D-Serine dehydratase from Escherichia coli. IV. Comparative sequences of pyridoxylpeptides derived from the active site and from an inhibitory site of the enzyme.

Abstract d-Serine dehydratase from Escherichia coli has two distinct binding sites for pyridoxal 5'-phosphate; a high affinity site (Kdiss 0.03 µm) which binds the catalytically essential coenzyme and a low affinity site (Kdiss near 1000 µm) which binds pyridoxal-P, pyridoxal, or 5-deoxypyridoxal approximately equally, with resultant inhibition of enzymatic activity. Following borohydride reduction, carboxymethylation and chymotryptic digestion of the holoenzyme, a single non-apeptide containing the N6-pyridoxyllysine residue, and thus corresponding to a portion of the high affinity coenzyme binding site was isolated by two different chromatographic procedures and shown to have the sequence: Ser-(Pxy)Lys-Gly-Arg-Ile-Asn-Lys-Ala-Thr. The peptide differs from the corresponding coenzyme-binding site peptides isolated from other pyridoxal-P-dependent enzymes in its highly basic and relatively hydrophilic character. When borohydride reduction of d-serine holodehydratase was carried out in the presence of an inhibitory concentration (7.5 mm) of pyridoxal-P, one additional major pyridoxylpeptide, corresponding to a portion of the low affinity pyridoxal-binding site was isolated. The sequence of this inhibitory site peptide was determined to be: Val-(Pxy)Lys-Ala-Gly-Ala-Phe.

in the immediate vicinity of the cofactor.
As part of such a study, the isolation and determination of the sequence of a nonapeptide containing the lysine residue which interacts with pyridoxal-P in n-serine dehydratase are described here.
At concentrations far higher than those required for holoenzyme formation pyridoxal-P inhibits n-serine dehydratase, indicating the presence of one or more binding sites with low affinity for pyridoxal-P (3). Reinvestigation of this inhibition shows that it results from binding of pyridoxal-P mostly at one site on the protein.
The sequence around this pyridoxal-P-binding inhibitory site is also reported here.
"Ultrapure" guanidine hydrochloride was from Mann Research Laboratories. All other chemicals were reagent grade and were used wit,hout further purification.

Methods
Preparation of Borohydride-reduced and Carboxymethylated D-Serine Dehydratase-For study of the high affinity binding site for pyridoxal-P, 300 mg of crystalline native n-serinc holodehydratase (specific activity 310) were dissolved in 15 ml of 0.1 M potassium phosphate buffer, pH 7.2, containing 1.0 mM dithiothreitol and 5 m pyridoxal-P. Then, 0.9 ml of a 3% solution of NaRH4 in distilled water was added dropwise over a period of 10 min, while the enzyme solution was gently stirred then allowed to stand at 0" with stirring for 30 min. The reduced holoenzyme preparation was dialyzed 24 hours against two 5liter changes of distilled water in the dark at 4".

7359
For study of the low affinity pyridoxal-P binding site, 300 mg of the pure dehydratase were dissolved in 150 ml of 0.1 M potassium phosphate-7.5 mM pyridoxal-P buffer, pH 7.8. After 30 min at 25" the solution was brought to 0" and 33 ml of a 3% solution of NaBHa in distilled water was added slowly, with gentle stirring, while the pH was controlled at 7.8 by the slow addition of 6 N HCl and the temperature was kept at 0". The solution was then allowed to stand at 0" with stirring for 30 min, then concentrated to about 30 ml by rotary evaporation under vacuum and dialyzed for 36 hours against three lo-liter changes of distilled water in the dark at 4".
These reduced and dialyzed samples were carboxymethylated at pH 8.5 by the procedure of Crestfield et al. (6), then dialyzed in the dark at 4" against several changes of distilled water for a total of about 40 hours.
The digestions were carried out with gentle stirring at 25" in the dark for 15 hours with the further addition of 3 mg of the chymotrypsin preparation after 3 hours. The digests were then adjusted to pH 3 with 2 N HCl and lyophilized.
Amino Acid Analysis-Samples were deaerated, hydrolyzed in sealed tubes with 6 PIT HCl for 20 to 50 hours at 110", then analyzed for amino acids by the method of Spackman et al. (7) with an automatic amino acid analyzer (Beckman model 120C). Ne-Pyridoxyllysine does not separate from histidine under the standard conditions used for amino acid analysis, but was completely separated from other amino acids by replacing the sodium citrate buffer (pH 5.28) usually used for the short column of the amino acid analyzer by the same buffer adjusted to pH 4.55 with concentrated HCI (1). N6-Pyridoxyllysine was eluted from the column between 100 and 110 min at a flow rate of 68 ml per hour. Histidine and ammonia, which were not well separated under these conditions, were eluted from the column between 65 and 75 min.
Seguentiat Edman L)egradation of Pep&-The phenylisothiocyanate procedure in its three-stage form was used (8,9). The liberated phenylthiohydantoin derivatives were identified by comparison with authentic standards after chromatography on Eastman chromatogram sheets No. 6060 (silica gel with fluorescent indicator) in Solvent Systems II and III of Brenner et OZ. (10). Phenylthiohydantoinamino acids were located by their fluorescence under ultraviolet light. The phenylthiohydantoin derived from arginine is soluble in water and was identified by use of the Sakaguchi reagent (11) following paper electrophoresis at pH 6.5 and 2000 volts for 1 hour with authentic standards.
Subtractive analysis for amino acids was also carried out at each degradation step (12).
Enzymatic Digestions of Peptides-Approximately 0.1 pmole of pyridoxylpeptide was treated at 37" with either 10 pg of leucine aminopeptidase (diisopropyl fluorophosphate-treated) in 0.2 ml of 0.05 M Tris-HC1 buffer (pH 8.5) containing 5 mM MgClt, or with 50 lg of diisopropyl fluorophosphate-treated carboxypeptidase A in 0.2 ml of 0.05 AI Tris-HCl buffer (pH 7.5). The latter enzyme (1 mg per ml) was dissolved in 10% (w/v) LiCl before use (13). These digestions were terminated by adding acetic acid until the pH was below 3; the mixtures were then lyophilized.
Amino acids released were determined quantitatively on the amino acid analyzer.
Control experiments were identical except that the peptide was omitted.

Isolation and Composition of High Afinity Coenzyme-binding
Site Peptide-All of the pyridoxal-P present in n-serine holodehydratase appears in chymotryptic digests of the reduced carboxymethylated enzyme as a single fluorescent zone, well separated from other peptides, following chromatography of the digest over Dowex 1 and peptide mapping of the combined fluorescent fractions (Fig. 1). The fluorescent peptide was first isolated on a preparative scale by the same procedure described in Fig. 1 by applying the peptide fraction (equivalent to 5 kmoles of pyridoxylpeptide as measured by absorbance at 325 nm) as a band on Whatman No 3MM paper for electrophoresis, then eluting the material in the fluorescent band, lyophilizing, and reapplying it as a band on a second paper for the chromatographic separation.
Elution from the paper in each case was with 30% acetic acid; the yield of purified peptide was about 16% of that present in the chymotryptic digest. A simpler procedure, applicable for preparation of larger amounts of the pyridoxylpeptide, was that developed by Fischer et al. (2,14), in which the P-pyridoxylpeptide fraction from a given column is modified by removing the phosphate residue by phosphatase digestion, and is then rechromatographed over the same column to separate contaminating peptides from bhe 0" -I

Chromatography -
FIG. 1. Peptide map of the fluorescent fractions isolated from a chymotryptie digest of reduced and carboxymethylated n-serine holodehydratase by chromatography over Dowex 1. The lyophilized chymot,ryptic digest from 300 mg of reduced and carboxymethylated n-serine holodehydratase was dissolved in 5 ml of 2% pyridime and 0.01% acetic acid (v/v), pH 7.3, centrifuged to remove traces of insoluble material, and applied to a column (I X 115 cm) of Dowex AGl-X2 (200 to 400 mesh, acetate form) previously equilibrated with the same buffer. The column was eluted with this same buffer at 25" and a flow rate of 20 ml per hour. Fractions (5 ml) were collected and monitored for ninhydrin color values at 570 nm (26) and for the covalently bound pyridoxyl group by their absorbance at 325 nm. The pyridoxylpeptides were not retained by this column and appeared in the first major peptide fraction to be eluted. A sample of this fraction (containing about 0.04 pmole of the pyridoxylpeptide) was chromatographed in the first dimension (solvent butanol-I-pyridine-glacial acetic acid-water, 90:60:18:72) followed by electrophoresis in the second. Hatched zones are fluorescent when viewed in ultraviolet light prior to ninhydrin treatment. newiy formed pyridosylpeptide.
This procedure, described in Fig. 2, gives pure peptide (as indicated by peptide mapping and analysis) in 409& yield in only two st,eps, and is clearly the isolation method of choice.
The amino acid composition of these independently prepared samples of the pyridoxylpeptide was identical within experimental error; a representative analysis is shown in Column 2 of Table I and corresponds to that of a nonapeptide.
The presence of N'j-pyridoxyllysine and the absence of histidine was verified by paper electrophoresis against appropriate standards. Sequence Analysis of High Afinity Coenzyme Binding Site Peptide-Extensive digestion of the fluorescent peptide with leucine aminopeptidase (15 hours, 37") liberated all of its component amino acids except aspartic acid. The amount of serine indicated by amino acid analysis was twice that obtained on acid hydrolysis, indicating the presence of asparagine (which elutes with serine under the standard conditions of analysis) instead of aspartic acid in the peptide.
Digestion of the peptide with leucine aminopeptidase for a limited time (  ably -(Lys,Ala)-Thr. Table I shows the results of amino acid analysis of the peptide remaining after each of five sequential Edman degradations. Subtractive analysis following further degradation was ambiguous, but the results clearly indicate the sequence of the first 5 residues to be Ser-(Pxy)Lys-Gly-Arg-Ile-.
To determine the sequence of the last four amino acids, the fluorescent peptide was subjected to trypsin digestion.
On paper electrophoresis at pH 3.6 and 2000 volts for 60 min, the tryptic digest showed three ninhydrin-reactive spots, only one of which exhibited fluorescence (Fig. 3). The peptides from the main  -, indicates results obtained by Edman degradation; ---+, those obtained by treatment with leucine aminopeptidase; and c ---, those obtained by carboxypeptidase A digestion.
T-l, T-2, and T-3 are peptides obtained on tryptic digestion of the high affinity pyridoxal-P-binding site peptide (cj. Fig. 3).
portion of the digest were then purified by preparative paper electrophoresis and analyzed after acid hydrolysis.
Peptide T-3 contained alanine, 1.03 (1) and threonine, 0.97 (I) ; and indicates that the COOH-terminal sequence is -Ala-Thr in terms of the results of carboxypeptidase A digestion.
Peptide T-2 released the phenylthiohydantoin derivatives of isoleucine and asparagine, respectively, in the first and second Edman degradation cycle; subtractive analysis following the first cycle (lysine, 1.05 (1) ; aspartic acid, 0.96 (1); isoleucine, 0.13) also showed that isoleucine had been removed, and only free lysine was present in the unhydrolyzed solution after the second degradative cycle. Peptide T-3 released phenylthiohydantoin-alanine after one reaction cycle and only free threonine was present in the unhydrolyzed solution.
These sequence studies are summarized in Fig. 4; there appear to be no ambiguities in assigning the sequence shown for the high affinity coenzyme-binding site of D-serine dehydratase. Inhibition of o-Serine Holodehydratase by High Concentrations of Pyridoxal Analogues-The inhibition data previously reported (3) were obtained by first incubating the enzyme with pyridoxal-P and then diluting 50-fold for immediate assay of enzymatic activity.
Since inhibition1 was slowly reversible on dilution, this procedure underestimated its actual extent.
To avoid the high blanks produced by high concentrations of pyridoxal analogues and thus permit assay of serine dehydratase at the same concentration of pyridoxal analogues used for its inactivation, the assay procedure was modified as described in Fig. 5, which shows about 90% inhibition of D-serine holodehydratase by 9 mM pyridoxal-P.
Pyridoxal and 5-deoxypyridoxal inhibit almost as effectively as pyridoxal-P; thus the 5'-phosphate residue does   (Table II), corresponding to fixation of the coenzyme to the high affinity binding site. Reduction in the presence of sufficient pyridoxal-1 to inhibit enzymatic activity by 90%, yields slightly over 2 moles of W-pyridoxyllysine per mole of enzyme (Table II) with a corresponding decrease in the amount of lysine.
Inhibition of the holoenzyme by pyridoxal-P is thus accompanied by the binding of 1 eq of pyridoxal-P per mole of enzyme as a Schiff's base to a lysine residue at one or more sites distinct from that occupied by the catalytically essential pyridoxal-P residue.
Isolation and Sequence Analysis of Low Afinity Inhibitory Site Peptide-Chymotryptic digests of preparations of r)-scrine dehydratase that had been reduced with borohydride in the presence of inhibitory concentrations of pyridosal-P then carbox.ymethylated, contained two pyridoxylpeptides in large amounts (Fig. 6). The second of these (Fractions 136 to 150, Fig. 6A)  Fig. 2A) and is identical with that peptide. The new pyridoxylpeptides in Fractions 10 to 35 arise as the result of the presence of inhibitory concentrations of pyridoxal-P during borohydride reduction, and will be referred to as the inhibitory site peptides. Fractions 10 to 22 were collected, lyophilized, and applied to a Dowex AG 1 column (Fig. 6B). One major and two minor peptides absorbing at 325 nm appeared.
It is not known whether the three pyridoxylpeptides result from combination of pyridoxal-P at independent low affinity sites, or whether they represent different stages in the chymotryptic digestion of a single sequence carrying the pyridoxyl group.
The major inhibitory site peptide, after digestion with phosphatase and reapplication to the same column ( Fig.   6C) was pure as indicated by peptide mapping; approximately 1 pmole of the peptide was isolated from 6 pmoles of n-serine dehydratase.
Its analysis (Column 2, Table III) corresponds to a hexapeptide.
Valine was identified as the NHe-terminal amino acid by thin layer chromatography of the phenylthiohydantoin derivative and also by subtractive analysis after Edman degradation (Table   III). Short term (10 min, 37") carboxypeptidase A digestion liberated phenylalanine (0.75 residue), alanine (0.5 residue), and glycine (0.08 residue) indicating a probable COOH-terminal sequence, -Ala-Phe.
From these results and those of subtractive Edman degradation (Table III) there appears no ambiguity in assigning the sequence Val-(Pxy)Lys-Ala-Gly-Ala-Phe to the principal inhibitory site peptide (Fig. 4B).

DISCUSSION
The structure of pyridoxylpeptides from the coenzyme-binding sites of various pyridoxal-P proteins is compared in Table IV. The active site peptide from n-serine dehydratase is easily the most polar and basic of the group, basic amino acids comprising one-third of its total residues, a frequency over 3 times that ex- petted from theaverageamino acid composition of theenzyme (1). It is also unique among the pyridoxylpeptides from E. coli enzymes so far studied (but resembles the mammalian transaminases and phosphorylase) in that the amino acid residue on the NHz-terminal side of the pyridoxyllysine residue is not basic. In particular, it differs markedly from the corresponding sequence in tryptophanase, which catalyzes cx ,&elimination reactions of several L-amino acids (including L-serine (16)) formallysimilar to the reaction catalyzed by D-S&De dehydratase (17). All of the enzymes of Table IV except phosphorylase catalyze reactions which, in either the forward or reverse sense, require labilization of a hydrogen atom on the carbon atom corresponding to the a-carbon atom of the substrate amino acid. If these active site peptides contain any common catalytic residues which assist in this act, the only likely candidate is the lysine residue which in these peptides carries the pyridoxyl group, since this is the only residue they share in common.
In the parent enzymes this lysine residue is present as an azomethine of pyridoxal-P, but according to current concepts (18) its e-amino group is freed when the enzyme-substrate complex is formed, and if it remains in close proximity to the amino acid (cf. 19,20) would be free to act as a proton acceptor or donor. In all of the decarboxylases, as well as in tryptophanase and tryptophan synthetase, a histidine or a lysine b The values listed for (Pxy)Lys are the sum of t.hose obtained for NG-pyridoxyllysine and for free lysine (15). Approximately 10 to 20% destruction of (PxyjLys occurs during hydrolysis (see text).
c In the last degradation cycle, the residue peptide was put on the amino acid analyzer column directly without acid hydrolysis. 7363 residue immediately precedes the pyridoxyllysine residue of these peptides, and it has been suggested that this residue could play a catalytic role (14,21). This possibility is not, of course, eliminated by the present results.
In space-filling models, however, these residues are so placed as to readily undergo ion pair formation with the phosphate group of pyridoxal-P during formation of holoenzyme from apoenzyme; they may thus contribute primarily to coenzyme binding, a possibility also recognized previously (2,21). In the apoenzyme, and also in the holoenzymesubstrate complex, such closely adjacent basic groupings may also aid in lowering the pK value of the e-amino group of the lysine residue that undergoes azomethine formation with pyridoxal-P.
Such a lowered pK value by increasing the concentration of the nucleophilic amino group would contribute both to apoenzyme-coenzyme interaction at low pH, and to the putative catalytic function of this group near the pH optimum of these enzymes.
Such speculations at present are useful primarily in providing concepts for further experimental testing, since it is not yet certain that any of the catalytic residues of the enzymes are present in these peptides.
n-Serine dehydratase contains a total of 18 lysine residues per peptide chain ((3) and Table II) The uniqueness of the lysine residue at the high affinity binding site for pyridoxal-P (Kdiss = 0.038 PM (3)) is emphasized by the finding (Fig. 5) that concentrations of pyridoxal-P some 25,000 times higher are required to inhibit enzymatic activity by 50%; this inhibition correlates with reaction of only 1 additional lysyl residue, that at the inhibitory site, with pyridoxal-P (Table II). The spatial relationship of the lysine residue of the inhibitory site to that of the ac-  tive site is unknown, as is the reason for inhibition. The peptide fragments from the two sites (Fig. 4) are markedly different, the active site peptide being highly basic and hydrophilic, the inhibitory site peptide being hydrophobic and containing only a single basic amino acid residue.