Isolation, Characterization, and Structure of a Mutant 89 Arg + Cys Bisphosphoglycerate Mutase IMPLICATION OF THE ACTIVE SITE IN THE MUTATION*

The phosphatase assay was performed as described earlier (27) and adapted to utilization of nonlabeled glycerate-2,3-Pz (28). The assay consisted of a 1-ml mixture containing 50 pmol of glycylglycine-KOH buffer, pH 7.5, 2.7 pmol of EDTA, 0.7 pmol of 2-mercaptoethanol, 5 pmol of glycerate-2,3-P2, and 35 pg of bisphosphoglycerate mutase. After incubation for 2 h at 37 "C, the reaction was stopped by deproteinization in 8% trichloroacetic acid solution. Estimation of inorganic phosphate was performed according to the Fiske and Sub-barrow method (29). For studies of the stimulation of the enzyme by glycolate-2-P, and in view of the high stimulatory effect of this compound on phospha- tase activity (27), a 4-fold diluted sample and a 10-min incubation were used. Assay of mutase activity was performed as already outlined (21). Brownlee with same gradient treated with 4-vinylpyridine Sequence Determination of Peptides-Amino acid microsequences of all tryptic peptides were determined using an Applied Biosystems 470-A microsequencer and a standard procedure (35). Mass spectrometry was performed on a HS-ZAB HS mass spec-trometer (VG-analytical, Manchester, United Kingdom) interfaced to a VG-2035 Data system via a PDR-8A computer (36). The FAB spectra were generated by a neutral xenon atom beam of 8 keV. The samples were dissolved in a 1:lO solution of acetic acid in water. One p1 of the dissolved sample was first placed on the sample probe tip; 1 pl of 5:l dithiothreitol/dithioerythritol was then mixed with the sample on the probe tip. Some 50 ng of isolated peptide was used in each run. About 20-40 plans were accumulated and stored using the multichannel mode of acquisition. Calibration was performed by reference to cesium iodide peaks.


IMPLICATION OF THE ACTIVE SITE IN THE MUTATION*
Raymonde Rosa$, Yves Blouquit, Marie-Claude Calvin, Daniele PromeQ, Jean-Claude PromeQ, and Jean Rosa From the Znstitut National de la Sante et de la Recherche Medicale U.91, the Centre National de la Recherche Scientifique UA 607 Hcjpital Henri Mondor, 94010, Criteil, France, and the §Centre National de la Recherche Scientifique, LD 8201,31000 Toulouse, France Bisphosphoglycerate mutase (EC 5.4.2.4.) is a trifunctional enzyme which displays synthase, mutase, and phosphatase activities. The purification, characterization, and structural study of an abnormal form of the enzyme, isolated from a patient which we reported earlier (Rosa, R., Prehu, M. o., Beuzard, Y., and Rosa, J. (1978) J. Clin. Invest. 62,[907][908][909][910][911][912][913][914][915], is described. The abnormal enzyme, present at 50% of the level of the normal enzyme as estimated by immunological methods, showed elevated electrophoretic mobility and hybridized with erythrocyte phosphoglycerate mutase (EC 5.4.2.1.) in the same manner as the normal control. The mutant enzyme was unstable at 55 OC and could be protected against thermal instability by 0.5 m M glycerate 2,3-bisphoshate but not by either glycerate 3-phosphate or glycolate 2-phosphate. Two of the three functions of the mutant enzyme were distinct from those of the normal protein. The specific activity of the synthase was 0.57% of normal and that of the mutase 4.1%. By contrast, the specific phosphatase activity was not affected by the mutation. However, the phosphatase activity of the mutated protein was markedly less stimulated by glycolate-2-phosphate than that of the control. High performance liquid chromatography analysis of tryptic peptides derived from the mutant enzyme showed an abnormal profile with the absence of two peaks normally containing the T12 and T13 peptides and without the appearance of a supplementary peak. Amino acid sequence and mass spectrometric analysis demonstrated the substitution of Arg + Cys residue in position 89 producing an uncleaved T12-Tl3 present in the same peak as the T6. Considered together, our data suggest that Arg-89 is located at or near the active site of bisphosphoglycerate mutase and that this residue is probably involved in the binding of monophosphoglycerates.
Bisphosphoglycerate mutase (EC 5.4.2.4.), also termed di-* This work was supported by grants from the Institut National de la Sante et de la Recherche Medicale, from the Centre National de la Recherche Scientifique, from the Association Francaise contre les Myopathies, and from the Universiti. de Paris XII. 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.
$ phosphoglycerate mutase or bisphosphoglycerate synthase, has been extensively studied. This enzyme is a homodimer (1, 2) whose amino acid and cDNA sequences (3-5) have been reported in addition to the chromosomal assignment of the gene (6). In red cells, bisphosphoglycerate mutase forms a hybrid with phosphoglycerate mutase (EC 5.4.2.1.) (71, a glycolytic enzyme which resembles bisphosphoglycerate mutase closely in its structure and function. The main function of bisphosphoglycerate mutase resides in its synthase activity (Equation 1) since this enzyme synthesizes glycerate-2,3-P~,' a potent inhibitor of the affinity of hemoglobin for oxygen. Bisphosphoglycerate mutase displays two other minor enzymic activities, i.e. that of a phosphatase (Equation 2) and of a mutase (Equation 3).

7837
in partially purified preparations of the enzyme (23). W e presently describe the characterization of the purified enzyme and the identification of the mutation in its primary structure.

EXPERIMENTAL PROCEDURES
Materials-All substrates and commercial enzymes were purchased from Boehringer Mannheim (German Federal Republic (G. F. R.)) except for NADH which was a product of Sigma. Buffer salts, EDTA, polyethylene glycol (M, 6000) and 2-mercaptoethanol were obtained from Merck, Darmstadt G. F. R.): CM-Sephadex C-50 and DEAE-Sephadex A-50 were provided by Pharmacia Fine Chemicals, Uppsala (Sweden). Acrylamide and bisacrylamide were supplied by Fluka (Switzerland), and Cellogel strips by Chemetron Corp., Milan (Italy). Isopropyl alcohol was obtained from Prolabo (Paris, France).
Purified normal red cell bisphosphoglycerate mutase (23) and rabbit antiserum (24) directed against this enzyme were prepared according to methods previously reported. Red cell phosphoglycerate mutase was purified as described earlier (25).
Enzyme Assays-Synthase activity was measured according to a technique which was previously reported (26) and subsequently modified (21). The assay mixture, in a 1-ml volume, contained 50 pmol of Tris-HC1 buffer, pH 8.0, 1 pmol of NAD, 7 pmol of fructose-1,6-Pz, 7 pmol of KHzP04, 2 pmol of glycerate-3-P, 0.18 IU of aldolase, 0.1 IU of triosephosphate isomerase, and 0.32 IU of glyceraldehydephosphate dehydrogenase. After g 5 min of preincubation at 37 "C, 0.05 ml of sample was added. Incubation was subsequently carried out for 10 min. Synthase activity was evaluated by determination of the NADH formed in the course of the glyceraldehydephosphate dehydrogenase reaction.
For studies of the stimulation of the enzyme by glycolate-2-P, and in view of the high stimulatory effect of this compound on phosphatase activity (27), a 4-fold diluted sample and a 10-min incubation were used.
Electrophoretic Procedures-Polyacrylamide gel electrophoresis was performed in the presence of sodium dodecyl sulfate according to Laemmli (30). The gel was stained with Coomassie Blue R-250. Cellogel electrophoresis was performed in a buffer containing 0.075 M Tris/citric acid and 4 mM EDTA, pH 8.0, at 220 V for 2.5 h, according to methods which we have previously described (9, 31, 32).
Measurements of Antibody Consumption-The mutant bisphosphoglycerate mutase was detected and quantitatively estimated by an antibody consumption method as reported earlier (23). A fixed amount of rabbit antiserum was incubated with progressively increasing amounts of inactive bisphosphoglycerate mutase in the presence of 5% polyethylene glycol (Mr 6000) for 30 min at 30 "C and then centrifuged. The excess of antibody in the supernatant was subsequently titrated with known amounts of the enzyme. The residual enzyme activity in the supernatant was inversely correlated with antibody concentration (in the second step), and was therefore proportional to the amount of enzyme complexed by the antibody in the first step of the assay.
Purification Procedure-All purification steps were performed at 4 "C, and all buffers contained 1 mM EDTA and 1 mM dithiothreitol. Bisphosphoglycerate mutase was detected by the immunoconsumption technique (23) as described above. At the end of each chromatographic step, the peak fractions containing bisphosphoglycerate mutase were concentrated in a Diaflo system (Amicon).
The two first steps of purification were previously described for the normal bisphosphoglycerate mutase (23) with the following modification for the CM-Sephadex step: the Tris-HC1 buffer was replaced by a 10 mM phosphate buffer at pH 6.4, which permitted direct application of the sample to the DEAE-Sephadex column without dialysis. The two last steps were performed by high pressure liquid chromatography (HPLC) systems, using hydrophobic interaction chromatography on a column of TSK gel Phenyl 5-PW (Tokyo, Japan), and anion exchange chromatography on a CK 30-S Mitshubishi Gel Column (DEAE-MCI) (Tokyo, Japan).
TSK Gel Chromatography-After concentration from DEAE-Sephadex, the sample (46 mg) was dialyzed against sodium phosphate 0.1 M, pH 7.0, and ammonium sulfate was then added to a final concentration of 1.0 M. The sample was then passed through a TSK gel phenyl 5-PW (21.5 X 150-mm) column equilibrated with buffer A (0.1 M sodium phosphate and 1.0 M ammonium sulfate at pH 7.0). The proteins were eluted from the hydrophobic column by application of a linear gradient of ammonium sulfate from 23 to 0% saturation using buffer B (0.1 M sodium phosphate and 5% isopropyl alcohol at pH 7.0) as the second eluent.
Ion Exchange Chromatography-After dialysis against 20 mM Tris-HCl at pH 8.0, the sample was applied to a DEAE-MCI (CK 30-S) column (7.5 X 75-mm) equilibrated with buffer A. Elution was performed by passage of a linear gradient from 20 mM Tris-HC1, pH 8.0, to 60% buffer B (200 mM Tris-HC1 at pH 7.4, at a flow rate of 1 ml/ min and for a period of 120 min. Isolation of Peptides-Pure protein (100 pg) was digested with chloro-3-(4-tosylamido)-4-phenyl-2 butanone trypsin (Worthington) in 25 mMNH, HCOI, pH 8.1, at a ratio of 4% (w/w) (33). Peptides were isolated by reverse-phase HPLC on a Beckman 343 GM liquid system as reported by Schroeder et al. (33). The separation was carried out on a Macherey-Nagel 5-pm Cla 300-A Nucleosil column (4.6 X 130-mm). Solvent A consisted of 0.1% trifluoroacetic acid in water, and solvent B consisted of 0.1% trifluoroacetic acid in 60% acetonitrile. The gradient was linear from 0 to 100% in 120 min. When necessary, peptides were repurified on a column of RP-300 Brownlee with the same gradient or treated with 4-vinylpyridine (34).
Sequence Determination of Peptides-Amino acid microsequences of all tryptic peptides were determined using an Applied Biosystems 470-A microsequencer and a standard procedure (35).
Mass spectrometry was performed on a HS-ZAB HS mass spectrometer (VG-analytical, Manchester, United Kingdom) interfaced to a VG-2035 Data system via a PDR-8A computer (36). The FAB spectra were generated by a neutral xenon atom beam of 8 keV. The samples were dissolved in a 1:lO solution of acetic acid in water. One p1 of the dissolved sample was first placed on the sample probe tip; 1 pl of 5:l dithiothreitol/dithioerythritol was then mixed with the sample on the probe tip. Some 50 ng of isolated peptide was used in each run. About 20-40 plans were accumulated and stored using the multichannel mode of acquisition. Calibration was performed by reference to cesium iodide peaks.

RESULTS
Complete purification of t h e mutant bisphosphoglycerate mutase could not be obtained with the procedure used for t h e normal enzyme. We therefore modified the last two steps of our purification method. In the last step (HPLC-MCI), the enzyme was eluted in a single asymmetric peak which was collected in two equal fractions (MCI-1 and MCI-2). Both were checked for purity by two electrophoretic methods.
Both fractions MCI-2 ( Fig. 1) and MCI-1 (not shown) presented a single band of the mutant enzyme upon sodium dodecyl sulfate-polyacrylamide gel electrophoresis; this band was situated in the same position as the normal protein, thereby indicating that its apparent molecular weight was unaltered. Upon electrophoresis o n Cellogel, on which proteins migrate according to their charge, fraction MCI-1 (not shown) appeared to be contaminated by an additional band upon staining for mutase activity and was therefore eliminated. Fraction MCI-2 showed a single band (Fig. 2) upon either protein staining (strip I ) or immunoblotting (32) (strip 11) or upon detection of mutase activity (strip 111). This band migrated further toward the anode than the control.
These results indicated that the enzyme was pure according to the criteria used and in particular was not contaminated by the phosphoglycerate mutase or by the hybrid form of the enzyme. In addition, the purified enzyme appeared more negatively charged than the normal enzyme. Fraction MCI-2 was thus considered as pure and was used for determination of the physicochemical and immunological properties of the enzyme and for structural studies.
Hybridization with Phosphoglycerute Mutase-In a previous report (7), we demonstrated that a spontaneous hybrid composed of bisphosphoglycerate mutase and phosphoglycerate mutase was present in red cells and that such a hybrid could be obtained in vitro. As can be seen in Fig. 3 of the present report, the electrophoretic pattern obtained in the control hemolysate (lane 7), as well as that observed after in vitro hybridization of normal bisphosphoglycerate mutase and phosphoglycerate mutase (lune 2) showed three bands which r"""l+ Hybridization of mutant bisphosphoglycerate mutase and purified phosphoglycerate mutase from normal red cells. Cellogel electrophoresis was carried out as described in the legend to Fig. 2  formed between the two enzymes. This figure also demonstrates that the mutant bisphosphoglycerate mutase hybridizes in vitro with phosphoglycerate mutase (channel 4 ) . However, the pattern given by the mutant hemolysate (lane 6) exhibited only the bands corresponding to the phosphoglycerate mutase and the hybrid, as the mutase activity of bisphosphoglycerate mutase is below the level of detection in the mutant red cells. Kinetic and Immunologic Properties-Upon evaluation of the kinetic characteristics of the purified mutant enzyme, we observed a marked decrease in the specific activities of the synthase and the mutase, which attained, respectively, 0.57 and 11.4% of those of the control (Table I). By contrast, the specific activity of the phosphatase was similar to that of the normal enzyme. In addition, glycolate-2-P, the most potent activator of the phosphatase (27), stimulated such activity in the mutant enzyme some 2-fold less than in the control. Such differences between the marked decrease in the synthase and mutase activities on the one hand and the unmodified phosphatase activity on the other, could be suggestive of the presence of an extrinsic phosphatase contaminant in the sample. Fig. 4 shows that the phosphatase activity of the mutant enzyme was neutralized by specific antibodies to bisphosphoglycerate mutase, while these antibodies neutralized both the phosphatase and the synthase activities of the normal enzyme. The neutralization curves resembled each other and no differences in antibody affinity for the mutant enzyme could be detected. Moreover, thermostability studies (Table 11) (23). Progressive amounts of antibisphosphoglycerate mutase serum were added to 21 pg of purified bisphosphoglycerate mutase in 0.14 M NaCl, 0.1 M sodium phosphate buffer, pH 7.5, and 5% polyethylene glycol. After incubation for 30 min at room temperature, the mixture was centrifuged and the supernatant assayed.

TABLE I1
Thermostability of bisphosphoglycerate mutase and protection by its substrates. Reaction mixtures were incubated at 55 "C for 60 min then centrifuged The supernatant was tested by the antibody consumption technique (see "Experimental Procedures."). The level of immunoreacting material was deduced from the amount of noncomplexed antibody titrated in the assay system. completely protected the mutant protein against heat instability, whereas only slight protection was obtained with either 4 mM glycerate-3-P or 4 mM glycolate-2-P. In a previous report (23), we observed that glycerate-3-P protected the mutant enzyme against heat instability in the hemolysate. Such a discrepancy with the results reported here could arise from a glycerate-3-P solution contaminated by glycerate-2,3-P2, a compound which protects the enzyme at very low concentrations. We evaluated this protective effect on a new hemolysate using glycerate-3-P free of glycerate-2,3-P2, and under these conditions did not find any protective effect of glycerate-3-P on the heated mutant bisphosphoglycerate mutase. Loss of both mutase and phosphatase activities occurred in parallel in the mutant enzyme (Table 111); such loss was prevented upon protection by glycerate-2,3-P2. Structural Studies- Fig. 5 shows the elution pattern of the tryptic digest of 300 pg of the abnormal bisphosphoglycerate mutase. Thirty-six major peaks were obtained, each of them containing one or more peptides. This pattern differed from that obtained under the same conditions from a tryptic hydrolysate of normal bisphosphoglycerate mutase by the ab- Comparative thermostability of the mutase and phosphatase activities of the mutant bisphosphoglycerate mutase The purified bisphosphoglycerate mutase was incubated at 55 "C for 30 min in glycylglycine/NaOH buffer containing 1 mM EDTA, 1 mM 2-mercaptoethanol, and 30% bovine serum albumin. The incubation was stopped by plunging the tubes into iced water. After centrifugation, the enzymatic activities were determined.

% of initial activity
Control without glycerate-2,3-P2 100 100 Mutant without glycerate-2,3-P2 0 0 Mutant with 0.5 mM glycerate-2,3-P2 100 100 sence of two peaks, whose positions are indicated on the figure by a dotted line; these peaks normally correspond to peptides T12 and T13. On the other hand, no supplementary peak was observed in the elution profile obtained from bisphosphoglycerate mutase Cr6teil. All the mutant peptides exhibited a normal sequence except for peptides T12 and T13. Microsequence determination on material in the peak fraction (indicated by an asterisk in Fig. 5) from the variant tryptic hydrolysate after repurification demonstrated the simultaneous presence of two sequences in this peak one of these was identical to that of a normal T6 peptide but lacked cysteinylphenylthiohydantoin at position 2, while the other corresponded to an uncleaved T12-13, the Arg at position 4 being absent. As cysteine residues in the protein had not been alkylated before microsequence determination, these results suggested the presence in this peak of a mixed disulfide T6-T12-13 in which Arg-89, the COOH-terminal residue of the normal peptide T12, would have been replaced by a Cys residue. Such a substitution is consistent with the lack of cleavage of the T12-13 sequence after tryptic hydrolysis.
To test this hypothesis, determination of the mass spectrum of the peak material was performed. The measured m/z value of the protonated molecular ion was found to be 2582.03 (Fig.  6), which is exactly as expected for a T6-T12-13 mixed disulfide given that the calculated monoisotopic value for such molecular species is 2581.22.
Finally, in order to obtain direct evidence, the content of the peak was applied to the trifluoroacetate-treated cartridge of the sequencer, directly derivatized with 4-vinylpyridin (34), and microsequenced The results confirmed the presence in positions 2 in the T6 and 4 in the T12-Tl3 of a pyridinilated cysteinyl-phenylthiohydantoin, thereby providing conclusive evidence that the abnormality in bisphosphoglycerate mutase Crbteil involves the substitution of an Arg residue at position 89 by a Cys.

DISCUSSION
A mutant form of bisphosphoglycerate mutase, an enzyme which normally displays three distinct activities, has been purified from the red cells of a patient deficient in glycerate-2,3-P2. Purification of this protein presented a number of difficulties in view of the small amounts of the mutant enzyme which were available to us, its instability and the absence of synthase activity. Nonetheless, the enzyme was judged to be pure and no protein contaminant could be detected on electrophoresis in sodium dodecyl sulfate-polyacrylamide gel and in Cellogel at pH 8.0. In addition, no contamination with another enzyme displaying mutase activity was detectable on Cellogel electrophoresis. These observations were particularly significant since red cell phosphoglycerate mutase closely resembles bisphosphoglycerate mutase in its structure and function (37). Consequently, the presence of slight phospho- glycerate mutase contamination could substantially perturb studies both of the kinetics of the mutant bisphosphoglycerate mutase and of its structure. For these studies, the small amounts of purified material prevented alkylation of the mutated protein prior to tryptic hydrolysis, resulting probably in the presence of several homogeneous and mixed disulfides in the tryptic hydrolysate. Since the disappearance of peaks normally containing the T12 and the T13 peptides was not accompanied by the occurrence of supplementary peaks, it was necessary to determine the amino acid sequence of material in each peak. Finally microsequencing of the putative pyridinilated mixed disulfide peptide and determination of its mass by spectrometry provided the primary structure and identified the substitution as Arg --., Cys. As previously shown in the determination of the sequence of erythrocyte phosphoglycerate mutase, association of these two methods constitutes an efficient manner in which to perform structural studies of proteins available in very small amounts (38).

Properties and Structure
The substitution of a cysteine for an arginine residue removes a positive charge, thereby explaining the electrophoretic mobility of the mutant enzyme and possibly its heat instability. It should be emphasized that the band of bisphos- phoglycerate mutase could not be detected in the hemolysate from the propositus upon staining for mutase activity (21), reflecting the decreased mutase activity and the 50% diminished content of the enzyme in the propositus' red cells (23). The greater anodal mobility of the mutant bisphosphoglycerate mutase was also demonstrable in the hybrid, as well as in the red cells. Indeed the mutant bisphosphoglycerate mutase, like the normal form (7), can hybridize with the red cell phosphoglycerate mutase. This hybrid is spontaneously formed in red cells and can easily be reproduced in uitro. The mutant hybrid probably exists as a dimer composed of one bisphosphoglycerate mutase subunit and one of phosphoglycerate mutase. As previously demonstrated for the normal protein (7), the mutase activity displayed by the hybrid most probably originates from the phosphoglycerate mutase subunit since the specific activity of the latter is at least 100-fold higher than that of the intrinsic mutase activity of bisphosphoglycerate mutase. These results indicate that the mutation does not affect subunit association and does not modify the active site of the phosphoglycerate mutase subunit to any marked degree.
A marked discrepancy was observed between the three enzymatic activities in the purified mutant bisphosphoglycerate mutase. Thus, synthase activity was almost absent and a marked decrease in mutase activity was displayed. Rather significantly, we observed that the specific activity of the phosphatase was identical to that of the normal enzyme. Similar results were found earlier for the normal bisphosphoglycerate mutase treated by sulfhydryl reagents (28), and we cannot exclude the possibility that data for the mutant are relevant to the presence of a cysteine residue at position 89. Evidence is provided that the phosphatase activity of the mutant protein is thus intrinsic to bisphosphoglycerate mutase, since this activity is unstable to heat in the same manner as that for the mutant enzyme; indeed the mutant form has previously been reported to be unstable after incubation for 30 min at 55 "C, whereas the normal enzyme is completely stable for at least 60 min at this temperature (23). In addition, the phosphatase activity of the mutant was neutralized by specific antibodies to bisphosphoglycerate mutase, which also inhibited the synthase and phosphatase activities of the normal enzyme. However, glycolate-2-P, the most potent activa- tor of the phosphatase activity (27), showed a 2-fold lesser stimulating effect on the mutant than on the normal enzyme. However, considered together, our results afford evidence that the mutation is situated at or near the active site of the enzyme. Moreover, the existence of normal glycerate-2,3-P2 phosphatase activity and the protection of the enzyme against heat instability by glycerate-2,3-P2 support the contention that the binding site for bisphosphoglycerates is not involved in the mutation. Indeed, a histidine residue has been demonstrated to be located at the active site of bisphosphoglycerate mutase; this residue is phosphorylated by bisphosphoglycerates (18,19) and may play a prominent role in the mechanism of the three enzymatic reactions (15-17). The histidine residue at position 8 of the mutant enzyme is not affected by the mutation since the microsequence of the corresponding peptide was normal. Consequently, it can be assumed that glycerate-1,3-P2, the substrate of the synthase activity can bind to the His-8 residue of the mutant. Labeling of the enzyme by incubation with 32P-labeled substrates would allow us to verify whether His-8 is phosphorylated. Unfortunately, we lacked additional samples on which to perform such experiments; the patient is now deceased. On the other hand, several lines of evidence suggest that the site which binds glycerate-3-P and glycolate-2-P is involved in the mutation, and these include (i) the absence of synthase activity requiring glycerate-3-P as a cofactor, (ii) the marked diminution in mutase activity, and (iii) the decreased stimulation of phosphatase by glycolate-2-P.
Mutated residue 89 has not been identified until now among the amino acid residues at the active site of the bisphosphoglycerate mutase. Nevertheless, it is notable that residue 89 is invariant in the bisphosphoglycerate mutase of all the species tested to date, and moreover, this residue is invariant in several monophosphoglycerate mutases, an enzyme whose structure is closely related to that of the red blood cell bisphosphoglycerate mutase (39). Moreover, and as represented in Fig. 7 , residue 89 is present within a large invariable portion of the molecule which, according to the three-dimensional model of Winn et al. (40) for the yeast phosphoglycerate mutase, is in close proximity to histidine 8 of the reactive site. Taking into account these considerations, we hypothesize that this portion of the molecule could be located at or near the site involved in the binding of glycerate monophosphates and glycolate-2-P. In view of the scarcity of naturally occurring variants of bisphosphoglycerate mutase, further studies of the relationship of structure to function in this protein will require the use of site-directed mutagenesis; such an experimental approach is possible since the cDNA structure of the bisphosphoglycerate mutase has been established (4, 41).
Biochemical investigations of a mutant enzyme thus constitute a valid model for the elucidation of the active site mechanism of the normal protein and may be compared with studies of the structure and function of mutant hemoglobins, the latter having provided substantial knowledge of normal hemoglobin function.