Labeling the peptidyltransferase center of the Escherichia coli ribosome with photoreactive tRNA(Phe) derivatives containing azidoadenosine at the 3' end of the acceptor arm: a model of the tRNA-ribosome complex.

Abstract

Photoreactive derivatives of yeast tRNAh containing 2-azidoadenosine (2N3A) at position 73 or 76 have been crosslinked to the peptidyl site of Escherichia coli ribosomes. Covalent tRNA-ribosome attachment was dependent upon the replacement of adenosine by 2N3A in the tRNA, irradiation with 300-nm light, and the presence of poly(U). In all cases, the modified tRNAs became crosslinked exclusively to 50S ribosomal subunits. While the tRNA derivative containing 2N3A at position 73 labeled only protein L27, that containing 2N3A at position 76 labeled proteins L15, L16, and L27 as well as a segment of the 23S rRNA. The site of crosslinking in the rRNA was identified as guanosine-1945, which lies within a highly conserved sequence adjacent to a number of modified bases and has not until now been identifiled at the peptidyl- The formation and analysis of intermolecular crosslinks provides an effective methodology for delineating the topography of tRNA binding sites on the ribosome (1, 2). The UV-induced attachment of tRNAvW' to the peptidyl (P) site of Escherichia coli ribosomes (3), for instance, led to the demonstration that the 5-carboxymethyoxyuridine at position 34 (cmo5U34), the 5' anticodon base of the tRNA, undergoes cycloaddition with, and is therefore within a few angstroms of, C1", a highly conserved nucleotide near the 3' end of 16S rRNA (4). Further, through the use of immuno electron miroscopy, the site of interaction was shown to lie in the cleft between the head and platform of the 30S ribosomal subunit (5). Precise localization of the decoding site within the ribosome was feasible in large part because of the short length of this crosslink. In contrast, contacts between the 3' end of tRNA and the ribosome at the peptidyltransferase center are still poorly defined. Thus, tRNAs containing chemically or photochemically reactive groups attached to the aminoacyl moiety have been reported to label proteins L2, 11, L14, L15, L16, L18, L23, and L27 of the 50S subunit, proteins S14 and S18 of the 30S subunit, and various segments of the 23S rRNA (2). However, immuno electron microscopic studies indicated that the proteins listed above are scattered over a large fraction of the ribosome surface (6) and clearly cannot all be in the vicinity of the tRNA 3' terminus. The diversity of proteins labeled no doubt derives, at least in part, from the size of the reactive substituents, as the crosslinks they establish are typically 10-20 A in length and may therefore extend to ribosomal components at a considerable distance from the site of tRNA binding.To develop a strategy that would yield short-range crosslinks between tRNA and the peptidyltransferase center approximating that between cmo5U34 in tRNAval and C'" in 16S rRNA, we have selectively incorporated photoreactive azidopurines into yeast tRNAPhC in place of their natural counterparts. In a previous report (7)

Calculation of the relative geometry of tRNAs in the ribosome from directed hydroxyl-radical probing data

“…Strand scission of tRNAs by hydroxyl radicals generated from 59-Fe(II)-tRNA+ tRNA f Met , complexes with tRNA f Met [ 32 P]pCp bound to the P site, and either tRNA Phe (GMP) or 59-Fe(II)-tRNA Phe (GMPS-Fe) bound to the A site+ tRNA Lys , complexes with tRNA Lys [ 32 P]pCp bound to the A site, and either tRNA Phe (GMP) or 59-Fe(II)-tRNA Phe (GMPS-Fe) bound to the P site+ Lanes U2, Phy M, and T1 are enzymatic sequencing of tRNA[ 32 P]pCp using the respective ribonucleases; Ak: alkaline hydrolysis of tRNA[ 32 P]pCp; Mock: reaction chemically treated as for the probing reaction, but in the absence of ribosome+ Positions of strand scission are indicated by the bars+ of the two tRNAs+ These two mutually exclusive arrangements have been designated as the R and S orientations, respectively (Lim et al+, 1992)+ Fluorescence resonance energy transfer experiments provided important constraints for modeling the mutual arrangement of tRNAs (Fairclough & Cantor, 1979b;Wells & Cantor, 1980;Johnson et al+, 1982;Paulsen et al+, 1983) + Paulsen et al+ (1983) proposed a model with a 60 6 308 angle between the A-and P-site tRNAs, but were unable to distinguish between the R and S orientations using FRET constraints+ Although the FRET data by themselves were insufficient to decide this issue, they served the useful purpose of eliminating models that were inconsistent with the measured distances+ Based on a stereochemical analysis of the peptidyl transferase reaction, Spirin and Lim (1986) proposed Figure 5+ an R orientation with a 1008 angle between the two tRNAs+ Ofengand et al+ (1986), based on tRNA crosslinks to 30S subunit protein S19 and other biochemical data, proposed a model with the tRNAs in the S orientation with an angle of 658 between the tRNA planes+ A detailed stereochemical model using the crystal structure of tRNA and incorporating the FRET data was proposed by McDonald and Rein (1987), who maintained the tRNA crystal structures as rigid bodies and manipulated the conformation of the mRNA+ Their model has an S orientation with an approximately 458 angle between the tRNA planes and a 338 kink between the two mRNA codons+ Models with mRNA fixed in an idealized A-form geometry, deforming the anticodon loops of the tRNAs from the crystal structure (Prabahakaran & Harvey, 1989), or where both the mRNA codons and tRNAanticodonloopshavealteredconformations (Easterwood et al+, 1994) were also proposed; the tRNAs in these models are in the S orientation with an angle of approximately 458+ While the S orientation is preferred in some of the models based on crosslinking and chemical protection studies (Stern et al+, 1988;Wower et al+, 1989Wower et al+, , 1993Noller et al+, 1990;Nagano et al+, 1991;Nagano & Nagano, 1997), Lim and coworkers invoked extensive tRNA, mRNA, and rRNA crosslinking data to derive a model with R ...…”

“…Transfer RNAs (tRNAs) are the substrates used by ribosomes for synthesizing proteins+ To understand the mechanism of translation at the molecular level, it is essential to understand how tRNAs interact with the ribosome (Rheinberger et al+, 1981;Grajevskaja et al+, 1982;Kirillov et al+, 1983;Lill et al+, 1984)+ The anticodon arm of P-site tRNA interacts with mRNA in the vicinity of the cleft formed by the platform and the head of the 30S subunit (Lake, 1980;Gornicki et al+, 1984)+ The acceptor arms of A-site and P-site tRNAs interact with the peptidyl transferase center, which lies close to the base of the central protuberance of the 50S subunit (Ofengand, 1980;Olson et al+, 1982;Wower et al+, 1989)+ During peptide bond formation, the CCA-39 termini of A-site and P-site tRNAs must be proximal to allow peptidyl transfer, while their anticodons interact with adjacent codons on mRNA (Fairclough & Cantor, 1979a;Matzke et al+, 1980)+ Several models have been proposed for the arrangement of A-site and P-site tRNAs in the ribosome based on stereochemical considerations (Fuller & Hodgson, 1967;Woese, 1970;Rich, 1974;Sundaralingam et al+, 1975;Lake, 1977;Spirin & Lim, 1986;McDonald & Rein, 1987;Prabahakaran & Harvey, 1989;Nagano et al+, 1991;Easterwood et al+, 1994), fluorescence resonance energy transfer (FRET) (Johnson et al+, 1982;Paulsen et al+, 1983), crosslinking studies (Ofengand, 1980, Ofengand et al+, 1981Wower et al+, 1989Wower et al+, ,hotra et al+, 1998, and other methods (Spirin, 1983;Smith & Yarus, 1989;Wagenknecht et al+, 1989;Nierhaus et al+, 1998)+ Most of these models can be assigned to either the R (Rich, 1974) or S (Sundaralingam et al+, 1975) orientations+ In the R orientation, the T loop of A-site tRNA ...…”

Peptidyl transferase antibiotics perturb the relative positioning of the 3′-terminal adenosine of P/P′-site-bound tRNA and 23S rRNA in the ribosome

“…First, the cross-linking levels of the P/P9-site-bound deacylated and N-Ac-Phe-(2N 3 A76)tRNA to fragments F19, F29, and F49 and to ribosomal proteins L27 and L33 were determined+ The latter were estimated after RNAse T 1 treatment of cross-linked ribosomes and polyacrylamide gel-SDS analysis (Wower et al+, 1989)+ Crosslinking yields to the individual rRNA fragments and the ribosomal proteins were closely similar for both tRNAs, indicating that they occupy the same P/P9 site (Table 2)+ These data were reinforced by estimating the molar ratios of total cross-linking to rRNA and ribosomal protein after treatment with and without proteinase K, respectively (Wower et al+, 1989)+ These estimates correlate well with the cross-linking yields to the rRNA fragments and ribosomal proteins (Table 2)+ Total cross-linking yields for the deacylated and N-AcPhe-(2N 3 A76)tRNA were about 4% and 5%, respec-FIGURE 1. Autoradiograms showing the sites of deacylated [ 32 P](2N 3 A76)tRNA cross-linking within the 23S rRNA using a primer extension approach+ A and B: Analysis of tRNA cross-links on whole 23S rRNA+ Deacylated [ 32 P](2N 3 A76)tRNA was complexed with 70S ribosomes, irradiated at 365 nm, and the extracted rRNA was analyzed by primer extension using primers EC2621 in A and primer EC2102 in B+ Lane 1: control ribosomes irradiated in the absence of deacylated [ 32 P](2N 3 A76)tRNA+ Lanes 2-4: deacylated [ 32 P](2N 3 A76)tRNA was bound under the following conditions+ Lane 2: 0 8C for 60 min; lane 3: 37 8C for 5 min; and lane 4: 37 8C for 30 min before irradiating+ The bands resulting from tRNA-rRNA cross-links are indicated by arrows, as is the tRNA band+ C: Analysis of tRNA cross-links on an isolated rRNA fragment, corresponding to F19 and F29, obtained by RNAse H cleavage using oligonucleotide EC2424 (approx+ positions 2440-2904)+ Primer extension was performed on both a control noncross-linked fragment (Co), and the cross-linked fragment (X) using primer EC2621 where G, A, U, and C represent sequencing tracks+ tively (Table 2)+ We are confident that these reflect functional ribosomal states of the tRNAs because we observed that the cross-linking yield was directly proportional to the amount of tRNA bound to ribosomes for complexes prepared at increasing tRNA:ribosome molar ratios of mixing in the range of 0+1-1+0 (data not shown), in agreement with the conclusions of an earlier study (Wower et al+, 1988)+ Next, we examined the effect of several antibiotics, each of which is known to inhibit peptide bond formation or perturb the nascent peptide in vitro (Table 1)+ A polyacrylamide gel is presented showing the relative crosslinking yields to fragments F19, F29, and F49 in the presence of N-Ac-Phe-(2N 3 A76)tRNA and the antibiotics (Fig+ 3A)+ The annealing sites of the oligonucleotides used to produce the 39 fragments, in a combined RNAse H digestion, are indicated on the rRNA secondary structure in Figure 3B+ The gel shows changes in the yields of rRNA fragments for each of the antibiotics and the quantified data are summarized in Table 3+ Control experiments were performed with preirradiated (2N 3 A76)tRNA with each antibiotic, and no tRNA-rRNA cross-linking bands were detected (data supplied to editor)+ In general, the data were reproducible for independently prepared tRNA-ribosome complexes provided (1) a single RNAse H digestion was made with a mixture of oligonucleotides that ensured 100% cleavage at all sites (Fig+ 3A) and (2) that fairly fresh stocks of antibiotics were used that had not undergone multiple freezing and thawing+ Care was also taken to ensure that the level of N-acetylation was high (.95%), although this produced...…”

“…Autoradiograms showing the sites of deacylated [ 32 P](2N 3 A76)tRNA cross-linking within the 23S rRNA using a primer extension approach+ A and B: Analysis of tRNA cross-links on whole 23S rRNA+ Deacylated [ 32 P](2N 3 A76)tRNA was complexed with 70S ribosomes, irradiated at 365 nm, and the extracted rRNA was analyzed by primer extension using primers EC2621 in A and primer EC2102 in B+ Lane 1: control ribosomes irradiated in the absence of deacylated [ 32 P](2N 3 A76)tRNA+ Lanes 2-4: deacylated [ 32 P](2N 3 A76)tRNA was bound under the following conditions+ Lane 2: 0 8C for 60 min; lane 3: 37 8C for 5 min; and lane 4: 37 8C for 30 min before irradiating+ The bands resulting from tRNA-rRNA cross-links are indicated by arrows, as is the tRNA band+ C: Analysis of tRNA cross-links on an isolated rRNA fragment, corresponding to F19 and F29, obtained by RNAse H cleavage using oligonucleotide EC2424 (approx+ positions 2440-2904)+ Primer extension was performed on both a control noncross-linked fragment (Co), and the cross-linked fragment (X) using primer EC2621 where G, A, U, and C represent sequencing tracks+ tively (Table 2)+ We are confident that these reflect functional ribosomal states of the tRNAs because we observed that the cross-linking yield was directly proportional to the amount of tRNA bound to ribosomes for complexes prepared at increasing tRNA:ribosome molar ratios of mixing in the range of 0+1-1+0 (data not shown), in agreement with the conclusions of an earlier study (Wower et al+, 1988)+ Next, we examined the effect of several antibiotics, each of which is known to inhibit peptide bond formation or perturb the nascent peptide in vitro (Table 1)+ A polyacrylamide gel is presented showing the relative crosslinking yields to fragments F19, F29, and F49 in the presence of N-Ac-Phe-(2N 3 A76)tRNA and the antibiotics (Fig+ 3A)+ The annealing sites of the oligonucleotides used to produce the 39 fragments, in a combined RNAse H digestion, are indicated on the rRNA secondary structure in Figure 3B+ The gel shows changes in the yields of rRNA fragments for each of the antibiotics and the quantified data are summarized in Table 3+ Control experiments were performed with preirradiated (2N 3 A76)tRNA with each antibiotic, and no tRNA-rRNA cross-linking bands were detected (data supplied to editor)+ In general, the data were reproducible for independently prepared tRNA-ribosome complexes provided (1) a single RNAse H digestion was made with a mixture of oligonucleotides that ensured 100% cleavage at all sites (Fig+ 3A) and (2) that fairly fresh stocks of antibiotics were used that had not undergone multiple freezing and thawing+ Care was also taken to ensure that the level of N-acetylation was high (.95%), although this produced only minimal differences from the results with the deacylated tRNA (Table 3)+ In some experiments with lincomycin, RNAse H inhibition and/or rRNA aggregation was observed and, therefore, only Data are presented as counts per minute of cross-linked (2N 3 A76) tRNA for one picomole of ribosome-bound (2N 3 A76)tRNA+ The specific activity of the (2N 3 A76)tRNA was normalized to an initial activity of 10,000 dpm/pmol or about 1,800 cpm/pmol estimated in an Instant Imager+ Eighteen counts per minute corresponds to a 1% crosslink of (2N 3 A76)tRNA at a given site+ Each experiment was repeated between four and seven times and the estimated error limits are given in parentheses+ two reliable sets of data were collected for this drug (Table 3)+ The cross-linking yields to ribosomal proteins L27 and L33 were determined as described and illustrated earlier (Wower et al+, 19...…”

“…Although the contribution of ribosomal components to the formation of peptide bonds is still not understood, it is known that the peptidyl transferase loop region of 23S-like rRNA plays an important role Khaitovich et al+, 1999)+ The results of diverse experimental approaches are consistent with this rRNA region contributing to the binding sites of the 39 termini of both donor and acceptor tRNA substrates and of many antibiotics that are known to inhibit peptide bond formation (reviewed in Kirillov et al+, 1997)+ The approaches include rRNA footprinting and damage-selection analyses of tRNA substrates and antibiotics bound to ribosomes (Moazed & Noller, 1989;Rodriguez-Fonseca et al+, 1995;Bocchetta et al+, 1998) and the characterization, within this rRNA region, of several single-site mutations that confer antibiotic resistance on mutant strains that lead to changes in tRNA substrate binding and/or a lowering of peptidyl transferase activity on isolated mutant ribosomes (e+g+, Porse et al+, 1996; Green et al+, 1997)+ An azido group can be introduced into the 2 position of the 39-terminal adenosine of tRNA such that when the modified tRNA is bound in the P/P9 site, which is identical to the earlier defined P site (Kirillov et al+, 1997), and irradiated with ultraviolet (UV) light at 365 nm, cross-links are generated with 23S rRNA and ribosomal proteins in the P9 site of the 50S subunit (Wower et al+, 1988(Wower et al+, , 1989(Wower et al+, , 1995)+ The rRNA cross-links have been localized within three fragments, F1 (nt 2570-2590), F2 (nt 2500-2520), and F4 (nt 2060-2080), all of which extend into the peptidyl transferase loop region of 23S RNA (Wower et al+, 1995)+ The main crosslinked proteins are L27 (strong) and L33 (weak) (Wower et al+, 1989), both of which assemble on domain V of 23S RNA that contains the peptidyl transferase loop Østergaard et al+, 1998)+ The former protein was also shown earlier to react preferentially with an affinity-labeled peptide moiety linked to a P/P9-site-bound tRNA (Eilat et al+, 1974)+ In this study, we first identified the cross-linked sites at a nucleotide level on the 23S rRNA of azidoderivatized deacylated (2N 3 A76)tRNA and N-Ac-Phe-(2N 3 A76)tRNA bound at the P/P9 site in the presence of poly(U)+ Then we chose several antibiotics that either inhibit peptide bond formation or perturb the nascent peptide on Escherichia coli ribosomes (Table 1) and examined their effects on the cross-linking yields of the azido-derivatized tRNAs to rRNA and ribosomal proteins+ Each antibiotic affected cross-linking yields almost exclusively at the rRNA sites+…”