Translational Incorporation of S-Nitrosohomocysteine into Protein*

The non-protein amino acid homocysteine (Hcy), owing to its structural similarity to the protein amino acids methionine, isoleucine, and leucine, enters first steps of protein synthesis and is activated by methionyl-, isoleucyl-, and leucyl-tRNA synthetases in vivo. However, translational incorporation of Hcy into protein is prevented by editing mechanisms of these synthetases, which convert misactivated Hcy into thiolactone. The lack of efficient interactions of the side chain of Hcy with the specificity subsite of the synthetic/editing active site is a prerequisite for editing of Hcy. Thus, if the side chain thiol of Hcy were reversibly modified with a small molecule that would enhance its binding to the specificity subsite and prevent editing, such modified Hcy is predicted to be transferred to tRNA and incorporated translationally into protein. Here I show thatS-nitroso-Hcy is in fact transferred to tRNA by methionyl-tRNA synthetase and incorporated into protein by the bacterium Escherichia coli. S-Nitroso-Hcy-tRNA also supports translation of mRNAs in a rabbit reticulocyte system. Removal of the nitroso group yields Hcy-tRNA and protein containing Hcy in peptide bonds. S-Nitrosylation-mediated translational incorporation of Hcy into protein may occur under natural conditions in cells and contribute to Hcy-induced pathogenesis in atherosclerosis.

MetRS directs methionine and Hcy into the synthetic and editing pathways, respectively, by employing the following three subsites of the synthetic/editing active site (3,13,14): (i) a subsite containing Asp-52 and Arg-233 that bind ␣-amino and -carboxyl groups, respectively, of the amino acid substrates; (ii) the specificity subsite, containing Trp-305, Phe-197, and Tyr-15, that preferentially binds the side chain of the cognate substrate methionine; (iii) the thiol-binding subsite that interacts with the side chain thiol of Hcy and facilitates thioester bond formation during editing; this site also accepts any thiol mimicking the side chain of Hcy. IleRS employs a similar thiol-binding subsite to facilitate editing of Hcy (15).
In the synthetic pathway, intermolecular reaction of the activated carboxyl group of methionine with the 2Ј-hydroxyl of the terminal adenosine of tRNA Met yields Met-tRNA Met . In the editing pathway, intramolecular reaction of the activated carboxyl group of Hcy with the sulfur of its side chain yields thiolactone. Whether an amino acid is transferred to tRNA or edited is determined by the competition for its activated carboxyl group between the side chain of the amino acid and the 3Ј-terminal adenosine of tRNA (13,14). MetRS transfers methionine to tRNA, because the side chain of methionine is firmly bound in the specificity subsite of the active site by hydrophobic and hydrogen bonding interactions with Trp-305, Phe-197, and Tyr-15, respectively. In contrast, the side chain of Hcy, missing the methyl group of methionine, cannot interact with the specificity subsite as strongly as the side chain of methionine does. This allows the side chain of Hcy to bind in the editing (or thiol-binding) subsite (3,14), which facilitates the intramolecular reaction between the activated carboxyl group and the side chain thiol of Hcy (Fig. 1). Thus, keeping the side chain of Hcy in the specificity subsite is predicted to prevent its editing and favor the transfer to tRNA. A possible way to achieve this is by utilizing S-nitrosothiol chemistry. Here I show that MetRS uses S-nitroso-Hcy as a substrate in the tRNA aminoacylation reaction, forming S-nitroso-Hcy-tRNA, and that S-nitroso-Hcy-tRNA transfers S-nitroso-Hcy into protein.

Plasmids and Strains-Plasmid-encoded gene for Escherichia coli
MetRS was overexpressed in E. coli strain JM101 as described before (16). E. coli strain CAG1849/pREP4/pQE15 (metE::Tn10/pREP4/pQE15) (17), which produces mouse DHFR protein upon induction with IPTG was kindly provided by K. Kiick and D. Tirrell (Division of Chemistry and Chemical Engineering, California Institute of Technology, Pasadena, CA). The strain requires methionine for growth because of the insertion of the transposon Tn10 into the metE gene, which is essential for the final step of methionine biosynthesis.
Preparation of L-[ 35 S]Hcy Thiolactone-The method of Baernstein (18), based on the digestion of methionine to Hcy thiolactone in the presence of hydriodic acid, was used with the following modifications. 250 nmol of L-Met was combined with 5 mCi of carrier-free L-[ 35 S]Met (0.3-ml aqueous solution containing 0.1% 2-mercaptoethanol, Amersham Pharmacia Biotech) in a 1-ml glass ampoule and lyophilized on a Speed-Vac. The residue was dissolved in 0.15 ml of 57% hydriodic acid containing 1% hypophosphorous acid (Aldrich). To serve as a refluxing tube, a 12-cm glass tube was attached, by flame-sealing, at the tip of the ampoule. The ampoule was then placed on a heating block, and the digestion was allowed to proceed for 4.5 h at 128°C. After digestion, the resulting crude preparation of L-[ 35 S]Hcy thiolactone was lyophilized, dissolved in 20 l of water, and purified by 2D TLC on cellulose plates (20 ϫ 10 cm) (Eastman Kodak Co.) using butanol/acetic acid/water (4:1:1, v/v) as the first-dimension solvent and 2-propanol/ethyl acetate/ water (5:5:1, v/v) as the second-dimension solvent (7,8). Thiolactone spot, localized under UV light and by autoradiography, was scraped off the plate and eluted with 2 mM HCl. The overall yield of the procedure was 65%. The preparation of L-[ 35 S]Hcy thiolactone was at least 96% pure on analytical 2D TLC. Maximum levels of contamination with Met, Hcy, and homocysteine were Ͻ1, Ͻ0.8, and Ͻ1%, respectively.
Preparation of L-[ 35  Preparation of S-Nitroso-Hcy-S-Nitrosylation of 1 mM L-[ 35 S]Hcy (20,000 Ci/mol) or 50 mM L-Hcy was carried out using equimolar amounts of NaNO 2 in 0.1 M HCl (19). The solution of L-[S-nitroso- 35 S]Hcy, neutralized with Na 2 HPO 4 , can be stored at Ϫ70°C for at least two weeks.
ATP-PP i Exchange and tRNA Aminoacylation Assays-All assays were performed at 37°C in a standard buffer containing 0.1 M potassium-HEPES, pH 7.4, 10 mM MgCl 2 , and 0.1 mM EDTA. Reactions were initiated by the addition of MetRS, and four time points were taken at time intervals from 1-8 min. ATP-PP i exchange assays (6,13) were performed in the presence of 1 mM [␥-32 P]ATP (100 Ci/mol, Amersham Pharmacia Biotech), 1 mM PP i , 0.03-1 mM S-nitroso-Hcy, and 2 M MetRS. Radiolabeled ATP and PPi, separated by TLC on polyethyleneimine-cellulose plates (Merck) using 3 M ammonium sulfate as a solvent, were quantitated by scintillation counting. Unless otherwise stated, tRNA aminoacylation assays (6,13)  Editing Activity Assay-Unless stated otherwise, editing activity was measured as amino acid-dependent AMP formation from 1 mM [␣-32 P]ATP (50 Ci/mol) in the presence of 1 mM amino acid and 0.2 M MetRS in the standard buffer (16).
Preparation of L-[S-Nitroso- 35  MetRS in the standard buffer, was monitored by trichloroacetic acid precipitation (14). First order rate constants, k, were calculated from reaction half-lives, t 0.5 , according to kϭln2/t 0.5 . Aliquots of reaction mixtures were also analyzed by TLC (14).
In Vitro Translation-Translation of luciferase and globin mRNAs (Promega) was carried out in methionine-free rabbit reticulocyte system according to the manufacturer's procedure (20)  SDS-PAGE Analyses-Products of in vivo and in vitro translation were analyzed on 13% denaturing polyacrylamide gels (21).

S-Nitroso-Hcy Is a Substrate for MetRS in the Adenylate Formation and tRNA Aminoacylation Reactions-S-Nitroso-
Hcy was a substrate for MetRS in the aminoacyl adenylate formation reaction (Table I). Catalytic efficiency for S-nitroso-Hcy was intermediate between those for methionine and Hcy. The k cat value for S-nitroso-Hcy (11.2 s Ϫ1 ) was 7-and 2-fold lower than the k cat values for methionine and Hcy, respectively. The K m value for S-nitroso-Hcy (0.37 mM) was 18.5-fold higher and 14.3-fold lower than the K m values for methionine and Hcy, respectively. The enzyme-bound S-nitroso-HcyϳAMP intermediate was relatively stable and, in contrast to the HcyϳAMP intermediate (6), was not edited, as determined by the lack of S-nitroso-Hcy-dependent ATP hydrolysis (Fig. 2B).
S-Nitroso-Hcy was also a substrate in the tRNA aminoacylation reaction catalyzed by MetRS ( Fig. 2A). Non-saturating kinetics were observed with 2-160 M S-nitroso-Hcy (not shown). Catalytic efficiencies (k cat /K m ) indicate that tRNA aminoacylation with S-nitroso-Hcy is about 80-fold less efficient than the tRNA aminoacylation with methionine (Table II). More than 50% of tRNA Met or tRNA fMet could be charged with S-nitroso-Hcy ( Fig. 2A). As expected (4, 5), Hcy was not incorporated into tRNA ( Fig. 2A). Active site mutations in MetRS, known to have detrimental effects on tRNA aminoacylation with methionine (13,14), had similar affects on tRNA aminoacylation with S-nitroso-Hcy (Table II), suggesting that S-nitroso-Hcy and methionine bind to the same active site (Fig. 3).
Properties of S-Nitroso-Hcy-tRNA-S-Nitroso-Hcy-tRNA, like Met-tRNA, can be isolated from reaction mixtures by phenol extraction. Like Met-tRNA, purified S-nitroso-Hcy-tRNA underwent relatively slow deacylation, both in the absence and presence of MetRS (Table III). In the presence of free thiols, S-nitrosothiols undergo fast transnitrosylation reactions (22). Similarly, S-nitroso-Hcy-tRNA was converted to Hcy-tRNA by treatment with an excess thiol. Hcy-tRNA was about 100 times less stable than S-nitroso-Hcy-tRNA (Table III). Deacylations of Hcy-tRNA and S-nitroso-Hcy-tRNA yielded Hcy thiolactone and S-nitroso-Hcy, respectively, as products (not shown).  Translational Incorporation of S-Nitroso-Hcy into Protein-Once an amino acid is attached to tRNA, it is destined to become incorporated translationally into protein (23). Thus, the facile aminoacylation of tRNA with S-nitroso-Hcy catalyzed by MetRS suggested that S-nitroso-Hcy could be incorporated into protein. To test this, E. coli strain CAG1849/pREP4/ pQE15 (17), which produces mouse DHFR protein upon induction with IPTG was employed. The strain is a methionine auxotroph, unable to metabolize Hcy to methionine because of the insertion of the transposon Tn10 into the metE gene. As shown in Fig. 4A, E. coli utilized S-nitroso-Hcy for incorporation into proteins and, after induction with IPTG, into recombinant mouse DHFR. S-Nitroso-Hcy-DHFR co-migrated with the Met-DHFR on denaturing polyacrylamide gels (Fig. 4B). When [S-nitroso- 35 (24,25). After reduction to Hcy, homocystine is subsequently converted to thiolactone by MetRS, IleRS, and LeuRS in E. coli (7,10). Because post-translational reaction of thiolactone with protein lysine residues may also lead to incorporation of Hcy into protein (26), control experiments in which E. coli cultures were incubated with Hcy or thiolactone were carried out. There was no incorporation of Hcy (Fig. 4B) or thiolactone (not shown) into bacterial proteins under these conditions. Taken together, these data indicate that S-nitroso-Hcy is incorporated translationally 2 H. Jakubowski, unpublished data.   (5,14) showed that Hcy thiolactone and S-nitroso-Hcy were products of deacylations of Hcy-tRNA and S-nitroso-tRNA, respectively. AA into proteins synthesized by E. coli.
As shown in Fig. 5, S-nitroso-Hcy-tRNA supported translation of globin mRNA and luciferase mRNA in a methionine-free rabbit reticulocyte translation system. Globin and luciferase proteins labeled with [S-nitroso- 35 S]Hcy were indistinguishable from the corresponding [ 35 S]Met-labeled proteins on polyacrylamide gels. The data thus indicate that S-nitrosylation provides a mechanism for the translational incorporation of Hcy into protein.
Conclusions-In cultured human cells, Hcy can be incorporated into proteins post-translationally in a two-step mechanism (11, 26 -28). In the first step, Hcy is converted into thiolactone as a result of an error-editing reaction of MetRS. In the second step, thiolactone acylates side chain amino groups of protein lysine residues. Protein damage caused by N-homocysteinylation may underlie involvement of Hcy in vascular diseases (11, 26 -28). Because translational incorporation of Snitroso-Hcy into protein is also likely to lead to protein damage, the present data illustrate an additional mechanism that may explain Hcy-induced pathology associated with atherosclerosis. Specifically, translational incorporation of S-nitroso-Hcy into protein would explain why atherosclerosis is originating mostly at branch points in arteries (29) that are subject to mechanical stress leading to increased production of nitric oxide (19). S-Nitroso-thiols are present in human blood (30), and S-nitroso-Hcy has been detected in cultures of human endothelial cells (19). It is likely that, in subjects with elevated serum Hcy levels, local concentrations of nitric oxide and S-nitroso-Hcy are higher at arterial branch points than elsewhere. Therefore, at branch points, the incorporation of S-nitroso-Hcy into endothelial cell proteins, and resulting protein damage, would be greater than at other points in arteries. Preliminary data sug-gest that Hcy is incorporated both translationally and posttranslationally into proteins from endothelial cell cultures maintained on Hcy in folate-limited media. One cycle of Edman degradation of proteins modified with Hcy thiolactone releases all incorporated Hcy (26). However, similar Edman degradation of Hcy-labeled proteins from endothelial cell cultures releases only 37% of total incorporated Hcy. 3 Thus, both translational (via S-nitroso-Hcy) and post-translational (via Hcy thiolactone) incorporation of Hcy into protein provide plausible mechanisms whereby Hcy can affect physiological function.