Specific interaction between the ribosome recycling factor and the elongation factor G from Mycobacterium tuberculosis mediates peptidyl-tRNA release and ribosome recycling in Escherichia coli

Authors


Abstract

Once the translating ribosomes reach a termination codon, the nascent polypeptide chain is released in a factor-dependent manner. However, the P-site-bound deacylated tRNA and the ribosomes themselves remain bound to the mRNA (post-termination complex). The ribosome recycling factor (RRF) plays a vital role in dissociating this complex. Here we show that the Mycobacterium tuberculosis RRF (MtuRRF) fails to rescue Escherichia coli LJ14, a strain temperature-sensitive for RRF (frrts). More interestingly, co-expression of M.tuberculosis elongation factor G (MtuEFG) with MtuRRF rescues the frrts strain of E.coli. The simultaneous expression of MtuEFG is also needed to cause an enhanced release of peptidyl-tRNAs in E.coli by MtuRRF. These observations provide the first genetic evidence for a functional interaction between RRF and EFG. Both the in vivo and in vitro analyses suggest that RRF does not distinguish between the translating and terminating ribosomes for their dissociation from mRNA. In addition, complementation of E.coli PEM100 (fusAts) with MtuEFG suggests that the mechanism of RRF function is independent of the translocation activity of EFG.

Introduction

In eubacteria, once the translating ribosomes reach a termination codon, binding of RF1 in response to UAA and UAG, or RF2 in response to UAA and UGA into the ribosomal A-site triggers hydrolysis of the P-site-bound peptidyl-tRNA culminating in the release of the nascent polypeptide chain (Hershey, 1983). Several studies have now shown that RF3 binding to the ribosomal A-site catalyses recycling of RF1 and RF2 (Buckingham et al., 1997; Pavlov et al., 1997b). However, subsequent to the release of polypeptide, the 70S ribosome and the P-site-bound deacylated tRNA still remain attached to the mRNA. The most important events in initiation of protein synthesis occur on the P-site of the 30S ribosomal subunits (Hershey, 1983; Kozak, 1983). Therefore, it is imperative that the terminating ribosomal complex dissociates before the start of a fresh round of polypeptide biosynthesis. Mechanistic details of this post-termination step (also termed as the ‘fourth step’ of protein synthesis) have remained largely unknown. The pioneering work of Kaji and others suggested the involvement of a new factor, ribosome recycling factor (RRF), in dissociation of the termination complex (Hirashima and Kaji, 1972; Subramanian and Davis, 1973). Using in vitro assays, it was also demonstrated that a highly purified preparation of RRF was inadequate to dissociate the model substrates. However, when the RRF preparation contained elongation factor G (EFG), it converted the polysomes into monosomes in the presence of GTP (Hirashima and Kaji, 1972). Subsequent work demonstrated that RRF is vital for cellular metabolism and, in Escherichia coli, it has already been shown to be an essential protein (Janosi et al., 1994). RRF appears to be unique to the prokaryotic world.

This observation has stimulated a great deal of research on RRF in recent times, and crystal structures of RRF from E.coli, Thermotoga maritima and Thermus thermophilus have been determined recently (Selmer et al., 1999; Kim et al., 2000; Toyoda et al., 2000). Interestingly, the crystal structures reveal that RRF is a remarkable mimic of tRNA. This observation as well as an in vitro study suggested that similarly to its action at the elongation step, EFG translocates RRF from the ribosomal A-site to the P-site, which in turn dissociates the post-termination complexes (Pavlov et al., 1997a; Selmeret al., 1999). On the contrary, a recent elegant biochemical study which utilized short synthetic mRNAs suggested that the EFG function at the post-termination step is distinct from its classical role at the translocation step (Karimi et al., 1999). Both EFG and RRF are essential proteins and, therefore, demonstration of the relevance of these in vitro studies and various other theoretical models (Janosi et al., 1996) of RRF function in the cellular context has remained a challenging task.

We are interested in exploring the mechanism of protein synthesis in E.coli and in Mycobacterium tuberculosis, a Gram-positive (acid-fast) eubacterium that afflicts nearly one-third of the world's population and is a major cause of human death worldwide. In this study, by simultaneous expression of RRF and EFG from M.tuberculosis in E.coli, we provide the first genetic evidence for the functional interaction between RRF and EFG in vivo. More importantly, this assay system has allowed us to explore the mechanism of RRF action in the cellular milieu and discriminate among the various models proposed for RRF action (Janosi et al., 1996).

Results

MtuRRF alone does not rescue the temperature-sensitive phenotype of E.coli LJ14 (frrts)

The comparison of the primary sequences of EcoRRF and MtuRRF (Figure 1A) shows an ∼40% identity and 55% similarity between the two proteins. However, when we carried out the complementation analysis in the E.coli LJ14 (frrts) strain using the MtuRRF gene cloned onto a medium copy expression vector (pTrcMtuRRF), it failed to complement the RRFts phenotype of the strain (Figure 1B, 42°C, sectors 3 and 4). As expected, under the same conditions, expression of EcoRRF (pTrcEcoRRF) complemented the strain (sectors 5 and 6). Immunoblot analysis of the cell-free extracts of the transformants grown at the permissive temperature (30°C) showed that MtuRRF was produced in the LJ14 strain (Figure 1C, lanes 1 and 2). This observation ruled out the lack of MtuRRF expression as a possible reason for the failure to restore the growth of the strain at the non-permissive temperature (42°C). To investigate further, we determined the growth rates of the various transformants in liquid culture at the permissive and non-permissive temperatures (Figure 1D). At the permissive temperature, transformants harbouring pTrcEcoRRF or pTrcMtuRRF grew at the same rate as those harbouring the vector alone, suggesting that the expression of these RRFs was not toxic to E.coli. However, at the non-permissive temperature (42°C), while the transformants harbouring pTrcEcoRRF grew well, those harbouring either the vector or pTrcMtuRRF did not. This observation confirms the results shown in Figure 1B, and suggests that MtuRRF alone does not complement the LJ14 strain to any detectable level.

Figure 1.

Figure 1.

(A) Comparison of the primary structures of E.coli RRF (EcoRRF) and M.tuberculosis RRF (MtuRRF). Shadings in black and grey indicate identical and similar residues, respectively. (B) Analysis of complementation of E.coli LJ14 (frrts) with MtuRRF. Various plasmids were introduced into E.coli LJ14 and the transformants were streaked on LB agar plates containing ampicillin and IPTG in duplicate. One plate was incubated at the permissive temperature (30°C) and the other at the non-permissive temperature (42°C). Sectors: 1 and 2, pTrc99C vector; 3 and 4, pTrcMtuRRF; 5 and 6, pTrcEcoRRF. (C) Detection of MtuRRF expression in E.coli LJ14 by immunoblotting using anti-MtuRRF antibodies. Protein extracts were prepared from transformants harbouring either the pTrcMtuRRF expression construct (lanes 1 and 2) or the pTrc99C vector (lanes 3 and 4). Cultures were grown in the presence (+) or absence (−) of IPTG. (D) Growth curves of various transformants of E.coli LJ14. Fresh cultures were started by inoculating with overnight cultures (0.06%) from 30°C, and grown at the permissive (30°C, open symbols) and non-permissive (42°C, filled symbols) temperatures. The OD at 600 nm (y-axis) was monitored at regular intervals (x-axis). The presence of pTrc99C or its recombinants is indicated in the inset.

Figure 1.

Figure 1.

(A) Comparison of the primary structures of E.coli RRF (EcoRRF) and M.tuberculosis RRF (MtuRRF). Shadings in black and grey indicate identical and similar residues, respectively. (B) Analysis of complementation of E.coli LJ14 (frrts) with MtuRRF. Various plasmids were introduced into E.coli LJ14 and the transformants were streaked on LB agar plates containing ampicillin and IPTG in duplicate. One plate was incubated at the permissive temperature (30°C) and the other at the non-permissive temperature (42°C). Sectors: 1 and 2, pTrc99C vector; 3 and 4, pTrcMtuRRF; 5 and 6, pTrcEcoRRF. (C) Detection of MtuRRF expression in E.coli LJ14 by immunoblotting using anti-MtuRRF antibodies. Protein extracts were prepared from transformants harbouring either the pTrcMtuRRF expression construct (lanes 1 and 2) or the pTrc99C vector (lanes 3 and 4). Cultures were grown in the presence (+) or absence (−) of IPTG. (D) Growth curves of various transformants of E.coli LJ14. Fresh cultures were started by inoculating with overnight cultures (0.06%) from 30°C, and grown at the permissive (30°C, open symbols) and non-permissive (42°C, filled symbols) temperatures. The OD at 600 nm (y-axis) was monitored at regular intervals (x-axis). The presence of pTrc99C or its recombinants is indicated in the inset.

Figure 1.

Figure 1.

(A) Comparison of the primary structures of E.coli RRF (EcoRRF) and M.tuberculosis RRF (MtuRRF). Shadings in black and grey indicate identical and similar residues, respectively. (B) Analysis of complementation of E.coli LJ14 (frrts) with MtuRRF. Various plasmids were introduced into E.coli LJ14 and the transformants were streaked on LB agar plates containing ampicillin and IPTG in duplicate. One plate was incubated at the permissive temperature (30°C) and the other at the non-permissive temperature (42°C). Sectors: 1 and 2, pTrc99C vector; 3 and 4, pTrcMtuRRF; 5 and 6, pTrcEcoRRF. (C) Detection of MtuRRF expression in E.coli LJ14 by immunoblotting using anti-MtuRRF antibodies. Protein extracts were prepared from transformants harbouring either the pTrcMtuRRF expression construct (lanes 1 and 2) or the pTrc99C vector (lanes 3 and 4). Cultures were grown in the presence (+) or absence (−) of IPTG. (D) Growth curves of various transformants of E.coli LJ14. Fresh cultures were started by inoculating with overnight cultures (0.06%) from 30°C, and grown at the permissive (30°C, open symbols) and non-permissive (42°C, filled symbols) temperatures. The OD at 600 nm (y-axis) was monitored at regular intervals (x-axis). The presence of pTrc99C or its recombinants is indicated in the inset.

Figure 1.

Figure 1.

(A) Comparison of the primary structures of E.coli RRF (EcoRRF) and M.tuberculosis RRF (MtuRRF). Shadings in black and grey indicate identical and similar residues, respectively. (B) Analysis of complementation of E.coli LJ14 (frrts) with MtuRRF. Various plasmids were introduced into E.coli LJ14 and the transformants were streaked on LB agar plates containing ampicillin and IPTG in duplicate. One plate was incubated at the permissive temperature (30°C) and the other at the non-permissive temperature (42°C). Sectors: 1 and 2, pTrc99C vector; 3 and 4, pTrcMtuRRF; 5 and 6, pTrcEcoRRF. (C) Detection of MtuRRF expression in E.coli LJ14 by immunoblotting using anti-MtuRRF antibodies. Protein extracts were prepared from transformants harbouring either the pTrcMtuRRF expression construct (lanes 1 and 2) or the pTrc99C vector (lanes 3 and 4). Cultures were grown in the presence (+) or absence (−) of IPTG. (D) Growth curves of various transformants of E.coli LJ14. Fresh cultures were started by inoculating with overnight cultures (0.06%) from 30°C, and grown at the permissive (30°C, open symbols) and non-permissive (42°C, filled symbols) temperatures. The OD at 600 nm (y-axis) was monitored at regular intervals (x-axis). The presence of pTrc99C or its recombinants is indicated in the inset.

MtuRRF requires MtuEFG to rescue the temperature-sensitive phenotype of E.coli LJ14 (frrts)

In vitro experiments performed with RRF have revealed that for its function it requires EFG (Janosi et al., 1996). Therefore, we reasoned that the presence of MtuEFG might facilitate complementation of the LJ14 strain with MtuRRF. Interestingly, when MtuRRF and MtuEFG were expressed simultaneously from the two compatible plasmids, pACDHMtuRRF and pTrcMtuEFG, in E.coli LJ14, the temperature-sensitive phenotype of the strain was rescued and it grew well at the non-permissive temperature (42°C) (Figure 2A, sectors 7 and 8). However, neither MtuRRF (sectors 3 and 4) nor MtuEFG (sectors 5 and 6) alone complemented the strain. Further, as shown in Figure 2B, overproduction of EcoEFG with MtuRRF (compare sectors 5 and 6, with sectors 3 and 4 at 42°C) also did not rescue the temperature-sensitive phenotype of the LJ14 strain. The growth curves (Figure 2C) showed that simultaneous expression of MtuRRF and MtuEFG completely alleviated the temperature-sensitive phenotype of E.coli LJ14. Thus, these observations provide genetic evidence for a functional interaction between RRF and EFG.

Figure 2.

Figure 2.

(A) Analysis of complementation of E.coli LJ14 (frrts) with MtuRRF. Various plasmids were introduced into E.coli LJ14 and the trans formants were streaked on LB agar plates containing ampicillin, tetracycline and IPTG in duplicate. One plate was incubated at the permissive temperature (30°C) and the other at the non-permissive temperature (42°C). Sectors: 1 and 2, pTrc99C and pACDH vectors; 3 and 4, pTrc99C and pACDHMtuRRF; 5 and 6, pTrcMtuEFG and pACDH; 7 and 8, pTrcMtuEFG and pACDHMtuRRF. (B) Analysis of complementation of E.coli LJ14 (frrts) with MtuRRF along with EcoEFG. The transformants were streaked on LB agar plates containing ampicillin, tetracycline and IPTG in duplicate. One plate was incubated at the permissive temperature (30°C) and the other at the non-permissive temperature (42°C). Sectors: 1 and 2, pTrc99C and pACDH vector control; 3 and 4, pTrcMtuEFG and pACDHMtuRRF; 5 and 6, pTrcEcoEFG and pACDHMtuRRF. (C) Growth of E.coli LJ14 (frrts) harbouring various plasmids at the permissive (30°C) and non-permissive (42°C) temperatures. Escherichia coli LJ14 harbouring various plasmid constructs were grown overnight at the permissive temperature (30°C). Fresh cultures in LB containing ampicillin, tetracycline and IPTG were started by inoculating (0.06%) with overnight cultures from 30°C, and grown at the permissive (30°C, open symbols) and non-permissive (42°C, filled symbols) temperatures. The OD at 600 nm (y-axis) was monitored at regular intervals (x-axis). The inset shows the proteins that were overproduced from the pACDH or pTrc99C vectors. pTrcMtuEFG and pACDH (open and closed circles), pTrcMtuEFG and pACDHMtuRRF (open and closed inverted triangles).

Figure 2.

Figure 2.

(A) Analysis of complementation of E.coli LJ14 (frrts) with MtuRRF. Various plasmids were introduced into E.coli LJ14 and the trans formants were streaked on LB agar plates containing ampicillin, tetracycline and IPTG in duplicate. One plate was incubated at the permissive temperature (30°C) and the other at the non-permissive temperature (42°C). Sectors: 1 and 2, pTrc99C and pACDH vectors; 3 and 4, pTrc99C and pACDHMtuRRF; 5 and 6, pTrcMtuEFG and pACDH; 7 and 8, pTrcMtuEFG and pACDHMtuRRF. (B) Analysis of complementation of E.coli LJ14 (frrts) with MtuRRF along with EcoEFG. The transformants were streaked on LB agar plates containing ampicillin, tetracycline and IPTG in duplicate. One plate was incubated at the permissive temperature (30°C) and the other at the non-permissive temperature (42°C). Sectors: 1 and 2, pTrc99C and pACDH vector control; 3 and 4, pTrcMtuEFG and pACDHMtuRRF; 5 and 6, pTrcEcoEFG and pACDHMtuRRF. (C) Growth of E.coli LJ14 (frrts) harbouring various plasmids at the permissive (30°C) and non-permissive (42°C) temperatures. Escherichia coli LJ14 harbouring various plasmid constructs were grown overnight at the permissive temperature (30°C). Fresh cultures in LB containing ampicillin, tetracycline and IPTG were started by inoculating (0.06%) with overnight cultures from 30°C, and grown at the permissive (30°C, open symbols) and non-permissive (42°C, filled symbols) temperatures. The OD at 600 nm (y-axis) was monitored at regular intervals (x-axis). The inset shows the proteins that were overproduced from the pACDH or pTrc99C vectors. pTrcMtuEFG and pACDH (open and closed circles), pTrcMtuEFG and pACDHMtuRRF (open and closed inverted triangles).

Figure 2.

Figure 2.

(A) Analysis of complementation of E.coli LJ14 (frrts) with MtuRRF. Various plasmids were introduced into E.coli LJ14 and the trans formants were streaked on LB agar plates containing ampicillin, tetracycline and IPTG in duplicate. One plate was incubated at the permissive temperature (30°C) and the other at the non-permissive temperature (42°C). Sectors: 1 and 2, pTrc99C and pACDH vectors; 3 and 4, pTrc99C and pACDHMtuRRF; 5 and 6, pTrcMtuEFG and pACDH; 7 and 8, pTrcMtuEFG and pACDHMtuRRF. (B) Analysis of complementation of E.coli LJ14 (frrts) with MtuRRF along with EcoEFG. The transformants were streaked on LB agar plates containing ampicillin, tetracycline and IPTG in duplicate. One plate was incubated at the permissive temperature (30°C) and the other at the non-permissive temperature (42°C). Sectors: 1 and 2, pTrc99C and pACDH vector control; 3 and 4, pTrcMtuEFG and pACDHMtuRRF; 5 and 6, pTrcEcoEFG and pACDHMtuRRF. (C) Growth of E.coli LJ14 (frrts) harbouring various plasmids at the permissive (30°C) and non-permissive (42°C) temperatures. Escherichia coli LJ14 harbouring various plasmid constructs were grown overnight at the permissive temperature (30°C). Fresh cultures in LB containing ampicillin, tetracycline and IPTG were started by inoculating (0.06%) with overnight cultures from 30°C, and grown at the permissive (30°C, open symbols) and non-permissive (42°C, filled symbols) temperatures. The OD at 600 nm (y-axis) was monitored at regular intervals (x-axis). The inset shows the proteins that were overproduced from the pACDH or pTrc99C vectors. pTrcMtuEFG and pACDH (open and closed circles), pTrcMtuEFG and pACDHMtuRRF (open and closed inverted triangles).

MtuEFG participates at the translocation step in E.coli

To investigate the mechanism of functional interaction between EFG and RRF, we introduced the MtuEFG expression vector into E.coli PEM100. This strain possesses a mutation in the fusA gene, which renders it temperature sensitive because the encoded EFG (G502D) is defective at the translocation step (Hou et al., 1994). All the transformants (pTrc99C, pTrcEcoEFG and pTrcMtuEFG) grew at the permissive temperature (Figure 3A, 30°C). Interestingly, both EcoEFG and MtuEFG rescued the temperature-sensitive phenotype of the strain (Figure 3A, sectors 2 and 3, 42°C).

Figure 3.

Figure 3.

(A) Analysis of complementation of E.coli PEM100 (fusAts) with MtuEFG. The PEM100 strain was transformed with various plasmids and the transformants streaked in duplicate on LB agar containing ampicillin and IPTG. One plate was incubated at the permissive temperature (30°C) and the other at the non-permissive temperature (42°C). Sectors: 1, pTrc99C; 2, pTrcMtuEFG; 3, pTrcEcoEFG. (B) Analysis of growth of E.coli TG1 containing various plasmids on LB agar containing ampicillin and IPTG alone (left) or with 25 μg/ml (centre) and 50 μg/ml (right) fusidic acid, respectively. Sectors: 1 and 2, pTrc99C; 3 and 4, pTrcMtuEFG; 5 and 6, pTrcEcoEFG.

Figure 3.

Figure 3.

(A) Analysis of complementation of E.coli PEM100 (fusAts) with MtuEFG. The PEM100 strain was transformed with various plasmids and the transformants streaked in duplicate on LB agar containing ampicillin and IPTG. One plate was incubated at the permissive temperature (30°C) and the other at the non-permissive temperature (42°C). Sectors: 1, pTrc99C; 2, pTrcMtuEFG; 3, pTrcEcoEFG. (B) Analysis of growth of E.coli TG1 containing various plasmids on LB agar containing ampicillin and IPTG alone (left) or with 25 μg/ml (centre) and 50 μg/ml (right) fusidic acid, respectively. Sectors: 1 and 2, pTrc99C; 3 and 4, pTrcMtuEFG; 5 and 6, pTrcEcoEFG.

In yet another experiment, we overproduced EcoEFG or MtuEFG in E.coli (wild-type for fusA) and checked if they resulted in a similar response to fusidic acid. Fusidic acid is a classical inhibitor, which affects the translocation function of EFG by freezing it onto ribosomes in its GDP-bound form (Rodnina and Wintermeyer 1998). Interestingly, when compared with the vector control, the overproduction of EcoEFG in cells made them hypersensitive to fusidic acid (Figure 3B, compare sectors 5 and 6 with sectors 1 and 2 in the panels on the left with no fusidic acid with the panel in the centre with 25 μg/ml fusidic acid or the panel on the right with 50 μg/ml fusidic acid). Further, similarly to EcoEFG, expression of MtuEFG also resulted in hypersensitivity of the host to fusidic acid (sectors 3 and 4). While the exact mechanism of hypersensitivity of E.coli to fusidic acid upon overproduction of EFGs is not clear, it could be that the high occupancy of ribosomes with the EFG-GDP complex (Agrawal et al., 1998) in the presence of the inhibitor causes toxicity to cells. Nevertheless, the observation that overproduction of either of the EFGs makes E.coli hypersensitive to fusidic acid further supports that the mycobacterial protein can substitute for the function of EcoEFG.

In vitro conversion of polysomes to monosomes by RRF and EFG

To further our understanding of the role of MtuRRF and MtuEFG in ribosome recycling, we used an established method described by Kaji and co-workers (Ohnishi et al., 1999) for the in vitro activity assay of RRF, wherein conversion of polysomes to monosomes was monitored (Figure 4A and B). Factor-free polysomes were prepared from E.coli MRE600 and, under the experimental conditions, the profile of the buffer control shows a clear separation between the monosomes and the higher forms of the polysomes in the preparation (panels marked a). The controls wherein the polysomes were treated with RRF or EFG alone also show a similar profile (panels c and d). When the polysomes were treated with EcoRRF and EcoEFG, the polysomes were converted into monosomes (Figure 4A, b). Similarly, as expected from the in vivo complementation data, treatment of E.coli polysomes with the M.tuberculosis proteins (RRF and EFG) also resulted in enrichment of the monosome peak (Figure 4B, b). Interestingly, neither of the heterologous combinations of proteins (EcoRRF with MtuEFG or MtuRRF with EcoEFG) showed any activity in the conversion of polysomes to monosomes (Figure 4B, e). The findings of this experiment corroborate those of the in vivo experiments (Figures 1 and 2) and, taken together, clearly demonstrate that RRF, for its function, requires EFG from homologous sources.

Figure 4.

RRF activity assays. Reactions were set up as described in Materials and methods. (A) (a) Polysomes alone; (b) EcoRRF and EcoEFG (15 μg each); (c) EcoEFG (30 μg); (d) EcoRRF (15 μg); (e) MtuRRF and EcoEFG (15 and 30 μg, respectively). (B) (a) Polysomes alone; (b) MtuRRF and MtuEFG (15 and 30 μg, respectively); (c) MtuEFG (30 μg); (d) MtuRRF (15 μg); (e) EcoRRF and MtuEFG (15 and 30 μg, respectively).

Role of RRF and EFG in peptidyl-tRNA release from the translating ribosomes in vivo

Not all the peptide chains that begin from the initiation codon of the mRNA give rise to a full-length protein. Even under normal physiological conditions, peptidyl-tRNAs are released from the ribosomes during chain elongation. As the accumulation of peptidyl-tRNAs is toxic, cells contain peptidyl-tRNA hydrolase (PTH), which hydrolyses the peptidyl-tRNA ester bond (Kossel and RajBhandary, 1968). Conditional mutants of E.coli (pthts) have been isolated, which grow at 30°C (permissive temperature) but not at 37°C (non-permissive temperature) (Atherly and Menninger, 1972; Menninger et al., 1973). Recently, in a genetic screen, it was found that the suppressors of pthts mapped to the promoter region of frr, resulting in low levels of RRF in the cell (Heurgue-Hamard et al., 1998). This suggests that RRF can somehow cause premature release of peptidyl-tRNAs from the ribosome.

In the experiment shown in Figure 5A, by making use of the acid urea gels (Varshney et al., 1991a), we examined the consequence of overproduction of EcoRRF and MtuRRF in E.coli AA7852 (pthts) on the accumulation of peptidyl-tRNAs. These gels allow direct analysis of the in vivo status of the tRNAs by separating the aminoacylated and uncharged forms. However, when peptides of heterogeneous length are attached to the tRNA, the aminoacylated form of the tRNA is converted into a smear. As expected, the transformants grown at 30°C (permissive temperature) do not show accumulation of peptidyl-tRNAs in the cell (lanes 1–5, absence of a smear above the Tyr-tRNATyr band). However, at 34°C, the temperature at which PTH is partially inactive, minimal accumulation of peptidyl-tRNA (as detected by the smear corresponding to peptidyl-Tyr-tRNATyr) was seen in the vector control upon overexposure of the autoradiogram (Figure 5A, lane 6, lower panel). Overproduction of EcoRRF resulted in the accumulation of peptidyl-tRNA at 34°C (lane 10). More importantly, while overproduction of MtuRRF alone (lane 8) or MtuEFG alone (lane 9) had no effect on accumulation of peptidyl-tRNA over and above the vector-alone background (lane 6), their simultaneous expression resulted in an increased accumulation of peptidyl-tRNA in cells (lane 7). Thus, in the context of E.coli ribosomes, MtuRRF participated even in enhanced release of peptidyl-tRNA and, at this step also, it required the presence of EFG from the homologous source. Taken together, these results point to a unified mechanism of RRF action both at the post-termination and the peptidyl-tRNA release steps. Furthermore, to examine the effect of the accumulation of peptidyl-tRNA, we determined the growth rates of the various transformants (Figure 5B). The cells harbouring the EcoRRF plasmid, as expected, grew slowly in the liquid medium. Similarly, the cells harbouring the EcoEFG plasmid also grew slowly. On the other hand, the transformants harbouring either the MtuRRF or the MtuEFG plasmids grew just as well as the vector control. More interestingly, the transformants harbouring both the MtuRRF and MtuEFG plasmids grew slowly, most probably as a consequence of the accumulation of the peptidyl-tRNA (Figure 5A, lane 7).

Figure 5.

Figure 5.

(A) Analysis of the accumulation of peptidyl-tRNA in E.coli AA7852 (pthts). The E.coli pthts (AA7852) was transformed with various plasmid constructs, grown in LB containing ampicillin, tetracycline and IPTG at the permissive (30°C) or partially non-permissive (34°C) temperature and used to prepare total tRNA under acidic phenol conditions. Total tRNA was fractionated on an acid urea gel, electroblotted onto a nytran membrane and hybridized with 5′-32P-end-labelled anti-tRNATyr probe. Photographs of autoradiographs with short (upper panel) and longer (lower panel) exposures. −, pACDH vector; M, pACDHMtuRRF; and E, pTrcEcoRRF in the RRFs row; and −, pTrc99C; and M, pTrcMtuEFG in the EFG row, respectively. Temperatures at which the cultures were grown are as indicated. (B) Analysis of growth of E.coli AA7852 (pthts) harbouring various plasmids in LB medium containing ampicillin, tetracycline and IPTG grown at the partially non-permissive temperature (35°C). pTrc99C and pACDH vectors (closed circles); pTrc99C and pACDMtuRRF (open circles); pTrcMtuEFG and pACDH (closed inverted triangles);pTrcEcoEFG and pACDH (open squares); pTrcEcoRRF and pACDH (open inverted triangles); and pTrcMtuEFG and pACDHMtuRRF (closed squares).

Figure 5.

Figure 5.

(A) Analysis of the accumulation of peptidyl-tRNA in E.coli AA7852 (pthts). The E.coli pthts (AA7852) was transformed with various plasmid constructs, grown in LB containing ampicillin, tetracycline and IPTG at the permissive (30°C) or partially non-permissive (34°C) temperature and used to prepare total tRNA under acidic phenol conditions. Total tRNA was fractionated on an acid urea gel, electroblotted onto a nytran membrane and hybridized with 5′-32P-end-labelled anti-tRNATyr probe. Photographs of autoradiographs with short (upper panel) and longer (lower panel) exposures. −, pACDH vector; M, pACDHMtuRRF; and E, pTrcEcoRRF in the RRFs row; and −, pTrc99C; and M, pTrcMtuEFG in the EFG row, respectively. Temperatures at which the cultures were grown are as indicated. (B) Analysis of growth of E.coli AA7852 (pthts) harbouring various plasmids in LB medium containing ampicillin, tetracycline and IPTG grown at the partially non-permissive temperature (35°C). pTrc99C and pACDH vectors (closed circles); pTrc99C and pACDMtuRRF (open circles); pTrcMtuEFG and pACDH (closed inverted triangles);pTrcEcoEFG and pACDH (open squares); pTrcEcoRRF and pACDH (open inverted triangles); and pTrcMtuEFG and pACDHMtuRRF (closed squares).

Discussion

Translation termination and the post-termination steps that lead to recycling of ribosomes remain one of the least understood aspects of protein biosynthesis (Janosi et al., 1996; Kisselev and Buckingham, 2000). However, the three-dimensional structure determination of RRF from T.maritima, T.thermophilus and E.coli, and recent biochemical studies have led to an unprecedented stimulation and advancement of research in these areas (Pavlov et al., 1997a,b; Janosi et al., 1998; Karimi et al., 1999; Selmer et al., 1999; Inokuchi et al., 2000; Kim et al., 2000; Toyoda et al., 2000). Accordingly, the current view of the post-termination events that lead to dissociation of the termination complex suggests that subsequent to the peptide chain release by RF1/RF2, RF3 mediates the eviction of these release factors in a GTP hydrolysis-dependent process, and the subsequent action of RRF and EFG leads to the dissociation of the post-termination complex. The in vitro experiments (Pavlov et al., 1997a) suggested that for ribosome recycling, EFG might be involved in translocation of RRF from the ribosomal A-site to the P-site (similar to the translocation of the peptidyl-tRNA). This hypothesis received a boost from the three-dimensional structure of RRF, which showed it to be a spectacular example of protein mimicry of a tRNA molecule (Selmer et al., 1999). However, recently, an elegant biochemical study suggested that the function of EFG at this step is distinct from that of its involvement at the translocation step (Karimi et al., 1999).

Recently, it was shown that the RRF from Pseudomonas aeruginosa complements E.coli LJ14 (Ohnishi et al., 1999). On the other hand, the RRF from T.thermophilus does not complement the temperature-sensitive phenotype of E.coli (frrts). However, a mutant that lacked five amino acids from the C-terminal end of the T.thermophilus RRF did complement the E.coli for frrts (Fujiwara et al., 1999; Toyoda et al., 2000). Surprisingly, in our studies with the M.tuberculosis RRF, neither the wild-type (Figure 1) nor a mutant protein that lacked the C-terminal five amino acids complemented the E.coli frrts strain (data not shown). More importantly, we discovered that simultaneous expression of both the MtuRRF and MtuEFG (but not EcoEFG) rescued the temperature-sensitive phenotype of E.coli (frrts). This observation has allowed us to use E.coli as a surrogate host to investigate the mechanism of action of RRF in vivo. As we failed to detect an interaction between MtuEFG and MtuRRF by surface plasmon resonance and far western analysis (data not shown), it seems unlikely that EFG escorts RRF to the ribosomes. Most probably, specific interactions between the two proteins are effected on the ribosomes, and the appropriateness of these interactions is necessary for their biological activity. Recently, it was shown that the overproduction of RRFs from heterologous sources resulted in toxicity to E.coli (Atarashi and Kaji, 2000). A possible explanation for the observed toxicity could be that these proteins, by virtue of their structural mimicry of tRNA, occupy the ribosomal A-site but fail to establish functional interaction with EcoEFG. The formation of such ‘dead’ complexes could be deleterious for the growth of E.coli. It would be interesting to study whether the simultaneous expression of EFGs from homologous sources could alleviate the toxic effects of the heterologous RRFs.

To gain further insights into the mechanism of EFG participation in ribosome recycling, we used E.coli PEM100 (fusAts). This strain carries a G502D mutation in the EFG, which makes it defective in translocation at the non-permissive temperature (Hou et al., 1994). Recently, several other mutants (H583K, H583R, etc.) in this region (domain 4) of EcoEFG were also demonstrated to be defective in translocation (Savelsbergh et al., 2000). However, the properties of these mutants in ribosome binding, interaction with the α-sarcin loop of the 23S rRNA, GTP hydrolysis and Pi release were unaffected (Savelsbergh et al., 2000). Interestingly, the MtuEFG complements E.coli PEM100, a fusAts strain, suggesting that the translocation function of the EcoEFG and MtuEFG is conserved. On the other hand, as shown by the polysome dissociation assay (Figure 4), MtuEFG fails to collaborate with EcoRRF and, since the ribosome recycling is an essential function, the survival of the PEM100 strain in the presence of MtuEFG at the non-permissive temperature raises an intriguing question. How do these cells carry out ribosome recycling? We can offer two possible explanations: (i) the resident EcoEFG (chromosomally encoded) must still be able to function at the ribosome recycling step by a translocation-independent mechanism; or (ii) RF3, a classical G-protein that shares homology with EFG but lacks domain 4 (important in translocation) and binds to ribosomes, may substitute for the function of EFG (Grentzmann et al., 1998). Therefore, in either case, the mechanism of ribosome dissociation has to be independent of the translocation-like function of EFG. Thus it is appealing to speculate that the ‘conformational tickling’ following the hydrolysis of the factor-bound GTP may constitute a critical aspect of RRF-mediated ribosome recycling.

Yet another aspect of RRF activity that we have investigated by using E.coli as a surrogate host is its involvement in the step of release of peptidyl-tRNAs from ribosomes. Although the accumulation of peptidyl-tRNAs is toxic to cells, for fidelity in translation the peptidyl-tRNA release from ribosomes constitutes an important aspect of protein biosynthesis. Therefore, the peptidyl-tRNA hydrolase, important in recycling of the tRNAs, is an essential enzyme. The observation that the suppressors of pthts map to the promoter region of frr (down-regulation) was of great significance in establishing a role for RRF at this step (Heurgue-Hamard et al., 1998). As expected, in our studies, overproduction of EcoRRF resulted in enhanced release of peptidyl-tRNA. We observed that overproduction of EcoEFG alone also resulted in increased accumulation of peptidyl-tRNA in the pthts strain (data not shown) and inhibited the growth of E.coli AA7852 (pthts) (Figure 5B), suggesting that they work together in peptidyl-tRNA release. Although not reported so far, this observation predicts that another category of pthts suppressors could be the one where EFG expression is down-regulated. More importantly, as was the case at the post-termination step, overexpression of neither MtuRRF nor MtuEFG alone caused increased release of peptidyl-tRNA in the pthts strain. However, when the MtuRRF and MtuEFG were expressed simultaneously, enhanced release of peptidyl-tRNA was evident. Taken together, these observations suggest that peptidyl-tRNA release from the translating ribosome must lead to the dissociation of the translating ribosomes in much the same way as the post-termination complexes. Also, these studies suggest that RRF must compete with the aminoacyl-tRNAs for the occupancy of the ribosomal A-site. If the ribosomal A-site remains unfilled with an aminoacyl-tRNA for more than a ‘threshold’ period, these translating ribosomes are most probably destined for dissociation by RRF.

In conclusion, we have demonstrated that E.coli can be used as a surrogate host to investigate the mechanism of action of M.tuberculosis proteins. Since both RRF and EFG are essential proteins, development of this surrogate system is a major step to gain insights into the mechanistic aspects of RRF action in the cellular milieu. Notably, M.tuberculosis is an important pathogen that continues to cause most casualties worldwide. Thus, our use of the mycobacterial proteins in this study is also an important step forward in understanding the mechanism of action of the M.tuberculosis RRF which possibly can be used as an important drug target to control mycobacterial growth.

Materials and methods

Plasmids, strains and growth conditions

The various plasmids and strains used in this study are listed in Table I. LB liquid or solid (with 1.5% agar) media (Sambrook et al., 1989) were used for growth. The medium was supplemented with various antibiotics at the following final concentrations: tetracycline 12.5 μg/ml and ampicillin 100 μg/ml, as required. Isopropyl-β-D-thiogalactopyranoside (IPTG) was added to a final concentration of 0.5 mM when needed. The growth rate experiments were conducted at different temperatures by recording densities of fresh cultures at 600 nm at different times. Fresh cultures were started with 0.06% (v/v) inoculum from the overnight cultures grown at 30°C.

Table 1. Plasmids and strains used
Plasmid/strainRelevant detailsReference
pACDHthis plasmid harbours the ACYC origin of replication, and is derived from pACD (Mangroo and RajBhandary, 1995) vector by mutating the HindIII site of the TetR marker to an NheI siteT.K.DineshKumar and U.Varshney (unpublished)
pACDHMtuRRFM.tuberculosis RRF cloned into pACDHthis work
pTrcEcoRRFE.coli RRF cloned into pTrc99Cthis work
pTrcMtuRRFM.tuberculosis RRF cloned into pTrc99Cthis work
pTrcEcoEFGE.coli EFG cloned into pTrc99Cthis work
pTrcMtuEFGM.tuberculosis EFG cloned into pTrc99Cthis work
E.coli AA7852F, arg, leu, thr, his, thi, pth1tsAtherly and Menninger (1972)
E.coli LJ14MC1061 containing the frr14ts alleleJanosi et al. (1998)
E.coli PEM100ara Δ(lac-proAB) φ80 lacZΔM15 fusA100tsHou et al. (1994)
E.coli CA274HfrH, lacZam, trpEamBrenner and Beckwith (1965)

Cloning of RRF (frr) from M.tuberculosis

Based on the putative frr sequence of M.tuberculosis (Cole et al., 1998), a forward (5′-GCGCCCATGGTTGATGAGGCTCTCTTC-3′) and a reverse (5′-AGCGAAGCTTGAGCCTCCAGCAGCTC-3′) primer containing NcoI and HindIII restriction sites, respectively, were designed to amplify the open reading frame (ORF) corresponding to RRF by PCR using chromosomal DNA from M.tuberculosis. PCR was performed using Pfu DNA polymerase (Promega) and the cycling conditions included an initial template denaturation step at 94°C for 5 min followed by 30 cycles of incubations at 94°C for 1 min, 45°C for 1 min and 68°C for 1 min. The PCR product of the expected size (555 bp) was obtained, digested with NcoI and HindIII and cloned into similarly digested pTrc99C. The authenticity of the clone was verified by DNA sequence analysis of the complete ORF. Subsequently, the RRF was also cloned into the pACDH vector between the NcoI and HindIII restriction sites.

Cloning of EFG (fusA) from M.tuberculosis

Using the available fusA sequence of M.tuberculosis (Cole et al., 1998), a forward primer (5′-GCACAGAAGGACGTGCTGAC-3′) starting from the second codon of the ORF and a reverse primer (5′-CTCGTAAGCTT GCGCTCACTC-3′) which flanked the termination codon and contained a HindIII restriction site were designed. PCR was carried out using Vent DNA polymerase (exo+; New England Biolabs) and the cycling conditions consisted of an initial template denaturation step at 94°C for 5 min followed by 30 cycles of incubations at 94°C for 1 min, 58°C for 1 min and 68°C for 5 min. The PCR product was digested with HindIII, cloned into NcoI (blunted) and HindIII sites of pTrc99C vector and verified by DNA sequence analysis.

Cloning of the E.coli RRF gene (frr)

A forward (5′-CGTACCATGGGTAGCGATATCA-3′) and a reverse (5′-GTCGCTGCAGAAATCAGAACTGC-3′) primer containing NcoI and PstI sites, respectively, were used to amplify the ORF corresponding to RRF (Ichikawa and Kaji, 1989) from E.coli genomic DNA using Pfu DNA polymerase. The reaction conditions were the same as detailed for the amplification of MtuRRF. The PCR product was digested with NcoI and PstI and cloned into similarly digested pTrc99C. Subsequently, EcoRRF was mobilized into pET11d using NcoI and HindIII restriction endonucleases.

Cloning of EFG from E.coli

PCR was done using E.coli CA274 genomic DNA and the EcoEFG gene (Zengel et al., 1984)-specific forward (5′-ATGGCTCGTACAACACCCATC-3′) and reverse (5′-GTATGGATCCTTAGGCTTATTTACC-3′) primers using Pfu DNA polymerase. After an initial template denaturation step at 95°C for 3 min, 30 cycles of incubations at 95°C for 1 min, 55°C for 1 min and 70°C for 5 min were carried out. The PCR product was digested with BamHI and cloned between end-filled NcoI (with the Klenow fragment of DNA polymerase I) and BamHI sites of pTrc99C vector.

Purification of RRF and EFG from E.coli and M.tuberculosis

Details of the purification and biochemical characterization of these proteins will be described elsewhere. In brief, the respective ORFs were subcloned into the T7 RNA polymerase-based expression vector, pET11d, and the recombinants transformed into E.coli BL21 (DE3). Fresh cultures of the transformants were induced with 0.5 mM IPTG, and the overproduced proteins were purified by gel filtration and ion exchange column chromatography. RRF proteins were purified to apparent homogeneity, whereas the EFG preparations contained a few detectable contaminating bands. Further characterization of the proteins was done by N-terminal sequence analysis and/or ESI MS.

Production of the anti-RRF polyclonal antibodies

Antibodies against RRF were raised in New Zealand rabbits by the standard method (Harlow and Lane, 1988) using Freund's incomplete adjuvant and purified on HiTrap protein G (Pharmacia Biotech) according to the manufacturer's protocol.

Preparation of cell-free extracts and immunoblotting

The cell-free extracts were prepared from log phase cultures (2 ml) of transformants exactly as described (Varshney et al., 1991b) and quantified using Bradford's dye binding assay for protein estimation using bovine serum albumin (BSA) as a standard (Sedmak and Grossberg, 1977). Total cellular proteins (10 μg) were electrophoresed on SDS–PAGE (15%) and electroblotted (Towbin et al., 1979) onto a nitrocellulose membrane (0.45 μm; Amersham) at 200 mA for 2 h. The membrane was blocked with 3% BSA in Tris–HCl-buffered saline (TBS; 20 mM Tris–HCl pH 7.5, 150 mM NaCl) for 2 h and then incubated for 1 h at room temperature with the purified anti-RRF antibodies (1:2000 dilution). After three washings with TBS, the blots were incubated with alkaline phosphatase-conjugated goat anti-rabbit IgG (Gibco-BRL) at a 1:2000 dilution for 1 h at room temperature. After three washings with TBS, the blot was developed with 12 μM p-nitroblue tetrazolium chloride and 6 μM 5-bromo-4-chloro-3-indolyl phosphate in 0.15 M Tris–HCl pH 9.5 containing 5 mM MgCl2 to visualize the bands.

Preparation of total tRNA and northern blot analysis

Total tRNA was prepared under acidic conditions at 4°C, separated on acid urea gels and electroblotted onto nytran membranes exactly as described before (Varshney et al., 1991a). The blot was probed with a 5′-32P-labelled oligonucleotide complementary to tRNATyr exactly as described before (Varshney et al., 1991a) except that the SET buffer was substituted with the SSC buffer.

Polysome isolation

Polysomes were prepared according to the method described before (Tai and Davis, 1979) at 4°C. Briefly, E.coli MRE600 culture (1.5 l) was grown to log phase, supplemented with 0.3 mM tetracycline and quick-chilled on an ice–salt mixture. The cells were harvested by centrifugation at 5000 r.p.m. for 3 min using a GSA rotor (Sorvall), and suspended in 10 ml of buffer A (25% sucrose in 250 mM Tris–HCl pH 8.0). Subsequently, 4 ml of the lysozyme-Tris-Na2EDTA solution (250 mM Tris–HCl, 8 mM Na2EDTA pH 8.0 containing 40 mg of lysozyme) was added and, after keeping for 10 min, it was supplemented with 10 ml of a solution containing 1.67% Brij 35, 0.1% deoxycholate, 90 mM NH4Cl, 35 mM magnesium acetate. After 5 min, 0.75 ml of 1 mg/ml DNase I was added and kept for 10–15 min. The lysate was centrifuged at 15 000 r.p.m. for 10 min using an SS34 rotor (Sorvall). The supernatant was stored and the pellet was resuspended in 5 ml of buffer B (10 mM Tris–HCl pH 7.8, 1 M NH4Cl, 40 mM magnesium acetate, 10 mM 2-mercaptoethanol, 2 mM Na2EDTA) and centrifuged at 15 000 r.p.m. for 10 min. The two supernatant fractions were pooled, layered onto a 60% sucrose solution in buffer B and subjected to ultracentrifugation for 8 h at 27 000 r.p.m. using an SW 28 rotor (Beckman). The pellet (polysomes) was suspended in RRF assay buffer and stored at −70°C.

In vitro assay for RRF activity

Various RRFs and EFGs were assayed using E.coli polysomes as substrates (Ohnishi et al., 1999). The reaction mixtures (250 μl) containing polysomes (3.0 A260) and various combinations of RRF/EFG were incubated at 32°C for 20 min in a buffer (10 mM Tris–HCl pH 7.4, 80 mM NH4Cl, 8.2 mM MgSO4, 1 mM dithiothreitol, 10 μM puromycin and 160 μM GTP) and fractionated on 15–30% sucrose gradients by centrifugation at 40 000 r.p.m. for 1 h using an SW50.1 Ti rotor (Beckman). The gradients were scanned from top to bottom at 254 nm using an ISCO gradient fraction collector.

Acknowledgements

We thank Dr U.L.RajBhandary for his suggestions on this manuscript, and Drs A.Kaji, University of Pennsylvania, Philadelphia, PA, P.E.March, University of New South Wales, Sydney, Australia, and J.R.Menninger, The University of Iowa, Iowa City, IA, for kindly providing us with the E.coli strains used in this study. We are grateful to Dr K.V.A.Ramaiah, University of Hyderabad, Hyderabad, India for his guidance and use of the equipment for the polysome profiles. This work was supported by research grants from the Department of Biotechnology, and the Department of Science and Technology, Government of India, New Delhi.

Ancillary