Activity of glycogen phosphorylase in the crystalline state.

Glutaraldehyde cross-linked crystals of muscle phosphorylase a and b (alpha-1,4-glucan:orthophosphate glucosyltransferase, EC 2.4.1.1) in the tetragonal form have been shown to be catalytically active in the direction of saccharide synthesis. Precession x-ray photographs at 5.5 A resolution of a single crystal of cross-linked phosphorylase b at the hol zone indicate little change in the diffraction pattern when compared to non-cross-linked phosphorylase b under similar conditions. Non-cross-linked crystals crack and dissolve in the presence of both substrates, maltopheptaose and glucose 1-phosphate, although they are stable in the presence of each individually. These phenomena are prevented by treatment with glutaraldehyde, which causes a marked increase in mechanical stability and completely suppresses solubilization of the enzyme under our assay conditions. Diffusion of substrates into cross-linked microcrystals does not appear to be rate-limiting and assays of such crystals are linear with respect to both time and enzyme concentration. Kinetic constants for both substrates are reported. The maximal velocities of phosphorlyase a are larger than those of phosphorylase b in both the soluble and crystalline states under our assay conditions, with the above substrates. It appears that crystallization (and cross-linking) reduces maximal velocities by about 11- to 50-fold in the case of phosphorylase b and 50- to 100-fold for phosphorylase a. Little or no differences were found between the Km values for maltoheptaose or glucose 1-phosphate in the soluble or crystalline states. Kinetic data suggest that substrate binding sites are similar in both states. Although loss of catalytic efficiency points to differences in the active site of the enzyme caused by crystallization, another explanation is that the crystal is restricting a conformational change that is an essential part of the catalytic cycle.


RESULTS
Glutaraldehyde cross-linking procedures have been described in detail for crystalline carboxypeptidase A (3, 4) and phosphorylase b in solution (13). Suspensions of tetragonal microcrystalline phosphorylase b in crystallization buffer were cross-linked with 0.03% glutaraldehyde, obtained from Sigma, for 1 h at 22". After exhaustive washing of the cross-linked microcrystals with crystallization buffer, they were fractionated into distinct groups by repeated centrifugation at very low speeds in a clinical centrifuge, equipped with a swinging bucket head.
The largest of these had average dimensions of about 10 x 10 x 70 pm and the smallest about 1 x 1 x 10 pm. Size was determined by direct measurement.
Phosphorylase a was prepared from phosphorylase b with phosphorylase b kinase as described by Krebs et al. (14). Crystallization of tetragonal microcrystals of phosphorylase a was carried out as previously described for phosphorylase b, except that IMP in the crystallization buffer was replaced by 50 mM glucose (11 Amino acid analyses were carried out on a Spinco automatic amino acid analyzer.
Crystal samples were hydrolyzed in 6 N HCl at 110" for 44 h in sealed evacuated tubes. Errors between crystal enzyme concentrations as determined by amino acid analysis and the protein methods described above were less than 3%.
Reaction rates were determined in the direction of saccharide synthesis by a modification of the method described by Hu and Gold (18). Reaction mixtures were 50 or 100 ~1 and contained 10 mM Bes (pH 6.7), 10 rnM magnesium acetate, 1 mM EDTA, and 5 mM dithiothreitol. When soluble enzymes were assayed 0.5 mg/ml of bovine serum albumin was present. Either 0.5 to 3 pg of soluble enzyme or 0.2 to 10 Kg of cross-linked crystals were used. Cross-linked microcrystals were added with Drummond microcaps as a suspension in crystallization buffer. Enzyme and maltoheptaose were preincubated for 15 min at 30" before initiating the reaction with glucose-l-P'. In all kinetic studies only the smallest microcrystals having dimensions about 2 x 2 x 10 pm or less were used unless otherwise noted. In preliminary studies, assays normally contained 12.5 mM glucose-l-P, 2 mM IMP, and 13.7 mM (as nonreducing chain termini) maltoheptaose for phosphory1ase.b and 50 mM glucose-l-P, 27.4 mM maltoheptaose without nucleotide for phosphorylase a. Crystal assays were normally carried out in triplicate and solution assays in duplicate.
Inorganic phosphate was determined by the method of Baginski et al. (19,20)  Details of the x-ray analysis will appear in a later publication.3 The average chain length of the pure preparation was determined from the ratio of the reducing end group to the total hexose content using maltose as the standard.
Hexose was determined by the method of Dishe as described by Ashwell (23). Reducing power was determined by the copper-neocuprione method of Dygert et al. (24). The product gave an average chain length of 7.05 glucose units per reducing group and was chromatographically pure as indicated by the method of Tetragonal microcrystals of phosphorylase a and b were prepared primarily to limit kinetic effects of diffusion of substrates and products in the crystal lattice as has been discussed by others (1,4,9). Assays of suspensions of the smallest cross-linked microcrystals of phosphorylase 6, whose average size was about 1 x 1 x 10 pm indicated that assays Activity of Crystalline Phosphorylase were linear with time (Fig. 2). However, specific activities of larger crystals, whose average minimum dimension was about 5 pm, showed slight curvature as a function of time (Fig. 2). Nevertheless, the specific activities of the two sizes of crystals, as indicated by the slopes of the lines in Fig. 2, are identical within experimental error. As pointed out by Quiocho and Richards (4) a study of enzyme activity on both a weight and crystal surface basis, when related to crystal size can indicate the influence of diffusion on activity measurements.
A constant specific activity as a function of crystal size indicates that activity measurements are free of diffusion effects. An inverse relationship between activity (on a surface area basis) and crystal size indicates an influence of diffusion on activity. In Table I we present studies of the effect of crystal size on activity. The relative activities on a weight basis were essentially identical even though the relative surface area increased lo-fold as a function of decreasing size. The results indicate no diffusion effects for crystals whose minimum dimension is less than 10 pm. The work on carboxypeptidase (4, 9) and papain (l), indicates that the activities of crystals whose minimum dimension is about 5 Mm or less are not affected by the diffusion of substrate. Furthermore, the results of studies of the reaction of azide with myoglobin crystals, gave no evidence of diffusion limitation with minimum crystal dimensions in the range 2 to 5 w (26).
Additional information which is of importance relative to the question of crystal surface activity is the finding that microcrystals exhibited significant activity with glycogen as the FIG. 1. A split ho1 precession photograph at 5.5 8, resolution for the native b (left) and glutaraldehyde cross-linked enzyme. saccharide acceptor. Quiocho and Richards (4) have, without success, attempted to estimate the contribution of the surface layer to the activity of crystals of carboxypeptidase, by the use of the synthetic block co-polymer of, glutamic acid and tyrosine, which is a potent inhibitor of the enzyme. With phosphorylase some of the nonreducing chains of glycogen are presumably able to bind and undergo chain extension at the crystal surface. In the presence of 27 mM maltoheptaose and 50 mM glucose-l-P in crystallization buffer, cross-linked microcrystals of phosphorylase b gave a specific activity of 2.34 (micromoles per h per mg); with 1% glycogen and 50 mM glucose-l-P, specific activity = 0.28; with all components, specific activity = 2.61. In the presence of glucose-l-P and either maltoheptaose or glycogen no detectable activity was present in the supernatant solution obtained after sedimenta-   5). Phosphorylase a, however, showed curvature at low concentrations of maltoheptaose (Fig. 6) while the Hill plot of this data indicated cooperativity in solution and slightly less cooperativity in the crystalline state (Fig. 7).   (Fig. 8). The kinetics observed for phosphorylase b and glucose-l-P were linear with both crystal and soluble enzymes (Fig. 9). Kinetic parameters for the various enzyme forms are presented in Table  II The same crystal preparation as i n Fig. 5