The Contribution of Adjacent Subunits to the Active Sites ofd-3-Phosphoglycerate Dehydrogenase*

d-3-Phosphoglycerate dehydrogenase (PGDH) from Escherichia coli is allosterically inhibited by l-serine, the end product of its metabolic pathway. Previous results have shown that inhibition by serine has a large effect on V max and only a small or negligible effect on K m . PGDH is thus classified as a V-type allosteric enzyme. In this study, the active site of PGDH has been studied by site-directed mutagenesis to assess the role of certain residues in substrate binding and catalysis. These consist of a group of cationic residues (Arg-240, Arg-60, Arg-62, Lys-39, and Lys-141′) that potentially form an electrostatic environment for the binding of the negatively charged substrate, as well as the only tryptophan residue found in PGDH and which fits into a hydrophobic pocket immediately adjacent to the active site histidine residue. Interestingly, Trp-139′ and Lys-141′ are part of the polypeptide chain of the subunit that is adjacent to the active site. The results of mutating these residues show that Arg-240, Arg-60, Arg-62, and Lys-141′ play distinct roles in the binding of the substrate to the active site. Mutants of Trp-139′ show that this residue may play a role in stabilizing the catalytic center of the enzyme. Furthermore, these mutants appear to have a significant effect on the cooperativity of serine inhibition and suggest a possible role for Trp-139′ in the cooperative interactions between subunits.

D-3-Phosphoglycerate dehydrogenase (PGDH) 1 (EC 1.1.1.95) from Escherichia coli is a homotetrameric enzyme that is inhibited in an allosteric manner by L-serine (1,2). The crystal structure (Ref. 3, Protein Data Bank, Brookhaven National Laboratory, code 1PSD) shows that each subunit is composed of three domains which are referred to as the substrate binding domain, the nucleotide binding domain, and the regulatory or serine binding domain. Each subunit interacts noncovalently with two adjacent subunits through contacts at their respective regulatory domains and their respective nucleotide binding domains (Fig. 1). L-Serine binds in the two interfaces formed between each pair of regulatory domains while the catalytically active site is in a cleft between the substrate binding domain and the nucleotide binding domain of each subunit. Serine binding induces a conformational change at the regulatory domain interfaces (4,5) that is subsequently transferred to the active sites to produce inhibition of catalytic activity. Moreover, the inhibition of catalytic activity by serine displays sigmoidal kinetics, indicating a cooperative effect between subunits. The nature of this cooperativity has not yet been established, but serine binding data (4) suggest that binding of serine to only two sites, of a total of four serine binding sites, is all that is necessary to inhibit the enzyme by more than ninety percent.
Inspection of the active site of PGDH from the crystal structure (3) of the inhibited enzyme (with L-serine bound) reveals several unique features that may be related to substrate binding and catalysis and perhaps the allosteric interaction between subunits. First, the active site of PGDH contains five positively charged residues whose side chains protrude into the solvent accessible space of the active site cleft. These are Lys-39, Arg-60, Arg-62, Arg-240, and Lys-141Ј. Because the substrate, phosphoglyceric acid, is a very negatively charged molecule, these positively charged residues may play a role in substrate binding. Interestingly, one of these basic residues, Lys-141Ј, is contributed by the adjacent subunit (3). That is, by the subunit adjacent to the subunit that contains the active site in question (Fig. 2). The charged groups of the side chains of both Lys-141Ј and Arg-60 are very near the binding site for the negatively charged heavy metal compounds that were used for multiple isomorphous replacement phasing in the crystallographic studies. This is thought to be the most probable binding site for the phosphate group of the substrate.
Second, the Glu-His pair, which makes up the so-called charge relay system common to many dehydrogenases and similar to that found in serine proteases as a His-Asp pair, rests at the top of a hydrophobic pocket into which is inserted Trp-139Ј of the adjacent subunit (Fig. 3). The suggestion has been made that residues at the active site that come from the subunit adjacent to that which forms that particular active site cleft may play a role in the cooperative nature of the inhibition kinetics. Preliminary evidence suggests that cooperativity is observed for both serine binding and for inhibition kinetics but that they are not identical. This manuscript explores the active site of PGDH by site-directed mutagenesis and presents data that may have implications for subunit cooperativity in inhibition of catalytic activity.

MATERIALS AND METHODS
PGDH was expressed and isolated as described previously (6,7). Activity was determined at a constant temperature in 20 mM Tris buffer at pH 7.5 using either ␣-ketoglutarate (8) or hydroxypyruvic acid phosphate (1, 2) as the substrate and by monitoring the decrease in absorbance of NADH at 340 nm (9). Protein concentration was determined by the Bradford method as described previously (4,10). All mutations are constructed in PGDH 4C/A which is a form of the enzyme where the four native cysteine residues in each subunit have been converted to alanine. This construct has been described previously (4) and is used here for consistency of comparison to past studies. Kinetically, native PGDH and PGDH 4C/A are very similar (4).
Mutagenesis was performed by PCR (11). All PCR reagents were * This work was supported by Grant GM 56676 from the National Institutes of Health. 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.
‡ To whom correspondence should be addressed: Dept. of Molecular Biology and Pharmacology, Box 8103, Washington University School of Medicine, 660 S. Euclid Ave., St. Louis, MO 63110. Fax: 314-362-4698; E-mail: ggrant@pharmsun.wustl.edu. 1 The abbreviations used are: PGDH, phosphoglycerate dehydrogenase; PCR, polymerase chain reaction. obtained from Perkin-Elmer, and PCR products were purified with a QIAquick PCR Purification kit (Qiagen Inc.) Restriction fragments were isolated from agarose gels with a QIAquick Gel Extraction Kit (Qiagen Inc.). Plasmids were isolated with a QIAprep Spin Miniprep Kit (Qiagen Inc.), and all mutations were confirmed by sequencing on an Applied Biosytems Model 373 automated DNA sequencer using Big Dye terminator chemistry. PCR products were placed into plasmids by way of flanking restriction sites, and the entire length of the restricted PCR insert was verified by sequence analysis.
Kinetic parameters were determined from direct linear plots. Coefficients of cooperativity for serine inhibition were determined by fitting the inhibition data to the Hill equation with the Kaleidograph (Synergy Software) curve-fitting program. Figs. 4 and 5 were also produced with Kaleidograph. Figs. 2 and 3 were produced using MOLSCRIPT (12).

RESULTS AND DISCUSSION
Description of the Active Site-The reaction catalyzed by PGDH involves hydride transfer and proton extraction at the C2 position of the substrate, interconverting phosphohydroxypyruvate and phosphoglycerate. This is mediated by NADH and His-292 (Fig. 2). In addition, Glu-269 acts in tandem with His-292 to form a proton shuttle as seen in many dehydrogenases and in serine proteases as a His-Asp pair (Fig. 2). Also common to dehydrogenases such as malate and lactate dehydrogenase, which have similar substrates and which share this type of mechanism, is a basic residue that serves to anchor the C1 carboxyl group. Arg-240 in PGDH appears to correspond to this residue based on its position relative to His-292 and NADH (Fig. 2). The additional basic residues, Arg-60, Arg-62, Lys-39, and Lys-141Ј, are unique to PGDH and may interact with the acidic group at the distal end of the substrate. The natural substrate for PGDH is phosphohydroxypyruvate, but ␣-ketoglutarate has also been found to be an effective substrate (8). Thus, this group would be a phosphate for phosphohydroxypyruvate and a carboxylate for ␣-ketoglutarate. The distance along the carbon chain between the two acidic groups in each substrate is nearly identical.
Arg-60 and Lys-141Ј are in closest proximity to His-292 with their protonatable groups in approximately the same plane as those of His-292 and Arg-240 (Fig. 2). In addition, the distance between Arg-240 and the cationic groups of Lys-141Ј and Arg-60 is approximately the same as the distance between the two anionic groups of the substrates. Arg-62 and Lys-39 are further removed, with Lys-39 being the furthest from Arg-240 and nearest the opening of the active site cleft. Thus, it appears from the structure that Arg-240, Arg-60, Lys-141Ј, and possibly Arg-62 are the most likely residues to interact with the substrate when it is bound at the active site. One must keep in mind, however, that the available structure is from the inhibited enzyme with bound L-serine. Although it is known that the substrate can bind to the inhibited enzyme, it is not known how serine inhibition affects the relative position of residues in the active site cleft.
Interestingly, Lys-141Ј is contributed by the adjacent subunit whose nucleotide binding domain overlaps with the active site cleft of its partner (Fig. 1). In addition to Lys-141Ј, Trp-139Ј also projects into the active site cleft of its adjacent subunit. In this case, however, Trp-139Ј is not solvent-exposed but rather fits into a hydrophobic pocket formed by residues of the adjacent subunit like a tab into a slot (Fig. 3). This pocket is made The diagram in panel B shows them as they would be seen by looking into the active site cleft from the side. This is relatively the same orientation of the cleft as shown in the diagram on the right of Fig. 1. The diagram in panel A is rotated approximately 90 degrees to the left from that of the diagram in panel B. The NADH cofactor is not shown for clarity, but its nicotinamide ring would be above the plane of the paper and in front of the side chain of Arg-240 in the diagram in panel B. Lys-141Ј is indicated with a prime to designate that it is contributed by the adjacent polypeptide chain. up largely of proline and phenylalanine residues that surround Trp-139Ј. Curiously, the His-Glu pair, which includes the active site histidine, is immediately adjacent to the aromatic ring of Trp-139Ј and appears to form the top of the pocket. This intimate association of active site catalytic residues with a hydrophobic residue from the adjacent subunit is curious, and its unique positioning suggests that it could play a role in stabilizing the structure of the active site.
Mutations of the Active Site Cationic Residues-The five cationic residues found in the active site cleft of PGDH likely make up a pre-organized electrostatic framework for the binding of the doubly negative-charged substrate. However, the spatial arrangement of these residues suggested that they may not all play a significant role in substrate binding. Thus, each residue was mutated in turn to an alanine side chain to assess the relative role of their cationic groups. The kinetic paramefilling model showing Trp-139Ј in white inserted into the hydrophobic pocket formed by residues of the adjacent polypeptide chain shown in shades of gray. The light gray residues are the His-292,Glu-269 catalytic pair, and the dark gray residues are proline side chains. B, ball and stick diagram in the same orientation as shown in the space filling model. All of the residues except Trp-139Ј are from the same polypeptide chain. Trp-139Ј is indicated with a prime and is colored white to designate that it is contributed by the adjacent polypeptide chain. C, the same as the diagram in panel B except that it is rotated 90 degrees to the left.  Table I. The coefficient of cooperativity is shown to the right for each curve.
ters determined for the mutant enzymes with either ␣-ketoglutarate or phosphohydroxypyruvate as substrate are presented in Table I. Data was determined for both substrates to assess if the additional oxygen function found on the phosphate group of phosphohydroxypyruvate demonstrated a differential interaction with one or more of the cationic residues.
These data indicate that the major effect of the phosphate group of phosphohydroxypyruvate compared with the carboxyl group of ␣-ketoglutarate appears to be in the K m , which shows a 15-fold difference for PGDH 4C/A , whereas the k cat appears largely unaffected. This relationship generally holds for the mutants as well. However, the difference in K m is smaller with some of the mutants, particularly for R240A, where only an approximately 2-fold difference is seen. This might initially suggest that the side chain of Arg-240 is involved in interacting with the phosphate but not the carboxylate. However, the enzyme crystal structure shows that Arg-240 is positioned very close to His-292 and the nicotinamide ring of NADH. This would indicate that Arg-240 interacts with the anionic group closest to the catalytic center of the substrate which is the carboxyl at C-1 in both substrates. In fact, the structure suggests that this is the only residue available for binding the C-1 end of the substrate to the active site. Thus, a more probable interpretation is that Arg-240 is critical for binding of either substrate, and its absence essentially overrides any incremental binding effect that may be contributed differentially by other cationic side chains binding with the phosphate or the carboxylate.
Mutation of Lys-39, which is the furthest removed from the catalytic center, has little effect on the K m and only a 2-fold effect on the k cat . Lys-39 would thus seem not to play a significant role in the interaction with the substrate or in the catalysis. Excepting Arg-240 for the reason discussed above, the most pronounced effect in k cat is seen for R60A and K141ЈA, whereas the most pronounced effect on K m is seen for R62A. In addition, the difference in K m for R62A and K141ЈA between the two substrates is about half of what it is for R60A. Taking the structure of the substrate and the geometry of these residues in the active site together with these data, they suggest that Arg-60, Arg-62, and Lys 141Ј all play a role in binding the end of the substrate furthest from the active center and that Arg-62 and Lys-141Ј may interact a little more strongly with the phosphate group of phosphohydroxypyruvate than does Arg-60. Interestingly, the most pronounced effect overall, again excluding Arg-240, is seen for Lys-141Ј, whose k cat /K m ratio is 2 orders of magnitude less than PGDH 4C/A and at least an order of magnitude less than R60A and R62A. This represents a major contribution to the active site from a residue that comes from the adjacent subunit.
The effect of the cationic residue mutations on the ability of serine to inhibit the enzyme is shown in Table I, which lists the concentration of serine at which 50% inhibition is achieved and the coefficient of cooperativity for inhibition. The coefficient of cooperativity for inhibition of native PGDH is approximately 1.7. This suggests that at least two of the subunits interact cooperatively in the inhibition. This is consistent with the preliminary observation that serine binding to only two of the four sites is required for inhibition. Furthermore, the symmetry of the tetramer suggests that subunit dimers may be the functional allosteric units. While Lys-141Ј seems to increase the enzyme's sensitivity to serine, it does it to no greater extent than do the other cationic mutants, and its coefficient of cooperativity is also essentially unchanged from the unmutated form. Thus, although Lys-141Ј contributes an electrostatic ligand to the adjacent subunit in PGDH and is critical for activity, it does not seem to be involved in the subunit cooperativity seen in the serine inhibition kinetics because its absence does not reduce the apparent cooperativity.
Mutations at Tryptophan 139Ј-Like Lys-141Ј, Trp-139Ј also appears to interact with the active site of the adjacent subunit. However, in this case it does so by filling a hydrophobic pocket at the base of two catalytically important residues, His-292 and Glu-269. The role of Trp-139Ј was investigated by mutating it to residues with hydrophobic side chains of decreasing length and volume. The kinetic parameters for these mutations are listed in Table I. All mutations tested lower the k cat /K m with the largest effect being on the k cat of the enzyme. Fig. 4 shows the relationship between the k cat /K m of these enzymes and the volume of the residue at position 139Ј. Substitution of tryptophan with phenylalanine is the most conservative and lowers the k cat /K m by approximately an order of magnitude. Substitution with leucine decreases this parameter by more than 2 orders of magnitude, and W139V, W139A, and W139G are over 3 orders of magnitude lower. The decrease in k cat /K m is approximately linear with respect to the volume of the side chain until a threshold is reached for the valine mutant. Smaller side chains produce very little additional effect. This suggests that the bulk of Trp-139Ј is necessary for stabilization of the pocket and thus the relationship of the His-Glu pair to each other and to the active site. As the volume of the side chain decreases, the pocket may become more and more unstable until a point is reached where the maximum effect is seen. Additional reduction in volume past this point no longer produces an additional effect on the activity. Unlike Lys-141Ј, mutation of Trp-139Ј appears to have a definite effect on the cooperativity of serine inhibition. Fig. 5 shows the serine inhibition data determined for Trp-139 mutants fitted to the Hill equation, and the coefficients of cooperativity determined from the fit are listed in Table I. The curve for W139F appears to retain a sigmoidal character similar to that for the unmutated enzyme, but the coefficient of cooperativity is somewhat lower than that for the unmutated enzyme. The inhibition curves for the other mutants all appear to be more hyperbolic in nature, and their coefficients are significantly reduced. It is possible that perhaps the leucine and valine side chains are causing a steric problem because their side chains might not fit as smoothly into the pocket as the aromatic residues. However, this is not the case for glycine, which has no side chain to get in the way. Yet, it produces the lowest degree of cooperativity of any of the mutants while maintaining an IC 50 very close to that of unmutated enzyme.
This data is consistent with the possibility that Trp-139Ј may, at least in part, contribute to the interaction between subunits and the mechanism of inhibition. The relatively larger effect of these mutations on k cat as compared with K m is also consistent with PGDH being a V-type enzyme. These data suggest that only a relatively small movement of Trp-139Ј away from the His-Glu pair would be enough to produce the desired effect and would be consistent with relatively small, subtle changes in conformation caused by serine binding. While this may be a rather simplified view, with the actual mechanism being more complex and involving other induced conformational changes at the active site, these studies provide the first suggestion of a possible molecular basis for the effect on the PGDH active site of effector binding to a remote location.