The influence of the 5′ codon context on translation termination in Bacillus subtilis and Escherichia coli is similar but different from Salmonella typhimurium

Gene 212 (1998) 189–196

The influence of the 5∞ codon context on translation termination in Bacillus subtilis and Escherichia coli is similar but different from Salmonella typhimurium Salim Mottagui-Tabar, Leif A. Isaksson * Department of Microbiology, Stockholm University, S-10691 Stockholm, Sweden Received 30 December 1997; received in revised form 9 March 1998; accepted 16 March 1998; Received by D. Schlessinger

Abstract The last two amino acids in the nascent peptide influence translation termination in E. coli (Mottagui-Tabar et al., 1994; Bjo¨rnsson et al., 1996). We have compared the effects on termination in Escherichia coli, Bacillus subtilis and Salmonella typhimurium obtained by varying the −1 and −2 codons upstream of the weak UGAA stop signal. The peptide effect from the penultimate amino acid on translation termination in B. subtilis is similar to that seen in E. coli (with 66.5% RF-2 amino acid sequence similarity), whereas the influence in S. typhimurium (with 95.3% similarity to E. coli) is weaker. The effect of changing the −1 codon (P-site) is weaker in S. typhimurium as compared to those in E. coli and B. subtilis. RF-2s from E. coli and S. typhimurium have a threonine or alanine at position 246, respectively. This amino acid exchange in RF-2 can explain the difference in efficiency and toxicity during overexpression when E. coli and S. typhimurium are compared ( Uno et al., 1996). However, B. subtilis RF-2 also has an alanine at that position, yet the sensitivity to the nascent peptide is similar to that in E. coli. Thus, the amino acid difference at position 246 in the RF-2 sequences cannot explain why termination in E. coli and B. subtilis is similar in peptide sensitivity while being different from that in S. typhimurium. Sequence alignments of RF-2 from the three bacteria show other regions of the molecule that could be involved in the functional interactions with the C-terminal end of the nascent peptide. © 1998 Elsevier Science B.V. All rights reserved. Keywords: Nascent peptide; Release factor RF-2; Stop codon; 5∞ Codon context

1. Introduction In bacteria, translation termination factor RF-1 recognizes the termination codons UAG and UAA, whereas RF-2 recognizes UGA and UAA (Nakamura et al., 1996). Both RF-1 (Ryde´n and Isaksson, 1984) and RF-2 (Mikuni et al., 1991) associate with a third factor, RF-3 (Grentzmann et al., 1994; Mikuni et al., 1994), for maximal efficiency to release the nascent peptide from the P-site tRNA (Matsumura et al., 1996). Efficiency of translation termination is influenced by the nature of the stop codon (Sharp and Burgess, 1992), with UAA and UAG usually being more efficient than * Corresponding author. Tel: +46 8 164197; Fax: +46 8 6129552; e-mail [email protected] Abbreviations: IPTG, isopropyl b--thiogalactopyranoside,; PAGE, polyacrylamide gel electrophoresis; PCR, polymerase chain reaction; RF, release factor; SDS, sodium dodecyl sulphate; TABA, tryptone blood agar. 0378-1119/98/$19.00 © 1998 Elsevier Science B.V. All rights reserved. PII: S0 3 7 8 -1 1 1 9 ( 9 8 ) 0 0 17 6 - 0

UGA. Termination is also dependent on the codon context (Buckingham, 1990; Tate and Brown, 1992), the last and the penultimate amino-acid residues in the nascent peptide (Mottagui-Tabar et al., 1994; Bjo¨rnsson et al., 1996), the 3∞ codon context (Pedersen and Curran, 1991; Bjo¨rnsson and Isaksson, 1993; Bonetti et al., 1995; Poole et al., 1995), the P-site tRNA and its interaction with its codon or the release factors (Zhang et al., 1996) and possibly also some modified bases in the P-site tRNAs ( Yarus and Curran, 1992). The nature of the competing near-cognate or suppressor tRNAs (Mottagui-Tabar et al., 1994) and the elongation factor Tu affects stop codon readthrough and, as a consequence, termination efficiency (Mottagui-Tabar and Isaksson, 1996). The influence of the N-terminus of the nascent peptide on translation has been investigated in prokaryotes and eukaryotes (Lovett and Rogers, 1996). In the chloramphenicol acetyltransferase gene cat-86 in B. subtilis, the N-terminal pentapeptide can inhibit peptidyltransferase

190

S. Mottagui-Tabar, L.A. Isaksson / Gene 212 (1998) 189–196

activity in the presence of chloramphenicol (Moffat et al., 1994; Lovett and Rogers, 1996). The 24-aminoacid-long peptide at the N-terminus of the E. coli tnaC gene in the tryptophanase (tna) operon, in the presence of tryptophan, interacts with the ribosome and perhaps other factor(s) and prevents the release of the peptide at the stop codon ( Konan and Yanofsky, 1997). The C-terminal end of the nascent protein seems to influence the termination process (Mottagui-Tabar et al., 1994; Bjo¨rnsson et al., 1996; Zhang et al., 1996) during expression of natural E. coli genes, since highly expressed genes terminating with the weak stop signal UGAA code for C-terminal amino acids that promote efficient termination (Bjo¨rnsson et al., 1996; Mottagui-Tabar and Isaksson, 1997). As compared to RF-2 in E. coli, overproduced RF-2 in S. typhimurium is not toxic for bacterial growth and is more efficient in the termination process. This difference has been attributed to the nature of the amino acid at position 246 (threonine for E. coli and alanine for S. typhimurium) ( Uno et al., 1996). We have here compared the effects on termination obtained by changing the −1 and −2 codons at the 5∞ side of the UGAA stop signal in E. coli, B. subtilis and S. typhimurium. This has been done by adapting a translation assay gene, originally constructed for E. coli and S. typhimurium, to the B. subtilis system. The influence of the UGAA 5∞ codon context on termination in an exponentially growing culture can then be measured in all three organisms using a similar system. The assay gene is plasmid-borne and codes for three repeats (3A∞) of a semi-synthetic Staphylococcus aureus protein A-derived domain (Bjo¨rnsson and Isaksson, 1988; Bjo¨rnsson et al., 1998). A stop codon is inserted between the second and the third A∞ coding region (Fig. 1) such that a short translation termination product (2A∞) and a long readthrough protein product (3A∞) are formed. Molar quantities of the two protein species can be used to measure the efficiency of termination. The results suggest that there is a 20-fold peptide influence by the penultimate amino acid in the nascent peptide on efficiency of termination at UGAA in B. subtilis, similar to that in E. coli, but only a fivefold effect on termination in S. typhimurium. This difference in sensitivity between E. coli and S. typhimurium cannot be explained solely by the amino acid at position 246 in RF-2, being threonine and alanine, respectively ( Uno et al., 1996), since B. subtilis also has an alanine at this position. Instead, a combination of differences in amino acid composition of the three RF-2 species could explain the nascent peptide sensitivity when the RF-2 sequences from the three bacterial species are compared. In addition, the effect on termination by changing the −1 codon at the 5∞ side of the stop signal is different in the three bacterial species.

2. Materials and methods 2.1. Bacterial strains and plasmid constructions The E. coli strain used for cloning purposes was MC1061 (araD139, D(lacIPOZY)74, galU, galK, rpsL, D(araABC-leu)7697, hsdR, mcrB) (Sambrook et al., 1989). The B. subtilis strain (the 3G18 derivative of strain 168) (ade, met, trpC2) used for transformation and protein A∞ assay was a gift from Dr Per Johansson, Lund University, Sweden. The S. typhimurium strain UX863 (metA22, metE551, galE496, xyl 404, nml, H1-b, H2enx, hsdL6, hsdSA29, ilv, zhc::Tn10, rpsL+), a derivative of DA653 (Hughes et al., 1987), was used. The shuttle vector pDG148, for cloning work in E. coli and expression in B. subtilis, was a generous gift from Dr H. Putzer (IBPC, Paris, France). A KpnI restriction site, along with an extended Shine–Dalgarno sequence (oligosequence 5∞-GGAGGAGGTACC-3∞), was inserted downstream of the promoter in pDC148, between the HindIII and SphI restriction sites, resulting in plasmid pBSM01. The 3A∞ gene was first subcloned by digesting the fragment between NcoI and ClaI from the pSMplasmid series (Mottagui-Tabar et al., 1994) and ligating into an intermediate vector pSMT113, which has a KpnI site upstream of the NcoI site. From this intermediate plasmid, the KpnI–SacI fragment containing the 3A∞ gene with the test stop codon context between the second and the third A∞ coding regions was subcloned into pBSM01. Fig. 1 shows the final derivative (pBSM02) with the 3A∞ gene cloned downstream of the IPTGinducible promoter spac. All contexts were subcloned from earlier described plasmids (Mottagui-Tabar et al., 1994; Bjo¨rnsson et al., 1996). 2.2. Protein A∞ assay in B. subtilis and S. typhimurium Transformation of B. subtilis 3G18 was performed according to the starvation method using Spizizen medium [0.14% K HPO 0.06% KH PO , 0.002% 2 4, 2 4 MgSO · 7H O, 0.02% (NH ) SO , 0.01% sodium cit4 2 42 4 rate, 0.05% glucose] without amino acids. Selection for transformants was on TABA plates with 10 mg/ml of kanamycin. Transformants were inoculated in 10 ml of Spizizen medium with all 20 amino acids at a final concentration of 50 mg/ml and an appropriate antibiotic, and the culture was incubated overnight at 37°C with agitation. The overnight culture was then diluted 50-fold in 50 ml of the same medium and incubated with agitation until the density reached an OD of 0.5. 540 The 3A∞ gene was induced with the addition of IPTG at a final concentration of 20 mM. Induction was allowed to proceed for 1.5 h, after which the cells were harvested and chilled. The pellet was resuspended in 10× TST (0.5 M Tris, 1.5 M NaCl, 0.5% Tween 20) and sonicated for 30 s (Rapidis A350G, Ultrasonics

S. Mottagui-Tabar, L.A. Isaksson / Gene 212 (1998) 189–196

191

Fig. 1. The plasmid pBSM02 with an ampicillin resistance gene, used for selection in E. coli (for cloning purposes), and the kanamycin resistance gene, for selection in B. subtilis (for assaying purposes). The origin of replication in E. coli is derived from pBR322 and that of B. subtilis from pTP1. There are approximately 20 copies of the plasmid in B. subtilis cells. The transcription terminator at the end of the protein A∞ genes is from the rrnB operon of E. coli. For details of construction, see Section 2.

Ltd, at power 50 and tuning 40 cycles). The resulting suspensions were then heated at 95°C for 10 min, followed by centrifugation for 5 min in the cold. The supernatant was loaded directly on to IgG-Sepharose columns and protein A∞ purification and quantification was carried out as described previously (Mottagui-Tabar et al., 1994). The E. coli plasmids were transformed into the restriction-defective Salmonella strain UX863. Protein A∞ assays were performed in minimal M9 medium as described earlier for E. coli (Mottagui-Tabar et al., 1994; Bjo¨rnsson et al., 1996).

3. Results An in-vivo assay system for translation termination has been used, based on a repeated semi-synthetic

domain (A∞) from the protein A in Staphylococcus aureus. By using constructs that had a stop codon between the second and third coding repeats, the relative molar quantities of the truncated (2A∞), resulting from termination, versus the full-length (3A∞) product, resulting from readthrough by a near-cognate tRNA, can be measured. The molar ratio [3A∞]/[2A∞] can be used to determine the transmission ( T ) value, which reflects the level of readthrough. The UGAA stop signal was initially studied in E. coli since, compared to UAGA and UAAA, it was a weaker stop signal. The resulting high readthrough allowed for adequate amounts of readthrough products to be synthesized for quantification using gel electrophoresis. For the purposes of consistency and direct comparison, the 3A∞ reporter gene was subcloned from constructs used in earlier work (MottaguiTabar et al., 1994; Bjo¨rnsson et al., 1996), which had different contexts at the 5∞ side of the stop codon in the

192

S. Mottagui-Tabar, L.A. Isaksson / Gene 212 (1998) 189–196

linker region. Plasmids with the 3A∞ gene derivatives were introduced into B. subtilis, S. typhimurium and E. coli, and readthrough was measured. Thus, the influence on termination efficiency by varying codons in the −2 and −1 positions immediately upstream of the stop codon has been tested. The protein A∞ reporter system does not carry any transcriptional signals in E. coli that could influence the readthrough measurements, and the protein is stable (Bjo¨rnsson et al., 1998). The total amount of protein A∞ product from equal volumes of cell culture was approximately the same for the three species, indicating that the protein was stable and that the influence from transcriptional polarity, if any, was negligible in all three species. Fig. 1 describes schematically the plasmid vector used for cloning and protein assays in B. subtilis as well as the arrangement of the A∞ repeats. Fig. 2 shows the results of SDS–PAGE, giving the 2A∞ and 3A∞ protein bands that are used for quantitative measurement of readthrough in B. subtilis 3G18. Table 1 shows the influence of the last amino acid in the nascent peptide and/or the P-site codons/tRNA on UGA codon readthrough for 12 different −1 codons as measured in E. coli, B. subtilis and S. typhimurium. The results suggest that translation termination in B. subtilis is inefficient (allowing for high readthrough) if the −1 codon codes for a proline, or to some extent a threonine, in analogy to the results with E. coli (Mottagui-Tabar et al., 1994; Bjo¨rnsson et al., 1996). Readthrough in E. coli and B. subtilis was similar for most contexts, the exception being UGG, and was generally higher than those observed for S. typhimurium for the same contexts. In an earlier study using an E. coli strain carrying a trpT(Su9) UGA suppressor tRNA, the transmission

Fig. 2. SDS–PAGE gel electrophoresis of readthrough products (3A∞) and termination products (2A∞) from B. subtilis 3G18, for constructs having different −1 and −2 amino acids in the nascent peptide. Lane 1 shows the low-molecular-weight protein marker, from MW 94 000 (phosphorylase B) to MW 14 400 (a-lactalbumin) followed by an empty lane. Lanes 2–9 show the difference in readthrough due to changes in the amino acid in position −2, and lanes 10–16 show the influences of different −1 codons on translation termination. The stop signal in all constructs is UGAA. For details of construction of plasmids and for protein assay and extraction, see the Section 2.

values associated with these 12 amino acids at the C-terminus of the nascent peptide gave a significant correlation with their propensity to appear in natural ahelix and b-strand structures (Bjo¨rnsson et al., 1996). In the present study, suppressor free strains of E. coli, B. subtilis and S. typhimurium have been used. In such strains, the readthrough values are too low to allow for an attempt to correlate readthrough with any physical or chemical property of the last amino acid in the nascent peptide. Table 2 shows the result of a test of all four glycine codons (GGC/U/A/G) in the −1 position immediately upstream of the stop codon, since earlier reports have indicated that glycine tRNAs ( Komine et al., 1990) are involved in the 5∞ influence on termination efficiency (Bjo¨rnsson et al., 1996; Zhang et al., 1996). In S. typhimurium, the readthrough values are very low, making any comparison between different glycine codons difficult. However, there is a difference of up to twofold between GGU/C and GGA/G, in agreement with the results observed for E. coli. In B. subtilis, the glycine codons do not give any difference in influence on readthrough ( Table 1; first four rows). A test for cooperativity between the effects of changed −1 and −2 codons on readthrough was carried out by replacing the codon for aspartic acid (GAC ) in the −2 position with an arginine codon (CGC ) in combination with different −1 codons ( Table 2). The constructs with CGC as the −2 codon showed similar low readthrough values when E. coli and B. subtilis were compared. However, the same four −1 codons, when preceded by a codon for an acidic amino acid ( last four lines), gave significantly increased readthrough values for E. coli, B. subtilis and S. typhimurium. Thus, the −1 and −2 codons influenced translation termination, in a cooperative manner, in all three species. Fig. 3 shows the results of correlating the isoelectric point of −2 amino acid residues in the nascent peptide to transmission values. For B. subtilis, acidic amino acids like glutamic acid allow for high readthrough and basic amino acids like arginine favor termination, thus decreasing readthrough at a stop codon. These results are similar to E. coli termination data (Mottagui-Tabar et al., 1994; Bjo¨rnsson et al., 1996), which are included for the sake of comparison in Fig. 3. Thus, the influence of the penultimate amino acid on translation termination in E. coli and B. subtilis is similar. Phenylalanine at position −2 provides an exception since it is associated with an extraordinarily high readthrough in B. subtilis. Readthrough in S. typhimurium is drastically different from that of E. coli and B. subtilis since the dependence on the penultimate amino acid is quite low. Thus, the translation termination in S. typhimurium is almost five times less sensitive to 5∞ codon context influences than it is in E. coli. The influence of S. typhimurium RF-2 on translation

193

S. Mottagui-Tabar, L.A. Isaksson / Gene 212 (1998) 189–196 Table 1 Influence of different −1 codons on UGAA decoding Codons upstream of UGAA

GAC GAC GAC GAC GAC GAC GAC GAC GAC GAC GAC GAC

CCA CGC AAG UGG GCA CAG UUC AGU ACU GUG CAC GGU

Amino acids context

Transmission ( T )

D–P D–R D–K D–W D–A D–Q D–F D–S D–T D–V D–H D–G

E. coli

B. subtilis

S. typhimurium

0.91±0.017 0.05±0.006 0.09±0.012 0.47±0.001 0.10±0.008 0.08±0.012 0.06±0.006 0.06±0.012 0.26±0.009 0.04±0.012 0.10±0.013 0.09±0.019

0.71±0.034 0.03±0.008 0.05±0.020 0.10±0.005 0.07±0.023 0.03±0.013 0.10±0.004 0.10±0.006 0.30±0.003 0.05±0.004 0.15±0.006 0.10±0.021

0.12±0.025 0.03±0.007 0.04±0.010 0.03±0.003 0.02±0.004 0.04±0.008 0.03±0.006 0.02±0.010 0.05±0.006 0.03±0.012 0.06±0.005 0.02±0.004

Influence of changed 5∞ neighbouring codons on translation termination efficiency in E. coli, B. subtilis and S. typhimurium. Suppressor-free strains representing all three species were used for protein A∞ assays. The T values are averages of four independent measurements. Standard errors are given.

Table 2 5∞ codon context influences on UGAA decoding Amino acids

D–G D–G D–G D–G R–V R–S R–P R–G D–V D–S D–P D–G

5∞ context

GAC GGC GAC GGU GAC GGA GAC GGG CGC GUG CGC AGU CGC CCA CGC GGU GAC GUG GAC AGU GAC CCA GAC GGU

E. coli

S. typhimuriuma

B. subtilis

Plasmid

T

T

Plasmid

T

pMB51 pMB50 pMB60 pMB61 pAB121 pAB115 pSM27 pMB76 pAB120 pAB109 pSM11 pMB50

0.07±0.006 0.09±0.019 0.21±0.019 0.16±0.013 0.02±0.005 0.02±0.009 0.03±0.006 0.03±0.007 0.04±0.012 0.06±0.012 0.91±0.017 0.09±0.019

0.015±0.002 0.016±0.004 0.030±0.007 0.024±0.002 n.d. n.d. 0.03±0.006 n.d. 0.03±0.012 0.02±0.010 0.12±0.025 0.016±0.004

pBSM44 pBSM26 pBSM33 pBSM45 pBSM28 pBSM38 pBSM05 pBSM48 pBSM47 pBSM42 pBSM04 pBSM26

0.12±0.010 0.10±0.021 0.11±0.031 0.12±0.015 0.02±0.004 0.03±0.002 0.03±0.006 0.02±0.005 0.05±0.004 0.10±0.006 0.71±0.034 0.10±0.021

5∞ codon context influence on termination efficiency. The last two amino acids in the nascent peptide and codon contexts immediately upstream of UGAA are shown. Cooperativity between the −1 and −2 codon alterations is shown by the constructs with CGC (arginine) or GAC (aspartic acid) at position −2 together with different −1 codons. Readthrough is given as transmission values [T=(3A∞)/(2A∞)]. n.d., not determined. Standard errors are given. aThe plasmids used in S. typhimurium are the same as in E. coli.

termination in E. coli has been reported earlier (Matsumura et al., 1996). According to those results, the plasmid borne S. typhimurium prfB+ gene product reduced the readthrough of stop codons in E. coli and complemented the temperature sensitive phenotype of a prfB mutant E. coli. The result presented here that readthrough at UGA in S. typhimurium is generally much lower than that in E.coli, is in agreement with the work of Matsumura and collaborators. Fig. 4 shows the alignment of the RF-2 amino acid sequences from the three bacterial species analysed here. The alignment shows that the residue at position 246 (bold letters) is the same for RF-2 from both S. typhimurium and B. subtilis. The figure also shows the residues

that are common to E. coli and B. subtilis (bold letters) but are different in S. typhimurium. The boxed residues are different in all three factors.

4. Discussion Extensive studies on 3∞ context effects on amber codons located at different positions in the lacI gene in E. coli and S. typhimurium suggest that the context influence on translation termination in the two organisms is similar (Sharp, 1991). Although no such genetic analysis has been made for the Gram-positive bacterium B. subtilis, statistical analysis (Brown et al., 1990) of

194

S. Mottagui-Tabar, L.A. Isaksson / Gene 212 (1998) 189–196

Fig. 3. Correlations between stop codon readthrough ( T values) and isoelectric points for amino acid residues (Hardy, 1985) in the penultimate position of the nascent peptide. Fifteen or 20 different amino acids were analyzed in B. subtilis and E. coli, respectively. Only the indicated charged amino acids and two neutral ones were analyzed for S. typhimurium. Transmission values for E. coli (0) and B. subtilis 3G18 ( ) and S. typhimurium (Ω) are plotted against the isoelectric point of the penultimate amino acid in the nascent peptide, and the linear correlation is indicated (E. coli, ———; B. subtilis, – – –; and S. typhimurium, - - -). Standard errors for the mean (SEM ) of E. coli determinations are <13% (Bjo¨rnsson et al., 1996; Mottagui-Tabar et al., 1994). The T values for B. subtilis and S. typhimurium are the average of four independent assays with a SEM of <20%.

codon frequencies on the 3∞ side of stop codons shows great similarities to E. coli. Stop codon usage is similar between the two organisms, although synonymous codon usage differs (Sharp and Bulmer, 1988; Brown et al., 1990; Sharp and Burgess, 1992). For example, the four glycine codons are used differently in highly and lowly expressed genes in the two bacteria. There is an approximately 500-fold over-representation of the codons GGC and GGU (decoded by tRNAgly3) compared to GGA (tRNAgly2) and GGG (tRNAgly1,2) ( Komine et al., 1990) in highly expressed genes in E. coli. The same codons in similar genes in B. subtilis have a difference in bias of only 20-fold Yarus and Curran, 1992). Evolutionary divergence, however, has placed E. coli and B. subtilis in separate phyla ( Woese, 1987). The differences have been reflected, in part, in differences in the translational apparatus. Bacillus may be devoid of S1 protein (Schnier and Faist, 1985), although recent reports indicate the possible presence of an S1 homologue (Sorokin et al., 1996). The S1 protein in E. coli is the largest ribosomal protein and is involved in the initiation of translation (Subramanian, 1984) and translational specificity (Roberts and Rabinowitz, 1989). Bacillus uses extended Shine–Dalgarno elements for efficient initiation (McLaughlin et al., 1981).

We have adapted the E. coli 3A∞ translation assay system to B. subtilis and measured the influence of the 5∞ codon context on UGAA decoding. We find that this influence in B. subtilis is similar to the earlier published results from E. coli (Mottagui-Tabar et al., 1994; Bjo¨rnsson et al., 1996). Some codons at the P-site in E. coli (position −1) give a high readthrough, like CCA, UGG, and to some extent ACU. Both CCA and ACU give a high readthrough in wild-type B. subtilis as well. However, the readthrough for these contexts is very low when measured in S. typhimurium. Thus, translation termination in S. typhimurium does not show the same sensitivity to the 5∞ codon context as that observed in E. coli and B. subtilis. In the two latter species, the −2 amino acids influence the readthrough depending on their charge ( Fig. 3; Mottagui-Tabar et al., 1994). A negatively charged amino acid gives a low RF2-mediated termination and positively charged side chains favor translation termination. Our data also suggest that the influence from different −1 and −2 codons is cooperative in B. subtilis, similar to that observed for E. coli, since they combine to give the 5∞ context effect on termination. Release factors RF-1 ( UAG, UAA) and RF-2 ( UGA, UAA) have common properties with regard to ribosome binding and peptidyl tRNA hydrolysis but exhibit differences in codon discrimination and decoding rates (Craigen et al., 1985; Nakamura et al., 1996). Termination at UGA (RF-2) is more sensitive to the charge of the penultimate amino acid in the nascent peptide than is termination at UAG (RF-1) (MottaguiTabar et al., 1994). The amino acid sequence similarity between the RF-1 and RF-2 can vary between 43% and 92% (Mikuni et al., 1991), depending on the region of the amino acid sequence being compared, with an overall similarity of 31%. Therefore, the highly conserved areas of the factors are probably involved in common activities like peptidyl-tRNA hydrolysis, and the dissimilar areas of the sequence would perhaps be responsible for the diverse activities like codon discrimination. The RF-2 amino acid sequence homology is 68% between E. coli and B. subtilis, indicating a certain conservation of function and structure. The RF-2 sequence homology between the two Gram-negative bacteria E. coli and S. typhimurium is 95.6%, leaving only 16 amino acids as possible candidates for the low nascent peptide sensitivity on termination in S. typhimurium when compared to E. coli and B. subtilis. As shown here, termination at UGAA in E. coli and B. subtilis RF-2 displays a high degree of functional similarity with regard to the 5∞ context influence, being different from S. typhimurium. The regions of the RF-2 molecules that could be sensitive to the 5∞ codon context are possibly located in the sequence that is 68% conserved in these two organisms. In S. typhimurium, termination is more accurate than in E. coli. Overproduction of RF-2 is

S. Mottagui-Tabar, L.A. Isaksson / Gene 212 (1998) 189–196

195

Fig. 4. Protein sequences for the factor RF-2 of E. coli, B. subtilis and S. typhimurium. Sequences were obtained from the GMBL database (Accession No. p07012 and p28353, respectively). The alignment was performed using the Pileup program provided by GCG Inc. The complete sequence of B. subtilis RF-2 was kindly provided by Dr V. Lazarevic (Inst. de Ge´ne´tique et de Biologie Microbiennes, Lausanne, Switzerland ). The boxed residues are different in all three sequences, and the bold amino acids are the same in E. coli and B. subtilis RF-2. The residues at position 246 in all three sequences are shown in bold. The acidic amino acids group together in three regions of the sequence marked (*).

toxic in E. coli but not in S. typhimurium ( Uno et al., 1996). The residue at position 246 (threonine) of the E. coli RF-2 has been reported to be solely responsible for both the difference in accuracy and toxicity. Thus, E. coli RF-2 has a threonine in position 246, giving it a low termination activity and high toxicity upon overproduction, whereas S. typhimurium has alanine at the same position making it more active and non-toxic at higher concentrations ( Uno et al., 1996). However, B. subtilis also has an alanine at position 246, like S. typhimurium. The residue in position 246 can thus only partly explain the difference in nascent peptide sensitivity between these species. If RF-2 is the target molecule, there might be other areas of RF-2 that confer the peptide sensitivity that are similar for B. subtilis and E. coli but different in S. typhimurium. In E. coli, an excess of RF-2 fails to reduce the aminoacid-dependent influence of the nascent peptide on termination. Also, a changed expression level of the 3A∞ gene with an internal stop codon does not affect the

level of readthrough (data not shown). Thus, possible differences in RF-2 or 3A∞ mRNA concentrations between the three species are unlikely to be a factor influencing peptide sensitivity.

Acknowledgement We thank Mounia Heddad and Maria Ohlsson for help with the subcloning of intermediate plasmids. The complete RF-2 sequence of B. subtilis has been reproduced with the permission of Dr V. Lazarevic, Inst. de Ge´ne´tique et de Biologie Microbiennes, Switzerland. We also thank Dr Harald Putzer, IBPC, Paris, for providing advice and material for Bacillus subtilis plasmids. We thank Steven Muir and Derek Logan for their comments on the manuscript. This work has been supported by grants from the Swedish Natural Science Foundation (NFR) to Dr Leif A. Isaksson.

196

S. Mottagui-Tabar, L.A. Isaksson / Gene 212 (1998) 189–196

References Bjo¨rnsson, A., Isaksson, L.A., 1988. Test system for measurement of translational activity in vivo. Nucleosides and Nucleotides 7, 565–569. 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. Bjo¨rnsson, A., Mottagui-Tabar, S., Isaksson, L.A., 1996. Structure of the C-terminal end of the nascent peptide influences translation termination. EMBO J. 15, 1696–1704. Bjo¨rnsson, A., Mottagui-Tabar, S., Isaksson, L.A., 1998. The analysis of translational activity using a reporter gene constructed from repeats of an antibody-binding domain from Protein A. Meth. Mol. Biol. 77, Chapter 6, in press. Bonetti, B., Fu, L., Moon, J., Bedwell, D.M., 1995. The efficiency of translation termination is determined by a synergistic interplay between upstream and downstream sequences in Saccharomyces cerevisiae. J. Mol. Biol. 251, 334–345. Brown, C.M., Stockwell, P.A., Trotman, C.N.A., Tate, W.P., 1990. The signal for the termination of protein synthesis in procaryotes. Nucleic Acids Res. 18, 2079–2086. Buckingham, R.H., 1990. Codon context. Experientia 46, 1126–1133. Craigen, J.W., Cook, R.G., Tate, W.P., Caskey, C.T., 1985. Bacterial peptide chain release factors: Conserved primary structure and possible frameshift regulation of release factor 2. Proc. Natl. Acad. Sci. 82, 3616–3620. Grentzmann, G., Brechemier-Baey, D., Heurgue, V., Mora, L., Buckingham, R.H., 1994. Localization and characterization of the gene encoding release factor RF3 in Escherichia coli. Proc. Natl. Acad. Sci. USA 91, 5848–5852. Hardy, P.M., 1985. The protein amino acids. In: Barrett, G.C. ( Ed.), Chemistry and Biochemistry of the Amino Acids. Chapman & Hall, London, pp. 7–24. Hughes, D., Atkins, J.F., Thompson, S., 1987. Mutants of elongation factor Tu promote ribosomal frameshifting and nonsense readthrough. EMBO J. 6, 4235–4239. Komine, Y., Adachi, T., Inokuchi, H., Ozeki, H., 1990. Genomic organization and physical mapping of the transfer RNA genes in Escherichia coli K12. J. Mol. Biol. 212, 576–598. Konan, K.V., Yanofsky, C., 1997. Regulation of the Escherichia coli tna operon: nascent leader peptide control at the tnaC stop codon. J. Bacteriol. 179, 1774–1779. Lovett, P.S., Rogers, E.J., 1996. Ribosome regulation by the nascent peptide. Microbiol. Rev. 60, 366 McLaughlin, J.R., Murray, C.L., Rabinowitz, J.C., 1981. Unique features of the ribosomal binding site sequence of the Gram positive Staphylococcus aureus b-lactamase gene. J. Biol. Chem. 256, 11283–11291. 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., Kawakami, K., Nakamura, Y., 1991. Sequence and functional analysis of mutations in the gene encoding peptidechain-release factor-2 of Escherichia coli. Biochimie 73, 1509–1516. Mikuni, O., Ito, K., Moffat, J., Matsumura, K., McCaughan, K., Nobukuni, T., Tate, W., Nakamura, Y., 1994. Identification of the prfC gene, which encodes peptide-chain-release factor 3 of Escherichia coli. Proc. Natl. Acad. Sci. USA 91, 5798–5802.

Moffat, J.G., Tate, W.P., Lovett, P.S., 1994. The leader peptides of attenuation-regulated chloramphenicol resistance genes inhibit translational termination. J. Bacteriol. 176, 7115–7117. Mottagui-Tabar, S., Bjo¨rnsson, A., Isaksson, L.A., 1994. The second to the last amino acid in the nascent peptide as a codon context determinant. EMBO J. 13, 249–257. Mottagui-Tabar, S., Isaksson, L.A., 1996. Influence of the last amino acid in the nascent peptide on EF-Tu during decoding. Biochimie 78, 953–958. Mottagui-Tabar, S., Isaksson, L.A., 1997. Only the last amino acids in the nascent peptide influence translation termination in Escherichia coli genes. FEBS Lett. 414, 165–170. Nakamura, Y., Ito, K., Isaksson, L.A., 1996. Emerging understanding of translation termination. Cell 87, 147–150. Pedersen, W.T., Curran, J.F., 1991. Effects of the nucleotide-3∞ to an amber codon on ribosomal selection rates of suppressor transfer RNA and release factor-1. J. Mol. Biol. 219, 231–241. Poole, E.S., Brown, C.M., Tate, W.P., 1995. The identity of the base following the stop codon determines the efficiency of in vivo translational termination in Escherichia coli. EMBO J. 14, 151–158. Roberts, M.W., Rabinowitz, J., 1989. The effect of E. coli ribosomal protein S1 on translational specificity of bacterial ribosomes. J. Biol. Chem. 264, 2228–2235. 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. Sambrook, J., Fritsch, E.F., Maniatis, T. ( Eds.), 1989. Molecular Cloning: A Laboratory Manual, 2nd ed. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Schnier, J., Faist, G., 1985. Comparative studies on the structural gene for the ribosomal protein S1 in ten bacterial species. Mol. Gen. Genet. 200, 476–481. Sharp, P.M., Bulmer, M., 1988. Selective differences among translation termination codons. Gene 63, 141–145. Sharp, P.M., 1991. Determinants of DNA sequence divergence between Escherichia coli and Salmonella typhimurium: codon usage, map position and concerted evolution. J. Mol. Evol. 33, 23 Sharp, P.M., Burgess, C.J., 1992. Selective use of termination codons and variations in codon choice. In: Hatfield, D.L., Lee, B.J., Pirtle, R.M. ( Eds.), Transfer RNA in Protein Synthesis. CRC Press, Boca Raton, FL, pp. 397–425. Sorokin, A., Serror, P., Pujic, P., Azevedo, V., Ehrlich, S.D., 1996. The Bacillus subtilis chromosome region encoding homologues of the Escherichia coli mssA and rpsA gene products. Microbiology 141, 311–319. Subramanian, A.R., 1984. Structure and function of the largest Escherichia coli ribosomal protein. Trends Biochem. Sci. 9, 491–494. 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, 935–944. Woese, C.R., 1987. Bacterial Evolution. Microbiol. Rev. 51, 221 Yarus, M., Curran, J., 1992. The translational context effect. In: Hatfield, D.L., Lee, B.J., Pirtle, R.M. ( Eds.), Transfer RNA in Protein Synthesis. CRC Press, Boca Raton, FL, pp. 319–365. Zhang, S.P., Ryde´n-Aulin, M., Isaksson, L.A., 1996. Functional interaction between release factor one and P-site peptidyl-tRNA on the ribosome. J. Mol. Biol. 261, 98–107.