Affinity Alkylation of the Trp-B4 Residue of the β-Subunit of the Glutaryl 7-Aminocephalosporanic Acid Acylase of Pseudomonas sp. 130*

Glutaryl 7-aminocephalosporanic acid acylase ofPseudomonas sp. 130 (C130) was irreversibly inhibited in a time-dependent manner by two substrate analogs bearing side chains of variable length, namely 7β-bromoacetyl aminocephalosporanic acid (BA-7-ACA) and 7β-3-bromopropionyl aminocephalosporanic acid (BP-7-ACA). The inhibition of the enzyme with BA-7-ACA was attributable to reaction with a single amino acid residue within the β-subunit proven by comparative matrix assisted laser desorption/ionization-time of flight mass spectrometry. Further mass spectrometric analysis demonstrated that the fourth tryptophan residue of the β-subunit, Trp-B4, was alkylated by BA-7-ACA. By 1H-13C HSQC spectroscopy of C130 labeled by BA-2-13C-7-ACA, it was shown that tryptophan residue(s) in the enzyme was alkylated, forming a carbon-carbon bond. Replacing Trp-B4 with other amino acid residues caused increases in K m , decreases ink cat, and instability of enzyme activity. None of the mutant enzymes except W-B4Y could be affinity-alkylated, but all were competitively inhibited by BA-7-ACA. Kinetic studies revealed that both BA-7-ACA and BP-7-ACA could specifically alkylate Trp-B4 of C130 as well as Tyr-B4 of the mutant W-B4Y. Because these alkylations were energy-requiring under physiological conditions, it is likely that the affinity labeling reactions were catalyzed by the C130 enzyme itself. The Trp-B4 residue is located in the middle of a characteristic αββα sandwich structure. Therefore, a large conformational alteration during inhibitor binding and transition state formation is likely and suggests that a major conformational change is induced by substrate binding during the course of catalysis.

intermediate in the production of many semisynthetic antibiotics, can be synthesized by the chemical deacylation of cephalosporin C (CPC) employing iminoethers, nitrosyl chloride, and methanol. Because these steps involve toxic compounds causing environmental contamination, enzymatic methods of CPC deacylation are of great interest. Since cephalosporin acylase (CA) has very low activity toward CPC, most of the bioprocess reported methods involve a two-step conversion of CPC to 7-ACA. First, CPC is oxidized by D-amino acid oxidase (1), and the product, glutaryl-7-ACA (GL-7-ACA), is then converted to 7-ACA by GL-7-ACA acylase (EC 3.5.1.11) (Fig. 1A).
The gene encoding GL-7-ACA acylase of Pseudomonas sp. 130 (C130) has been cloned (2) and overexpressed in Escherichia coli (3). The active enzyme consists of two dissimilar subunits, a ␤-subunit (58 kDa) and an ␣-subunit (18 kDa). The subunits are derived from one precursor polypeptide. The catalytic mechanism of CA was deduced from its sequence and that of the functionally analogous penicillin G acylase (4). The latter enzyme is a well known serine proteinase that carries out an N-terminal nucleophilic (Ntn) attack (5) via a single serine as its catalytic center (6).
The first crystal structure of a CA, from Pseudomonas diminuta (CAD), was solved at a resolution of 2.0 Å (7). A similar crystal structure of C130 at a resolution of 2.4 Å, was solved independently. 2 The sequence identity between CAD and C130 is Ͼ98%. The crystal structures revealed some similarities between the active centers of the CAs and penicillin G acylase, although their overall structures are quite different, as expected from the low sequence similarities of the two enzymes. The structure of CAD (and C130) bears a characteristic ␣␤␤␣ motif. The N-terminal residue Ser-B1 of the ␤-subunit is the catalytic center, indicating that CA belongs to the Ntn hydrolase superfamily, as defined by the Structural Classification of Proteins data base (8). Recently, structures of the binary complexes of CAD with GL-7-ACA and glutarate were solved at 2.5-and 2.6-Å resolutions, respectively (9). Although the site of initial substrate binding can be inferred from these studies, this information may not reflect the dynamic status of the enzyme-substrate complex during catalysis.
In this study, a series of substrate analogs with side chains of variable length were synthesized. Two compounds, 7␤-bromoacetyl aminocephalosporanic acid (BA-7-ACA) and 7␤-3-bromopropionyl aminocephalosporanic acid (BP-7-ACA), could specifically function as affinity reagents. The reaction site was the indole ring of Trp-B4 as determined by LC/MS peptide mapping, tandem MS spectra analysis, and NMR. These results provide useful insight into the dynamic status of substrate binding and the catalytic mechanism of the C130 acylase.

Synthesis of Substrate Analogues
Substrate analogs were synthesized by acylation of 7-ACA with the relevant acyl chloride. These structures were verified by MS and NMR.

Enzyme Purification
The crude enzymes were prepared as described (3). They were further purified on a Sepharose Q F (Amersham Biosciences, Inc.) column equilibrated with 10 mM sodium phosphate buffer (pH 8.0) and eluted with a gradient of NaCl ranging from 0 to 500 mM on fast protein liquid chromatography (Amersham Biosciences). The enzyme peak was collected and concentrated to Ͼ1 mg/ml with the Molecut II LGC TM (Millipore Corp.) spin column. Further purification was achieved via size exclusion chromatography. Purified C130 was more than 90% homogenous, as determined by densitometric scanning after Coomassie Blue staining of SDS-PAGE gels using an LKB Ultrascan XL densitometer. In those cases where only partial purification was achieved (i.e. some mutant enzymes), quantitation was performed by SDS-PAGE, followed by densitometric scanning, using various amounts of purified C130 protein as standard.

Enzyme Activity Assay
Enzyme activity was determined as previously described (12). One unit of activity is that amount of enzyme that produces 1 mol of 7-ACA/min (3). The specific activity of the purified C130 was 6.2 mol/min/mg.

MALDI-TOF Mass Spectrometric Studies
All experiments were performed on a Bruker REFLEX mass spectrometer. Enzyme samples (1 mg/ml) were prepared using ␣-cyano-4hydroxycinnamic acid as matrix in solution with 50% acetonitrile, 0.1% trifluoroacetic acid in water. 1 l of sample solution was mixed with 1 l of matrix solution, and 1 l of the mixture was applied to the SCOUT 384 target.

Site-directed Mutagenesis
The W-B4L mutant was conducted by PCR using plasmid pKKCA1 as template and the primer, 5Ј-TCC AAC TCC TTG GCG GTG GCC CCG-3Ј. This oligonucleotide corresponds to the coding sequence from Ser-B1 to Pro-B8 with the Trp-B4 codon replaced by a Leu codon (underlined). This mutated gene retained the characteristic BamHI site at the Ser-B1 codon and created a SmaI site by altering the Ala-B7 codon. Mutants of P-B8G, W-B4A, W-B4T, W-B4F, W-B4H, and W-B4Y were obtained by replacing this BamHI-SmaI fragment using oligodeoxyribonucleotides that replaced Pro-B8 with Gly and Trp-B4 with Ala, Thr, Phe, His, and Tyr, respectively. Correctness of the mutant clones was confirmed by DNA sequencing.

Preparation of Labeled Enzymes
The acylases (0.5 mg/ml) were incubated with excess BA-7-ACA (8 mM) in 100 mM sodium phosphate buffer (pH 8.0) for 1 h at 37°C. The course of the labeling reaction was followed by measuring residual enzymatic activity assay (lower than 1.0% of the original enzymatic activity, indicating the completion of labeling). After concentration, the reaction mixture was subjected to gel filtration (Superdex 75; Amersham Biosciences). The fractions containing modified enzyme were collected and concentrated to Ͼ1 mg/ml with Molecut II LGC TM (Millipore Corp.).

Digestion of Denatured Enzymes
Both the original enzyme (2 mg/ml) and the labeled enzymes (2 mg/ml) were denatured at 65°C in 8 M urea for 45 min. The denatured enzymes were digested either with trypsin or Lys-endoproteinase (Roche Molecular Biochemicals) according to the protocol of the supplier.

HPLC and MS-MS Analysis for Oligopeptides
On-line HPLC separation was performed on a Hewlett-Packard model 1100 system at a flow rate of 0.05 ml/min. Solvent A was 0.01% trifluoroacetic acid, 1% HOAc in water, and solvent B was 0.01% trifluoroacetic acid, 1% HOAc in acetonitrile. A 2.1 ϫ 30-mm C8 reverse phase column (Applied Biosystems) was used for oligopeptide separation after proteolytic digestion. The gradient was 0 -100% solvent B over 60 min. The mass spectra were obtained on a Finnigan LCQ ion trap mass spectrometer (ThermoQuest, San Jose, CA) equipped with an electrospray ionization source. The MS spray voltage was 4.25 kV. The heated capillary was maintained at temperature of 200°C. The spectrum was scanned from 400 to 2000 m/z. The ZoomScan, tandem MS (MS/MS) was performed in the data-dependent mode. The MS/MS collision energy value was 42%.

NMR Study
Heteronuclear correlation 1 H-13 C HSQC (13) spectrum was recorded on a Varian Inova 750-MHz spectrometer equipped with a z axis shielded triple resonance probe. The sample of C130 modified by BA-2-13 C-7-ACA was dissolved in 20 mM sodium phosphate buffer (pH 8.0) at a concentration of 40 mg/ml. BA-2-13 C-7-ACA was synthesized as described previously.

C130 Undergoes Stoichiometrically Labeling by BA-7-ACA in a Time-dependent
Manner-Based on the structure of GL-7-ACA, we designed and synthesized a series of substrate analogs by replacing the side chain of GL-7-ACA with different length of bromoacyl moieties (Fig. 1B). Although these compounds are structurally similar to the native substrate, GL-7-ACA, only marginal hydrolytic activities remained after incubating with C130 (Table I). Only BA-7-ACA and BP-7-ACA could inhibit the activity after prolonged incubation with C130. In addition, the inactivation was irreversible, because reactivated C130 could not be recovered even if the inhibitors were removed (data not shown). The other analogs and bromoacetic acid itself did not affect the activity of C130. Because BA-7-ACA is a much more effective inhibitor than BP-7-ACA (see Tables I and III), we focused our attention on the inhibition of C130 caused by BA-7-ACA.
The inhibition of C130 by BA-7-ACA proceeded in a time-dependent manner under physiological conditions. The kinetics of inhibition was pseudo-first-order. Each individual product of the reaction, glutaric acid or 7-ACA, slowed down the rate of inactivation (Fig. 2). A semilogarithmic plot of inhibition rate k versus the concentration of BA-7-ACA yielded a straight line with a slope of 0.91. This implies that BA-7-ACA was functioning as an affinity labeling reagent, resulting in single hit modification of C130. In addition, far UV circular dichroism spectra of the original and modified C130 were almost identical (data not shown), indicating that this modification does not lead to a gross change in protein structure.
To confirm the equimolecular character of the alkylation of C130, a comparison of the native and labeled C130 was made by MALDI-TOF mass spectrometry to determine the stoichiometry of inhibitors binding. Unambiguously, the signals at m/z 17474.75 and m/z 17474.94 (Fig. 3, A and C) corresponding to the ␣-subunits (original and labeled) indicated that there was no labeling of the ␣-subunit of C130. On the other hand, signals corresponding to the ␤-subunits at m/z 58,315.54 ( The T1 Tryptic Fragment of the C130 ␤-Subunit Contains the Alkylation Site-Theoretically, complete digestion of C130 with trypsin should produce 65 peptides. A precise comparison of the LC/MS tryptic fingerprints of native C130 and the alkylated enzyme was made. More than 30 peptides in each of the two digests, covering about 80% of the total amino acid residues, were identified and confirmed by tandem mass spectroscopy (data not shown). Among the peptides identified, we found only one [M ϩ 2] 2ϩ signal at m/z 509 corresponding to SNS-WAVAPGK (T1) from C130 that was not present in the peptide spectrum of the labeled C130. In the peptides from the alkylated species, a new [M ϩ 2] 2ϩ signal at m/z 665 (T1-R) appeared. The T1 tryptic fragment originates from the N terminus of the C130 ␤-subunit. Within this segment, the first serine residue (Ser-B1) is known to be the Ntn catalytic center (4, 7). The additional molecular mass of 312 corresponds to that of the inhibitor residue (-R). Thus, the alkylation of the C130 ␤-subunit by BA-7-ACA is likely to form a stable enzyme-CH 2 CO-7-ACA complex, and the labeling site is likely to lie within the T1 peptide.
Trp-B4 of the ␤-Subunit of C130 Is the Alkylation Site-Further investigation of MS/MS spectra of both T1 and T1-R was carried out to elucidate the sequence of T1 and to identify the chemically modified amino acid residue within T1-R. As shown in Fig. 4, most of the B ions and Y ions (14) from the MS/MS spectrum of T1 were identified and matched well with the proposed sequence of SNSWAVAPGK. However, the pres- ence of the labeled inhibitor residue -R made it more complicated to clarify the daughter ions in the MS/MS spectrum of T1-R (Fig. 5). Under the conditions of the tandem MS ion scan mode, the -R residue may not be stable enough to maintain its structure unchanged. A certain percentage of the -R residues may be further cleaved to produce two possible different subresidues, -R 1 and -R 2 , with molecular weight of 253 and 42, respectively (Fig. 6) 2ϩ . Therefore, the parental ion T1-R may first be cleaved to produce two subparental ions, T1-R 1 and T1-R 2 , and then most of the daughter ions may come from the three parental ions. Although a complete series of Y ions was not observed, probably due to the multiple internal cleavages rather than sequential cleavages, subseries ions, signals at m/z 683, m/z 754, m/z 853, and m/z 924, strongly indicated that the labeling site was on the small peptide fragment NSW. This hypothesis was substantiated by the presence of the B4 -60 ion (SNSW ϩ R 1 ). The signal at m/z 835 corresponding to WAVAPG plus R 1 (782 ϩ 253) allowed an assignment of Trp (Trp-B4) as the target residue within the T1-R peptide fragment.
A similar investigation of the BA-7-ACA-modified P-B8G mutant further supported the above conclusion. It has been reported that the proline imide bond (15) might interfere with the normal collision-induced dissociation in MS. In addition, based on our observations, replacing Pro-B8 with glycine is likely to be a stable mutant maintaining most of the original enzymatic properties. 3 Thus, labeled P-B8G was digested by Lys-endoproteinase, and the first peptide fragment of the ␤-subunit, from Ser-B1 to Lys-B10, containing the -R residue within it, was analyzed by LC/MS, giving a [M ϩ 2] 2ϩ signal at m/z 645. The signals at m/z 439 and m/z 510 in the MS/MS spectrum, corresponding to Trp-R 1 and [Trp-Ala]-R 1 , respectively, directly identified the labeling site of BA-7-ACA.

Alkylation of the Indole Ring of Tryptophan Forms a Carbon-
Carbon Bond-To confirm that a chemical reaction involving the bromoacetyl group of BA-7-ACA and Trp-B4 of C130 had taken place, a 13 C-labeled inhibitor, BA-2-13 C-7-ACA, was synthesized and used in the affinity labeling of C130. The 1 H-13 C COSY spectrum of BA-7-ACA (see "Experimental Procedures") indicated that the 13 C chemical shift of its C-2Ј was assigned at ␦ C 30.1 ppm in PBS, pH 8.0 (or ␦ C 28.5 ppm in CD 3 OD), correlated to the two C-2Ј protons (at ␦ 3.90 and 3.98 ppm, in PBS, pH 8.0). Only one signal of 1 H-13 C correlation resonance was clearly presented in the two-dimensional HSQC spectroscopy of the 13 C-labeled enzyme-inhibitor complex (Fig. 7A). Thus, only one kind of amino acid residue specifically alkylated. Meanwhile, this alkylation was confirmed by the obvious change of the 13 C chemical shift of the 13 C labeled carbon, which was assigned at ␦ C 43.4 ppm. From the known increments of substituted alkanes (16), we can predict the 13 C chemical shift of the 13 C labeled carbon when an amino acid residue is alkylated by BA-2-13 C-7-ACA. For tryptophan, it is believed that C-3 of the indole ring is initially attacked by the bromide and then a C-2 alkylated product would be formed after rearrangement (17,18). Therefore, C-2 of the indole ring is considered to be the only possible position to be alkylated (19) (Fig. 7B). Here, we predict the chemical shift by calculation with the following increments: ␦ C ϭ Ϫ2.3 (CH 4 ) ϩ 22 (-CONH 2 ) ϩ 23 (phenyl) ϭ 42.7 (ppm), where the -CONH 2 and phenyl groups are approximate substitutes for -CO-7-ACA and Trp residue, respectively. This predicted shift is close to the measured shift (␦ C 43.4 ppm) assigned in Fig. 7A. With a similar chemical environment, the C-2Ј of penicillin G (Fig. 7B) gives a measured shift at ␦ C 42.8 ppm (20), very close to the predicted one. Since there are no cysteine residues in C130, other possible alkylated functional groups on the residues are -OH, -COOH, and -NH 2 (or ϭNH), giving predicted shifts at ␦ C of 76.7, 71.7, and 56.2 (60.2) ppm, respectively. These shifts are significantly different from those observed, excluding the possible alkylation of Ser, Tyr, Thr, Asp, Glu, His, Lys, Arg, Asn, and Gln residues, as well as N-1 of the indole ring of Trp. Therefore, C-2 of the indole ring is the 3  most likely position for alkylation with BA-7-ACA; this event would lead to the formation of a carbon-carbon bond between C130 and the inhibitor. This covalent bond was stable enough to generate useful information from MS/MS spectra for the determination of modification of Trp-B4 (Fig. 5).
Analysis of Trp-B4 Mutants Indicates Possible Mechanism of Affinity Labeling-In order to further confirm the specific reaction of BA-7-ACA with Trp-B4 of the C130 ␤-subunit, six site-directed mutants were generated, replacing Trp-B4 with Ala (W-B4A), Leu (W-B4L), Thr (W-B4T), Phe (W-B4F), His (W-B4H), and Tyr (W-B4Y) respectively. Each of the mutant proteins retained significant activity toward its natural substrate, GL-7-ACA, with moderately decreased values of k cat /K m (Table II), indicating that the Trp-B4 of the C130 ␤-subunit is essential for neither substrate binding nor catalytic activity, despite the fact that this residue can be specifically alkylated by BA-7-ACA.
Both BA-7-ACA and BP-7-ACA irreversibly inhibited W-B4Y but not the other mutants in a time-dependent manner. On the other hand, the activities of all of the mutants were readily inhibited by BA-7-ACA (Tables II and III and Fig. 8). It is obvious that the irreversible inhibition of W-B4Y is attributable to the presence of the nucleophilic phenolic group, while the lack of such groups on the mutated amino acid residues makes W-B4A, W-B4L, and W-B4F resistant to modification by these inhibitors. Unlike the alkylation of tryptophan, where an electrophilic addition reaction on C-3 of indole ring is the first reaction step, the modification of other amino acid residues substituted with an active proton is considered to be an S N 2 nucleophilic substitution reaction. For an S N 2 reaction, deprotonation of the substituted residue is essential for anion formation. The pK a values of the relevant residues are as follows: Thr-OH (ϳ18) Ͼ Trp-NH (16) Ͼ His-NH 2 ϩ -NH (7.0, 14.4) Ͼ Tyr-OH (10.2).
The histidine side chain is readily protonated, with a pK a value near 7, but it is difficult to deprotonate with an apparent pK a value of about 14.4. Obviously, nucleophilic substitution at the N-H of the Trp indole ring is also impossible. Therefore, the inability to covalently modify W-B4H and W-B4T can be explained by the much higher pK a values of their active protons.  R (acyl-7-ACA). The residue R may undergo two cleavage paths resulting in detectable m/z ion peaks for the subresidues R 1 and R 2 in the MS/MS spectra (Fig. 5).

FIG. 7. 1 H-13 C NMR HSQC spectrum of C130 labeled by BA-2-13 C-7-ACA (A) and structural comparison of predicted Trp-R and penicillin G (B).
A, only one peak assigned to the isotopic carbon in the enzyme-13 CH 2 CO-7-ACA complex, corresponds well to the structure of Trp-R predicted in Fig. 7B (described under the section "Alkalation of the Indole Ring of Tryptophan Forms a Carbon-Carbon Bond"). B, the 2Ј-C on both compounds has a similar chemical environment indicated by the 13 C chemical shifts measured, validating the predicted structure of Trp-R, which represents the Trp-B4 alkylated by BA-7-ACA.
Considering the characteristics of BA-7-ACA labeling or inhibition of C130 and its missense mutants, it is further confirmed that Trp-B4 in C130 is a site of alkylation for BA-7-ACA.
According to Huang and Colman (21), an affinity labeling event can be expressed as follows, where E represents the free enzyme, I represents the inhibitor, EI* is the reversible enzyme-inhibitor complex, and EI is the covalently modified enzyme. The observed rate constant (k) at a particular concentration of the inhibitor is described as follows, where K i ϭ (k Ϫ1 ϩ k 2 )/k 1 . This represents the concentration of inhibitor giving a half-maximal rate of inactivation, and k max is the maximum rate of modification at saturating concentrations of the inhibitor. Accordingly, the inhibition kinetics of BA-7-ACA and BP-7-ACA on C130 and W-B4Y were analyzed by a double reciprocal plot of the pseudo-first-order rate constant (k) against the concentration of inhibitors ([I]). All four affinity alkylation reactions gave straight lines (Fig. 8, A and B). The respective K i and k max values were determined (Table III). As an affinity labeling reagent, BA-7-ACA is much better than BP-7-ACA for either enzyme, with lower K i values and higher k max values, indicating its higher affinity and reactivity. The relatively poor affinity of BP-7-ACA may well be due to its longer side chain, while the lower k max is certainly due to its poor electrophilicity. With the same inhibitor, the K i values of the two enzymes are quite different, while the k max values are almost the same. This result indicates that the fourth amino acid residue of the ␤-subunit has a stronger influence upon the binding of the inhibitor rather than its reactivity. This argument can be substantiated by the characteristics of the other five mutant enzymes, W-B4A, W-B4L, W-B4T, W-B4F, and W-B4H, which failed to react with BA-7-ACA but were inhibited reversibly and competitively by the inhibitor, with different competitive inhibition constants K i (Table II). Significantly, none of the mutant enzymes with one residue replaced were stable under normal enzyme assay conditions, although the mutant W-B4H was much more stable than the others (Table II). This indicates that the Trp-B4 is essential for the enzymatic stability and may thus play an important role in catalysis.

DISCUSSION
To investigate the substrate binding site of GL-7-ACA acylase C130, we synthesized a series of bromide-substituted substrates with varying length of acyl-7-ACA. Only analogs with shorter acyl-side chains, BA-7-ACA and BP-7-ACA, were able to specifically bind and label the C130 ␤-subunit as affinity reagents. The labeling mechanism of BA-7-ACA was shown to occur via the alkylation of the C-2 on indole ring of Trp-B4. Mutational analysis of the site-directed mutants of the 4th residue of the C130 ␤-subunit further supported this conclusion.
Some of the above findings were unexpected. First, 7␤-4-bromobutyrylamidocephalosporanic acid, 7␤-5-bromopentanoylamidocephalosporanic acid, and 7␤-6-bromohexanoylamidocephalosporanic acid failed to affinity-modify C130 or to function as competitive inhibitors despite the fact that these chemicals, particularly the 7␤-5-bromopentanoyl aminocephalosporanic acid, seem to be better substrate analogs than BA-7-ACA. Second, there is steric hindrance of the 4th residue of the ␤-subunit, as shown in the C130 crystal structure. A significant energy barrier to the observed reaction makes the Trp-B4 labeling site unexpected. In addition, these results must be further considered with respect to their effect on the binding and hydrolysis of the native substrate, GL-7-ACA.
The affinity labeling of Trp-B4 by BA-7-ACA and BP-7-ACA described in this paper are unique cases of the specific alkylation of a tryptophan residue under physiological conditions. The activation energy that must overcome during this reaction TABLE II Kinetic parameters of C130 ␤ subunit mutants and half-life of their enzymatic activities Wild type and mutant enzymes were partially purified (Ͼ50%). The concentration of enzyme was determined on the basis of chromatic scanning of Coomassie Blue-stained SDS-PAGE gels. Enzyme kinetics employed standard procedures (see "Experimental Procedures") with approximately 10 g/ml enzyme and various concentrations of GL-7-ACA. The kinetic constants K m and k cat were determined from Lineweaver-Burk plots. Inhibition kinetics employed the same procedure with the addition of various concentrations of BA-7-ACA (2.0, 4.0, and 8.0 mM), giving competitive patterns. The K i values were determined as competitive inhibition constants by standard methods. The loss of enzyme activity was assumed to proceed via a first-order reaction. Enzymes (1 mg/ml) were incubated in sodium phosphate buffer (pH 8.0) at 37°C. At suitable time intervals, 20 l of the enzyme was withdrawn, and the residual enzymatic activity was immediately determined under standard assay conditions. The t 1/2 values were measured from the plot of ln(residual activity) versus time (min). All of the experiments were repeated three times, and the S.D. values were calculated with n ϭ 3.  is substantial, because the C-Br bonds of these two analogs are much more stable than those of other affinity reagents, such as the oxidant N-bromosuccinimide (22) and the alkylation reagent 2-hydroxy-5-nitrobenzyl bromide (23)(24)(25)(26). It follows that the affinity alkylations of C130 and W-B4Y are more likely to be suicide reaction-catalyzed by the enzyme rather than normal chemical modifications. In addition, the activation energy of catalysis is likely to be supplied by a dramatic conformational change associated with binding (see below). With the length of the side chain of the inhibitor increased by one carbon, the K i of BP-7-ACA was greatly increased compared with that of BA-7-ACA (Table III). The other substrate analogs containing longer side chains (4 -6 carbons) could neither inhibit nor modify C130 or its mutant W-B4Y. Because BA-7-ACA has the shortest side chain of all the analogs tested, BA-7-ACA is likely to be the analog of the core backbone of the substrate, the -CO-7-ACA moiety rather than the complete substrate (GL-7-ACA) or the side chain of the substrate (the glutaryl moiety).
The affinity labeling of BA-7-ACA on C130 was protected by glutaric acid and 7-ACA, which implies the existence of overlapping binding sites. In addition, the competitive inhibitory effects of BA-7-ACA on mutant enzymes having amino acid switches at position 4 and the decrease in their K m values toward the natural substrate further support the contention that Trp-B4 is likely to be involved in substrate binding, although it is not essential. In fact, the involvement of a Trp residue in substrate binding for GL-7-ACA acylase was first indicated by the oxidation of a Trp residue of CA GK16 by N-bromosuccinimide, causing a moderate increase in its K m (22). Although the position of this Trp residue remains to be determined (22), we suggest that the modified tryptophan is Trp-B4, because this CA is almost identical in sequence with C130 (27). It is noteworthy that Trp-B4 is conserved in all of the known CAs and the majority of the PAGs.
Recently, Kim and Hol (9) solved the structures of binary complexes of CAD with GL-7-ACA and glutarate at 2.5-and 2.6-Å resolutions, respectively. Models of enzyme-ligand complex were inferred from these studies. Because these complex structures were determined by co-crystallizing ligands with the enzyme, no significant conformational changes were observed between the structure of apoenzyme and the complexes. In addition, based on the enzyme-substrate complex model, the distance between the O␥ of Ser-B1 and the target carbonyl carbon of the substrate glutaryl side chain is 3.4 Å, which is too great to form a transition state tetrad. Therefore, these models may only reflect the initial stage of substrate binding but not the details of enzyme-substrate interaction during catalysis.
As expected, Trp-B4 was not involved in an obvious manner in either of the two ligand binding models. Because Trp-B4 is located in the first ␤ sheet of the ␣␤␤␣ sandwich structure with its indole ring facing the first ␣ helix layer (Fig. 9), there is no space to accommodate ligand (Fig. 10) without a conformational alteration that would enlarge the space between the first ␣␤ structure. Therefore, we propose that the binding of BA-7- The pseudo-first-order reaction rate constant k was plotted against different concentrations of the inhibitors in double reciprocal plots. The enzymes (0.25 mg/ml) were incubated in sodium phosphate buffer (pH 8.9) at 37°C with various concentrations of BA-7-ACA (0.5, 1.0, 1.5, 2.0, and 2.5 mM) or BP-7-ACA (4.0, 2.0, 6.7, 10, and 20 mM). The inhibition rate k was calculated as described in Fig. 2. Parameters (listed in Table III) were calculated according to the method described (21).