Lysine 207 as the site of cross-linking between the 3'-end of Escherichia coli initiator tRNA and methionyl-tRNA formyltransferase.

The specific formylation of initiator methionyl-tRNA by methionyl-tRNA formyltransferase (MTF) is important for initiation of protein synthesis in Escherichia coli. In attempts to identify regions of MTF that come close to the 3′-end of the tRNA, we oxidized 32P-3′-end-labeled E. coli initiator methionine tRNA with sodium metaperiodate and cross-linked it to MTF. The cross-linked MTF was separated from uncross-linked MTF by DEAE-cellulose chromatography, and the tRNA in the cross-linked MTF was hydrolyzed with nuclease P1 and RNase T1, leaving behind an oxidized fragment of [32P]AMP attached to MTF. Trypsin digestion of the cross-linked MTF followed by high pressure liquid chromatography of the digest yielded two peaks of radioactive peptides, I* and II*. These peptides were characterized by N- and/or C-terminal sequencing and by matrix-assisted laser desorption ionization mass spectroscopy. Peptide I* contained amino acids Gln186-Lys210 with Lys207 as the site of the cross-link. Peptide II*, a partial digestion product, contained amino acids Gln186-Arg214 also with Lys207 as the site of the cross-link. The molecular masses of peptides I* and II* indicate that the final product of the cross-linking reaction between the periodate-oxidized AMP moiety of the tRNA and Lys207 is most likely a morpholino derivative rather than a reduced Schiff's base.

From assembly and packaging of RNA viruses (1) to mRNA localization during development (2), the specific recognition of RNAs by proteins plays an important role in many biological processes. Examples of these biological processes include RNA processing, RNA splicing, RNA transport, ribosome assembly, translation, and translational regulation (3). As molecules that interact with a variety of different proteins, tRNAs provide an excellent system for studying the molecular basis of specificity in recognition of RNAs by different proteins (4).
We are studying the specific recognition of Escherichia coli initiator methionyl-tRNA (Met-tRNA), 1 during its formylation to formylmethionyl-tRNA by the enzyme methionyl-tRNA formyltransferase (MTF, EC 2.1.2.9; 10-formyltetrahydrofolate:L-methionyl-tRNA N-formyltransferase). Formylation of initiator Met-tRNA is important for initiation of protein synthesis in eubacteria and in eukaryotic organelles such as mitochondria and chloroplasts (5)(6)(7)(8)(9). In E. coli, formylation provides a positive determinant for allowing the initiation factor IF2 to select the initiator tRNA from other tRNAs (10,11) and a second negative determinant for blocking the binding of elongation factor EF-Tu to the initiator tRNA (12)(13)(14). The formylation reaction is highly specific; the enzyme formylates the initiator Met-tRNA but not the elongator Met-tRNA or any other aminoacyl-tRNA (15). Previous studies have shown that most of the determinants on the initiator tRNA important for its formylation by MTF are clustered in the acceptor stem (16 -19) (Fig. 1). However, although the protein sequence of MTF is known (20), little is known about the amino acid residues in MTF that are important for this recognition and the molecular basis of the specificity in recognition.
As a first step in identifying regions of MTF that come close to the acceptor stem of the tRNA, we have cross-linked periodate-oxidized tRNA to MTF and have analyzed the site(s) of cross-linking. We show that Lys 207 in the sequence KLSKE (207-211) of MTF is the site of cross-link to the 3Ј terminus of the tRNA. 2 The cross-linking of periodate-oxidized E. coli tRNA fMet to MTF has been studied before by Blanquet and co-workers (21). However, the sites of cross-linking were not analyzed in the previous work.

MATERIALS AND METHODS
Chemicals, Enzymes, and Radioisotopes-Sodium metaperiodate, folinic acid, carboxypeptidase Y (sequencing grade), and methionine were obtained from Sigma. Sodium borohydride, sodium cyanoborohydride, and ␣-cyano-4-hydroxycinnamic acid were purchased from Aldrich. E. coli tRNA nucleotidyltransferase and E. coli methionyl-tRNA synthetase were purified in our laboratory by Mike Dyson. Nuclease P1, modified trypsin (sequencing grade), and chymotrypsin (sequencing grade) were obtained from Boehringer Mannheim. RNase T1 was from Sankyo Chemical Company Ltd. (Tokyo, Japan). [ 35 S]Methionine (specific activity ϭ 1175 Ci/mmol) and [␣-32 P]ATP (specific activity ϭ 3000 Ci/mmol) were purchased from DuPont NEN. All of the solvents used for HPLC were HPLC grade and procured from EM Science. All other routinely used chemicals were of the highest purity grade available.
Assay for Aminoacylation of tRNA-The reaction was carried out at 37°C. The incubation mixture (20 l) contained 20 mM imidazole-HCl buffer (pH 7.5), 0.1 mM EDTA, 2 mM ATP, 150 mM NH 4 Cl, 10 g/ml bovine serum albumin, 4 mM MgCl 2 , 25 M [ 35 S]methionine (specific activity ϭ 5,000 -10,000 cpm/pmol), tRNA (approximately 0.1 A 260 of total tRNA or 0.01 A 260 of pure tRNA fMet ) and saturating amounts of purified methionyl-tRNA synthetase. Aliquots (5 l) were withdrawn at 5-min intervals and spotted onto 3 MM Whatman paper discs (pre-* This work was supported by National Institutes of Health Grant GM17151. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
We dedicate this paper to Professor Nelson J. Leonard on the occasion of his eightieth birthday.
Assay for Formylation of Met-tRNA fMet -The incubation (20 l), carried out at 37°C, contained 10 l of the above aminoacylation reaction mixture (which had been preincubated for 30 min), 0.3 mM N 10formyltetrahydrofolate, and appropriate amounts of MTF (depending on the purity and specific activity of preparation). The reaction was allowed to proceed for 15 min and was terminated by the addition of 20 l of 0.36 M CuSO 4 in 1.1 M Tris-HCl (pH 7.3) and incubated further for 3 min at room temperature (20). Acid-precipitable radioactivity was measured as described for the aminoacylation assay.
Purification of MTF-Purification of MTF was carried out using a procedure developed by Dr. D. Mangroo. 3 The specific activity of the purified enzyme was 2.69 ϫ 10 5 pmol of formyl group incorporated into Met-tRNA fMet /min/g of protein. The purified enzyme migrated as a single band on SDS-polyacrylamide gels.
Analytical Methods-Protein concentrations were estimated by the modified Lowry DC method as described by the supplier (Bio-Rad) using IgG as a standard. The concentration of purified MTF was also determined using an absorbance of 1.39 at 280 nm for a solution containing 1 mg/ml of MTF and with a light path of 1 cm (23). This value agreed well with the value obtained from the Bio-Rad protein assay. SDSpolyacrylamide gel electrophoresis (SDS-PAGE) was carried out according to Laemmli (24) after boiling the protein samples for 5 min in 62.5 mM Tris-Cl, pH 6.8, containing 4% (w/v) SDS, 10% glycerol, 10% ␤-mercaptoethanol, and 0.02% bromphenol blue.
Purification and 3Ј-End Labeling of tRNA fMet -Total tRNA was isolated by phenol extraction from E. coli cells (17 g, wet weight) overproducing tRNA 2 fMet from the plasmid pUC13 trnfM (which carries the tRNA 2 fMet gene). The crude tRNA (100 A 260 units) was fractionated on a 10% native polyacrylamide gel, and tRNAs were visualized by UV shadowing (25). The gel slice containing tRNA fMet was cut, and tRNA fMet was eluted from the gel in 20 mM Tris-HCl buffer, pH 8.0, containing 1 mM EDTA (TE buffer). The purity of gel-eluted tRNA fMet , estimated by polyacrylamide gel electrophoresis and by aminoacylation assay, was Ͼ95% (22). 32 P-3Ј-End labeling of tRNA fMet was carried out using tRNA nucleotidyltransferase. First, the 3Ј-terminal A76 was removed by Whitfeld degradation by treatment of the tRNA with sodium periodate and subsequently with aniline (26). The 3Ј-terminal phosphate thus generated was removed by treatment with alkaline phosphatase. The reaction mixture for 3Ј labeling contained 50 mM glycine-NaOH (pH 9.2), 10 mM Mg(OAc) 2 , 10 mM glutathione-SH, 10 g/ml bovine serum albumin, 50 M [␣-32 P]ATP, 6 M tRNA fMet (ϪA76) and 2 units of E. coli tRNA nucleotidyltransferase. Incubation was at 37°C for 20 min. One unit of tRNA nucleotidyltransferase is the amount of enzyme that catalyzes the incorporation of 1 mol of AMP/h (27). Unreacted ATP was removed by spin column chromatography through a 1-ml Sephadex G-25 column. The 32 P-3Ј-end-labeled tRNA fMet was recovered by ethanol precipitation and dissolved in 20 l of TE buffer (pH 8.0). The labeled tRNA fMet was subsequently purified by denaturing polyacrylamide gel electrophoresis (8 M urea-PAGE), eluted from gel, and precipitated with ethanol. The pellet was dried and dissolved in distilled water and stored at Ϫ20°C.
Preparation of Periodate-oxidized tRNA fMet -The reaction mixture (40 l) contained 1.0 A 260 unit of 3Ј-end labeled [ 32 P]tRNA fMet (30,000 cpm), 2.5 mM sodium periodate, and 100 mM sodium acetate, pH 5.2. The reaction was carried out for 30 min at 0°C in the dark. Excess periodate was destroyed by the addition of ethylene glycol to 5 mM (final concentration). The oxidized tRNA fMet was ethanol-precipitated, the pellet was dissolved in water, 0.1 volume of 3 M sodium acetate buffer, pH 5.0, was added to it, and the oxidized tRNA fMet was reprecipitated with ethanol. The pellet was collected by centrifugation, excess ethanol was removed, and the pellet was dried and dissolved in 50 l of distilled water and used immediately for the cross-linking reaction. The extent of oxidation of tRNA fMet was checked by determining the residual methionine acceptance activity (22) of oxidized tRNA and was found to be essentially complete (Ͼ96%).
Cross-linking of Periodate-oxidized tRNA fMet to MTF-For most experiments, the periodate-oxidized tRNA fMet (10 pmol) was incubated in a total volume of 10 l with MTF (0.066 g, 2 pmol) in 20 mM imidazole-HCl, pH 7.5, 10 mM MgCl 2 , and 0.1 mM EDTA at 37°C for 1 h in the presence of 1 mM sodium cyanoborohydride (21). The reaction was quenched by the addition of sodium borohydride (1 mM final concentration), and the extent of cross-linking was determined by SDS-PAGE followed by autoradiography and quantitation by PhosphorImager analysis. The effect of the addition of either excess substrate (tRNA fMet ) or cofactor (N 10 -formyltetrahydrofolate) to the reaction mixture on the cross-linking efficiency was determined by adding appropriate amounts of these ligands to the reaction mixture followed by incubation for 10 min at 37°C prior to the addition of periodate-oxidized tRNA fMet . The effect of cross-linking of periodate-oxidized tRNA fMet to MTF on the enzymatic activity of MTF was evaluated by withdrawing (5 l) aliquots of a large scale reaction mixture (60 pmol of MTF and 300 pmol of either 32 P-labeled unoxidized initiator tRNA or oxidized initiator tRNA) at various times (every 5 min for 30 min) and estimating the residual enzyme activity after diluting the reaction mixture 500-fold with 20 mM imidazole buffer, pH 7.5, containing 10 mM MgCl 2 and 0.1 mg/ml bovine serum albumin. For isolation and analysis of cross-linked MTF, a large scale reaction was carried out using 20 g of MTF (600 pmol) and 1800 pmol of periodate-oxidized tRNA fMet (3-fold molar excess), and the reaction time was increased to 2 h.
Rate of Cross-linking of Various tRNAs to MTF-MTF (90 pmol) in 40 l of 20 mM imidazole-HCl, pH 7.5, 10 mM MgCl 2 , and 0.1 mM EDTA was incubated with 450 pmol of oxidized tRNAs in the presence of 1 mM sodium cyanoborohydride at 37°C. Aliquots (4 l) were removed at the indicated time intervals, and the reaction was quenched by the addition of 1 l of sodium borohydride (final concentration of 1 mM). The extent of cross-linking in each case was determined by SDS-PAGE followed by autoradiography and by PhosphorImager analysis.
Separation of the Cross-linked MTF from MTF and Digestion of the tRNA Moiety in the Cross-linked MTF-MTF cross-linked to tRNA was separated from uncross-linked MTF by DEAE-cellulose chromatography. Following the cross-linking reaction, the reaction mixture was diluted to 1 ml with 0.1 M Tris-HCl, pH 7.5, containing 0.2 M LiCl and loaded onto a DEAE-cellulose column (1 ml) preequilibrated with the same buffer. After extensive washing of the column to remove MTF, the cross-linked MTF was eluted with 0.1 M Tris-HCl, pH 7.5, containing 1 M LiCl. Fractions (0.3 ml) were collected. The fractions showing high radioactivity were pooled and subjected to Centricon-10 ultrafiltration for desalting. After centrifugation, excess tRNA and tRNA cross-linked to MTF was hydrolyzed to nucleotides (28) by incubating the crosslinked MTF with 0.3 unit of P1 nuclease and 5 units of RNase T1 at 45°C for 4 h. The extent of hydrolysis of tRNA from cross-linked MTF was analyzed by SDS-PAGE and autoradiography.
To remove free nucleotides from cross-linked MTF, the reaction mixture with nuclease P1 and RNase T1 was subjected to gel filtration on a Sephadex G-50 column (0.8 ϫ 14 cm). Fractions containing MTF (void volume) were concentrated and used for proteolytic digestion. Alternatively, centrifugation through the Centricon-10 membrane was used to remove the released nucleotides from cross-linked MTF.
Trypsin Digestion of Cross-linked MTF and Separation of Peptides by HPLC-The cross-linked MTF (20 g and ϳ10,000 cpm of Cerenkov radiation) in 100 l of 100 mM Tris-HCl, pH 7.5, was treated with 2 g 3 D. Mangroo and U. L. RajBhandary, unpublished observations. of sequencing grade modified trypsin (10%, w/w) at 37°C for 12 h. The tryptic peptides were separated by HPLC on a Hewlett-Packard HP-1090 HPLC system equipped with a reverse phase Vydac C 18 column (0.46 ϫ 25 cm). The trypsin digest was diluted with 0.1% trifluoroacetic acid in water to 250 l, centrifuged at 14,000 rpm, and injected into the column preequilibrated with 90% solvent A (0.1% trifluoroacetic acid in water) and 10% solvent B (0.1% trifluoroacetic acid in 100% acetonitrile). The elution rate was 0.5 ml/min in 90% solvent A and 10% solvent B from 0 to 10 min followed by a linear gradient from 10% solvent B to 20% solvent B between 10 and 60 min (0.2% per min) and then 20 -27% B between 60 and 160 min (0.07% per min) and then 27-100% solvent B in an additional 40 min (1.82% per min). The effluent was continuously monitored at 210 nm, and fractions of 0.5 ml were collected. The radioactivity in each fraction was determined by Cerenkov counting for 32 P in an LKB-1217 RACKBETA liquid scintillation counter.
Chymotrypsin Digestion of Peak II* and HPLC-Fractions containing peak II* (obtained from the HPLC of trypsin digestion of crosslinked MTF above) were lyophilized and dissolved in 100 l of 100 mM Tris-HCl, pH 7.5. This mixture was then treated with 2 g of chymotrypsin for 5 h at 37°C and subjected to reverse phase HPLC on a C 18 column. The column was preequilibrated with 100% solvent A. The elution rate was 0.5 ml/min in 100% solvent A from 0 to 10 min followed by a linear gradient from 0 to 25% Solvent B between 10 and 135 min (0.2% per min) and then 25-100% Solvent B for an additional 35 min. The effluent was continuously monitored at 210 nm, and fractions of 0.5 ml were collected. The radioactivity in each fraction was determined by Cerenkov counting for 32 P as above.
Mass Spectroscopic Analysis and Amino Acid Sequencing of Radioactive Peptides-The mass of the cross-linked peptides was determined on a MALDI mass spectrophotometer (PerSeptive Biosystems Voyager mass spectrophotometer) using delayed extraction technology. For the mass analysis of the peptides, approximately 1 pmol of peptide (based on radioactivity) was mixed with 1 l of ␣-cyano-4-hydroxycinnamic acid (matrix). For calibration of the mass spectrophotometer, oxidized insulin chain B (Sigma) was used as a standard. C-terminal amino acid sequencing was also carried out using limited digestion of peptide with carboxypeptidase Y followed by mass spectroscopic analysis. The Nterminal amino acid sequence of isolated peptides was determined using an automated gas phase protein/peptide sequence analyzer from Applied Biosystems (model 470A) equipped with an on-line phenylthiohydantoin analyzer (model 120) and computer (model 900A).

Cross-linking of MTF to Periodate-oxidized tRNA 2
fMet -Periodate-oxidized nucleotides and tRNAs have often been used as affinity reagents for cross-linking to proteins (29). Oxidation of tRNA with sodium periodate converts the terminal ribose moiety in the A76 of the tRNA to a 2Ј,3Ј-dialdehyde. The dialdehyde then reacts primarily with the ⑀-amino group of lysine residues in a protein that comes close to the 3Ј-end of tRNA.
For following the cross-linking of MTF to the tRNA, 32 P-3Јend-labeled tRNA 2 fMet was incubated with MTF in the presence of sodium cyanoborohydride, which reduces the intermediate Schiff's base and prevents reversal of the reaction (21). The reaction was quenched after 1 h by the addition of sodium borohydride, and the products were analyzed by SDS-polyacrylamide gel electrophoresis followed by autoradiography of the gel. Fig. 2, lane 2, shows that in an incubation mixture containing periodate-oxidized tRNA, most of the 32 P radioactivity migrated more slowly than tRNA at a position expected for that of a tRNA cross-linked to a protein.
Formation of the tRNA-protein cross-link requires oxidation of the tRNA with periodate. When 32 P-3Ј-end-labeled tRNA 2 fMet without periodate oxidation was incubated with MTF under identical conditions, there was no cross-link between the tRNA and the protein (Fig. 2, lane 1). The presence of sodium cyanoborohydride during the incubation was found to be important for maximal cross-linking (data not shown). This result indicates that the cross-linking of periodate-oxidized tRNA 2 fMet to MTF is saturable and that it is quite specific. Further evidence for the specificity of cross-linking was derived from an experiment in which increasing amounts of unoxidized tRNA 2 fMet was added to the cross-linking reaction containing a fixed amount of periodate-oxidized tRNA. It was found that there was a gradual decrease in the extent of cross-linking dependent upon the concentration of unoxidized tRNA 2 fMet in the reaction (data not shown). This result suggests that unoxidized tRNA and oxidized tRNA are competing for the same site in MTF. In contrast, the addition of N 10 -formyltetrahydrofolate, another substrate of MTF, had no significant effect on the extent of cross-linking (data not shown).
The cross-linking of MTF to periodate-oxidized tRNA 2 fMet leads to loss of enzymatic activity of MTF. Incubation of MTF with periodate-oxidized tRNA 2 fMet leads to a time-dependent inactivation of MTF (data not shown). This inactivation re- . After 1 h, an aliquot (5 l) was removed, the reaction was quenched by the addition of NaBH 4 , and the cross-linking mixture was subjected to SDS-PAGE. The gel was dried, and radioactivity was detected by autoradiography.
FIG. 3. Effect of initiator tRNA concentration on the extent of cross-linking to MTF. MTF (2 pmol) was incubated with increasing amounts of 32 P-3Ј-end-labeled periodate-oxidized initiator tRNA (0 -90 pmol). The reaction was carried out under standard cross-linking conditions as described under "Materials and Methods." Following incubation, the extent of cross-linking was determined by subjecting the crosslinking mixture to SDS-PAGE. The quantitation of radioactivity in the band corresponding to cross-linked MTF was carried out using Phos-phorImager analysis.
quires the presence of periodate-oxidized tRNA 2 fMet . The presence of a 10-fold excess of unoxidized tRNA 2 fMet over the oxidized tRNA, which competes against cross-linking of MTF to the oxidized tRNA, also protects MTF from inactivation (data not shown). A similar effect of cross-linking of oxidized tRNA fMet on enzymatic activity of MTF was described before by Hountondji et al. (21).
Identification of Amino Acid(s) in MTF Cross-linked to tRNA 2 fMet -The following strategy was used to isolate and analyze the peptide(s) attached to the tRNA. Following the crosslinking reaction, cross-linked MTF was separated from uncross-linked MTF by DEAE-cellulose chromatography (see "Materials and Methods"). The excess of 32 P-3Ј-end-labeled tRNA 2 fMet and most of the tRNA moiety in the cross-linked MTF were hydrolyzed to nucleotides using a mixture of nuclease P1 and RNase T1. A mixture of the two nucleases was used to ensure complete hydrolysis of the tRNA (28). This treatment should yield MTF carrying only a 32 P-labeled fragment of AMP attached to it. The cross-linked MTF carrying the 32 P label was separated from the mononucleotides and digested with trypsin, and the tryptic peptides were separated by high pressure liquid chromatography (HPLC). A similar tryptic digest of MTF that had not been cross-linked to tRNA was also subjected to HPLC. Fig. 4 shows the tryptic peptide and 32 P radioactivity profiles of MTF and MTF cross-linked to [ 32 P]AMP moiety of the tRNA. A peptide peak (designated peptide I) eluting at 66 min in digests of MTF (Fig. 4A), is absent in digests of cross-linked MTF (Fig. 4B) and is replaced in the latter by a peak (designated peptide I*) eluting at 104 min. This peptide I*, which is absent in digests of uncross-linked MTF, also contains 32 P radioactivity (Fig. 4C), suggesting that it is derived from the peptide eluting at 66 min in digests of uncross-linked MTF.
Another peak designated peptide II* also contains 32 P radioactivity. This peak contains a mixture of two peptides, one of which is radioactive. We show below that the radioactive peptide II* is a partial digestion product and is the same as peptide I* but with an extension of four amino acids at the C terminus.
Characterization of Peptides I and I*-A combination of MALDI mass spectroscopy (30), N-terminal sequencing using Edman degradation (31), and C-terminal sequencing using partial digestion with carboxypeptidase Y (32) followed by MALDI mass spectroscopic analysis was used to establish the sequence of these peptides. The following lines of evidence summarized in Table I show that peptide I* has the sequence 186 QLADG-TAKPEVQDETLVTYAEKLSK 210 , in which Lys 207 is linked to the [ 32 P]AMP moiety of the tRNA.
(i) MALDI mass spectroscopy of peptide I yielded two peaks with molecular masses of 2406.95 and 2390.59 Da. The only tryptic peptide of MTF that fits this is Gln 186 -Lys 207 . The difference of 16.36 Da in the two molecular masses is most probably due to the well known conversion of glutamine at the N terminus of peptides or proteins to pyroglutamic acid (33). The complete absence of a peak corresponding to peptide I in digests of cross-linked MTF (Fig. 4B) suggests that one of the two lysine residues in the peptide is linked to the AMP moiety of the tRNA. The most likely possibility is that Lys 207 is crosslinked to the AMP moiety, thereby making the peptide bond involving this lysine residue resistant to cleavage by trypsin in the cross-linked MTF.
(ii) N-terminal sequencing of peptide I* yielded the sequence QLADGTAKPE, which is the same as that of peptide I. This result confirms the suggestion above that peptide I* is derived from peptide I. Furthermore, the clear identification of lysine as amino acid number 8 of peptide I* shows that lysine 193 is not cross-linked in MTF and supports the conclusion above that Lys 207 is the one most probably cross-linked. While the Nterminal sequence data were unambiguous, the yields of phenylthiohydantoin-derivatives in the first and successive cycles were much lower than expected (ϳ35%), based on the 32 P radioactivity present in peptide I*. This is due to the fact that most of the glutamine at the N terminus is converted to pyroglutamic acid, which is inert to the reagents used for N-terminal sequencing.
(iii) Partial digestion with carboxypeptidase Y of peptide I* in situ on the sample plate used for MALDI mass spectroscopy, followed by mass spectroscopic analysis of the partial digestion products, indicated the loss successively of the amino acids lysine, serine, and leucine from the C terminus. Thus, the sequence at the C terminus of peptide I* is -LSK. These are the amino acids that immediately follow Lys 207 in the sequence of MTF. This result supports the conclusion above that peptide I* is derived from peptide I and shows that it has an extension of LSK at the C terminus beyond Lys 207 .
(iv) Final evidence for the sequence of peptide I* was derived by MALDI mass spectroscopic analysis. Mass spectroscopy yielded four peaks with molecular masses in decreasing order of 3052.23, 3035.31, 2917.37, and 2901.09 daltons (Fig. 5). The molecular mass of 3052.23 Da is very close to that expected for the cross-linked peptide Gln 186 -Lys 210 , in which Lys 207 is cross-linked to the AMP moiety of the tRNA (Table I) The average molecular mass of 3050.7 Da thus obtained for peptide I* differs by only about 2.2 Da from the molecular mass expected for a cross-linked product, which contains a morpholino type of linkage (Fig. 6, III) between the peptide and the ribose moiety of AMP rather than a reduced Schiff's base (Fig. 6, I).
Characterization of Peptide II*-MALDI mass spectroscopic analysis of peptide II* showed that it contained a mixture of peptides with molecular masses of 2641.47 and 3536.71 Da. The peptide with a molecular mass of 2641.47 Da was shown to have the sequence 19 HLDALLSSGHNVVGVFTQPDRPAGR 43 by N-and C-terminal sequence analysis. This peptide is also present in digests of uncross-linked MTF (Fig. 4A). The radioactive peptide II* with a molecular mass of 3536.71 Da was established by the following three lines of evidence to have the same sequence as peptide I* except that it has an extension at the C terminus of EEAR.
(i) Peptide II* was treated with chymotrypsin, and the digest was subjected to HPLC chromatography (Fig. 7). A predominant radioactive peak that coeluted with a peptide peak was obtained. This peptide was then used for sequence analysis using Edman degradation, the products of each cycle being also monitored for release of 32 P radioactivity.
(ii) Fig. 8 shows that the sequence of the chymotryptic peptide is AEXLSKEEAR, in which X in the third cycle (corresponding to Lys 207 in uncross-linked MTF) contains most of the 32 P radioactivity. This peptide sequence overlaps with the Cterminal sequence of peptide I*. These results establish that Lys 207 is the site of cross-linking of MTF to the 3Ј-end of tRNA 2 fMet . The yield of amino acid in the first and most of the successive cycles was at least 50% of that expected on the basis of 32 P radioactivity present.
(iii) Finally, as for peptide I*, the observed molecular mass of 3536.71 Da for peptide II* differs by only ϳ3 Da from the mass of 3533.66 Da expected for a peptide Gln 186 -Arg 214 , in which Lys 207 is attached through a morpholino type of linkage (Fig. 6, III) to the AMP moiety derived from the 3Ј-end of the tRNA. Cross-linking of Mutant tRNA fMet and Yeast tRNA Phe to MTF-Although MTF formylates only the initiator methionyl-tRNA species (15), previous studies have shown that it will nevertheless bind to other tRNAs almost as well as to the initiator tRNA (35). Therefore, we have investigated whether the periodate-oxidized G72 mutant of tRNA 2 fMet , which is a very poor substrate for MTF (16 -18), and yeast tRNA Phe , which is not a substrate for MTF, will cross-link to MTF. Fig. 9 shows the results. While all three tRNAs react with MTF, the rate of cross-linking with tRNA 2 fMet is significantly higher than with the mutant tRNA 2 fMet or yeast tRNA Phe . The reaction with tRNA 2 fMet reaches a plateau after 1 h, whereas even after 2 h  (Fig. 6, structure I). c Lysine linked to AMP moiety. d Molecular mass based on hydroxymorpholine derivative (Fig. 6, structure II). e Molecular mass based on morpholine derivative (Fig. 6, structure III).
FIG. 5. Mass spectrum of peptide I*. 1 l of peptide (1-2 pmol; based on the specific radioactivity) was mixed with 1 l of matrix, loaded on the sample plate, and allowed to dry completely. Prior to peptide mass determination, the mass spectrophotometer was calibrated using insulin chain B as a standard. the reactions with the other tRNAs have not reached the level attained with tRNA 2 fMet . These results are in agreement with those of Hountondji et al. (21), who showed that the rate of reaction of periodate-oxidized yeast tRNA Phe with MTF, as followed by inactivation of MTF, was about 6-fold lower than that of tRNA fMet .
The site(s) of cross-linking of the G72 mutant of tRNA 2 fMet and yeast tRNA Phe were also analyzed by HPLC of tryptic digests of MTF cross-linked to these tRNAs. The same two radioactive peaks as seen above for cross-linking of wild type tRNA 2 fMet to MTF were found (data not shown). Thus, the mutant tRNA 2 fMet and the yeast tRNA Phe also cross-link to Lys 207 in MTF. DISCUSSION As a first step in studies on the topology of interaction of E. coli initiator tRNA with MTF, we have shown that reaction of periodate-oxidized tRNA with MTF leads to cross-linking of the tRNA specifically to Lys 207 of the enzyme. This suggests that Lys 207 comes close to the 3Ј-end of the tRNA. Since MTF formylates the amino group of methionine attached to the 3Ј-end of the tRNA, Lys 207 is likely to be within or near the active site of the enzyme.
The Lys 207 of MTF is part of the sequence 207 KLSKE 211 . This sequence is related to a similar sequence, KMSKS or KLSKS, found in virtually all class I aminoacyl-tRNA synthetases (36). The significance, if any, of this similarity in sequences is not known. In two of the E. coli class I aminoacyl-tRNA synthetases, tyrosyl-tRNA synthetase and methionyl-tRNA synthetase, one or the other of the lysine residues in the KMSKS or KLSKS sequence has been shown to cross-link to the 3Ј-end of the corresponding periodate-oxidized tRNA (37,38). These lysine residues are also functionally important for stabilization of the ground state and/or the transition state during the formation of aminoacyl-adenylate (36,39,40). In a class II aminoacyl-tRNA synthetase also, a lysine residue that cross-links to the 3Ј-end of periodate-oxidized tRNA is part of a conserved motif (motif 2) important for aminoacyl-adenylate formation and for transfer of the amino acid to the tRNA (41,42). Whether Lys 207 or Lys 210 plays a functional role in MTF is not known. Although these amino acid residues lie within a conserved region in the six MTF protein sequences deduced on the basis of DNA sequences (Fig. 10), Lys 207 is present in E. coli, Hemophilus influenzae and Rickettsia prowazekii MTF but not in Thermus thermophilus, Mycoplasma genitalium, or yeast mitochondrial MTF.
Another region of very strong sequence conservation in MTF includes amino acids 83-150. This region contains amino acids Asn 109 , His 111 , and Asp 147 , thought to be involved in catalysis (43,44) in E. coli glycinamide ribonucleotide formyltransferase, another enzyme that, like MTF, transfers a formyl group using N 10 -formyltetrahydrofolate as a cofactor (45). These three amino acids are conserved in all six of the MTF sequences known so far. The strict conservation of these amino acid residues in MTF and the very strong homology in this region among MTF, the glycinamide ribonucleotide formyltransferases, and amino imidazole carboxamide ribonucleotide formyltransferases (46,47) from a number of sources suggest that these amino acids also play a similar catalytic role in MTF. If, as stated above, Lys 207 is at or near the active site of the MTF, it might be close to amino acids 109, 111, and 147 in the three-dimensional structure.
Reaction of periodate-oxidized G72 mutant initiator tRNA, which is a very poor substrate for MTF and yeast tRNA Phe , also led to cross-linking to MTF, although at a slower rate compared with the wild type initiator tRNA. This result agrees with the previous observation that although MTF formylates only the initiator methionyl-tRNA species, it binds almost as well to other tRNAs with dissociation constants in the micromolar range. The site of cross-linking to the mutant initiator tRNA and to yeast tRNA Phe is also Lys 207 . It would thus appear that MTF has a binding pocket for the 3Ј-end of the tRNA and that all of these tRNAs bind initially to MTF in a similar manner. The differences in rates of reaction of MTF with cognate versus noncognate tRNA or the G72 mutant initiator tRNA could be due to a conformational change, subsequent to binding, of the MTF-cognate tRNA complex, which places Lys 207 of MTF in a favorable position for reaction with the 3Ј-end of the tRNA in the cognate complex. Evidence for a possible conformational change of the cognate tRNA upon binding to MTF has been obtained by NMR analysis of MTF complexed to initiator and elongator species of methionine tRNA (35). NMR analysis showed a general broadening and loss of intensity of resonances assigned to G:C base pairs in the acceptor stem of the initiator tRNA species but not of the elongator tRNA species. The notion of a conformational change in the MTF-initiator tRNA complex subsequent to binding is similar to the situation with aminoacyl-tRNA synthetase-tRNA complexes in which a conformational change of the complex is triggered by cognate tRNAs but not by noncognate tRNAs (48).
Another explanation for the cross-linking of all three tRNAs to Lys 207 of MTF is that Lys 207 is just a particularly reactive lysine residue in MTF. We consider this unlikely. First, the periodate-oxidized initiator tRNA does not react with just any protein. For example, it does not react with bovine serum albumin (data not shown). Second, we have also cross-linked periodate-oxidized E. coli 5 S rRNA carrying a 3Ј-terminal 32 P-pC extension to MTF and analyzed the tryptic digest of the cross-linked MTF by HPLC. In contrast to the specific crosslinking of the 3Ј-ends of tRNAs to Lys 207 , there were several other radioactive peptides in the tryptic digest of MTF crosslinked to 5 S RNA (data not shown).
Mass spectral analysis of peptides derived from MTF that were cross-linked to the AMP moiety of the tRNA has proven extremely useful in the characterization of the cross-linked peptides and the nature of the cross-link formed. Earlier reports on the reaction of 2Ј,3Ј-dialdehydes derived by periodate oxidation of tRNA, nucleosides, or 5Ј-mononucleotides with primary amines indicated the production of either a Schiff's base (Fig. 6, I) or a morpholino derivative in which one of the carbon atoms in the morpholine ring carries a hydroxyl group (49) (Fig. 6, II). The molecular masses that we have obtained for the cross-linked peptides are, however, more consistent with a morpholino derivative described by Brown and Read (50) in which none of the carbon atoms in the morpholine ring carries a hydroxyl group (Fig. 6, III). A possible mechanism for the formation of this is that the secondary amine formed by reduction of the Schiff's base intermediate with sodium cyanoborohydride attacks the neighboring carbonyl group, leading to the formation of a morpholine ring. Loss of water followed by reduction with sodium cyanoborohydride or sodium borohydride would generate the morpholino derivative III shown in Fig. 6. Finally, the work described here has provided the first indication of the amino acid residue in MTF that comes close to the 3Ј-end of the tRNA. In parallel work, we are attempting isolation and identification of suppressor mutations in MTF that compensate for formylation defects of mutant initiator tRNA. The information derived from these experiments, along with knowledge of the amino acid residues in MTF that are highly conserved (Fig. 10), can be used for the selection of amino acid residues in MTF for site-specific mutagenesis and structurefunction relationship studies.