Simultaneous binding of trigger factor and signal recognition particle to the E. coli ribosome

Biochimie 86 (2004) 495–500 www.elsevier.com/locate/biochi

Simultaneous binding of trigger factor and signal recognition particle to the E. coli ribosome Amanda Raine a, Natalia Ivanova b, Jarl E.S. Wikberg a, Måns Ehrenberg b,* a b

Department of Pharmaceutical Biosciences, Biomedical Centre, Box 591, Uppsala University, 751 24 Uppsala, Sweden Department of Cell and Molecular Biology, Biomedical Centre, Uppsala University, Box 596, 751 24 Uppsala, Sweden Received 16 March 2004; accepted 7 May 2004 Available online 07 June 2004

Abstract Signal recognition particle (SRP) and trigger factor (TF) both bind to ribosomal protein L23 at the peptide exit area on the 50S subunit of the E. coli ribosome. In this study, we have developed a spin-down assay and used it to estimate KD values and the corresponding enthalpies for the binding of radio-labelled SRP and TF to naked ribosomes and to ribosomes carrying a tetrapeptidyl-tRNA in the P site. At 20 °C, the KD value for TF binding is 2 µM and for SRP it is 170 nM for naked as well as for translating ribosomes. At 4 °C, the KD values for TF and SRP binding are 1.1 µM and 90 nM, respectively. Competition binding experiments reveal that SRP and TF bind simultaneously to the ribosome with little affinity interference, showing that the factors have separate binding sites on L23. This makes an alternating binding mode for TF and SRP less plausible. © 2004 Elsevier SAS. All rights reserved. Keywords: Trigger factor; Signal recognition particle; Ribosome; Exit tunnel; L23; Protein synthesis; Protein export

1. Introduction Proteins L23 and L29 are neighbours on the surface of the 50S subunit of the E. coli ribosome; near the peptide exit region [1]. Recent experimental data suggest that L23 is the major attachment point for the signal recognition particle (SRP) and the chaperone trigger factor (TF) [2–4], which both bind co-translationally to nascent chains on the ribosome [5,6]. In E. coli, the SRP is composed of a 48 kDa GTPase protein (Ffh) and a 4.5S RNA. SRP binds to ribosomenascent-chain complexes exposing hydrophobic signal sequences and secures targeting of inner membrane proteins to the cell membrane [6]. The delivery of the SRP-ribosomenascent-chain complex to the translocon in the inner membrane is mediated by the interaction of SRP with its receptor FtsY, according to a mechanism that has remained obscure.

Abbreviations: TF, trigger factor; SRP, signal recognition particle; Ffh, fifty-four-homologue; BSA, bovine serum albumine. * Corresponding author. Tel.: +46-18-4714213; fax: +46-18-4714262. E-mail address: [email protected] (M. Ehrenberg). 0300-9084/$ - see front matter © 2004 Elsevier SAS. All rights reserved. doi:10.1016/j.biochi.2004.05.004

TF is a ribosome-associated chaperone with peptidylprolyl isomerase (PPiase) activity that assists in the folding of nascent chains [5], but its PPiase function is not essential for its protein folding activity [7]. In fact, the TF can bind to proline lacking nascent chains, with a preference for short peptide sequences with aromatic and basic residues [8]. In addition to its L23 binding, the TF also interacts with ribosomal protein L29, although this contact is dispensable for its binding to the ribosome [2]. Eukaryotic SRP54 binds to the eukaryotic equivalents of L23 and L29 and its cross-linking pattern is shifted towards L29 upon complex formation with the SRP receptor [9]. However, no interaction between E. coli SRP and L29 have been detected so far in cross-linking studies [3,4], either in the absence of the SRP receptor FtsY, or upon complex formation between SRP and FtsY on the ribosome. In this study we have developed a spin-down assay and used it to estimate the dissociation constants and the corresponding enthalpies for the binding of radio-labelled SRP and TF to naked ribosomes and to ribosomes that carry a tetrapeptidyl-tRNA in the P site. The KD value for TF binding to the ribosome is close to an earlier estimate based on fluorescence spectroscopy [10]. The KD value for the binding

496

A. Raine et al. / Biochimie 86 (2004) 495–500

of SRP to the E. coli ribosome has not been reported before and is shown to be close to the KD value for the binding of SRP to the 80S ribosome in a eukaryotic system [11]. Since TF and SRP share L23 as their major attachment site on the ribosome [2–4,9], it has been suggested that their binding sites are overlapping and that they cannot be present on the ribosome at the same time [3]. Here, we have found that TF and SRP bind simultaneously to the naked ribosome with little binding interference, suggesting separate binding sites on L23 for the two factors.

2. Experimental procedures 2.1. Chemicals, enzymes, E. coli strains and plasmids Radioactive compounds, enzymes and nucleotides were from Amersham Pharmacia (Uppsala, Sweden). All other chemicals were from Sigma. Ribosomes were purified from the E. coli strain MRE600. The plasmid pET21Ffhhis6 was used for expression of Ffh in E. coli strain BL21(DE3). 2.2. SRP, TF and ribosomes His tagged Ffh was expressed and purified as described [12]. [14C]-labelled 4.5S RNA was transcribed in vitro using T7 polymerase. SRP was reconstituted by incubating Ffh and 4.5S RNA in a 1:2 molar ratio in polymix buffer [13] at 20 °C for 5 min. His tagged TF was overexpressed and purified according to Hesterkamp et al. [14]. For [35S]-methionine labelling, cells were grown in 1 l of M9 media supplemented with 0.05% casamino acids and 0.08 mg/ml each of proline, leucine and tryptophane until OD600 reached 0.6. Then, 5 mCi [35S]-methionine was added and TF expression was induced with 1 mM IPTG. The [35S]-TF was purified to >95% purity as described above. The concentrations of non-labelled and [35S]-TF were determined spectrophotometrically using the extinction coefficient 15930 M–1 cm–1 at 280 nm. The 70S ribosomes were prepared by zonal centrifugation [15]. Ribosomal complexes (RCs) containing fMet-Phe-Thr-IletRNAIle in the P site were prepared in an E. coli system for mRNA translation with pure components [16]. The RCs were purified from other components by gel filtration through a Sephacryl S300 column [17] or by ultra centrifugation through 3 ml 1.1 M sucrose in polymix for 1.5 h at 230 000g. In both cases, from RCs with [14C]-Ile residues we found that more than 85% of the ribosomes carried fMet-Phe-Thr-IletRNAIle. 2.3. Ribosome binding assay Binding of SRP or TF to ribosomes was quantified by, first, incubating either 50 nM [14C]-SRP or 250 nM [35S]-TF with varying concentrations of ribosomes. The reactions were carried out in polymix buffer [13] in 100 µl volumes in

the presence of 0.2 mg/ml BSA, either at 4 or 25 °C during 10 min. Ribosomes, free or in complex with TF or SRP, were then pelleted by ultracentrifugation (Sorvall RCM150GX, 315 000g, S100-AT3 rotor) in 10 min, a time sufficient for pelleting all ribosomes, and at the same temperature (4 or 20 °C) as during the previous incubation step. Control samples without ribosomes, taken in parallel, showed insignificant background levels for TF or SRP. The supernatants were quickly removed with a vacuum devise and the pellets were dissolved in H2O for spotting on Whatman GF/C filters, that were dried and counted in a ProteinReady scint coctail (Beckmann Coulter) in a BeckmannCoulter LC6500 scintillaton counter. To obtain the dissociation constants (KD), the expression: 关 RS 兴 =

关 R 兴 · 关 S0 兴 共 关 R 兴 + KD 兲

was fitted to the experimental data with non-linear regression (Origin 7 software, OriginLab Corporation) by variation of the parameter KD. [RS] is the concentration of ligand (SRP or TF) in complex with ribosomes as measured from the pellet. [R] is the concentration of free ribosomes, [S0] is the total concentration of either ligand. The method measures the true KD value at the temperature of the centrifuge, provided that complex incorporation in the pellet is irreversible, and that ribosomes are spun down faster than the ligands. Equilibrium conditions pertain when sedimentation is much faster than diffusion, since then the amounts of ribosome complex and free ribosome that enter a small volume during sedimentation are exactly equal to the amounts that leave this volume, so that all equilibrium relations are maintained throughout a centrifugation run. The KD values and standard deviations were determined from a minimum of three independent titrations. Competition experiments were performed by (i) incubating 3 µM [35S]-TF and 0.5 µM 70S ribosomes with varying concentrations of non-labelled TF or SRP and (ii) by incubating 0.3 µM [14C]-SRP and 0.4 µM ribosomes with varying concentrations of non-labelled TF. Mixes, containing 0.5 µM SRP, 3 µM [35S]-TF, 0.5 µM ribosomes and 100 µM GTP, were titrated with unlabeled SRP receptor (FtsY). Bound and unbound SRP and TF were separated from ribosomes and analysed as described above.

3. Results 3.1. Dissociation constants for the binding of SRP and TF to naked ribosomes To determine the dissociation constants (KD) for the binding of SRP or TF to ribosomes, we used radio-labelled factors in a spin down assay. [35S]-methionine labelled E. coli TF ([35S]-TF) was expressed and purified (Experimental procedures). [35S]-TF at 250 nM concentration was titrated with naked ribosomes from zero to 5 µM concentra-

A. Raine et al. / Biochimie 86 (2004) 495–500

tion, incubated during 10 min at either 4 or 20 °C and spun down in a centrifuge, that held the same temperature as used in the previous incubation step. From the binding curves (Fig. 1A), the KD for the binding of TF to naked ribosomes was estimated by non-linear regression to be 1.1 µM at 4 and 2 µM at 20 °C (Table 1). The enthalpy of binding (DH = Hcomplex – Hfree) is estimated from these data to be 25.2 kJ/mole. Non-labelled TF displaced [35S]-TF from the ribosome (Fig. 1B), suggesting specific ribosomal binding of the TF. The spin down method developed here is an authentic equilibrium method (Experimental procedures), in line with the correspondence between our estimate of the KD value, and a previous estimate of KD = 1 µM at 20 °C based on fluorescence techniques [10]. Radio-labelled SRP ([14C]-SRP) was reconstituted from Ffh and a 4.5S RNA that had been uniformly labelled from [14C]-ATP. In the absence of Ffh, the [14C]-4.5S RNA had no detectable affinity to naked ribosomes (data not shown).

497

Table 1 The dissociation constants for SRP and TF binding to ribosomes at 4 and 20 °C Complex SRP–70S SRP–GTP–70S SRP–GDP–70S SRP–MFTI-RNC TF–70S

KD (µM) 4°C 0.09 ± 0.02 0.11 ± 0.02 0.09 ± 0.02 1.1 ± 0.3

DH (kJ/mole) 20°C 0.17 ± 0.02

26.8

0.18 ± 0.03 2.0 ± 0.3

25.2

All KD values were calculated from at least three independent experiments.

[14C]-SRP at 50 nM concentration was titrated with ribosomes from 0 to 1.5 µM concentration at 4 or 20 °C. From the binding curves (Fig. 2A), the KD value for the binding of SRP to naked ribosomes was estimated to be 90 nM at 4 °C and 170 nM at 20 °C. These results, estimating the enthalpy of the binding to be 26.8 kJ/mole are in line with a previous estimate of 70 nM for the dissociation constant for the binding of eukaryotic SRP to 80S ribosomes at 4 °C [11]. The same type of experiments were also performed in the presence of either 100 µM GTP or 100 µM GDP, and identical binding curves and dissociation constants were obtained (Table 1). 3.2. SRP binding to translating ribosomes It has been observed that binding of mammalian SRP to wheatgerm ribosomes, that contain a peptidyl-tRNA with a peptide that is either too short to traverse the 60S subunit or that traverses the tunnel but lacks signal sequence, is an order of magnitude stronger than the binding of SRP to naked 80S ribosomes [11]. This could mean that ribosomes in elongation phase have a conformation with higher affinity to SRP than have naked ribosomes. To test if this is the case in E. coli, 70S ribosome complexes carrying fMet-Phe-Thr-Ile-tRNAIle in the P site were prepared, and their affinity to SRP was compared to that of naked ribosomes. The binding affinities of translating and naked ribosomes to SRP were not significantly different (Fig. 2B and Table 1), suggesting that, in E. coli, translation per se does not bring about a conformational change of the ribosome that affects the strength of the binding of SRP at the peptide tunnel exit. 3.3. Simultaneous binding of TF and SRP to ribosomes

Fig. 1. TF binding to naked ribosomes at 4 °C as determined by a spin down assay. (A) Titration of radio-labelled TF (250 nM) with increasing concentrations of naked ribosomes (x-axis). The amount of TF–ribosome complexes (y-axis) is expressed as % bound TF. The KD value was determined to 1.1 µM by non-linear regression. (B) Titration of radio-labelled TF and naked ribosomes with non-labelled TF (x-axis). The amount of radiolabelled TF–ribosome complexes (y-axis) is expressed as the % of radiolabelled TF remaining ribosome bound.

The TF and SRP both have ribosomal protein L23 as major attachment site [2–4,9]. This could mean that the binding sites of TF and SRP overlap [3,4], so that they cannot simultaneously be present on a ribosome particle and that they alternate in their binding to the peptide exit site. To test if SRP and TF exclude each other’s binding to the naked ribosome, we performed competition experiments. First, [14C]-SRP at 0.3 µM and ribosomes at 0.4 µM were titrated with non-labelled TF from 0 to 30 µM concentration (Fig. 3B). The KD values for SRP and TF binding to the naked ribosome are 0.09 and 1.1 µM, respectively (Table 1).

498

A. Raine et al. / Biochimie 86 (2004) 495–500

Fig. 2. SRP binding to naked and translating ribosomes at 20 °C. (A) Titration of radio-labelled SRP (50 nM) with increasing concentrations of naked ribosomes (x-axis). The amount of SRP–ribosome is expressed as % bound SRP (y-axis). The KD value for SRP binding (170 nM) was obtained by non-linear regression. (B) Radio-labelled SRP was titrated with ribosomal complexes containing fMet-Phe-Thr-Ile-tRNA in the P site (x-axis). The KD value for SRP binding (180 nM) was obtained by non-linear regression.

From these values one can predict from the exact solution of the equilibrium binding relations that if TF and SRP binding sites were mutually exclusive, there would be a 5.5-fold reduction in the amount of [14C]-SRP in complex with ribosomes at the highest TF concentration. In contrast, we observed only a 25% reduction of SRP binding (Fig. 3B) The small decrease in SRP binding to the ribosome that was observed, could possibly be related to the propensity of TF to form homodimers in solution [10]. That is, if TF binds to the ribosome in dimeric form [18], this could sterically destabilise the binding of SRP to L23. Consistent with the finding that TF cannot displace SRP on the ribosome we observed that TF in complex with ribosomes cannot be displaced by SRP. In this reverse experiment, 3 µM [35S]-TF and 0.5 µM ribosomes were titrated with SRP from 0 to 2 µM concentration. If the factors had mutually exclusive binding sites on the ribosome, solving the

Fig. 3. Simultaneous binding of TF and SRP to the E. coli ribosome (4 °C). (A) TF binding (y-axis) plotted as a function of SRP concentration (x-axis). Increasing concentrations of non-labelled SRP were incubated with radiolabelled TF (3 µM) and ribosomes (0.5 µM). The amount of TF–ribosome complex is expressed as % of the binding in the absence of SRP. The plot shows that SRP does not displace TF from the ribosome. (B) SRP binding (y-axis) plotted as a function of TF concentration (x-axis). Increasing concentrations of non-labelled TF were incubated with radio-labelled SRP (0.3 µM) and ribosomes (0.4 µM). The amount of SRP–ribosome complex is expressed as % of the binding in the absence of TF.

equilibrium binding relations shows that there would have been a 7-fold decrease in the amount of ribosome bound TF in this experiment, but no significant reduction could be seen (Fig. 3A). The conclusion is that SRP and TF can bind simultaneously to the naked ribosome and that, accordingly, they must have separate binding sites on L23. When eukaryotic SRP interacts with its receptor protein SRa, the position of SRP is apparently shifted closer to ribosomal L35, the homologue of the eubacterial ribosomal protein L29, as judged from the results of cross-linking experiments [9]. If a similar scenario occurs in E. coli, i.e. that binding of FtsY (the eubacterial analogue of SRa) to ribosomes in complex with SRP brings the SRP closer to L29, this could have consequences for the interaction between SRP and TF on the ribosome. The reason is that TF is likely to be positioned close to L29 [2], and movement of

A. Raine et al. / Biochimie 86 (2004) 495–500

Fig. 4. FtsY does not displace TF binding on TF–SRP–ribosome complexes. Increasing concentrations of FtsY (x-axis) incubated with SRP (0.5 µM), GTP (100 µM), ribosomes (0.5 µM) and radio-labelled TF (3 µM). The amount of TF–ribosome complexes (y-axis) is expressed as % of the binding in the absence of FtsY.

SRP into this area by the action of FtsY could lead to a sterical clash between SRP and TF. To test this, ribosome complexes containing [35S ]-TF, unlabelled SRP and GTP were titrated with FtsY from 0 to 10 µM concentration. The results were, that if formation of the SRP–FtsY complex occurs on naked ribosomes it does not significantly destabilize the ribosomal binding of [35S ]-TF (Fig. 4).

4. Discussion In this study we used an E. coli system for protein synthesis with pure components to determine the binding affinities of SRP and TF to naked as well as to translating ribosomes carrying a tetrapeptidyl-tRNA in the P site. Measurements were performed at 4 and 20 °C with a spin-down assay for the pelleting of ribosomes in complex with radio-labelled SRP or TF. There is a weak temperature dependence of the dissociation constants for both factors. At 4 °C the KD value for TF is 1.1 µM and for SRP it is 90 nM. At 20 °C, the KD values for TF and SRP are 2 µM and 170 nM, respectively (Table 1). Our estimate of the KD value for TF binding to naked ribosomes (1.1 µM at 20 °C and 2 µM at 20 °C), based on the spin down assay, is close to a previous estimate of 1 µM at 4 °C, based on fluorescence spectroscopy [10]. The binding enthalpies determined in this study may be used to extrapolate the binding affinities at 4 and 20 °C to a more physiological relevant temperature. Accordingly, at 37 °C the KD values are estimated to be 313 nM and 3.5 µM for SRP and TF, respectively. The KD value and the corresponding enthalpy reported here for the E. coli SRP–70S ribosome complex are, to our knowledge, novel results. The dissociation constant is close to a KD value estimated to be 70 nM for the mammalian SRP–80S complex at 4 °C [11], suggesting that E. coli and eukaryotic SRPs bind to naked ribosomes with approximately the same affinity. We find that neither GTP nor GDP

499

affects the affinity of SRP to naked ribosomes (Table 1). These results, in conjunction with a detailed balance argument [20], suggest that neither GTP nor GDP should affect the affinity of SRP to the naked ribosome, in line with previous results obtained by Wintermeyer and coworkers [19]. We have found that E. coli SRP has the same affinity to naked ribosomes as to ribosomes in elongation phase with a tetrapeptidyl-tRNA in the P site. This means that ribosomal binding of SRP is not stimulated by a putative conformational change of the ribosome as it enters protein elongation. We expect, instead, that the affinity of SRP will increase when a nascent chain with a signal peptide sequence at the N-terminal emerges from the exit tunnel of the large subunit. This affinity increase could, furthermore, depend on the presence of GTP. These findings are in apparent contradiction to results from eukaryotic systems, for which a conformational difference between translating and the non translating 80S ribosome has been postulated to explain why mammalian SRP has an order of magnitude higher affinity to translating than to naked ribosomes [11]. The reason for this difference between ribosomes from the two kingdoms could be that the structure and function of eukaryotic SRP are more complex than for their E. coli counter part. Eukaryotic SRP contains the Alu domain [6], which is responsible for the arrest of protein elongation, while bacterial SRP lacks this domain [6] and does not arrest protein elongation [12]. The binding of both TF and SRP to the E. coli ribosome has been localised to ribosomal protein L23. It has been proposed that TF and SRP alternate in their binding to the ribosome [3,4]. In line with this, data from cross-linking of TF and SRP to naked ribosomes and subsequent immunoprecipitation with antibodies towards L23 suggest that there is competition between TF and SRP for binding to L23 [3]. In contrast, the present data suggest that SRP and TF can bind simultaneously to the naked ribosome with little interference. A possible explanation for the apparent discrepancy between these results is that the efficiency of cross-linking to L23 is reduced by the presence of both factors on the ribosome, although their respective affinities to the ribosome are the same as when they bind alone. The crystal structure of the N-terminal domain of the TF shows that its ribosome binding motif resides in a flexible tip, that could serve as the ribosome “anchor” [20]. SRP lacks a clear anchor candidate but the four helix bundle N-domain of SRP has been suggested to be the ribosome binding part of the factor [4], which would suggest that the TF and SRP binding interactions on L23 are distinct from each other [20]. Moreover, TF could be crosslinked both to L23 and the neighbour protein L29 [2], while E. coli SRP did not cross-link to L29 [3,4], suggesting that SRP binds to the side of L23 that is distal to L29 on the naked ribosome. The evidence for an interaction between SRP and L23 is further strengthened by a recent cryo-electron microscopy (cryo-EM) structure of the SRP–80S complex, showing that the tip of the NG domain of SRP54 binds to the ribosome in the L25 (L29 in prokaryotes) area [21]. Two additional

500

A. Raine et al. / Biochimie 86 (2004) 495–500

contacts were identified between the SRP54 M domain and 25S ribosomal RNA and another contact between the 7S RNA in the signal sequence binding domain and 25S RNA [21]. It is possible that the conformation of SRP on the naked ribosome allows only the contact between the tip of the NG domain of SRP and L23/L25 and that the additional contacts seen in cryo-EM are established with SRP in a different conformation due to the presence of a signal sequence. In eukaryotes, SRP cross-links are shifted towards L29 upon interaction with the SRP receptor [9]. This may suggest that SRP is repositioned on the ribosome after complex formation with its receptor. Interestingly, no such shift has been observed in E. coli [3,4], in line with our observation that TF cannot be displaced from the naked ribosome by SRP even in the presence of both FtsY and GTP.

[7]

[8]

[9]

[10] [11]

[12]

Acknowledgments We thank Dr W. Wintermeyer and Dr E. Deuerling for strains overexpressing SRP and TF and M. Lovmar for assistance with data calculations. The study was supported by the Swedish Research Council.

[13]

References

[15]

[1]

[2]

[3]

[4]

[5]

[6]

P. Nissen, J. Hansen, N. Ban, P.B. Moore, T.A. Steitz, The structural basis of ribosome activity in peptide bond synthesis, Science 289 (2000) 920–930. G. Kramer, T. Rauch, W. Rist, S. Vorderwulbecke, H. Patzelt, A. Schulze-Specking, N. Ban, E. Deuerling, B. Bukau, L23 protein functions as a chaperone docking site on the ribosome, Nature 419 (2002) 171–174. R.S. Ullers, E.N. Houben, A. Raine, C.M. ten Hagen-Jongman, M. Ehrenberg, J. Brunner, B. Oudega, N. Harms, J. Luirink, Interplay of signal recognition particle and trigger factor at L23 near the nascent chain exit site on the Escherichia coli ribosome, J. Cell. Biol. 161 (2003) 679–684. S.Q. Gu, F. Peske, H.J. Wieden, M.V. Rodnina, W. Wintermeyer, The signal recognition particle binds to protein L23 at the peptide exit of the Escherichia coli ribosome, RNA 9 (2003) 566–573. E.A. Craig, H.C. Eisenman, H.A. Hundley, Ribosome-tethered molecular chaperones: the first line of defense against protein misfolding? Curr. Opin. Microbiol. 6 (2003) 157–162. R.J. Keenan, D.M. Freymann, R.M. Stroud, P. Walter, The signal recognition particle, Annu. Rev. Biochem. 70 (2001) 755–775.

[14]

[16]

[17]

[18]

[19]

[20] [21]

G. Kramer, H. Patzelt, T. Rauch, T.A. Kurz, S. Vorderwulbecke, B. Bukau, E. Deuerling, Trigger factor peptidyl-prolyl cis/trans isomerase activity is not essential for the folding of cytosolic proteins in Escherichia coli, J. Biol. Chem. 279 (2004) 14165–14170. H. Patzelt, S. Rudiger, D. Brehmer, G. Kramer, S. Vorderwulbecke, E. Schaffitzel, A. Waitz, T. Hesterkamp, L. Dong, J. SchneiderMergener, B. Bukau, E. Deuerling, Binding specificity of Escherichia coli trigger factor, Proc. Natl. Acad. Sci. USA 98 (2001) 14244– 14249. M.R. Pool, J. Stumm, T.A. Fulga, I. Sinning, B. Dobberstein, Distinct modes of signal recognition particle interaction with the ribosome, Science 297 (2002) 1345–1348. R. Maier, B. Eckert, C. Scholz, H. Lilie, F.X. Schmid, Interaction of trigger factor with the ribosome, J. Mol. Biol. 326 (2003) 585–592. J.J. Flanagan, J.C. Chen, Y. Miao, Y. Shao, J. Lin, P.E. Bock, A.E. Johnson, Signal recognition particle binds to ribosome-bound signal sequences with fluorescence-detected sub-nanomolar affinity that does not diminish as the nascent chain lengthens, J. Biol. Chem. 278 (2003) 18628–18637. A. Raine, R. Ullers, M. Pavlov, J. Luirink, J.E. Wikberg, M. Ehrenberg, Targeting and insertion of heterologous membrane proteins in E. coli, Biochimie 85 (2003) 659–668. P.C. Jelenc, C.G. Kurland, Nucleoside triphosphate regeneration decreases the frequency of translation errors, Proc. Natl. Acad. Sci. USA 76 (1979) 3174–3178. T. Hesterkamp, E. Deuerling, B. Bukau, The amino-terminal 118 amino acids of Escherichia coli trigger factor constitute a domain that is necessary and sufficient for binding to ribosomes, J. Biol. Chem. 272 (1997) 21865–21871. M.V. Rodnina, W. Wintermeyer, GTP consumption of elongation factor Tu during translation of heteropolymeric mRNAs, Proc. Natl. Acad. Sci. USA 92 (1995) 1945–1949. K. Pedersen, A.V. Zavialov, M.Y. Pavlov, J. Elf, K. Gerdes, M. Ehrenberg, The bacterial toxin RelE displays codon-specific cleavage of mRNAs in the ribosomal A site, Cell 112 (2003) 131–140. D.V. Freistroffer, M.Y. Pavlov, J. MacDougall, R.H. Buckingham, M. Ehrenberg, 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 (1997) 4126–4133. G. Blaha, D.N. Wilson, G. Stoller, G. Fischer, R. Willumeit, K.H. Nierhaus, Localization of the trigger factor binding site on the ribosomal 50S subunit, J. Mol. Biol. 326 (2003) 887–897. J.R. Jagath, M.V. Rodnina, G. Lentzen, W. Wintermeyer, Interaction of guanine nucleotides with the signal recognition particle from Escherichia coli, Biochemistry 37 (1998) 15408–15413. O. Kristensen, M. Gajhede, Chaperone binding at the ribosomal exit tunnel, Structure (Camb) 11 (2003) 1547–1556. M. Halic, T. Becker, M.R. Pool, C.M. Spahn, R.A. Grassucci, J. Frank, R. Beckmann, Structure of the signal recognition particle interacting with the elongation-arrested ribosome, Nature 427 (2004) 808–814.