Articles Catalytic Cycling in â-Phosphoglucomutase: A Kinetic and Structural Analysis†,‡

Lactococcus lactis â-phosphoglucomutase (â-PGM) catalyzes the interconversion of â-D-glucose 1-phosphate (â-G1P) and â-D-glucose 6-phosphate (G6P), forming â-D-glucose 1,6-(bis)phosphate (âG16P) as an intermediate. â-PGM conserves the core domain catalytic scaffold of the phosphatase branch of the HAD (haloalkanoic acid dehalogenase) enzyme superfamily, yet it has evolved to function as a mutase rather than as a phosphatase. This work was carried out to identify the structural basis underlying this diversification of function. In this paper, we examine â-PGM activation by the Mg2+ cofactor, â-PGM activation by Asp8 phosphorylation, and the role of cap domain closure in substrate discrimination. First, the 1.90 Å resolution X-ray crystal structure of the Mg2+-â-PGM complex is examined in the context of previously reported structures of the Mg2+-R-D-galactose-1-phosphate-â-PGM, Mg2+-phospho-â-PGM, and Mg2+-â-glucose-6-phosphate-1-phosphorane-â-PGM complexes to identify conformational changes that occur during catalytic turnover. The essential role of Asp8 in nucleophilic catalysis was confirmed by demonstrating that the D8A and D8E mutants are devoid of catalytic activity. Comparison of the ligands to Mg2+ in the different complexes shows that a single Mg2+ coordination site must alternatively accommodate water, phosphate, and the phosphorane intermediate during catalytic turnover. Limited involvement of the HAD family metal-binding loop in Mg2+ anchoring in â-PGM is consistent with the relatively loose binding indicated by the large Km for Mg2+ activation (270 ( 20 μM) and with the retention of activity found in the E169A/D170A double loop mutant. Comparison of the relative positions of cap and core domains in the different complexes indicated that interaction of cap domain Arg49 with the “nontransferring” phosphoryl group of the substrate ligand might stabilize the cap-closed conformation, as required for active site desolvation and alignment of Asp10 for acid-base catalysis. Kinetic analyses of the specificity of â-PGM toward phosphoryl group donors and the specificity of phospho-â-PGM toward phosphoryl group acceptors were carried out. The results support a substrate induced-fit mechanism of â-PGM catalysis, which allows phosphomutase activity to dominate over the intrinsic phosphatase activity. Last, we present evidence that the autophosphorylation of â-PGM by the substrate â-G1P accounts for the origin of phospho-â-PGM in the cell. Phosphoglucomutases catalyze the interconversion of D-glucose 1-phosphate (G1P) and D-glucose 6-phosphate (G6P).1 Operating in the forward G6P-forming direction, this reaction links polysaccharide phosphorolysis to glycolysis. In the reverse direction, the reaction provides G1P for the biosynthesis of exo-polysaccharides (8). There are two classes of phosphoglucomutases, the R-phosphoglucomutases (R-PGM, EC 5.4.2.2), ubiquitous among eukaryotes and prokaryotes, and the â-phosphoglucomutases (â-PGM, EC 5.4.2.6), present in certain bacteria and protists. The two classes of mutases are distinguished by their specificity for Rand â-D-glucose phosphates and by their protein fold † This work was supported by NIH Grant GM61099 (to D.D.-M. and K.N.A.). ‡ The X-ray coordinates are listed in the Protein Data Bank as entry 1ZOL. * To whom correspondence should be addressed. D.D.-M.: telephone, (505) 277-3383; fax (505) 277-6202; e-mail, [email protected]. K.N.A.: telephone, (617) 638-4398; fax, (617) 638-4285; e-mail, [email protected]. 1 Abbreviations: R-PGM, R-phosphoglucomutase; R-PGM/PMM, dual-specificity R-phosphoglucomutase/R-phosphomannomutase; â-PGM, â-phosphoglucomutase; E, â-PGM-Mg2+; E-P, phospho-â-PGMMg2+; E-P-GP, â-PGM-â-glucose-6-phosphate-1-phosphoraneMg2+; â-G1P, â-D-glucose 1-phosphate; â-G16P, â-D-glucose 1,6(bis)phosphate; R-G1P, R-D-glucose 1-phosphate; R-G16P, R-D-glucose 1,6-(bis)phosphate; R-F16P, R-D-fructose 1,6-(bis)phosphate; pNPP, p-nitrophenyl phosphate; G6P, Rand/or â-D-glucose 6-phosphate; NADP, adenine dinucleotide 3′-phosphate; K+Hepes, potassium salt of 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid; K+Ches, potassium salt of N-cyclohexyl-2-aminoethanesulfonic acid; K+Mes, potassium salt of 2-(N-morpholino)ethanesulfonic acid; DTT, dithiothreitol; EDTA, ethylenediaminetetraacetic acid; HPLC, high-performance liquid chromatography; SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis; SA, specific radioactivity; MLF, maximum likelihood target using amplitudes. 9404 Biochemistry 2005, 44, 9404-9416 10.1021/bi050558p CCC: $30.25 © 2005 American Chemical Society Published on Web 06/17/2005 family. R-PGM belongs to the phosphohexomutase superfamily (9), while â-PGM belongs to the haloalkanoic acid (HAD) enzyme superfamily (10). The four-domain R-PGMs (∼50 kDa) (11-14) are approximately twice the size of the two-domain â-PGMs (∼25 kDa) (15). The common denominator between the mutases is catalysis via a glucose 1,6-(bis)phosphate (G16P) intermediate. Whereas R-PGM catalysis had been studied in depth (16-31), comparatively little is known about â-PGM catalysis, the topic of this paper. The catalytic cycles of two divergent R-PGMs, one from rabbit muscle (11, 12, 28-30) that functions in energy metabolism and the other from Pseudomonas aeruginosa [the duel specificity R-phosphoglucomutase/R-phosphomannomutase (R-PGM/PMM)] (13, 14, 32, 33) that functions in alginate and lipopolysaccharide biosynthesis (34-36), have been shown to be the same. The cycle consists of phosphorylation of R-G1P by a phosphorylated active site serine residue to form R-G16P as an intermediate, reorientation of the intermediate in the active site, transfer of the C(1) phosphoryl group to the active site serine, and finally release of G6P from the active site of the phosphorylated enzyme (25, 33) (Figure 1). Phosphorylated R-PGM is chemically stable in water (half-life of ∼7 years) yet transfers its phosphoryl group to R-G1P at a rate of 1000 s-1 (37). It has been presumed that in vivo R-PGM is phosphorylated by R-G16P [Km for activation ) 0.1 μM for P. aeruginosa R-PGM/PMM (32); Kd ) 0.02 μM for rabbit muscle R-PGM (37, 38)], even though the source of R-G16P in the cell has not yet been established. An alternate proposal is that R-PGM is autophosphorylated by ATP, an activity that has been observed in vitro (22). â-PGM isolated from Lactococcus lactis (39), Bacillus subtilis (40), Neisseria perflaVa (41), and Euglena gracilis (42) is, like R-PGM, known to utilize G16P as a phosphoryl donor and Mg2+ as a cofactor. However, there are distinct differences between the two catalytic scaffolds, which suggests that the respective phosphoenzymes are generated and stabilized by different mechanisms. X-ray crystallographic analysis of L. lactis â-PGM has provided three snapshots of the enzyme bound with its cofactor (Mg2+) and a phosphorylated ligand (43, 44, 58). These structures differ in the identity of the ligand and the solvent accessibility of the active site. The initial structure pictured the phosphoenzyme in an active site open conformation; the second structure showed the enzyme in an active site closed conformation with the Asp8 carboxylate oxygen forming a covalent bond to the phosphorus of the â-glucose-6-phosphate1-phosphorane, and the third structure revealed the enzyme in an active site closed conformation, with the substrate analogue R-D-galactose 1-phosphate oriented to place its phosphate group at the distal phosphate binding site and its C(6)OH group near the Asp8 nucleophile. The first two structures represent reaction intermediates stabilized in the crystalline state and the third an inhibitory, dead-end complex of substrate bound to the dephosphoenzyme. In this paper, we report the structure of a fourth complex, the “free” holoenzyme bound to its Mg2+ cofactor. Within the catalytic cycle, this enzyme form binds and is phosphorylated by G16P. Herein, the four structures are compared, and kinetic studies probing structural requirements for â-PGM phosphorylation and dephosphorylation are reported, to provide insight into how the phosphatase branch of the HAD family (45) has acquired phosphomutase function, and how the â-PGM catalytic strategy differs from that used by R-PGM. FIGURE 1: Sequence of reaction steps involved in a single catalytic cycle in enzymes R-PGM and â-PGM. For R-PGM, the substrate is RG1P, the intermediate is R-G16P, and the active site nucleophile (X) is Ser108. For â-PGM, the substrate is â-G1P, the intermediate is â-G16P, and the active site nucleophile (X) is Asp8. â-Phosphoglucomutase from Lactococcus lactis Biochemistry, Vol. 44, No. 27, 2005 9405