Evidence for Structural Homology between Human Red Cell Phosphoglycerate Mutase and 2,3=Bisphosphoglycerate Synthase*

Previous reports have suggested the possibility of extensive structural homology between human erythrocyte bisphosphoglycerate synthase (glycerate-1,3-P, -+ glycerate-2,3-P,) and phosphoglycerate mutase (glycerate-3-P e glycerate-2-P). This study lends credence to that conjecture through comparative physicochemical investigations involv-ing peptide mapping, circular dichroism, and immunological techniques. The data indicate that despite differences in function, both enzymes apparently manifest a high degree of similarity in primary, secondary, and tertiary structure. Mapping data also indicate that each protein is comprised of two apparently identical subunits.

Through purification of the aforementioned erythrocyte enzymes it was possible to establish unequivocally that each catalyst is multifunctional (3-6). Thus, it was found that both catalysts possess intrinsic 2,3-bisphosphoglycerate phosphatase activity (glycerate-2,3-P, + P-glycerate + P,). In human erythrocytes, most of the glycerate-2,3-P, phosphatase activity was shown to be associated with the bisphosphoglycerate synthase molecule, leading to the concept that the phosphoglycerate bypass as proposed by Rapoport and Luebering (7,8) is under the control of a single enzyme (9). Assays -Bisphosphoglycerate synthase was assayed essentially by the method described previously (5) except that the final concentrations of the ingredients in the reaction mixture were modified as follows: 25 rnM glycylglycine (pH 7.8), 1.5 rnM dithiothreitol, 1 rnM K,HPO,, 1 mM or,-glyceraldehyde-3-P, 1 rnM NAD+, and 1 mM glycerate-3-P.
The conditions cited give maximal initial velocities.
Phosphoglycerate mutase activity was assayed as previously described (6 [@I = 0 ObSd x (mean residue wt) The mean residue weight of each protein was calculated from its amino acid composition.
The fractions (f, etc.) of a, p, and random structures were calculated by the method of Reed et al. (13) from the relationships: and LOI = f,[a3 + fura + f,[a?
(2) f,+fi+fR=1  Fig. 3. Except for slight differences in the minima above 220 nm and the extents of ellipticity, both spectra are essentially equivalent. The occurrence of minima at 208 and 223 to 225 nm indicates the presence of a significant amount of (Y helix in each protein (17). This is corroborated by the calculations from CD data (Table I) of a! helix Cfti), /3pleated sheet (f,), and random structure (f,).
The most significant aspect of this study, however, is the apparently close agreement in secondary structure between bisphosphoglycerate synthase and phosphoglycerate mutase . . ---IO  I  I  I  I  I  I  200  220  240 Table I Reed et al. (13) are also in good agreement. This is particularly true in light of the report by the latter investi-  Fig. 3B (mutase) and Fig. 3C (synthase) show close coincidence between calculated points and the experimentally determined CD spectra. In both cases, however, slight discrepancies are found in the regions of the maxima (i.e. at 215 nm for the mutase and at 217 nm for the synthase).
Immunology -Rabbit antiserum generated against bisphosphoglycerate synthase was tested for cross-reactivity with human erythrocyte and rabbit muscle phosphoglycerate mutase by the double diffusion technique of Ouchterlony (15). As illustrated in Fig. 4, a positive precipitin test was obtained only with the human mutase. Moreover, because of the lack of "spurring," the pattern obtained is one of identity rather than cross-reactivity.
Thus, our immunological results not only complement but supplement the peptide mapping data discussed above by indicating that homologous determinants occur near the surface of each enzyme in its native conformation. It appears, therefore, that both proteins are similar in tertiary structure. The extent of this similarity is indicated in Fig. 5 which shows the incongruities in antigen-antibody equivalence points as determined by turbidimetric procedures (20). As illustrated, the response obtained with the mutase is considerably weaker than that of the synthase, indicating that each enzyme manifests a different binding capacity for anti-bisphosphoglycerate synthase. The precise reason for this binding phenomenon is not apparent, but Fig. 4 indicates that the phenomenon is definitely not related to differences in antigenic character.
The negative precipitin test obtained with rabbit muscle phosphoglycerate mutase (Fig. 4)  should contain common, reactive haptenic groups. Furthermore, precedents have been established for the termination of natural or acquired immunological tolerance by exposing animals to heterologous, cross-reacting antigens (21,22). At this point, however, it seems apropos to state that there is evidence which indicates that each mutase does not manifest exactly equivalent catalytic capacities4 This finding suggests that, despite structural homology, both enzymes are sufficiently different so that identical interaction with anti-bisphosphoglycerate synthase is precluded. It is possible, however, that under our experimental conditions the tolerance of the rabbit to its own phosphoglycerate mutase was maintained.

DISCUSSION
The data presented here and elsewhere (1) give evidence relating to extensive structural homology between two functionally distinct enzymes. Moreover, both of these enzymes have been obtained from the same species. This work, therefore, is unlike other numerous investigations which have concerned themselves with the characterization of functionally identical proteins from different species.
Despite considerable foundation in reality, all of the evidence in favor of structural homology between human red cell bisphosphoglycerate synthase and phosphoglycerate mutase is, nevertheless, of an indirect nature. Thus, additional experimentation will be required to lend more substance to our hypothesis. Areas for further investigation that immediately suggest themselves are sequence analysis and x-ray crystallography. With regard to the latter, we must point out that McPherson (23) has already crystallized red cell bisphosphoglycerate synthase from solutions of polyethylene glycol.
If our hypothesis concerning bisphosphoglycerate synthase and phosphoglycerate mutase is indeed correct, it would be of definite interest to determine the nature of those structural features which are responsible for the elicitation of one catalytic activity versus another. It would also be of interest to obtain clues leading to an explanation for the degrees of overlapping polyfunctionality exhibited by each enzyme.