Histidine 167 Is the Phosphate Acceptor in Glucose-6-phosphatase-β Forming a Phosphohistidine Enzyme Intermediate during Catalysis*

The glucose-6-phosphatase (Glc-6-Pase) family comprises two active endoplasmic reticulum (ER)-associated isozymes: the liver/kidney/intestine Glc-6-Pase-α and the ubiquitous Glc-6-Pase-β. Both share similar kinetic properties. Sequence alignments predict the two proteins are structurally similar. During glucose 6-phosphate (Glc-6-P) hydrolysis, Glc-6-Pase-α, a nine-transmembrane domain protein, forms a covalently bound phosphoryl enzyme intermediate through His176, which lies on the lumenal side of the ER membrane. We showed that Glc-6-Pase-β is also a nine-transmembrane domain protein that forms a covalently bound phosphoryl enzyme intermediate during Glc-6-P hydrolysis. However, the intermediate was not detectable in Glc-6-Pase-β active site mutants R79A, H114A, and H167A. Using [32P]Glc-6-P coupled with cyanogen bromide mapping, we demonstrated that the phosphate acceptor in Glc-6-Pase-β is His167 and that it lies inside the ER lumen with the active site residues, Arg79 and His114. Therefore Glc-6-Pase-α and Glc-6-Pase-β share a similar active site structure, topology, and mechanism of action.

The prototype of the family, Glc-6-Pase-␣, is a 357-amino acid, nine-transmembrane domain, endoplasmic reticulum (ER)-associated protein (10,11), which is expressed primarily in the liver, kidney, and intestine (12,13). Glc-6-Pase-␣ catalyzes the hydrolysis of glucose 6-phosphate (Glc-6-P) to glucose in the terminal step of gluconeogenesis and glycogenolysis (13). Between meals, the resulting release of glucose to the blood maintains glucose homeostasis. Naturally occurring loss of function mutations in Glc-6-Pase-␣ cause glycogen storage disease type Ia, a disorder that is characterized by loss of blood glucose homeostasis and disorders of glycogen and lipid metabolism (reviewed in Refs. 14 and 15).
The active site of Glc-6-Pase-␣ was originally identified by the presence of a conserved phosphatase signature motif found in lipid phosphatases, acid phosphatases, and vanadium haloperoxidases (16,17). This motif was shown to contribute to the active site of a vanadium chloroperoxidase by both x-ray crystal structure analysis (18) and mutational studies (19,20) and has since been shown to have a similar role in Glc-6-Pase-␣ (21,22). Hydrolysis of Glc-6-P to glucose and phosphate via a covalent phosphohistidine-Glc-6-Pase-␣ intermediate was first proposed by Nordlie and Lygre (23) based on pH kinetic studies of Glc-6-Pase-␣ catalysis, which was confirmed by the identification of an enzyme-bound 32 P-labeled histidine after incubating rat liver microsomes with [ 32 P]Glc-6-P (24 -26). Critical residues for Glc-6-Pase-␣ catalysis in the active site motif include: Arg 83 , which donates hydrogen ions to the phosphate and stabilizes the transition state; His 119 , which provides a proton to liberate the glucose moiety; and His 176 , which undertakes a nucleophilic attack on the phosphate to form a covalently bound phosphoryl enzyme intermediate (21,27). In Glc-6-Pase-␣, all of these residues lie together on the lumenal side of the ER membrane (10,11).
Given the sequence similarity between Glc-6-Pase-␣ and Glc-6-Pase-␤, a similar active site structure would be anticipated. Sequence alignments predict that His 167 is the residue that forms a covalent bond with phosphate to create a phosphoryl-Glc-6-Pase-␤ intermediate, whereas Arg 79 and His 114 are the hydrogen donors. These alignments also suggest that Glc-6-Pase-␤ contains nine putative transmembrane helices and that Arg 79 , His 114 , and His 167 lie inside the ER lumen.
In an effort to understand the role of this ubiquitous Glc-6-Pase activity, we have undertaken a study of the active site of Glc-6-Pase-␤ and confirmed that it is similar to Glc-6-Pase-␣. This information can now be used to design experiments that disrupt the activity of Glc-6-Pase-␤ and investigate the biological importance of a Glc-6-Pase activity that is not limited in expression to the liver, kidney, and intestine, the organs pre-viously considered to be the only ones capable of contributing to interprandial glucose homeostasis (13).
Protease Protection Assays and Western Blot Analysis-COS-1 cells in 25-cm 2 flasks were transfected with 10 g of Glc-6-Pase-␤-5FLAG, Glc-6-Pase-␤-3FLAG, Glc-6-Pase-␤-5fXaFLAG, or Glc-6-Pase-␤-3fX-aFLAG in the pSVL vector as described previously (10). After incubation at 37°C for 2 days, cell homogenates in 100 l of Buffer A (0.25 M sucrose and 5 mM HEPES, pH 7.4) were treated with 250 g/ml DNase I for 30 min at 30°C. The cell homogenates from Glc-6-Pase-␤-5FLAG or Glc-6-Pase-␤-3FLAG transfected cells were then treated with trypsin (Type XIII, 500 g/mg protein) for 30 min at room temperature followed by 5 mM phenylmethylsulfonyl fluoride and trypsin inhibitor (6 mg/mg protein) to inactivate trypsin. The cell homogenates from Glc-6-Pase-␤-5fXaFLAG or Glc-6-Pase-␤-3fXaFLAG transfected cells were treated with 15 g of factor Xa protease (New England Biolabs, Beverly, MA) for 2 h at room temperature in a 75-l reaction mixture containing 25 mM Tris-HCl, pH 7.4, 100 mM NaCl, and 2.5 mM CaCl 2 . The factor Xa was then inactivated by the addition of Dansyl-Glu-Gly-Arg-chloromethylketone (Calbiochem) to a final concentration of 10 M. The resultant trypsin-or factor Xa-treated cell homogenates were diluted 100-fold to 10 ml with cold Buffer A and centrifuged at 100,000 ϫ g for 1 h at 4°C to pellet the microsomes. The microsomal pellets were resuspended in Buffer A and used for Western blot analysis. The cell homogenates treated with 0.5% deoxycholate to disrupt the microsomes, followed by trypsin or factor Xa treatment, were used as controls.
For Western blot analysis, COS-1 lysates were separated by electrophoresis through a 12% SDS-polyacrylamide gel blotted onto polyvinylidene fluoride membranes (Millipore Co.) and incubated first with a monoclonal antibody against the FLAG epitope (Scientific Imaging Systems, Eastman Kodak) and then with goat anti-mouse IgG antibody (Kirkegarrd & Perry Laboratories, Gaithersburg, MD). The immunocomplex was detected with the horseradish peroxidase-linked chemiluminescent system containing the SuperSignal West Pico chemiluminescent substrate obtained from Pierce.
Identification of the [ 32 P]Phosphoryl-Glc-6-Pase-␤ Intermediate-32 P-Labeled Glc-6-P was synthesized from glucose and 5Ј-[␥-32 P]ATP by yeast hexokinase (Sigma) and partially purified by a Dowex Ag-1Xformate column as described previously (21). Microsomes were prepared from COS-1 cells infected with a recombinant adenovirus at a multiplicity of infection of 100 plaque-forming units/cell after incubation at 37°C for 48 h as described previously (6). For active site labeling, microsomes (100 g) were pre-incubated at 30°C for 10 min in a reaction mixture (100 l) containing 40 mM sodium cacodylate buffer, pH 6.5, 2 mM EDTA, and 100 M glucose 1-phosphate and then radiolabeled by the addition of ϳ8 ϫ 10 6 cpm of [ 32 P]Glc-6-P to a final volume of 130 l. After incubation for 10 s at 37°C, the reaction was terminated by the addition of an equal volume of ice-cold 20% trichloroacetic acid followed by incubation on ice for 30 min. The labeled proteins were then pelleted by centrifugation at 12,000 ϫ g for 10 min at 4°C and washed twice with 100 l of ice-cold 0.5 M Tris-HCl, pH 8.0, and once with 100 l of cold water. The pellet was then finally solubilized in a triple lysis buffer containing 50 mM Tris-HCl, pH 8.0, 150 mM NaCl, 0.1% SDS, 1% Nonidet P-40, 0.5% Na-deoxycholate, 0.02% NaN 3 , and the complete EDTA-free protease inhibitor mixture (Roche Diagnostics). The [ 32 P]phosphoryl enzyme intermediate was isolated by immunoprecipitation using a monoclonal anti-FLAG antibody in conjunction with protein A-agarose (Sigma) and identified by 12% SDS-polyacrylamide gel electrophoresis and autoradiography.
Purification of [ 32 P]Phosphoryl-Glc-6-Pase-␤ Intermediate-The 32 Plabeled phosphoryl microsomal proteins from Ad-Glc-6-Pase-␤-His 10infected cells were precipitated by trichloroacetic acid as described above. The pellet was washed twice with 100 l of ice-cold 0.5 M Tris-HCl, pH 8.0, and once with 100 l of cold water and finally dissolved in 400 l of Solution A (25 mM Tris-HCl, pH 8.0, 1% Triton X-100, 150 mM NaCl) containing the complete EDTA-free protease inhibitor mixture. The solubilized proteins were bound to a nickel chelate resin (His MicroSpin Purification Module, Amersham Biosciences) pre-equilibrated with Solution A. The resin was washed according to the manufacturer's protocol, and the bound protein was eluted with 30 l of SDS sample buffer (62.5 mM Tris-HCl, pH 6.8, 2% SDS, 5% 2-mercaptoethanol, and 10% glycerol) at 46°C for 30 min. The resulting proteins were fractionated by electrophoresis through a 12% SDS-polyacrylamide gel and the [ 32 P]phosphoryl-Glc-6-Pase-␤-His 10 intermediate visualized by autoradiography. A gel slice containing the radiolabeled protein was excised, eluted with water by incubating the mashed gel slice in water at 50°C for 1 h, lyophilized, and used for cyanogen bromide digestion.

RESULTS
Membrane Topography of Glc-6-Pase-␤ -Glc-6-Pase-␣ is embedded in the ER membrane by nine-transmembrane domains (10,11) oriented with the N terminus in the ER lumen and the C terminus in the cytoplasm. Hydropathy analysis of the Glc-6-Pase-␤ amino acid sequence using the TMpred program (32) predicts that Glc-6-Pase-␤ is also anchored in the ER by nine putative transmembrane helices (Fig. 1). If this is correct, the N and C termini of the protein will be on opposite sides of the ER membrane. To determine the subcellular localization of the termini of Glc-6-Pase-␤, we undertook protease protection assays using N-and C-terminal FLAG-tagged Glc-6-Pase-␤ constructs. Intact microsomes from COS-1 cells transfected with Glc-6-Pase-␤-5FLAG or Glc-6-Pase-␤-3FLAG were digested with trypsin, and the protein fragments were analyzed by Western blot analysis ( Fig. 2A). Microsomes expressing Glc-6-Pase-␤-3FLAG yielded no FLAG-tagged peptides upon digestion consistent with an exposed cytoplasmic C terminus. In contrast, microsomes expressing Glc-6-Pase-␤-5FLAG were more resistant to digestion. However, the trypsin-resistant FLAG-tagged product is not full-length (35 kDa) but a 14-kDa polypeptide ( Fig. 2A). In control experiments using microsomes permeabilized by detergent, both N-and C-terminal FLAG tags are cleaved by trypsin ( Fig. 2A).
To purify the [ 32 P]phosphoryl-Glc-6-Pase-␤ intermediate, we generated a Glc-6-Pase-␤ construct carrying a His 10 (32), was used to identify the transmembrane helices. Larger ovals outline the amino acid residues predicted to comprise the active center. The methionine (M) residues susceptible to cleavage by cyanogen bromide are denoted by black ovals, and the arginine/lysine residues susceptible to trypsin digestion are denoted by shaded ovals. fractionated through a NuPAGE Bis-Tris gel. A major band with an apparent molecular mass of 4.4 kDa and minor bands with apparent molecular masses of 18, 16, and 10 kDa were identified (Fig. 4A). The 4.4-kDa peptide corresponds to the expected molecular weight of the His 167 -containing peptide of 41 amino acids (residues 146 -186).
To further demonstrate that the His 167 -containing peptide of 4.4-kDa carries the [ 32 P]phosphoryl moiety, the 4.4-kDa band was eluted from the gel and analyzed by gel electrophoresis and isoelectric focusing. The purified 32 P-labeled peptide migrated as a 4.4-kDa band through a 10% NuPAGE Bis-Tris gel (Fig.  4B) with an isoelectric point above 8.3 (Fig. 4C). The results firmly established that in Glc-6-Pase-␤, His 167 is the amino acid that covalently binds the phosphoryl moiety during Glc-6-Pase-␤ catalysis.

DISCUSSION
Blood glucose homeostasis between meals depends on gluconeogenesis and glycogenolysis. In the terminal step common to both pathways, Glc-6-P is hydrolyzed to glucose by the Glc-6-Pase complex, which is composed of two functionally coupled proteins, the Glc-6-P transporter and a Glc-6-P phosphohydrolase (reviewed in Ref. 14). Until recently, only one functional Glc-6-Pase had been reported. This enzyme, previously called Glc-6-Pase, is expressed in the liver/kidney/intestine consistent with its role in glucose homeostasis (reviewed in Ref. 13). More recently, a ubiquitously expressed Glc-6-Pase-related protein, which was previously called UGRP and thought to be inactive (5), was demonstrated to be catalytically active (6,7). As a result, the proteins have been renamed Glc-6-Pase-␣ (liver/kidney/intestine Glc-6-Pase) and Glc-6-Pase-␤ (UGRP). Glc-6-Pase-␣ is well characterized, but little beyond the kinetic activity (6) and ER localization (6) of Glc-6-Pase-␤ is known. As a first step to characterizing the biological relevance of Glc-6-Pase-␤, we have sought to examine whether it is structurally similar to Glc-6-Pase-␣, a hydrophobic protein embedded in the ER by nine-transmembrane helical domains (10,11). The key active site residues in Glc-6-Pase-␣ have been characterized (21,22) and shown to lie facing into the lumen of the ER. During catalysis, Glc-6-Pase-␣ forms a phosphoryl enzyme intermediate (23)(24)(25)(26) by the formation of a covalent bond between the phosphoryl group of Glc-6-P and His 176 in Glc-6-Pase-␣ (21).
Glc-6-Pase-␤ lacks any apparent ER transmembrane protein retention motif, but double immunofluorescence microscopy studies have shown that it is localized in the ER (6). Theoretical modeling of the topography of Glc-6-Pase-␤ using the TMpred program (32) predicts that Glc-6-Pase-␤ is anchored in the membrane by nine helical domains. Using epitope-tagged Glc-6-Pase-␤ constructs and a protease protection assay by limited trypsin digestion, we found that the C terminus of Glc-6-Pase-␤ is exposed to the cellular cytoplasm but that the N terminus appears protected. Cleavage of the N-terminally FLAG-tagged protein yielded a 14-kDa peptide fragment that is consistent with a nine-transmembrane domain topography. A prediction of tryptic cleavage sites in the FLAG-tagged Glc-6-Pase-␤ by the Ensembl (www.ensembl.org) Peptide Cutter server reveals 20 potential cleavage sites (Fig. 1). Cleavage between residues 136 and 137 (. .VAT(R/A)RSR. .) would yield the closest theoretical fragment (15.5 kDa) to that observed. Given the increased electrophoretic mobility expected from the negatively charged FLAG epitope, this suggests that the FLAG epitope lies on a fragment representing the first 136 residues of Glc-6-Pase-␤. Sequence alignment to the nine-transmembrane domain model of Glc-6-Pase-␣ (10,11) corresponds to the first cytoplasmically exposed cleavage site that lies in the putative second cytoplasmic loop (Fig. 1). There are five other trypsin sites between residues 1 and 136 in Glc-6-Pase-␤, but by this model the first site is predicted to lie within the first transmembrane helix, the second site lies at the junction between cytoplasmic loop 1 and transmembrane helix 2, the next two sites lie within the second transmembrane helix, and the fifth site lies at the beginning of the first luminal loop. When the protease protection assays were repeated using constructs engineered to contain a unique protease cleavage site the results were consistent with a cytoplasmic C terminus and lumenal N terminus. Taken together, Glc-6-Pase-␤ has been shown to have an odd number of transmembrane domains, and proteolytic digests are consistent with the nine-transmembrane topography predicted both by hydropathy analysis (32) and sequence alignment to Glc-6-Pase-␣. In an earlier study, we showed that Glc-6-Pase-␤ couples with the Glc-6-P transporter to form an active Glc-6-Pase complex (6). This is also consistent with the nine-transmembrane domain model that puts the active site residues Arg 79 , His 114 , and His 167 on the lumenal side of the membrane.
To further identify His 167 as the phosphate acceptor, we purified the [ 32 P]phosphoryl-Glc-6-Pase-␤ intermediate by affinity chromatography and gel electrophoresis. Complete cleavage of the intermediate with cyanogen bromide is predicted to create 10 peptide fragments. The His 167 is predicted to lie on a 4.4-kDa peptide with a theoretical pI of 8.6. When cleaved, the intermediate yielded a predominant fragment of 4.4 kDa having a pI above 8.3, which is consistent with His 167 as the phosphate acceptor. In addition, several minor bands of 10, 16, and 18 kDa were also detected bound to the [ 32 P]phosphoryl moiety. Complete cyanogen bromide digestion yielded no fragment greater than 13.2 kDa, whereas the next largest peptides were 9.2 and 5.0 kDa. The minor bands are, however, consistent with partial digests that contain His 167 . The 18-kDa band is consistent with the peptide containing amino acids 127-291, the 16-kDa band is consistent with the peptide containing amino acids 146 -291, and the 10-kDa band is consistent with the peptide containing amino acids 127-192. Our results firmly established that His 167 acts as the nucleophile forming the phosphohistidine enzyme intermediate during Glc-6-Pase-␤ catalysis.
In summary, we showed that the ubiquitous Glc-6-Pase-␤ is anchored in the ER by nine-transmembrane helices oriented with its active site inside the lumen, like the liver/kidney/ intestine Glc-6-Pase-␣. The similarity between Glc-6-Pase-␤ and Glc-6-Pase-␣ also extended to the nature of the active site residues Arg 79 , His 114 , and His 167 and the formation of a covalently bound phosphoryl enzyme intermediate during catalysis. We are now in a position to study the effect of inactivation of Glc-6-Pase-␤ and assess its importance to blood glucose homeostasis and metabolism in general.