Substrate-assisted catalysis of peptide bond formation by the ribosome

24 downloads 0 Views 542KB Size Report
Oct 10, 2004 - plays an essential role in peptide bond formation. ...... Bass, B.L. & Cech, T.R. Ribozyme inhibitors: deoxyguanosine and dideoxyguanosine.
© 2004 Nature Publishing Group http://www.nature.com/nsmb

ARTICLES

Substrate-assisted catalysis of peptide bond formation by the ribosome Joshua S Weinger1,3, K Mark Parnell1,3, Silke Dorner2, Rachel Green2 & Scott A Strobel1 The ribosome accelerates the rate of peptide bond formation by at least 107-fold, but the catalytic mechanism remains controversial. Here we report evidence that a functional group on one of the tRNA substrates plays an essential catalytic role in the reaction. Substitution of the P-site tRNA A76 2′ OH with 2′ H or 2′ F results in at least a 106-fold reduction in the rate of peptide bond formation, but does not affect binding of the modified substrates. Such substrate-assisted catalysis is relatively uncommon among modern protein enzymes, but it is a property predicted to be essential for the evolution of enzymatic function. These results suggest that substrate assistance has been retained as a catalytic strategy during the evolution of the prebiotic peptidyl transferase center into the modern ribosome.

The ribosome is the molecular machine responsible for protein synthesis in all living cells. It catalyzes peptide bond formation between the α-amino group of an aminoacyl tRNA (A-site substrate) and the C-terminal carboxyl carbon of a peptidyl tRNA (P-site substrate) (Fig. 1a)1. The nascent peptide and the incoming amino acid are both covalently connected to the last nucleotide (A76) of their respective tRNAs via ester linkages to the O3′ oxygen1,2. This linkage is adjacent to the vicinal 2′ OH group on the ribose sugar, which increases the lability of the ester bond in solution by 40-fold3. The reaction is expected to proceed through a tetrahedral intermediate that collapses to form the products, a deacylated P-site tRNA and an A-site tRNA containing the peptide lengthened by one amino acid1. The active site of the ribosome is made up entirely of RNA, with no protein within ∼20 Å of the site of nucleophilic attack, as established by the crystal structure of the Haloarcula marismortui 50S subunit in complex with a peptidyl transfer transition state analog4. Despite the availability of high-resolution structural information, the mechanism of ribosomal catalysis remains uncertain. An initial structure-based model ascribed catalytic roles to active site rRNA bases4,5; however, many of the nucleotides that are reasonably positioned in the active site for involvement in catalysis can be mutated with minimal effect on the reaction rate6–9. This suggests either that there is no chemical catalysis by the ribosome, or that the functional groups involved, such as backbone phosphates, are unchanged by mutagenesis. A third possibility is that the catalytic functional groups are provided by the substrates. A recent report attributes the majority of the ribosome’s catalytic power to entropic catalysis, arguing that substrate binding and positioning are sufficient to explain the observed rate enhancement10. This catalytic strategy would not require the involvement of specific functional groups in a direct chemical catalytic role.

Here we present evidence that the A76 2′ OH of the P-site tRNA plays an essential role in peptide bond formation. Two earlier reports explored this functional group in peptidyl transfer assays with distinctly different conclusions. One group reported that tRNAPhe lacking the A76 2′ OH are active as A-site substrates, but inactive as peptide donors11. The other found that 2′-deoxyadenosine (dA)substituted tRNALys (dA76-substituted tRNALys) is unable to participate in the in vitro poly(A)-mRNA-dependent synthesis of poly(Lys). However, in contrast with the results of the first group, the second group found that dA substitution does not affect P-site donor activity when the tRNA is nonenzymatically loaded directly into the P site12. This led them to conclude that the inactivity of the dA76-substituted tRNALys in poly(Lys) synthesis resulted from failure of the substituted tRNA to translocate from the A site to the P site13. Thus, there is a substantial discrepancy in the literature regarding the role of this functional group. Two other studies found that minimal P-site substrates lacking the A76 2′ OH are inactive as peptide donors14,15, but no subsequent investigations of the P-site A76 2′ OH have been conducted using full-size tRNAs and a complete, well-defined translation system proceeding at a physiologically relevant rate. Although the biochemistry is ambiguous, models of the peptidyl transfer reaction suggest an important role for the A76 2′ OH16–19. Forty years ago, researchers used aminolysis reactions with various acyl nucleosides to propose that the 2′ OH group participates in peptidyl transfer16. In the crystal structure of a P-site bound substrate, a tRNA fragment in which the acetyl group is free to regioisomerize between the 2′ and 3′ hydroxyls, the peptidyl moiety is unambiguously attached to the A76 O3′, not the O2′ (ref. 17). This suggests that binding to the ribosome drives the isomerization equilibrium toward the O3′ linkage. A model of the ribosomal active site containing both

1Department

of Molecular Biophysics and Biochemistry, Yale University, 260 Whitney Avenue, New Haven, Connecticut 06520-8114, USA. 2Department of Molecular Biology and Genetics, Johns Hopkins University Medical School, Howard Hughes Medical Institute, Baltimore, Maryland 21205, USA. 3These authors contributed equally to this work. Correspondence should be addressed to S.A.S. ([email protected]). Published online 10 October 2004; doi:10.1038/nsmb841

NATURE STRUCTURAL & MOLECULAR BIOLOGY VOLUME 11 NUMBER 11 NOVEMBER 2004

1101

ARTICLES

© 2004 Nature Publishing Group http://www.nature.com/nsmb

Figure 1 The peptidyl transferase reaction and characterization of modified tRNALys containing A76, dA76 or fA76. (a) The peptidyl transfer reaction occurs when the α-amino group of the A-site substrate nucleophilically attacks the carbonyl carbon of the P-site substrate. Ribosomal bases critical in positioning the CCA end of the tRNA are indicated (E. coli numbering). The 2′ OH of the P-site tRNA A76 is in the immediate vicinity of the reaction and could be positioned to play a catalytic role. (b) Modified tRNALys molecules possessing 3′-terminal adenosine (A76), 2′-deoxyadenosine (dA76) or 2′-deoxy,2′-fluoroadenosine (fA76) were 5′-radiolabeled with [γ-32P]ATP and treated with periodate and aniline (+) or aniline only (–) to assay for extent of A76 present at the 3′ end. The tRNAs were separated by 10% (w/v) PAGE.

relevant rate20. Our results indicate that in the context of full-size tRNAs, the A76 2′ OH of the P-site tRNA is essential for donor activity. On the basis of these observations, we present possible models that may explain its critical role in substrate-assisted catalysis.

A-site and P-site substrates was constructed in silico from crystal structures of each substrate solved separately17. In the composite model the adjacent 2′ OH is in position to accept a hydrogen bond from the incoming α-amino group. The positioning of this functional group between the nucleophile and the O3′ leaving group further implicates it in a catalytic role, as it is one of very few functional groups sufficiently close to the nascent peptide bond to contribute directly to catalysis17. Another modeling study involving energy minimization of the tetrahedral intermediate, which was carried out before the structural determination of the ribosome, suggested that the A76 2′ OH may function in the transfer of protons during the reaction18. Because of the potential importance of the P-site tRNA A76 2′ OH in catalysis, we re-evaluated its role in the peptidyl transfer reaction. In addition to yielding conflicting results, previous studies of modified full-size tRNAs were done in minimal peptidyl transfer systems, and did not measure rates of peptidyl transfer. Instead, these studies monitored the presence or absence of product at a single time point11,12; thus large changes in the rate of chemistry upon the modification of A76 2′ OH may have gone undetected. To explore the different potential roles for the P-site tRNA A76 2′ OH, we modified the 2′ positions of full-size tRNA molecules using both dA and 2′-deoxy-2′-fluoro (fA) substitutions. The effects of these substitutions were characterized using the recently developed peptidyl transfer assay in which chemistry is rate-limiting and the reaction proceeds at a physiologically

RESULTS tRNAs with 2′ modifications at A76 We tested the role of the A76 2′-OH in the A-site and P-site tRNAs by replacing the terminal nucleotide of tRNALys with dA76 or fA76. A76 was quantitatively removed by treating native tRNALys twice with periodate, aniline and polynucleotide kinase21. The adenosine derivatives (A, dA or fA) were incorporated using the CCA-adding enzyme in the presence of CTP and the triphosphate of the desired adenosine analog. The quantitative incorporation of dA or fA at A76 was confirmed by the resistance of the resulting dA and fA containing tRNAs to periodate treatment, which requires a cis diol for reactivity (Fig. 1b). Peptidyl transfer activity of modified tRNAs The peptidyl transfer activity of the modified tRNAs was tested in a previously described rapid kinetics peptidyl transfer assay20. This system uses Escherichia coli 70S ribosomes in the presence of an RNA message, and is dependent on initiation and elongation factors for loading and translocation of the tRNAs (Fig. 2). Initially f-[35S]MettRNAfMet was enzymatically loaded into the P site by initiation factors IF1, IF2 and IF3. Lys-tRNALys, which corresponds to the second codon on the mRNA, was loaded into the A site by EF-Tu. Functionality of the tRNALys in the A site was determined by its ability to accept the fMet and form the dipeptide fMet-Lys. EF-G catalyzes the translocation of the fMet-Lys-tRNALys from the A site to the P site. P-site donor activity was measured by adding puromycin, a small A-site substrate,

Figure 2 The rapid kinetic peptidyl transfer assay. E. coli 70S ribosomes are provided with an mRNA encoding a peptide starting with the three amino acids formyl methionine, lysine and phenylalanine. f-[35S]Met-tRNAfMet is loaded into the P site with the assistance of initiation factors IF1, IF2 and IF3. Lys-tRNALys can then be loaded by EF-Tu into the A site and acceptor activity can be inferred from its ability to accept formyl methionine from f-[35S]Met-tRNAfMet to form the dipeptide fMetLys. After EF-G-catalyzed translocation to the P site, donor activity can be measured by monitoring the formation of the tripeptide fMetLys-puromycin. Puromycin binds rapidly in the A site, and chemistry is rate-limiting for the reaction.

1102

VOLUME 11 NUMBER 11 NOVEMBER 2004 NATURE STRUCTURAL & MOLECULAR BIOLOGY

© 2004 Nature Publishing Group http://www.nature.com/nsmb

ARTICLES Figure 3 Peptidyl donor activity of modified tRNAs. (a) Thin-layer electrophoresis showing the time courses for reactants and products of peptidyl transfer using modified tRNAs as P-site substrates. The tRNAs tested are indicated: A76 tRNALys, dA76 tRNALys, fA76 tRNALys, and none (a control in which no tRNALys was added). Pmn (–) indicates no Pmn added, whereas Pmn (+) indicates that Pmn was added to a final concentration of 10 mM, with incubation times of 10 s, 1 min, 6 min, 1 h and 24 h. Arrows indicate the spots corresponding to f-[35S]Met, f-[35S]Met-Lys, f-[35S]MetLys-Pmn, and the origin. Initial conversion of f-[35S]Met to f-[35S]Met-Lys indicates activity as an A-site substrate. Subsequent conversion to f[35S]Met-Lys-Pmn upon addition of Pmn indicates activity as a P-site substrate. (b) Time courses (10 s to 24 h) for peptidyl transfer reaction using Phe-tRNAPhe as the A-site substrate and assaying for the conversion of f-[35S]Met-Lys to f-[35S]Met-Lys-Phe. (c) Time course of peptidyl transfer reactions containing N-acetylated [14C]Lys-tRNALys tRNAs (A76, dA76 or fA76) directly loaded into the P site, with 10 mM Pmn as the A-site substrate. Fraction reacted represents the conversion from (N-Ac)2-[14C]Lys to (N-Ac)2-[14C]Lys-Pmn, as measured by phosphorimager analysis of substrate and product separated by thin-layer electrophoresis. Reaction rates were generated by fitting data to an exponential function.

and monitoring the formation of the tripeptide fMet-Lys-puromycin. This assay is particularly useful for evaluating P-site substrates because puromycin binds very rapidly in the A site, so chemistry is expected to be rate-limiting at saturating concentrations of puromycin10,20. tRNAs containing dA76 or fA76 substitutions were active as A-site substrates. Modified tRNALys was loaded into the A site in the presence of P-site loaded f-[35S]Met-tRNAfMet and the reaction assayed for the formation of the dipeptide f-[35S]Met-Lys. All three of the tRNAs (A76, dA76 and fA76) were capable of efficiently forming the dipeptide f-[35S]Met-Lys (Fig. 3a, compare lanes 1, 7 and 13 to lane 19). This indicates that deletion of the A76 2′ OH does not severely affect the tRNA’s ability to serve as an A-site substrate, consistent with a previous report12. The modified tRNAs were inactive when tested for their ability to function as P-site substrates (Fig. 3a, for example compare lanes 2, 8 and 14). After dipeptide formation and translocation to the P-site, the A76 tRNA reacted efficiently with puromycin and had reached the reaction endpoint (∼50% reacted) before the first time point (10 s; the rate of this reaction is ∼10 s–1)7,20. In contrast, dA76- and fA76-substituted tRNAs showed no reactivity, even after incubation to the reaction endpoint of 24 h (Fig. 3a, lanes 12 and 18). A slight amount of product (∼2%) appeared after 10 s (Fig. 3a, lanes 7 and 13), but this seems to result from a small amount of contaminating A76 tRNALys in one of the reaction components. The amount of this product did not increase with longer incubations, and it was less than that observed in the control to which no modified tRNALys was added (Fig. 3a, compare lanes 12, 18 and 24), suggesting that the supplied tRNALys derivative competed directly with this contaminant. The high efficiency of tRNA charging (∼60%) and the stoichiometry of Lys-tRNALys make it unlikely that a high concentration of deacylated tRNA can account for the observed defects. Given the extremely slow rate of reaction by the modified tRNA substrates, ribosomal inactivation was a significant competing reaction. This inactivation rate was measured by incubating the fMet-LystRNALys-70S ribosome complexes (A76) in reaction buffer for times ranging from a few minutes to 24 h, at which point puromycin was added. Peptidyl A76 tRNAs in active complexes were converted to tripeptide, whereas the inactive complexes remained as dipeptides. On the basis of this analysis, the inactivation of ribosomes containing A76 peptidyl tRNAs proceeded at the rate of (1.7 ± 0.6) × 10–4 s–1. This inactivation is most likely due to hydrolysis of the fMet-Lys-tRNALys, though ribosome decomposition and irreversible release of the tRNA

are also possibilities. Given two competing and irreversible reactions, the maximum estimate for the rate of tripeptide formation can be calculated using equation F = k1 / (k1 + k2) where k1 is the rate of tripeptide formation, k2 is the rate of inactivation of P-site complex, and F represents the fraction of dipeptide converted to tripeptide at the ribosomal inactivation endpoint (t = 24 h). On the basis of the observed inactivation rate of 1.6 × 10–4 s–1 and conservatively estimating the observed reacted fraction at the endpoint (24 h) to be ≤5%, the maximum rate of tripeptide formation for the dA76 and fA76 tRNAs is no greater than 9 × 10–6 s–1. If hydrolysis is the cause of ribosomal inactivation, k2 and k1 would be even slower for the dA76 and fA76 tRNAs owing to the reduced rate of hydrolysis of the dipeptide from the 2′-deoxyadenosine and 2′-fluoroadenosine3. Thus, ∼10–5 s–1 is a conservative estimate of the maximal peptidyl transferase rate for the modified tRNAs. Previous studies have shown that defects in peptidyl transferase activity resulting from active site mutations are observed using Pmn as an A-site substrate, but are largely eliminated when assayed with an intact aminoacyl A-site tRNA7. To test whether this was the case for these modified tRNAs, we repeated the assay using intact Phe-tRNAPhe as the A-site substrate (instead of puromycin) and monitored the production of the fMet-Lys-Phe tripeptide. Again, A76 tRNA reacted to its endpoint in