Skip to main content
NCBI home page
RNA logoLink to RNA
. 2008 Jul;14(7):1390–1398. doi: 10.1261/rna.1020708

Initiating translation with D-amino acids

Yuki Goto 1,2, Hiroshi Murakami 1, Hiroaki Suga 1,2,3
PMCID: PMC2441986  PMID: 18515548

Abstract

Here we report experimental evidence that the translation initiation apparatus accepts D-amino acids (Daa), as opposed to only L-methionine, as initiators. Nineteen Daa, as the stereoisomers to their natural L-amino acids, were charged onto initiator tRNAfMet CAU using flexizyme technology and tested for initiation in a reconstituted Escherichia coli translation system lacking methionine, i.e., the initiator was reprogrammed from methionine to Daa. Remarkably, all Daa could initiate translation while the efficiency of initiation depends upon the type of side chain. The peptide product initiated with Daa was generally in a nonformylated form, indicating that methionyl-tRNA formyltransferase poorly formylated the corresponding Daa-tRNAfMet CAU. Although the inefficient formylation of Daa-tRNAfMet CAU resulted in modest expression of the corresponding peptide, preacetylation of Daa-tRNAfMet CAU dramatically increased expression level, implying that the formylation efficiency is one of the critical determinants of initiation efficiency with Daa. Our findings provide not only the experimental evidence that translation initiation tolerates Daa, but also a new means for the mRNA-directed synthesis of peptides capped with Daa or acyl-Daa at the N terminus.

Keywords: D-amino acid, translation, initiation, genetic code reprogramming, flexizyme

INTRODUCTION

The translation machinery polymerizes α-amino acids according to the sequence information encoded in the open reading frame of the mRNA, designating the length and sequence of the synthesized polypeptide composed of 20 proteinogenic α-amino acids with L-stereo configuration (Laa). The main player that governs the strict use of Laa is aminoacyl-tRNA synthetases (aaRSs) that are able to discriminate cognate Laa against not only noncognate proteinogenic Laa but also nonproteinogenic ones including D-amino acids (Daa); thus aaRSs play a central role in refusing noncognate amino acids from the incorporating elements (Söll 1990; Sankaranarayanan and Moras 2001). However, even if the tRNA aminoacylation step is circumvented, Daa cannot be efficiently incorporated into the nascent peptide chain during elongation. For instance, a variety of Daa precharged onto an “amber” suppressor tRNACUA have been examined for elongation; they are either modestly or in many cases not at all incorporated into the nascent peptide chain (Roesser et al. 1989; Bain et al. 1991; Ellman et al. 1992; Starck et al. 2003; Tan et al. 2004; Murakami et al. 2006). This has been attributed to the fact that either elongation factor (EF-Tu) or ribosome (or possibly both) does not allow Daa-tRNACUA to read the amber stop codon, resulting in the undesired termination of peptide synthesis executed by release factor. A more recent attempt to use ribosome mutants has given a modest increase in efficiency for the incorporation of DMet and DPhe (Dedkova et al. 2003, 2006), but it is yet unclear how generally this approach is applicable to a variety of Daa.

Like elongation, the initiation event is also strictly governed by MetRS and initiation factors (IFs) (Kozak 1983; Gold 1988; Gualerzi and Pon 1990). In the prokaryotic translation system, peptide synthesis is exclusively initiated with N a-formyl methionine (f-LMet) (Kozak 1983). To circumvent this limitation, we have recently shown that upon using precharged aminoacyl-tRNAfMet CAU in a reconstituted Escherichia coli cell-free translation system (more details are discussed below), initiator Met can be reassigned to other proteinogenic Laa and peptide synthesis successfully initiated (Goto et al. 2008). This suggests that the initiation governance can be overridden by such a genetic code reprogramming strategy. On the other hand, we have no knowledge of whether or not Daa can adapt to the initiation event.

Here we report for the first time, to the best of our knowledge, that translation can be initiated with Daa. We have demonstrated that the translation apparatus tolerates any of 19 Daa at initiation with efficiencies depending upon the type of their side chains. The most intriguing discovery is that peptides initiated with Daa are not formylated in most cases. This result is in sharp contrast to the fact that peptides initiated with Laa are generally formylated. Moreover, the use of pre-N-acetylated Daa (Ac-Daa) significantly enhances the expression of peptides initiated with Daa, suggesting that formylation efficiency is one of the critical determinants of expression yield initiated with Daa. Our study represents the first evidence that translation initiation tolerates a variety of amino acids independent of stereochemistry and also offers a new means to generate various peptides containing Daa at the N terminus by translation.

RESULTS

DMet acts as an initiator upon the reprogrammed initiation

To facilitate the reprogramming of the translation initiation, we integrated two systems, PURE and flexizyme. The former system is a reconstituted E. coli cell-free translation system (PURE stands for protein synthesis using recombinant elements) (Shimizu et al. 2001). Using this system, we are able to withdraw methionine (Met) from the translation components (referred to as wPURE system), making the initiation codon (AUG) vacant (Fig. 1). The flexizyme system consists of artificially evolved ribozymes, enabling us to charge a wide variety of amino acids onto any desired tRNAs, including the initiator tRNAfMet CAU (Murakami et al. 2003a,b, 2006). Indeed, we have previously reported that various Laa-tRNAfMet CAU molecules prepared by the flexizyme system can function as an initiator in the Met-withdrawn wPURE system (Goto et al. 2008). Importantly, the flexizyme system tolerates not only Laa with nonproteinogenic side chains but also Daa with a variety of side chains (Murakami et al. 2006). Thus, the integration of these two systems facilitates the reassignment of the vacant initiation codon to Daa (Fig. 1).

FIGURE 1.

FIGURE 1.

Open in a new tab

Genetic code reprogramming of the initiation event with Daa. The initiator can be reassigned to Daa instead of LMet by the integration of wPURE system and flexizyme system. Initiation with Daa results in the nascent peptide containing Daa at the N terminus.

To examine whether the Daa-tRNAfMet CAU molecules prepared by the flexizyme system can initiate the translation reaction in the wPURE system, we designed a mRNA sequence for the expression of a 14-mer peptide (Fig. 2A). The flag peptide sequence (DYKDDDDK; D = Asp, Y = Tyr, K = Lys) (Brizzard et al. 1994) included in this peptide acted as a [14C]-Asp-labeling tag for detecting the expression level in tricine-SDS PAGE upon addition of [14C]-Asp and also as a purification tag for isolating the full-length peptide for MALDI-TOF analysis.

FIGURE 2.

FIGURE 2.

Open in a new tab

Tolerance of various Daa in initiation. (A) The mRNA sequence that expresses peptides initiated by various Daa. Flag in parentheses indicates the RNA sequence encoding the Flag peptide sequence (DYKDDDDK). (B) Tricine-SDS PAGE analysis of the translation products. (Lane 1) expression of wild type; (lane 2) in the absence of Met; (lane 3) reprogrammed initiation with LMet; (lanes 4–22) reprogrammed initiation with various Daa. Each expression level relative to the wild type is determined by a mean score of duplicates or more. The Daa giving >25%, 10%–25%, 1%–10%, and <1% of the wild-type expression level are highlighted in orange, pink, cyan, and gray, respectively. The band indicated by an asterisk corresponds to free [14C]-Asp that remained unincorporated into the Flag peptide.

Prior to examining Daa-initiation, we performed control experiments to ensure that the reprogrammed initiation would work as planned. Tricine-SDS PAGE analysis (Schagger and von Jagow 1987) of the peptide expressed in the ordinary PURE system yielded a single evident band, whereas the same assay using the wPURE system yielded only negligible background bands originated from “in-frame” misinitiations (see below). Most importantly, no band corresponding to the full-length peptide appeared in this analysis, indicating that LMet was in fact depleted in the wPURE system (Fig. 2B, lanes 1,2). On the other hand, when LMet-tRNAfMet CAU was added to the wPURE system, an evident band appeared with the same mobility and intensity as that observed in lane 1 (Fig. 2B, lane 3). MALDI-TOF analysis of the flag-purified cold (nonradiolabeled) peptide expressed under the same conditions as lane 3 showed the N-terminal formylated peptide with the expected molecular weight (Fig. 3, LMet), implying that Nα-formylation of LMet-tRNAfMet CAU occurred by methionyl-tRNA formyltransferase (MTF) present in the translation system, and the resulting f-LMet-tRNAfMet CAU exclusively initiated the translation. These control results were consistent with our previously reported results (Goto et al. 2008).

FIGURE 3.

FIGURE 3.

Open in a new tab

MALDI-TOF mass spectra of the translated peptides. * and † indicate a peak corresponding to the H- and f-peptide, respectively. “BG” indicates peaks corresponding to the background expression originated in-frame misinitiation. ‡ indicates the peak corresponding to f-DSer-peptide proposed in this work. The calculated mass (C) and observed mass (O) are shown in each spectrum.

Next, the same DNA template was translated with the wPURE system in the presence of DMet-tRNAfMet CAU (Fig. 2B, lane 4). Tricine-SDS PAGE analysis of the product gave a single evident band, but the band moved slightly faster than that produced by LMet-initiation (Fig. 2B, cf. lanes 4 and 3). Its molecular weight analysis by MALDI-TOF mass spectrometry indicated two peaks consistent with those of the nonformylated peptide (H-peptide) and formylated peptide (f-peptide) in an ∼1:1 ratio (Fig. 3, DMet; * and † indicate H- and f-peptides, respectively). Because LMet-initiation yielded the expected peptide as a fully formylated form, this result left us with two questions: (1) Was the partial formylation observed in the peptide caused by an incomplete formylation of DMet-tRNAfMet CAU catalyzed by MTF, resulting in the mixture of H-DMet-peptide and f-DMet-peptide? As an alternative scenario, (2) was DMet-tRNAfMet CAU racemized during the translation, yielding the mixture of H-DMet-peptide and f-LMet-peptide (and possibly f-DMet-peptide)?

Translation can be initiated with various kinds of Daa

Before addressing the above questions by additional sets of experiments, we decided to survey the expressions of the same DNA template initiated with the rest of 18 types of Daa-tRNAfMet CAU prepared by the flexizyme system (Fig. 2B, lanes 5–22). We found an apparent band of the product initiated with three Daa (Fig. 2B, DTyr, DCys, and DPhe [highlighted in orange]) with >25% initiation efficiency compared with that observed in wild-type expression (f-LMet-peptide) and a faint (10%–25%) yet evident band of the product initiated with four Daa (Fig. 2B, DSer, DThr, DTrp, and DHis [highlighted in pink]). Moreover, the expression initiated with six Daa (Fig. 2B, DPro, DGln, DArg, DVal, DAla, and DLeu [highlighted in cyan]) yielded a very faint (1–10%) yet clearly observable band. We isolated the full-length peptides from all translation samples by flag purification, and their molecular weight was analyzed by MALDI-TOF mass spectrometry. To our surprise, all samples gave the molecular weight corresponding to the respective peptide initiated with the expected amino acid (Fig. 3), including those that did not yield a detectable band on tricine-SDS PAGE analysis (Fig. 2B, DAsn, DAsp, DGlu, DIle, and DLys [highlighted in gray]). Even more surprisingly, the N terminus of the majority of peptides was the nonformylated form, i.e., H-peptides. Two exceptions were, however, observed; initiation with DSer yielded a mixture of H-Ser- and f-Ser-peptides (Ser denotes a single D-stereoisomer or a mixture of D- and L-stereoisomers), while DCys yielded three peaks, two of which were consistent with H-Cys- and f-Cys-peptides.

The MALDI-TOF analysis of Daa-initiated peptides that did not yield a clear band (Fig. 2B, highlighted in gray) generally gave a poor signal/noise ratio with one or occasionally two peaks originating from the background expression of mRNA by “in-frame” misinitiations (see Fig. 3, DIle; Supplemental Fig. S1). However, the most important outcome was that 16 out of 19 Daa-tRNAfMet CAU initiated peptide synthesis without formyl modification on their N terminus, giving the corresponding H-peptides only. This was in sharp contrast to our previously reported result that Laa-initiated peptides were generally formylated at their N terminus (Goto et al. 2008). This unmistakable difference in occurrence of the N-terminal formylation of peptide initiated with Laa or Daa suggested that Daa-tRNAfMet CAU in most cases was a poor substrate for MTF due to the D-stereochemistry of its α-carbon, so that it initiated the translation without Nα-formylation. This in turn suggested that Daa-tRNAfMet CAU that gave the H-peptide in MALDI-TOF analysis (Fig. 3, peaks labeled with *) was unlikely to have racemized during the translation; consequently, the observed respective H-peptide is H-Daa-peptide.

No racemization of Daa-tRNAfMet CAU occurs during translation

In order to solidify the above idea, we set competition experiments of initiation using Daa-tRNAfMet CAU against Laa-tRNAfMet CAU to see how much contamination of Laa-tRNAfMet CAU over Daa-tRNAfMet CAU would result in visualizing f-Laa-peptide. This experiment aimed at mimicking the situation in which the initiator Daa partially racemized to Laa during the translation reaction. We predicted that the competition efficiency would depend upon the type of Daa. Therefore, we chose three Daa (DTyr, DTrp, and DLeu) giving a high, moderate, and low efficiency, respectively, observed in the tricine-SDS PAGE analysis (Fig. 2B). Three different ratios, 99:1, 95:5, and 80:20, of the mixtures of the Daa-tRNAfMet CAU and Laa-tRNAfMet CAU were prepared for the expression of peptides and each translation product was analyzed by MALDI-TOF mass spectrometry (Fig. 4). A 1% L-contamination to the sample of DTyr and DTrp gave a single major peak corresponding to H-peptide, but a very tiny peak of f-peptide was accompanied in the case of DTrp-initiation. On the other hand, even with a 1% contamination of LLeu, DLeu suffered from a nearly 1:1 mixture of intense peaks of H- and f-peptides. A 5% L-contamination gave a significant increase in the relative intensity of the peaks corresponding to f-peptides in the cases of both DTrp and DLeu, while DTyr-initiation yielded only a minor peak of f-peptide. In the case of 20% L-contamination, f-LTrp- and f-LLeu-initiations almost competed out the corresponding Daa-initiation, and even for DTyr the peak ratio of H- and f-peptides became nearly 1:1. These results clearly indicated that even if a small amount, as little as 1%, of Laa-tRNAfMet CAU is contaminated in Daa-tRNAfMet CAU, the peptide product would suffer from the formation of f-peptide. Since we observed a single peak of H-peptide in the 16 cases in Figure 3, these products should be assigned to H-Daa-peptides. We thus concluded that the racemization of Daa-tRNAfMet CAU did not occur during the translation in these cases; even if it occurred, its degree should be far less than 1%, which would be negligible.

FIGURE 4.

FIGURE 4.

Open in a new tab

MALDI-TOF mass spectra of the peptides initiated with the mixture of Daa-tRNAfMet CAU and Laa-tRNAfMet CAU. * and † indicate a peak corresponding to the H- and f-peptide, respectively.

DMet-, DSer-, and DCys-tRNAfMet CAU act as modest substrates of MTF

The above studies so far supported that racemization of Daa-tRNAfMet CAU does not generally occur during the translation. However, in the case of initiation with DMet-, DCys-, or DSer-tRNAfMet CAU we did observe two peaks corresponding to H-peptide and f-peptide (for DCys there was an additional peak). Therefore, it was still necessary to verify the possibility of their racemization by another method. Accordingly, we decided to use the well-known fact that Daa are generally poor substrates in the elongation event using the amber suppression method (Noren et al. 1989; Bain et al. 1991; Ellman et al. 1992; Starck et al. 2003; Tan et al. 2004; Murakami et al. 2006). Since the flexizyme system is able to afford the Daa-tRNA molecules for initiation and elongation (charged onto tRNAfMet CAU and tRNAAsnE-2 CUA [Ohta et al. 2007], respectively) with exactly the same quality, we should be able to verify the occurrence of racemization by running the peptide synthesis with the Daa- or Laa-tRNAfMet CAU initiation and Daa- or Laa-tRNAAsnE-2 CUA elongation side by side. Thus, in parallel to the translation of the mRNA previously designed for initiating with the Daa-tRNAfMet CAU, we designed an mRNA template containing the amber codon and expressed the peptide in the presence of Daa-tRNAAsnE-2 CUA (Fig. 5A).

FIGURE 5.

FIGURE 5.

Open in a new tab

Determination of the configuration of the N-terminal residue. (A) The mRNA sequence used for the expression with amber suppression. UAG codon indicated in bold was suppressed with various amino acids. Flag in parentheses indicates the RNA sequence encoding the Flag peptide sequence (DYKDDDDK). (B) Tricine-SDS PAGE analysis of the translation products. (Lanes 1,5,9) reprogrammed initiation with Laa; (lanes 2,6,10) reprogrammed initiation with Daa; (lanes 3,7,11) amber suppression with Laa; (lanes 4,8,12) amber suppression with Daa. The types of side chain are Met in lanes 1–4, Cys in lanes 5–8, and Ser in lanes 9–12. The band indicated by an asterisk corresponds to free [14C]-Asp that remained unincorporated into the Flag peptide.

Both LMet-tRNAfMet CAU and LMet-tRNAAsnE-2 CUA functioned as the translation initiator and elongator, respectively, for the cognate mRNA templates, giving the expected full-length peptides (Fig. 5B, lanes 1,3). On the other hand, DMet-tRNAfMet CAU could initiate the translation whereas DMet-tRNAAsnE-2 CUA could not suppress the amber codon (Fig. 5B, lanes 2,4). This clearly indicated that DMet-tRNAAsnE-2 CUA was not racemized during the translation, and therefore it was reasonable to assume that DMet-tRNAfMet CAU was not either. Likewise, DCys was not incorporated into the nascent peptide chain by amber suppression (Fig. 5B, lanes 5–8), suggesting that DCys-tRNAAsnE-2 CUA was not racemized. This result also indicated that DCys-tRNAfMet CAU was not racemized and primed the translation with DCys. To our surprise, the amber suppression by DSer-tRNAAsnE-2 CUA gave an evident band as intense as that by LSer-tRNAAsnE-2 CUA; however, their mobility were slightly different (Fig. 5B, cf. lanes 11 and 12). MALDI-TOF analysis of the respective peptide products expressed in the presence of DSer-tRNAAsnE-2 CUA and LSer-tRNAAsnE-2 CUA showed the same expected molecular weight, implying that these peptides have the same compositions of sequence (Supplemental Fig. S2). These results suggested that the observed difference in mobility between lanes 11 and 12 (Fig. 5B) could be attributed to the difference in chirality of the Ser residue, i.e., DSer-tRNAAsnE-2 CUA elongated the peptide chain without racemization. This result assured that DSer-tRNAfMet CAU initiated the translation without racemization.

To this end, we concluded that neither DMet-, DCys-, nor DSer-tRNAfMet CAU were racemized, but they were partially formylated by MTF and competitively initiated the translation. Even though the f-Daa-tRNAfMet CAU was formed presumably in only a small amount, its higher initiation efficiency compared to Daa-tRNAfMet CAU could affect the outcome of translation. Consequently, we obtained a mixture of the respective H-Daa-peptide and f-Daa-peptide.

MALDI-TOF analysis of the DCys-primed peptides showed two peaks, of which the observed molecular weights were consistent with H-DCys- and f-DCys-peptides like DMet and DSer. However, there is an additional peak with the molecular weight of 1715.20 Da (Fig. 3, DCys, see peak indicated by ‡). Although it is difficult to define what exactly this peptide is, the molecular weight is consistent with f-DSer-peptide (the calculated molecular weight is 1714.81 Da). We therefore suggest the following mechanism as a possible scenario to convert DCys-primed peptide to f-DSer-peptide (Supplemental Fig. S3). The α-amino group of DCys-tRNAfMet CAU could be formylated by MTF similar to DMet and DSer, while the sulfhydryl side-chain group could also be formylated because of its inherent low pKa and high nucleophilicity. The formylated sulfhydryl group then might become a good leaving group so that the oxygen of N α-formyl group likely could attack the β-carbon of the side chain to yield an oxazoline-containing peptide. Hydrolysis of the oxazoline ring consequently might yield f-DSer-peptide. Note that this unusual event occurred in only the case of initiation with DCys, not LCys. Therefore, the D-configuration of DCys likely played a critical role in processing this unusual and interesting chemistry.

Preacylation of the α-amino group on Daa enhances initiation efficiency

In our previous work on reprogramming the initiation using various Laa, we have found that preacylation on Laa-tRNAfMet CAU, e.g., Nα-acetylation, significantly enhances the initiation efficiency (Goto et al. 2008). Therefore, we wondered whether the same trend could be observed for Daa-tRNAfMet CAU initiation. Three Daa with moderate initiation efficiencies (DMet, DTrp, and DPhe; 28%, 19%, and 38%, respectively) and two Daa with low efficiencies (DAsn and DAla; 0.4% and 3%, respectively) were chosen for the synthesis of the corresponding Ac-Daa substrates and charged onto tRNAfMet CAU using the flexizyme system. The resulting Ac-Daa-tRNAfMet CAU was used to initiate the translation to produce each peptide in parallel to the initiation with the corresponding Daa-tRNAfMet CAU (Fig. 6).

FIGURE 6.

FIGURE 6.

Open in a new tab

Tricine PAGE analysis of the translation products with Nα-acylated initiators. (Lane 1) expression of wild type; (lanes 2–11) reprogrammed initiation with Na-free amino acids or Na-acetyl amino acids. Each expression level relative to wild type is determined by a mean score of duplicates or more. The band indicated by an asterisk corresponds to free [14C]-Asp that remained unincorporated into the Flag peptide.

In all cases, the initiation with Ac-Daa-tRNAfMet CAU dramatically increased the expression level compared to that with Daa-tRNAfMet CAU; particularly, the preacetylation of DAsn and DAla increased the expression level from nearly invisible band intensity to a clearly visible intensity (Fig. 6, lanes 8–11, 0.4%–31% and 3%–56%, respectively). MALDI-TOF analyses of the respective peptides were also consistent with the expected molecular weights of the Ac-Daa-peptides (Supplemental Fig. S4), indicating that Ac-Daa-tRNAfMet CAU exclusively initiated the translation.

DISCUSSION

By means of the genetic code reprogramming for initiation, we have shown that the translation apparatus is able to use Daa in the initiation event even though its efficiency varies depending upon the type of side chain. The most intriguing finding is that in most cases the peptide product had a free N terminus, i.e., without a formyl group, indicating that MTF does not generally formylate the Daa-tRNAfMet CAU (Fig. 7). Moreover, the use of Ac-Daa-tRNAfMet CAU, an analog of formylated initiator, is able to significantly increase the initiation efficiency, giving the Ac-Daa-peptide with a higher yield than that initiated with Daa-tRNAfMet CAU. This suggests that the preacylation of Daa-tRNAfMet CAU is able to overcome the inherent modest initiation efficiency of Daa controlled by MTF.

FIGURE 7.

FIGURE 7.

Open in a new tab

Initiation versus elongation with Laa and Daa. (A) Laa charged onto tRNAfMet CAU is formylated by MTF to give f-Laa-tRNAfMet CAU; then IF2 brings it to the P site of the ribosome initiation complex and the translation is started to yield f-Laa-peptide. On the other hand, Daa charged onto tRNAfMet CAU is not generally formylated, i.e., it bypasses the formylation step; yet IF2 is able to bind and bring Daa-tRNAfMet CAU to the P site of the ribosome initiation complex; thus the translation reaction can be initiated with Daa, generally giving H-Daa-peptide. (B) Laa charged onto amber tRNACUA binds to EF-Tu and goes to the ribosome P site, and Laa is incorporated into the peptide nascent chain. Daa charged onto amber tRNACUA presumably binds to EF-Tu poorly and also is incompatible with the ribosome A site and in most cases fails to suppress the amber codon.

We have shown that the translation apparatus tolerates Daa for initiation to afford H-Daa-peptide or Ac-Daa-peptide when Daa-tRNAfMet CAU or Ac-Daa-tRNAfMet CAU is given to the Met-withdrawn wPURE system. This is in sharp contrast to the little success achieved in the incorporation of Daa into the nascent peptide chain via amber-suppression elongation (Fig. 7). Why so? To provide a definitive answer(s) to this question, more detailed mechanistic investigations are certainly required; but at the present stage we are able to suggest the following three potential reasons.

The first reason can be attributed to the nature of the mechanism for the fidelity control, which relies upon the function of initiation factor 2 (IF-2). Despite the fact that MetRS charges LMet onto both initiator tRNAfMet CAU and elongator tRNAMet CAU, IF-2 recognizes only f-LMet-tRNAfMet CAU over LMet-tRNAMet CAU (Schmitt et al. 1996; Boelens and Gualerzi 2002). The differences between these two LMet-tRNAs lies in the tRNA's body sequence (particularly that the 5′-terminal nucleotide is unpaired in the initiator whereas it is paired in the elongator) (Mayer et al. 2001) and the formyl group on LMet. Our previous and current studies of the reprogrammed initiation have shown that the formylated aminoacyl-tRNAfMet CAU very likely recruits IF-2 more efficiently than nonformylated, yet the formyl group is not the essential selection element. Apparently, a more critical selection element is the structural features of the initiator tRNAfMet CAU distinct from the elongator tRNAMet CAU (Mayer et al. 2003). Our results show that IF-2 likely tolerates not only various Laa but also even Daa (particularly when N α-acylated) as far as they are charged onto tRNAfMet CAU, suggesting that its recognition of side chains and the chirality of amino acids are less strict compared with that of the tRNA's body sequence. On the other hand, in the elongation event, it has been firmly established that EF-Tu discriminates cognate pairs of Laa and elongator tRNA against noncognate pairs upon recruiting them to the ribosome elongation complex with the strict control of a nearly uniform affinity toward the cognate pairs (Stanzel et al. 1994; Ibba and Söll 1999; LaRiviere et al. 2001). This mechanism is also applicable to pairs of nonproteinogenic amino acids and amber or other possible elongator tRNAs; particularly, the opposite chirality in Daa is forcefully rejected by this EF-Tu selection filter, resulting in the observed poor incorporation efficiency of Daa-elongation in general.

As the second reason, we suggest that the recognition of initiator Daa-tRNAfMet CAU by P site of ribosome is less strict than that of elongator Daa-tRNAs by the A site. Presumably, since the P site needs to accommodate peptidyl-tRNA, it has a more spacious pocket than the A site (Noller et al. 2005). In fact, our recent studies using various peptidyl-tRNAs for initiation imply that the P site has a surprising tolerance toward the nonproteinogenic peptidyl groups (Y. Goto, H. Murakami, and H. Suga, in prep.). Moreover, in the P site, the aminoacyl carbonyl carbon of initiator f-LMet-tRNAfMet CAU or peptidyl-tRNAs acts as an electrophile to the α-amino group of elongator Laa-tRNAs, so that it is only necessary to set its corresponding carbonyl group at the appropriate position; this positioning would be unlikely to be influenced by the chirality of the α-carbon of electrophile amino acids. In the A site, on the other hand, in order to efficiently process the nucleophilic attack of the α-amino group of elongator Laa-tRNA to the carbonyl group of the initiator or peptidyl-tRNAs, precise positioning of the α-amino group is critical. This positioning should be largely influenced by the chirality of the α-carbon of nucleophile amino acids, in contrast to that of electrophile in the P site. We thus propose that the spacious and functional differences in ribosome's P and A sites play a critical role in accepting and discriminating Daa in the initiation and elongation events, respectively.

The last reason involves a technical issue; since the background initiation was nearly completely suppressed by depleting Met in the wPURE system, undesired competition of f-Met-tRNAfMet CAU against Daa-tRNAfMet CAU does not occur, leading to the exclusive initiation of Daa-tRNAfMet CAU. On the other hand, the background of the amber suppression cannot be completely repressed due to the inclusion of RFs in the present methodology. This may suggest that if the background level of competing elongations were controlled by genetic code reprogramming, Daa can be incorporated into nascent peptide chain.

Although, as stated earlier, knowing the exact mechanism in the Daa-initiation requires more detailed investigations, particularly at the molecular level, the present work has given us an intriguing question for a general mechanism of the fidelity controls at initiation. It has been established that some aaRSs, such as TyrRS, TrpRS, and AspRS, mischarge the corresponding Daa onto the cognate elongator tRNAs (Calendar and Berg 1966; Soutourina et al. 2000). This mischarging event gives a negative impact on cell growth via two possible mechanisms. A small portion of such a mischarged Daa may be incorporated into a nascent peptide chain even though Daa is generally a poor substrate for elongation (Roesser et al. 1989; Bain et al. 1991; Ellman et al. 1992; Dedkova et al. 2003; Murakami et al. 2006). Alternatively, undesirable generation of the mischarged elongator tRNAs may decrease the concentration of available elongator tRNAs for the innate function. At least in bacteria and yeast, it has been found that DTyr-tRNA deacylase plays a role in discharging such mischarged Daa-tRNAs, and thus the harmful accumulation of Daa-tRNAs are avoided (Calendar and Berg 1967; Soutourina et al. 2004). On the other hand, we found in the present work that DMet-tRNAfMet CAU could be formylated by MTF and the resulting f-DMet-tRNAfMet CAU could rather efficiently initiate the translation. Therefore, even if MetRS mischarged only a small fraction of tRNAfMet CAU with Daa, the resulting f-DMet-tRNAfMet CAU would compete with f-LMet-tRNAfMet CAU at the initiation of protein synthesis. Hence, it is intriguing to study whether MetRS mischarges DMet onto tRNAfMet CAU. If so, it would be important to test whether DTyr-tRNA deacylase is able to discharge DMet-tRNAfMet CAU, since it has been shown that this enzyme is able to discharge not only DTyr-tRNATyr but also DTrp-tRNATrp and DAsp-tRNAAsp (Soutourina et al. 2000). If MetRS is able to mischarge DMet and if DTyr-tRNA deacylase is unable to discharge DMet-tRNAfMet CAU, it will be interesting to investigate the mechanism of how the generation of f-DMet-peptide is avoided or how f-DMet-peptide is processed, e.g., whether peptide deformylase and methionine aminopeptidase are able to remove the f-DMet group from the f-DMet-peptide.

Nevertheless, the technical merit of our work is apparent. The reprogrammed initiation with Daa and acyl-Daa in translation enables us to synthesize a variety of H-Daa and acyl-Daa-peptides, respectively. The Daa-capping would grant resistance against proteolytic degradation to the peptide as demonstrated in previous works using chemical synthesis (Hong et al. 1999; Tugyi et al. 2005). We have reported here a new method for the ribosomal synthesis of Daa-capped or acyl-Daa-capped peptides. The mRNA-programmed synthesis of such peptide libraries should provide a new avenue to discover novel physiologically stable peptidic drug candidates against various therapeutic targets, and such investigations are underway.

MATERIALS AND METHODS

Materials

Chemicals were purchased from Watanabe Chemical Industries, Nacalai Tesque, Kanto Chemical, Sigma-Aldrich Japan, or Wako Pure Chemical Industries unless noted otherwise and used without further purification. All oligonucleotides were purchased from Operon Biotechnologies. Flexizyme and tRNA molecules were synthesized using the same procedure as previously described (Murakami et al. 2006; Ohta et al. 2007; Goto et al. 2008).

General protocol of translation

Daa-tRNAfMet CAUU was prepared by the following procedure. We heated 40 μM tRNAfMet CAU in 0.2 M HEPES-K buffer (pH 7.5), 0.2 M KCl (7.5 μL) at 95°C for 3 min and cooled to 25°C for 5 min. MgCl2 (3 M, 3 μL) and flexizyme (dFx or eFx; see Murakami et al. 2006) (200 μM, 1.5 μL) were added and the mixture was incubated at 25°C for 5 min. The reaction was initiated by addition of 3 μL of 25 mM Daa substrate in DMSO and incubated on ice for the optimized times, generally 2–6 h (Murakami et al. 2006). After acylation, the reaction was stopped by addition of 45 μL of 0.6 M sodium acetate at pH 5, and the RNA was recovered by ethanol precipitation. The pellet was rinsed twice with 70% ethanol with 0.1 M sodium acetate at pH 5.0, and once with 70% ethanol. The Xaa-tRNAfMet CAU was dissolved in 0.5 μL of 1 mM sodium acetate just before adding to translation mixture.

The wPURE system containing all necessary components for translation except for all 20 standard amino acids was used in this study. Translation was carried out using wPURE system with 0.04 μM mDNA1 containing 200 μM each Thr, Tyr, Lys, 50 μM [14C]-Asp, and 120 μM of various Daa-tRNAfMet CAU molecules. The wild-type expression was carried out with wPURE system with 0.04 μM mDNA1 containing 200 μM each Met, Thr, Tyr, Lys and 50 μM [14C]-Asp. The translation mixture (2.5 μL) was incubated at 37°C for 1 h and analyzed by Tricine-SDS PAGE and autoradiography (FLA-5100, Fuji).

Analysis of peptides by MALDI-TOF

For MALDI-TOF analysis, the translation reaction (5 μL) was performed in the presence of Asp, instead of [14C]-Asp. The translation product from mDNA1 was immobilized with FLAG-M2 agarose (Sigma). After washing the resin with 30 μL of W buffer (50 mM Tris-HCl at pH 8.0, 150 mM NaCl), the immobilized peptides were eluted with 10 μL of 0.2% TFA. The purified peptide was desalted with ZipTipμ-C18 (Millipore), and eluted with 1 μL of a 50% MeCN, 0.1% TFA solution saturated with the matrix (R)-cyano-4-hydroxycinnamic acid. MALDI-TOF mass spectrometry was performed using an autoflex II TOF/TOF (BRUKER DALTONICS) under the linear/positive mode and externally calibrated with peptide calibration standard II (BRUKER DALTONICS).

Competition of Daa-tRNAfMet CAU and Laa-tRNAfMet CAU

Daa-tRNAfMet CAU and Laa-tRNAfMet CAU (aa = Tyr, Trp, and Leu) were prepared by the flexizyme system with the same procedure as described in the general protocol of translation. After completing the flexizyme reaction, the reaction mixture for Daa-tRNAfMet CAU and that for the corresponding Laa-tRNAfMet CAU were mixed with three different ratio (D:L = 99:1, 95:5, and 80:20). Then the reaction was stopped by addition of 45 μL of 0.6 M sodium acetateat pH 5, and the RNA mixture was recovered by ethanol precipitation. The pellet was rinsed twice with 70% ethanol with 0.1 M sodium acetate (pH 5.0), and once with 70% ethanol. The Xaa-tRNAfMet CAU was dissolved in 0.5 μL of 1 mM sodium acetate just before adding to translation mixture. Translation reaction was carried out with the same procedure as described above, except for adding the mixture of Daa-tRNAfMet CAU and Laa-tRNAfMet CAU to be 120 μM as total concentration of tRNAs in the translation mixture. MALDI-TOF analysis of the products was carried out in the same manner as above.

SUPPLEMENTAL DATA

Supplemental material can be found at http://www.rnajournal.org.

ACKNOWLEDGMENTS

We thank Dr. Patrick C. Reid for proofreading. This work was supported by grants from the Japan Society for the Promotion of Science Grants-in-Aid for Scientific Research (S) (16101007) to H.S., Grants-in-Aid for JSPS Fellows (18-10526) to Y.G., a research and development projects of the Industrial Science and Technology Program in the New Energy and Industrial Technology Development Organization (NEDO) to H.S., and the Industrial Technology Research Grant Program in NEDO (05A02513a) to H.M.

Footnotes

Article published online ahead of print. Article and publication date are at http://www.rnajournal.org/cgi/doi/10.1261/rna.1020708.

REFERENCES

  1. Bain J.D., Diala E.S., Glabe C.G., Wacker D.A., Lyttle M.H., Dix T.A., Chamberlin A.R. Site-specific incorporation of nonnatural residues during in vitro protein biosynthesis with semisynthetic aminoacyl-tRNAs. Biochemistry. 1991;30:5411–5421. doi: 10.1021/bi00236a013. [DOI] [PubMed] [Google Scholar]
  2. Boelens R., Gualerzi C.O. Structure and function of bacterial initiation factors. Curr. Protein Pept. Sci. 2002;3:107–119. doi: 10.2174/1389203023380765. [DOI] [PubMed] [Google Scholar]
  3. Brizzard B.L., Chubet R.G., Vizard D.L. Immunoaffinity purification of FLAG epitope-tagged bacterial alkaline phosphatase using a novel monoclonal antibody and peptide elution. Biotechniques. 1994;16:730–735. [PubMed] [Google Scholar]
  4. Calendar R., Berg P. The catalytic properties of tyrosyl ribonucleic acid synthetases from Escherichia coli and Bacillus subtilis . Biochemistry. 1966;5:1690–1695. doi: 10.1021/bi00869a034. [DOI] [PubMed] [Google Scholar]
  5. Calendar R., Berg P. D-Tyrosyl RNA: Formation, hydrolysis and utilization for protein synthesis. J. Mol. Biol. 1967;26:39–54. doi: 10.1016/0022-2836(67)90259-8. [DOI] [PubMed] [Google Scholar]
  6. Dedkova L.M., Fahmi N.E., Golovine S.Y., Hecht S.M. Enhanced D-amino acid incorporation into protein by modified ribosomes. J. Am. Chem. Soc. 2003;125:6616–6617. doi: 10.1021/ja035141q. [DOI] [PubMed] [Google Scholar]
  7. Dedkova L.M., Fahmi N.E., Golovine S.Y., Hecht S.M. Construction of modified ribosomes for incorporation of d-amino acids into proteins. Biochemistry. 2006;45:15541–15551. doi: 10.1021/bi060986a. [DOI] [PubMed] [Google Scholar]
  8. Ellman J.A., Mendel D., Schultz P.G. Site-specific incorporation of novel backbone structures into proteins. Science. 1992;255:197–200. doi: 10.1126/science.1553546. [DOI] [PubMed] [Google Scholar]
  9. Gold L. Posttranscriptional regulatory mechanisms in Escherichia coli . Annu. Rev. Biochem. 1988;57:199–233. doi: 10.1146/annurev.bi.57.070188.001215. [DOI] [PubMed] [Google Scholar]
  10. Goto Y., Ohta A., Sako Y., Yamagishi Y., Murakami H., Suga H. Reprogramming the translation initiation for the synthesis of physiologically stable cyclic peptides. ACS Chem. Biol. 2008;3:120–129. doi: 10.1021/cb700233t. [DOI] [PubMed] [Google Scholar]
  11. Gualerzi C.O., Pon C.L. Initiation of mRNA translation in prokaryotes. Biochemistry. 1990;29:5881–5889. doi: 10.1021/bi00477a001. [DOI] [PubMed] [Google Scholar]
  12. Hong S.Y., Oh J.E., Lee K.H. Effect of D-amino acid substitution on the stability, the secondary structure, and the activity of membrane-active peptide. Biochem. Pharmacol. 1999;58:1775–1780. doi: 10.1016/s0006-2952(99)00259-2. [DOI] [PubMed] [Google Scholar]
  13. Ibba M., Söll D. Quality control mechanisms during translation. Science. 1999;286:1893–1897. doi: 10.1126/science.286.5446.1893. [DOI] [PubMed] [Google Scholar]
  14. Kozak M. Comparison of initiation of protein synthesis in procaryotes, eucaryotes, and organelles. Microbiol. Rev. 1983;47:1–45. doi: 10.1128/mr.47.1.1-45.1983. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. LaRiviere F.J., Wolfson A.D., Uhlenbeck O.C. Uniform binding of aminoacyl-tRNAs to elongation factor Tu by thermodynamic compensation. Science. 2001;294:165–168. doi: 10.1126/science.1064242. [DOI] [PubMed] [Google Scholar]
  16. Mayer C., Stortchevoi A., Kohrer C., Varshney U., RajBhandary U.L. Initiator tRNA and its role in initiation of protein synthesis. Cold Spring Harb. Symp. Quant. Biol. 2001;66:195–206. doi: 10.1101/sqb.2001.66.195. [DOI] [PubMed] [Google Scholar]
  17. Mayer C., Kohrer C., Kenny E., Prusko C., RajBhandary U.L. Anticodon sequence mutants of Escherichia coli initiator tRNA: Effects of overproduction of aminoacyl-tRNA synthetases, methionyl-tRNA formyltransferase, and initiation factor 2 on activity in initiation. Biochemistry. 2003;42:4787–4799. doi: 10.1021/bi034011r. [DOI] [PubMed] [Google Scholar]
  18. Murakami H., Kourouklis D., Suga H. Using a solid-phase ribozyme aminoacylation system to reprogram the genetic code. Chem. Biol. 2003a;10:1077–1084. doi: 10.1016/j.chembiol.2003.10.010. [DOI] [PubMed] [Google Scholar]
  19. Murakami H., Saito H., Suga H. A versatile tRNA aminoacylation catalyst based on RNA. Chem. Biol. 2003b;10:655–662. doi: 10.1016/s1074-5521(03)00145-5. [DOI] [PubMed] [Google Scholar]
  20. Murakami H., Ohta A., Ashigai H., Suga H. A highly flexible tRNA acylation method for nonnatural polypeptide synthesis. Nat. Methods. 2006;3:357–359. doi: 10.1038/nmeth877. [DOI] [PubMed] [Google Scholar]
  21. Noller H.F., Hoang L., Fredrick K. The 30S ribosomal P site: A function of 16S rRNA. FEBS Lett. 2005;579:855–858. doi: 10.1016/j.febslet.2004.11.026. [DOI] [PubMed] [Google Scholar]
  22. Noren C.J., Anthony-Cahill S.J., Griffith M.C., Schultz P.G. A general method for site-specific incorporation of unnatural amino acids into proteins. Science. 1989;244:182–188. doi: 10.1126/science.2649980. [DOI] [PubMed] [Google Scholar]
  23. Ohta A., Murakami H., Higashimura E., Suga H. Synthesis of polyester by means of genetic code reprogramming. Chem. Biol. 2007;14:1315–1322. doi: 10.1016/j.chembiol.2007.10.015. [DOI] [PubMed] [Google Scholar]
  24. Roesser J.R., Xu C., Payne R.C., Surratt C.K., Hecht S.M. Preparation of misacylated aminoacyl-tRNAPhe’s useful as probes of the ribosomal acceptor site. Biochemistry. 1989;28:5185–5195. doi: 10.1021/bi00438a041. [DOI] [PubMed] [Google Scholar]
  25. Sankaranarayanan R., Moras D. The fidelity of the translation of the genetic code. Acta Biochim. Pol. 2001;48:323–335. [PubMed] [Google Scholar]
  26. Schagger H., von Jagow G. Tricine-sodium dodecyl sulfate-polyacrylamide gel electrophoresis for the separation of proteins in the range from 1 to 100 kDa. Anal. Biochem. 1987;166:368–379. doi: 10.1016/0003-2697(87)90587-2. [DOI] [PubMed] [Google Scholar]
  27. Schmitt E., Guillon J.M., Meinnel T., Mechulam Y., Dardel F., Blanquet S. Molecular recognition governing the initiation of translation in Escherichia coli. A review. Biochimie. 1996;78:543–554. doi: 10.1016/s0300-9084(96)80001-0. [DOI] [PubMed] [Google Scholar]
  28. Shimizu Y., Inoue A., Tomari Y., Suzuki T., Yokogawa T., Nishikawa K., Ueda T. Cell-free translation reconstituted with purified components. Nat. Biotechnol. 2001;19:751–755. doi: 10.1038/90802. [DOI] [PubMed] [Google Scholar]
  29. Söll D. The accuracy of aminoacylation—Ensuring the fidelity of the genetic code. Experientia. 1990;46:1089–1096. doi: 10.1007/BF01936918. [DOI] [PubMed] [Google Scholar]
  30. Soutourina J., Plateau P., Blanquet S. Metabolism of D-aminoacyl-tRNAs in Escherichia coli and Saccharomyces cerevisiae cells. J. Biol. Chem. 2000;275:32535–32542. doi: 10.1074/jbc.M005166200. [DOI] [PubMed] [Google Scholar]
  31. Soutourina O., Soutourina J., Blanquet S., Plateau P. Formation of D-tyrosyl-tRNATyr accounts for the toxicity of D-tyrosine toward Escherichia coli . J. Biol. Chem. 2004;279:42560–42565. doi: 10.1074/jbc.M402931200. [DOI] [PubMed] [Google Scholar]
  32. Stanzel M., Schon A., Sprinzl M. Discrimination against misacylated tRNA by chloroplast elongation factor Tu. Eur. J. Biochem. 1994;219:435–439. doi: 10.1111/j.1432-1033.1994.tb19956.x. [DOI] [PubMed] [Google Scholar]
  33. Starck S.R., Qi X., Olsen B.N., Roberts R.W. The puromycin route to assess stereo- and regiochemical constraints on peptide bond formation in eukaryotic ribosomes. J. Am. Chem. Soc. 2003;125:8090–8091. doi: 10.1021/ja034817e. [DOI] [PubMed] [Google Scholar]
  34. Tan Z., Forster A.C., Blacklow S.C., Cornish V.W. Amino acid backbone specificity of the Escherichia coli translation machinery. J. Am. Chem. Soc. 2004;126:12752–12753. doi: 10.1021/ja0472174. [DOI] [PubMed] [Google Scholar]
  35. Tugyi R., Uray K., Ivan D., Fellinger E., Perkins A., Hudecz F. Partial D-amino acid substitution: Improved enzymatic stability and preserved Ab recognition of a MUC2 epitope peptide. Proc. Natl. Acad. Sci. 2005;102:413–418. doi: 10.1073/pnas.0407677102. [DOI] [PMC free article] [PubMed] [Google Scholar]

Articles from RNA are provided here courtesy of The RNA Society

RESOURCES