Novel Catalytic Mechanism of Nucleophilic Substitution by Asparagine Residue Involving Cyanoalanine Intermediate Revealed by Mass Spectrometric Monitoring of an Enzyme Reaction*

l-2-Haloacid dehalogenase fromPseudomonas sp. YL catalyzes the hydrolytic dehalogenation, in which Asp10 acts as a nucleophile to attack the α-carbon of l-2-haloalkanoates to form an ester intermediate, which is subsequently hydrolyzed to produced-2-hydroxyalkanoates. Surprisingly, replacement of the catalytic residue, Asp10, by Asn did not result in total inactivation of the enzyme (Kurihara, T., Liu, J.-Q., Nardi-Dei, V., Koshikawa, H., Esaki, N., and Soda, K. (1995) J. Biochem. 117, 1317–1322). In this study, we monitored the D10N mutant enzyme reaction by ion-spray mass spectrometry, and found that the enzyme shows a unique structural change when it was incubated with the substrate, l-2-chloropropionate. LC/MS and tandem MS/MS analyses revealed that Asn10 attacks the substrate to form an imidate, and a proton and d-lactic acid are eliminated to produce a nitrile (β-cyanoalanine residue), followed by hydrolysis to reproduce Asn10. This is the first report of the function of Asn to catalyze nucleophilic substitution through its conversion to β-cyanoalanine residue as an intermediate structure. Also, these results demonstrate that mass spectrometry is remarkably useful in monitoring enzyme reactions.

We have studied intensely the reaction mechanism and structure of L-DEX YL, and revealed that the reaction begins with nucleophilic attack of Asp 10 on the ␣-carbon atom of L-2haloalkanoates, causing the release of a halide ion and the formation of an ester intermediate (Scheme 1, I), which is subsequently hydrolyzed (3,4). In the x-ray structure, there is a water molecule close to Asp 10 , Ser 175 , Asn 177 , and Asp 180 , the latter three of which are supposed to enhance the nucleophilicity of the water molecule for hydrolysis of the ester intermediate (5) (Scheme 1). These enzymological studies of L-DEX YL have provided critical information on the structures and functions of various hydrolases including P-type ATPases, which have a significant sequence similarity with L-DEX YL and are proposed to have the haloacid dehalogenase fold (6 -9). Recent studies on the crystal structure of Ca 2ϩ -ATPase from sarcoplasmic reticulum revealed that this ATPase has a high threedimensional structural similarity to that of L-DEX YL and many essential amino acid residues are conserved between these two proteins (10).
In the course of the studies of L-DEX YL, we surprisingly found that replacement of the catalytic residue, Asp 10 , by Asn did not completely inactivate the enzyme, whereas replacement by Ala, Gly, Ser, or Glu resulted in total inactivation (11). One possible explanation is that the activity of the D10N preparation is attributed to the post-translational production of the wild-type enzyme by deamidation of Asn 10 : deamidation of Asn to produce Asp has been shown to occur in a variety of proteins (12,13). However, it is also possible that the D10N enzyme itself catalyzes the hydrolysis of the substrate.
In the present study, we analyzed the mechanism of dehalogenation catalyzed by the L-DEX YL D10N preparation by ion-spray mass spectrometry (MS), and found that the D10N enzyme itself, in addition to the wild-type enzyme produced by deamidation of the D10N enzyme, catalyzes dehalogenation. Asn 10 was shown to undergo a unique structural change in the course of the catalysis: it is converted into a ␤-cyanoalanine (␤-CNAla) residue via an imidate, and then regenerated in the catalytic cycle. This is the first report of the formation of a ␤-CNAla residue as an enzyme reaction intermediate. Also, the results demonstrate the validity of mass spectrometry in enzyme dynamics.
DNA Techniques-Replacement of amino acid residue(s) was carried out by the method of Kunkel et al. (14). Mutant genes for D10N and D180N were constructed as reported previously (11). The synthetic mutagenic primer designed for the D10N/L11K double mutant was as follows (the underlines indicate the mutagenized nucleotides): 5Ј-CAGCGTACCGTACTTGTTGAAGGCAATACC. The substitutions were confirmed with a Shimadzu PPSQ-10 protein sequencer (Kyoto, Japan), an Applied Biosystem 373A DNA sequencer (Foster City, CA), and an ion-spray mass spectrometer PE-Sciex API 300 (Sciex, Thornhill, Ontario, Canada). Mutant enzymes were produced in Escherichia coli JM109.
Enzyme Purification-All operations were performed at 4°C, and 50 mM potassium phosphate (pH 7.5) was used as the standard buffer unless otherwise stated. The recombinant E. coli cells producing the enzyme were grown aerobically at 37°C for 14 h in LB medium containing 200 g/ml ampicillin and 0.2 mM isopropyl-1-thio-␤-D-galactoside. The cells harvested from a 4-liter culture were disrupted by sonication. The supernatant was fractionated with ammonium sulfate. A fraction of 40 -70% saturation was dissolved in the standard buffer, and then applied to a Butyl-Toyopearl 650 column (3 ϫ 25 cm) (TOSOH, Tokyo). The column was washed with 500 ml of the buffer supplemented with 30% (w/v) ammonium sulfate, and the enzyme was eluted with a linear gradient of 30Ϫ20% ammonium sulfate in the buffer with a total volume of 1 liter. The fractions containing the mutant enzyme were identified by sodium dodecyl sulfate-polyacrylamide gel electrophoresis, dialyzed against 10 mM potassium phosphate (pH 7.5), and applied to a DEAE-Toyopearl 650M column (1.7 ϫ 5 cm) (TOSOH, Tokyo) equilibrated with the same buffer. The enzyme was eluted with a linear gradient of 10Ϫ50 mM potassium phosphate (pH 7.5) with a total volume of 200 ml. The fractions containing the mutant enzyme were pooled and used as a purified preparation of the mutant enzyme.
Determination of the Enzyme Activity and Protein-L-DEX YL was routinely assayed by determination of chloride ions produced from L-CPA by the method of Iwasaki et al. (15). The standard assay mixture (100 l) contained 2.5 mol of L-CPA, 10 mol of Tris sulfate buffer (pH 9.5), and the enzyme. The reaction was terminated by addition of 10 l of 1.5 M sulfuric acid after incubation at 30°C for 10 min. One unit of the enzyme was defined as the amount of enzyme that catalyzes dehalogenation of 1 mol of L-CPA per min. Protein concentration was determined with a Bio-Rad protein assay kit (Hercules, CA).
MS Analysis of Wild-Type L-DEX YL, D10N, D10N/L11K, and D10N/D180N Incubated with L-CPA-Wild-type L-DEX YL, the D10N mutant enzyme, the D10N/L11K double mutant enzyme, or the D10N/ D180N double mutant enzyme (10 nmol each) was incubated with L-CPA in the standard assay mixture. The reaction was terminated with 10 l of 20% (v/v) formic acid, followed by centrifugation at 8200 ϫ g for 1 min. The enzyme was deionized with Jasco HPLC system (Tokyo) with a C 18 reverse phase column (Shiseido Capcell Pak, SG 300 Å 5 m, 4.6 ϫ 250 mm, Tokyo) at room temperature, employing linear gradients of solvent A (0.1% formic acid) and solvent B (acetonitrile containing 0.1% formic acid): sample injection; 3 min, 20% B; 12 min, 20Ϫ100% B; 3 min, 100% B; 2 min, 100Ϫ20% B (flow rate: 1 ml/min). The deionized enzyme was lyophilized, dissolved in 50% acetonitrile containing 0.05% formic acid, and introduced into the mass spectrometer, PE-Sciex API 300 or API 365 at a flow rate of 2.5 l/min. The quadrupole was scanned from 300Ϫ800 to 2,000 atomic mass units with a step size of 0.1Ϫ0.2 atomic mass unit and with a dwell time of 0.2-1 ms/step. The orifice potential was set at 30Ϫ60 V. The molecular mass was calculated using the Bio-MultiView software supplied by PE-Sciex.
LC/MS Analysis of the Peptides Containing Asn 10 -The D10N/L11K mutant enzyme incubated and deionized as described above was lyophilized, dissolved in 50 l of 8 M urea in 200 mM Tris sulfate buffer (pH 7.2), and incubated at 37°C for 1 h. The enzyme was then digested by addition of 82.5 pmol of Lys-C dissolved in 75 l of 1 M Tris sulfate buffer (pH 9.0) at 37°C for 12 h. The proteolysates (95-100 l) were loaded onto the C 18 Shiseido Capcell Pak column connected to the mass spectrometer, PE-Sciex API 300 or API 3000. Elution was carried out with 2% acetonitrile containing 0.1% formic acid for 5 min, followed by a linear gradient of 2Ϫ82% acetonitrile containing 0.1% formic acid over 40 min at a flow rate of 40 l/min. The quadrupole was scanned from 150Ϫ200 to 1,000 with a step size of 0.2 and with a dwell time of 0.5 ms/step. The orifice potential was set at 50 V. For detailed analysis of the peptides containing Asn 10 , the high performance liquid chromatography (HPLC) fractions were collected, concentrated, and injected into the mass spectrometer for tandem MS/MS analysis described below.
Tandem MS/MS Analysis of the Peptides Containing Asn 10 -The daughter ion spectra were obtained in the triple-quadrupole daughter ion scan mode by introducing the peptides containing Asn 10 from Q1 into a collision cell (Q2) and observing the daughter ions in Q3.  (16,17). Authentic Ala (1.2 mg), Asn (2.6 mg), Asp (1.6 mg), and ␤-CNAla (1.7 mg) were separately dissolved in 200 l of 60% pyridine containing 1.5% dimethylallylamine, mixed with 20 l of phenylisothiocyanate, and incubated anaerobically at 45°C for 40 min. Remaining phenylisothiocyanate was removed with benzene, and the resultant phenylthiocarbamyl derivatives were lyophilized. The lyophilized samples were mixed with 20 l of trifluoroacetic acid and incubated anaerobically at 45°C for 15 min. The resultant 2-anilino-5-thiazolinone derivatives were extracted with ethyl acetate, dried, and converted to the phenylthiohydantoin (PTH)-derivatives by incubating with 1 N hydrochloric acid at 80°C for 8 min. The PTH-derivatives were extracted with ethyl acetate, dried, and used for identification and quantification. The dried PTH-derivatives were dissolved in 40% acetonitrile containing 10 mM acetic acid, and loaded onto a reverse phase column for separation of PTH-derivatives (Wakopak WS-PTH column, 4.6 ϫ 250 mm) connected to the mass spectrometer, PE-Sciex API 365. Elution was carried out with 40% acetonitrile containing 0.1% formic acid at a flow rate of 45 l/min. The total ion current chromatogram was recorded in the singlequadrupole mode. The quadrupole was scanned from 200 to 1,000 with a step size of 0.1 and with a dwell time of 0.5 ms/step. The ion-spray voltage was set at 5 kV, and the orifice potential was 50 V. Deamidation of D10N-We found that the D10N preparation showed slight dehalogenation activity (relative specific activity, about 1% of that of the wild-type enzyme when determined immediately after purification). The activity increased in a time-and temperature-dependent manner. The D10N preparation showed 3.6 and 6.2% of the wild-type enzyme activity after storage at 4°C for 2 and 3 months, respectively; when stored at room temperature for about a month, it showed 26% activity of the wild-type enzyme activity. Amino acid sequencing of the D10N preparation showing 26% of the wild-type enzyme activity revealed the occurrence of Asp at the position of Asn 10 (data not shown). These results indicate that the side chain amide of Asn 10 is slowly deamidated to produce carboxylate.

MS Analysis of Wild-Type
MS Analysis of D10N-Although the D10N preparation contained the wild-type enzyme produced by the deamidation of Asn 10 as described above, its amount in the fresh preparation is considered to be at most 1% of the total enzyme judging from the specific activity of the preparation. Thus it is possible to monitor the structural change of D10N itself by MS, because the small amount of the wild-type enzyme does not interfere with the mass spectra of D10N, which is present abundantly. We determined the molecular masses of D10N incubated with L-CPA for the indicated periods shown in Fig. 2 Two probable mechanisms can be proposed for the structural change of D10N (Schemes 2 and 3). The first one is shown in Scheme 2: Asn 10 attacks L-CPA to form an asparagine ␤-imidate 1-carboxyethyl ester (II, M ϩ 72), and a proton and D-lactic acid are eliminated from the imidate to produce a nitrile (III, M Ϫ 18), which is subsequently hydrolyzed to reproduce Asn 10 (M). The difference between the measured (ϩ73) and estimated (ϩ72) mass increments can be attributed to a measurement error or to the conversion of the imidate to an aspartate 1-carboxyethyl ester as shown in Scheme 4 (V, M ϩ 73) upon denaturation of the enzyme with formic acid (18) (discussed below). In the second probable mechanism shown in Scheme 3, the imidate intermediate is produced in the same manner as shown in Scheme 2, followed by formation of an intramolecular crosslink (IV, M Ϫ 18) caused by nucleophilic attack by the hydroxyl group of an amino acid residue positioned in the vicinity of the imidate. The x-ray crystallographic analysis indicates that Ser 175 , Ser 176 , and Thr 14 possibly play this role (5).
Determination of the Structure Causing the 73-Da Increase-To examine if Asn 10 was modified during the incubation with L-CPA, we carried out amino acid sequencing. Native D10N gave the following sequence (the boldface indicates the 10th amino acid residue): Met-Asp-Tyr-Ile-Lys-Gly-Ile-Ala-Phe-Asn-Leu, which is identical to that predicted from the nucleotide sequence. Sequencing analysis of the M ϩ 73 variant gave the following result: Met-Asp-Tyr-Ile-Lys-Gly-Ile-Ala-Phe-X-Leu, in which X is an amino acid residue whose retention time is different from that of every common amino acid residue including Asn, indicating that Asn 10 is modified in the M ϩ 73 variant.
Leu 11 of the D10N enzyme was replaced by Lys to create D10N/L11K as described under "Experimental Procedures," which was used to examine the modification of Asn 10 by MS: Lys-C treatment of D10N/L11K produces a short peptide fragment containing Asn 10 , which is small enough for determination of molecular mass accurately. The mass spectrum of D10N/ L11K showed a peak at 26,195 Da (MЈ), which is virtually identical to the predicted value, 26,193 Da. It was confirmed that the MЈ ϩ 73 (26,268 Da) and the MЈ Ϫ 18 (26,177 Da) variants were formed when D10N/L11K was incubated with L-CPA for 10 s and 8 min, respectively (data not shown). Thus the Lys residue introduced next to the C-terminal side of Asn 10 has little effect on the reactivity of the enzyme.
D10N/L11K incubated with (or without) L-CPA was digested with Lys-C, and the resultant peptides were analyzed by LC/ MS. Mass spectrum of a peptide derived from native D10N/ L11K showed a peak at m/z 649.5, which was assigned as a positive monovalent ion of the hexapeptide Gly 6 -Lys 11 (Fig. 3A,  MЈ). The hexapeptide isolated from the enzyme incubated with L-CPA for 10 s and denatured with acid showed a new peak at m/z 722.6, which is about 73 Da higher than that derived from the unmodified one (Fig. 3B, MЈ ϩ 73). To determine the amino acid residue where the mass increment had occurred, tandem MS/MS analysis of the MЈ ϩ 73 peptide was carried out. The product ions produced from the unmodified and the MЈ ϩ 73 peptides are shown in Figs. 4, A and B, respectively. The Y series ions at m/z 649.3, 479.0, 408.1, and 261.2 derived from the unmodified peptide were assigned as Gly-Ile-Ala-Phe-Asn-Lys, Ala-Phe-Asn-Lys, Phe-Asn-Lys, and Asn-Lys, respectively (Fig. 4A). The ions produced from the MЈ ϩ 73 peptide at m/z 722.6, 552.5, 481.2, and 334.4 were about 73 Da higher than the corresponding ions derived from the unmodified one, respectively (Fig. 4B). However, the molecular mass of the fragment ion for the C-terminal Lys derived from the MЈ ϩ 73 peptide (m/z 147.1) was virtually identical to that from the unmodified one (m/z 147.0). This result indicates that the modification causing the 73-Da increase occurred at Asn 10 .
Because a 1-Da difference is significant in the mass range shown in Figs. 3 and 4, the exact mass increase in the MЈ ϩ 73 peptide at Asn 10 is 73 Da, not 72 Da. This indicates the presence of the ester, not the imidate, in the peptide. However, this result does not exclude the possibility that an imidate intermediate was actually produced in the reaction, because an imidate is readily hydrolyzed to an ester under a low-pH condition (18) (Scheme 4), which was employed in the present experiment to terminate the enzyme reaction. An 18 O atom is expected to be incorporated in the ester intermediate by addition of H 2 18 O upon acidification, if the presumed imidate intermediate is indeed hydrolyzed spontaneously by acidification. To examine if the observed ester was produced from the imidate, H 2 18 O was added to the reaction mixture at the termination of incubation. The mass spectrum showed a new peak at m/z 724.5 (Fig. 3C,  MЈ ϩ 75), indicating that an oxygen atom of a water molecule was introduced into the hexapeptide 6 -11. Tandem MS/MS analysis of the MЈ ϩ 75 peptide revealed that a 2-Da mass increase occurred at Asn 10 (data not shown). No MЈ ϩ 75 peptide was produced when H 2 18 O was added to the reaction mixture at 1 min after the termination of the reaction (data not shown), indicating that there is no exchange of an 18   crease-We carried out the Edman reaction with free ␤-CNAla as a substrate, and analyzed the product by MS. It was revealed that the Edman reaction gave PTH-Asn as the most predominant product (data not shown). A nitrile structure undergoes acid-catalyzed nucleophilic addition of water to produce an amide (19). Edman reaction employs strong acids such as trifluoroacetic acid and hydrochloric acid in the reaction cycle. Therefore the nitrile structure of ␤-CNAla is converted to an amide in the course of the reaction. Thus, even if a ␤-CNAla residue is formed in a peptide, it is converted to PTH-Asn during Edman degradation.
We carried out amino acid sequencing of the M Ϫ 18 variant of D10N. It gave the following sequence: Met-Asp-Tyr-Ile-Lys-Gly-Ile-Ala-Phe-Asn-Leu. Although this result may indicate that Asn 10 is intact in the M Ϫ 18 variant, it is also consistent with the mechanism of ␤-CNAla formation at residue number 10 followed by hydrolysis during Edman degradation.
To confirm the ␤-CNAla formation, the MЈ Ϫ 18 variant of D10N/L11K was digested, and the resultant peptides were analyzed by LC/MS. We obtained a peptide with an m/z value of 631.5 (Fig. 3D, MЈ Ϫ 18), and found by tandem MS/MS analysis that Asn 10 was specifically modified in such a way as its molecular mass becomes 18 Da lower than the original value (Fig. 4, A and C). Accordingly, the most probable structure formed at residue number 10 in the M Ϫ 18 (or MЈ Ϫ 18) variant is a ␤-CNAla residue. Since nitrile can hardly be produced from the cross-link structure in the process of MS analysis, the mechanism shown in Scheme 3 is excluded.
Reconversion of ␤-CNAla into Asn 10 -To examine if the final step of the reaction is the reconversion of ␤-CNAla into Asn as shown in Scheme 2, D10N/L11K was incubated in H 2 18 O in the presence (or absence) of L-CPA for 60 min. If ␤-CNAla is reconverted to Asn as shown in Scheme 2, an 18 O atom is expected to be incorporated in the side chain amide of Asn 10 when the mutant enzyme is incubated with the substrate in H 2 18 O. After incubation, D10N/L11K was denatured and digested with Lys-C. Since the C-terminal carboxyl group produced by proteolysis contains two 18 O atoms when the proteolysis is performed in H 2 18 O (20), the peptide 6 -11 reconverted from the MЈ Ϫ 18 peptide in H 2 18 O is expected to contain three 18 O atoms, two of which are in the carboxyl group of C-terminal Lys and the rest of which is in the side chain amido group of Asn 10 . The hexapeptide isolated from D10N/L11K incubated without L-CPA and digested in H 2 18 O showed a new peak at m/z 653.4, which is about 4 Da higher than that derived from the unmodified one (Fig. 3E, MЈ ϩ 4). When D10N/L11K was incubated in the presence of L-CPA, the mass spectrum showed a new peak at m/z 655.5, which is 6 Da higher than that derived from the unmodified one (Fig. 3F, MЈ ϩ 6). We found by tandem MS/MS analysis that two 18 O atoms were incorporated in the C-terminal Lys in both the MЈ ϩ 4 and MЈ ϩ 6 peptides, and that an 18 O atom was incorporated in Asn 10 specifically in the MЈ ϩ 6 peptide (data not shown). These results indicate that the nitrile undergoes the nucleophilic attack of a water molecule to produce the side chain amide of Asn 10 .
In conclusion, the unique structural change of D10N occurs through the mechanism shown in Scheme 2: Asn 10 attacks the substrate to form the imidate, and a proton and D-lactic acid are eliminated to produce the nitrile. This is the first report showing that Asn functions as a catalytic nucleophile in enzymatic hydrolysis where ␤-cyanoalanine residue is produced as a reaction intermediate. Also, the results demonstrate that mass spectrometry is remarkably useful in monitoring enzyme reactions.
placement of the mercapto group in cysteine by cyanide (25), whereas the ␤-CNAla residue is formed by ␣,␤-elimination of the imidate as shown in Scheme 2.
Possible Roles and Occurrence of the ␤-CNAla Residue in a Protein-Roy et al. (26) have reported that replacement of the side chain amide of Asn 5 of oxytocin by a nitrile produced an analog with very weak activity, whereas a similar substitution in glycinamide 9 produced a highly active analog. Thus replacement of Asn with a ␤-CNAla residue can alter the functions of a protein. We reported here that the side chain amide of Asn can be converted to a nitrile. These results raise the possibility that an Asn residue is post-translationally converted into a ␤-CNAla residue in vivo to modify the function of proteins. Although there has been no other report on a naturally produced ␤-CNAla residue in a protein so far, the presence of a ␤-CNAla residue has perhaps been unrecognized because of the lack of an appropriate analysis method for amino acid residues constituting proteins: Edman degradation, the most widely used method for amino acid sequencing, results in the conversion of the ␤-CNAla residue into PTH-Asn, thereby making it impossible to distinguish a ␤-CNAla residue from an Asn residue. MS analysis of proteins may reveal the presence of a ␤-CNAla residue in other proteins.