Tate, W.P. & Brown, C.M. (1992) Translational termination: ‘stop’ for protein synthesis or ‘pause’ for regulation of gene expression. Biochemistry 31, 2443–2450. Uno, M., Ito, K. & Nakamura, Y. (1996) Functional specificity of amino acid at position 246 in the tRNA mimicry domain of bacterial release factor 2. Biochimie 78, 935–943. Yanofsky, C., Horn, V. & Nakamura, Y. (1996) Loss or overproduction of polypeptide release factor 3 influences expression of the tryptophanase operon of Escherichia coli. J. Bacteriol. 178, 3755–3762.
Received: 17 March 1999 Accepted: 7 April 1999
q Blackwell Science Limited
