Deletion of the Gene Encoding the Ubiquitously Expressed Glucose-6-phosphatase Catalytic Subunit-related Protein (UGRP)/Glucose-6-phosphatase Catalytic Subunit-β Results in Lowered Plasma Cholesterol and Elevated Glucagon*

In liver, glucose-6-phosphatase catalyzes the hydrolysis of glucose-6-phosphate (G6P) to glucose and inorganic phosphate, the final step in the gluconeogenic and glycogenolytic pathways. Mutations in the glucose-6-phosphatase catalytic subunit (G6Pase) give rise to glycogen storage disease (GSD) type 1a, which is characterized in part by hypoglycemia, growth retardation, hypertriglyceridemia, hypercholesterolemia, and hepatic glycogen accumulation. Recently, a novel G6Pase isoform was identified, designated UGRP/G6Pase-β. The activity of UGRP relative to G6Pase in vitro is disputed, raising the question as to whether G6P is a physiologically important substrate for this protein. To address this issue we have characterized the phenotype of UGRP knock-out mice. G6P hydrolytic activity was decreased by ∼50% in homogenates of UGRP–/– mouse brain relative to wild type tissue, consistent with the ability of UGRP to hydrolyze G6P. In addition, female, but not male, UGRP–/– mice exhibit growth retardation as do G6Pase–/– mice and patients with GSD type 1a. However, in contrast to G6Pase–/– mice and patients with GSD type 1a, UGRP–/– mice exhibit no change in hepatic glycogen content, blood glucose, or triglyceride levels. Although UGRP–/– mice are not hypoglycemic, female UGRP–/– mice have elevated (∼60%) plasma glucagon and reduced (∼20%) plasma cholesterol. We hypothesize that the hyperglucagonemia prevents hypoglycemia and that the hypocholesterolemia is secondary to the hyperglucagonemia. As such, the phenotype of UGRP–/– mice is mild, indicating that G6Pase is the major glucose-6-phosphatase of physiological importance for glucose homeostasis in vivo.

In liver, glucose-6-phosphatase catalyzes the hydrolysis of glucose-6-phosphate (G6P) 2 to glucose and inorganic phosphate, the final step in the gluconeogenic and glycogenolytic pathways (1,2). Glucose-6-phosphatase is located in the endoplasmic reticulum membrane and is postulated to exist as a multi-component enzyme system in which a glucose-6-phosphatase catalytic subunit has its catalytic site directed toward the lumen of the endoplasmic reticulum, and a G6P transporter serves to deliver G6P from the cytosol to the active site of the catalytic subunit (3)(4)(5)(6). It also postulates the existence of transporters for inorganic phosphate and glucose that return the reaction products back to the cytosol, but neither transporter has yet been identified. In contrast, the glucose-6-phosphatase catalytic subunit (G6Pase or G6PC) (7,8) and a G6P transporter (9,10) have been well characterized, although their stoichiometry and topological relationships remain unclear (3)(4)(5)(6).
Mutations within G6Pase cause glycogen storage disease (GSD) type 1a, which is characterized by severe hypoglycemia in the post-absorptive state, hepatomegaly associated with excessive glycogen deposition, growth retardation, hepatic adenomas, hyperuricemia, anemia, proteinuria or microalbuminuria, kidney calcifications, osteopenia, increased alkaline phosphatase and ␥-glutamyltransferase activities, and increased serum cholesterol and triglyceride levels (11,12). Interestingly, the spectrum of clinical problems that arise varies between GSD type 1a patients, indicating the influence of modifier genes that affect the actual phenotype (12). Mutations in the G6P transporter give rise to GSD type 1b, which has a similar phenotype to that of GSD type 1a but also additional complications possibly related to independent functions of the G6P transporter in other tissues (11).
Although extracts from all tissues can hydrolyze G6P, enzymatic activity that is selective for G6P and that shows a latency consistent with an endoplasmic reticulum location is highest in liver but is also detected in kidney, intestine, brain, and islets (3)(4)(5)(6). Although glucose-6-phosphatase activity and G6Pase mRNA are detectable in islets, the activity displays distinct kinetic behavior and inhibitor profiles compared with that assayed in hepatic extracts, an observation that might be explained by the existence of a distinct G6Pase isoform in islets, present in addition to the known G6Pase isoform (13). Indeed, we have previously identified two novel cDNAs and genes (13)(14)(15)(16)(17) that encode G6Pase-related proteins that are expressed in islets. The first, islet-specific G6Pase-related protein (IGRP or G6PC2) (13)(14)(15) is expressed specifically in islet ␤ cells (18). The second, ubiquitously expressed G6Pase catalytic subunitrelated protein (UGRP) (16,17), also known as G6Pase-␤ (19) or G6PC3 (20), is expressed in all tissues examined to date. The human IGRP and UGRP genes encode proteins that share a similar size (355 amino acids, 40,552 daltons and 346 amino acids, 38,709 daltons, respectively) and sequence (50 and 36% identity, respectively) to human G6Pase (357 amino acids, 40,487 daltons). The homology extends over the full length of these molecules including the putative transmembrane domains. The exon/intron structures of all three genes are also similar, suggesting that they are evolutionarily related (16,17).
The biological functions of IGRP and UGRP are currently controversial. Overexpression of IGRP by transient transfection of fibroblast or endocrine cell lines has been reported either to not affect the rate of G6P hydrolysis in tissue homogenates (13,15,21,22) or to result in G6P hydrolysis at an extremely low rate (23). Similarly UGRP is reported either to be inactive (16,17) or modestly active (19, 20, 24 -26), exhibiting G6P hydrolytic rates significantly less than that of G6Pase. Questions thus arise as to whether the physiologically important substrate that UGRP hydrolyzes is something other than G6P or whether a low rate of G6P hydrolysis by UGRP might nevertheless influence glucose production in vivo because of the wide tissue distribution of the enzyme. To address these possibilities, this study describes the generation and characterization of UGRP knock-out mice. The results demonstrate that, although deletion of UGRP reduces glucose-6-phosphatase activity in tissue extracts, the phenotype of UGRP knock-out mice is mild, indicating that G6Pase is the major glucose-6phosphatase of physiological importance for glucose homeostasis in vivo.

EXPERIMENTAL PROCEDURES
Animal Care-The animal facilities at Vanderbilt University and Lexicon Genetics meet the standards published by the American Association for the Accreditation of Laboratory Animal Care, and all protocols were approved by the Vanderbilt University Medical Center or Lexicon Genetics Animal Care and Use Committees.
Generation of the UGRP Targeting Vector-The ugrp targeting vector was constructed by using the knock-out shuttle (KOS) system as described previously (27). The Lambda KOS phage library, arrayed into 96 superpools, was screened by PCR using exon 1-specific primers (UGRP-1, 5Ј-AAGGTTCCTC-GACTTGCACC, and UGRP-2, 5Ј-CTCAGCAATGAAGGT-GATCC). The PCR-positive phage superpools were plated and screened by filter hybridization using the 253-bp amplicon derived from primers UGRP-1 and UGRP-2 as a probe. A genomic clone, pKOS18, was isolated from the library screening and confirmed by sequence and restriction analysis. Gene-specific arms (UGRP-3, 5Ј-CCCTACAACGGGC-TCCAGGCTAGAGGACTCCTGATTCACC, and UGRP-4, 5Ј-CCCAAAGGGATAAGGGGACCAGGATGCCCGAAG-GCTTCTC) were appended by PCR to a yeast selection cassette containing the URA3 marker. The yeast selection cassette and pKOS18 were co-transformed into yeast, and clones that had undergone homologous recombination to replace a 228-bp region containing the 218-bp exon 1 and 10 bp of the first intron (17) with the yeast selection cassette were isolated. The yeast cassette was subsequently replaced with a LacZ/Neo selection cassette, by utilizing an SfiI site engineered to the ends of the yeast cassette and the LacZ/Neo cassette, to complete the ugrp targeting vector. The design of this targeting vector is such that the UGRP promoter drives LacZ expression, whereas the polyoma enhancer/herpes simplex virus thymidine kinase promoter drives Neo expression.
Generation of UGRP Knock-out Mice-The targeting vector (25 g) was linearized with NotI and electroporated into 10 7 129SvEv Brd -derived embryonic stem (ES) cells. ES clones that were G418-and FIAU-resistant were isolated, amplified, and screened for homologous recombination events by Southern blot analysis. Briefly, EcoRI-digested DNA was hybridized to a 3Ј external probe corresponding to a 364-bp region downstream of the 3Ј homologous arm. This probe was generated using wild type ES cell genomic DNA, designated Lex2 (see Fig.  1), as the template with primers UTT012-17 (5Ј-GGCCACTG-CATGATCACAG) and UTT012-18 (5Ј-ACTTGGTGAGG-GAAATGTGC). The Southern blot of genomic DNA digested with EcoRI produces a 20-kbp wild type band and a 4-kbp targeted band. The single targeting event was also confirmed by hybridizing SpeI-digested ES DNA with a neo probe (data not shown). Cells from the targeted clone, designated 1B2, were microinjected into C57BL/6J albino blastocysts, resulting in the generation of chimeric mice.
Breeding Strategy for UGRP Knock-out Mice-F1 chimeric (129SvEv Brd ϫ C57BL/6J) mice were interbred to generate F2 wild type, heterozygous, and homozygous knock-out mice. The F2 heterozygous mice were then bred with F1 hybrid (C57BL/ 6J ϫ 129SvEv Brd ) mice. Two male and two female heterozygous mice from this breeding, along with their offspring, were then interbred to generate a mouse colony used in the phenotypic characterization of the effect of UGRP gene deletion.
PCR Genotyping of UGRP Knock-out Mice-Mouse tail DNA was genotyped using PCR to distinguish the wild type and targeted alleles. Primers UTT012-1 (5Ј-AAGGTTC-CTCGACTTGCACC) and UTT012-23 (5Ј-TGAGGCG-GAGACTGTATTTG) were used to amplify a 497-bp product from the wild type allele, whereas primers UTT012-23 and Neo3a (5Ј-GCAGCGCATCGCCTTCTATC) were used to amplify a 376-bp product from the targeted allele. Tail DNA was isolated and purified by standard procedures (28). The wild type and targeted allele fragments were amplified using 100 ng of genomic DNA and AmpliTaq DNA polymerase (Applied Biosystems) with a PCR buffer containing 1.5 mM MgCl 2 (Applied Biosystems) in a final volume of 100 l under the following reaction conditions: 95°C for 30 s, 60°C for 30 s, and 72°C for 30 s for 30 cycles. PCR products were visualized by electrophoresis on 1.5% agarose gels containing ethidium bromide.
UGRP and G6Pase mRNA Expression Analyses-UGRP and G6Pase gene expression were analyzed by RNA blotting. Liver and kidney tissue was immediately frozen in liquid nitrogen after isolation and stored at Ϫ80°C. Total RNA was then isolated using the ToTALLY RNA kit (Ambion) according to the manufacturer's instructions. The samples (20 g) were separated on 1% agarose gels containing 18% formaldehyde and then transferred to Zeta-Probe membranes (Bio-Rad) by capillary transfer using 10ϫ SSC buffer. The blots were hybridized for 16 h at 65°C in ExpressHyb TM hybridization solution (Clontech) with a 32 P-radiolabeled randomly primed probe corresponding to either the full-length mouse UGRP open reading frame (ORF) or the full-length mouse G6Pase ORF. The UGRP ORF was isolated from a previously described pcDNA3.1/V5-His-TOPO plasmid (17) as a BamHI-NotI fragment. A mouse G6Pase cDNA in the pCMV-SPORT6 vector (Image clone number 4237115) was purchased from Open Biosystems. The G6Pase ORF was isolated as an XhoI-EcoRV fragment. The UGRP and G6Pase ORFs (25 ng) were labeled by random oligonucleotide priming with [␣-32 P]dATP using the Stratagene Prime-It II random primer labeling kit according to the manufacturer's instructions (final specific radioactivity, ϳ0.6 Ci⅐nmol Ϫ1 ). After hybridization, the blots were washed for 40 min at 55°C in 2ϫ SSC, 0.05% SDS and then in 0.1ϫ SSC, 0.1% SDS at 55°C for up to 2 h prior to visualization by autoradiography. The blots were then stripped by washing in 0.1ϫ SSC, 0.1% SDS at 100°C for 20 min and reprobed with a labeled mouse ␤-actin probe to correct for variation in mRNA loading. A mouse ␤-actin cDNA in the pCMV-SPORT6 vector (Image clone number 2812162) was purchased from Open Biosystems. The full-length ␤-actin ORF was isolated as a HindIII-EcoRI fragment and labeled as described above. The data were quantitated through the use of a Packard Instant Imager. For some experiments ϳ3 g of poly(A) ϩ RNA was isolated from 400 g of kidney total RNA using the Ambion MicroPoly(A) kit; 0.75 g of purified RNA was then separated by electrophoresis as described above.
Phenotypic Analysis of UGRP Knock-out Mice-The mice were fed a standard rodent chow (LabDiet 5001; 23% protein and 4.5% fat; PMI Nutrition International) and were analyzed at 4 months of age. After a 5 h fast mice were weighed. One hour later mice were anesthetized using isoflurane, and blood samples (ϳ200 l) were isolated from the retro-orbital venous plexus. Blood glucose concentrations were determined using 3 l of whole blood and an Accu-Check Advantage monitor (Roche Applied Science). EDTA (5 l; 0.5 M) was added to the remaining samples that were then spun to isolate plasma. Trasylol (aprotinin; 5 l; Bayer Health Care) was added to the plasma to prevent proteolysis of glucagon. Cholesterol and fatty acids were assayed using a cholesterol reagent kit (Raichem) and a nonesterified fatty acid C Kit (Wako Chemicals, Neuss, Germany), respectively, according to the manufacturer's instructions. Triglyceride and glycerol were assayed using a serum triglyceride determination kit (Sigma) according to the manufacturer's instructions. Insulin and glucagon levels were quantitated using radioimmunoassays (29) by the Vanderbilt Diabetes Center Hormone Assay Core. Glycogen content was determined using the enzymatic method of Chan and Exton (30).
Immunohistochemical Staining-Pancreatic tissue dissected from female wild type and UGRP knock-out mice was fixed in 4% (w/v) paraformaldehyde in phosphate-buffered saline (PBS) for 1 h at 4°C. The tissues were rinsed in PBS for 15 min at 4°C and then stored overnight in PBS at 4°C. The tissues were then sequentially equilibrated with 30% (w/v) sucrose in PBS for 3h, 30% (w/v) sucrose in PBS/(OCT) embedding media (1:1) for 2 h and then embedded in OCT on dry ice for cyrosectioning (7 m). The slides were baked at 37°C for 10 min and hydrated in PBS for 10 min, and antigen retrieval was performed using 10 mM citrate buffer, pH 6.0, for 2 min in a microwave oven. The sections were then washed in PBS and blocked for 1 h with 5% normal goat serum in PBS at room temperature. Double immunofluoresence labeling of insulin and glucagon was undertaken with guinea pig insulin and mouse glucagon antibodies (Sigma) at a dilution of 1:100 with overnight incubation at 4°C in a humid chamber. Subsequently, the sections were washed in PBS and incubated with Cy2 anti-guinea pig and Cy3 anti-mouse antibodies (Jackson Immunolabs, CA). The nuclei were stained with Hoechst 33258. The slides were rinsed and mounted in 30% glycerol in PBS for visualization.
At least 10 islets from a single or 20th consecutive pancreatic section were examined and scored from all groups of mice (n ϭ 6). Images of individual islets were recorded with an Olympus BX51 microscope using a Pixera 600 digital color camera and analyzed with Image-Pro Plus software (Media Cybernetics, Silver Spring, MD). Individual islet area was determined, and the nuclei stained with Hoechst 33258 were counted manually. Subsequently, the Cy3 stained glucagonpositive cells were counted in a double blind manner by two independent observers.
Statistical Analyses-The data were analyzed for differences from the control values, as specified in the figure legends. The statistical comparisons were calculated using a Student's t test: two samples assuming equal variance. The level of significance was as indicated in the figure legends (two-sided test).

RESULTS
Generation of UGRP Knock-out Mice-A modified mouse UGRP allele in which exon 1 and 10 bp of the first intron (17) were replaced by a LacZ/Neo cassette was created by gene targeting in ES cells (Fig. 1A). The design of the targeting vector was such that the UGRP promoter drives LacZ expression, whereas the polyoma enhancer/herpes simplex virus thymidine kinase promoter drives Neo expression. Correct gene targeting was confirmed by Southern blot (data not shown) and PCR (Fig.  1B) analysis. ES cells containing this targeted allele were then used to generate mice.
Genotype analysis of 194 3-week old pups generated by cross-breeding heterozygous UGRP Ϫ/ϩ mice demonstrated that 43 mice were UGRP ϩ/ϩ , 101 were UGRP Ϫ/ϩ , and 50 were UGRP Ϫ/Ϫ , consistent with a pattern of Mendelian inheritance. The ratio of male to female mice was 88:106. Cross-breeding experiments revealed that male and female homozygous UGRP Ϫ/Ϫ mice are fertile, indicating that UGRP expression in testis, prostate, and ovary (17) is not essential for reproduction.
Biochemical analyses confirmed that the UGRP gene was not expressed in UGRP Ϫ/Ϫ mice. Thus, RNA blotting demonstrated the absence of UGRP mRNA (Fig. 1C) and Western blot analyses of brain and testis extracts from wild type and UGRP Ϫ/Ϫ mice, using an anti-UGRP 13-amino acid C-terminal peptide antibody (17), showed the loss of a 30-kDa component (data not shown) that corresponds in size to UGRP as previously demonstrated in transfected COS 7 cells (17). Phosphatase activity was assessed in brain, a tissue in which mouse UGRP is relatively abundantly expressed at the mRNA level (17). UGRP activity was calculated as the difference between activities obtained with the wild type and knock-out mouse tissue. Deletion of UGRP resulted in a highly significant decrease in glucose-6-phosphate hydrolytic activity (3.8 Ϯ 0.6 versus 1.4 Ϯ 0.5 nmol/min/mg, n ϭ 4; p Ͻ 0.01) consistent with the ability of UGRP to hydrolyze G6P (19,20). In contrast, assays of transfected COS cell extracts in the same experiments showed little change in G6P hydrolytic activity in UGRP transfectants compared with the vector-only control (4.0 Ϯ 0.6 versus 2.8 Ϯ 0.8 nmol/min/mg, n ϭ 4). Under the same conditions, when G6Pase was expressed at levels equivalent to that of UGRP (data not shown), G6Pase transfectants showed an activity of 373.4 Ϯ 14 nmol/min/mg. The reason for this discrepancy is unclear; it may indicate that UGRP does not fold correctly when expressed in COS 7 cells or that these cells lack a co-factor required for glucose-6-phosphatase activity.
Phenotypic Analysis of UGRP Knock-out Mice-UGRP Ϫ/Ϫ mice exhibited normal behavior at birth and up to 1 year in age, and no morphological changes were apparent. However, at 4 months of age female UGRP Ϫ/ϩ mice (Table 1) and UGRP Ϫ/Ϫ mice (Table 1 and Fig. 2A) were both shorter than female wild type mice. In addition, female UGRP Ϫ/Ϫ mice weighed less than female wild type mice (Table 1 and Fig. 2B). No differences were seen in the weights or lengths of male UGRP Ϫ/Ϫ and wild type mice (Table 1).
To determine the effect of removing UGRP on metabolism, blood glucose, and plasma cholesterol, triglyceride, glycerol, nonesterified free fatty acid, insulin, and glucagon concentrations were assayed at 4 months of age following a 6-h fast. No differences were apparent in any of these parameters in comparisons of male UGRP Ϫ/Ϫ and wild type mice (Table 1). Similarly, there were no differences in blood glucose or plasma triglyceride, glycerol, nonesterified free fatty acid, and insulin concentrations between female UGRP Ϫ/Ϫ and wild type mice (Table 1). However, the plasma cholesterol concentration was decreased by ϳ20% in female UGRP Ϫ/Ϫ mice relative to wild type female mice (Table 1 and Fig. 2C). In contrast, the plasma glucagon concentration was increased ϳ60% in female UGRP Ϫ/Ϫ mice relative to wild type female mice (Table 1 and Fig. 2D). Glucagon is known to suppress plasma cholesterol levels through a variety of mechanisms (31,32). Therefore, we hypothesize that the hypocholesterolemia detected is secondary to the hyperglucagonemia. Fig. 2 shows that the values for length, weight, cholesterol, and glucagon are quite variable between individual animals, which presumably reflects the fact that the mice are a mixed 129SvEv Brd ϫ C57BL/6J background, and there are multiple other genes that affect these parameters.
The growth retardation seen in female UGRP Ϫ/Ϫ mice is also a key characteristic of patients with GSD type 1a (11,12) and of G6Pase Ϫ/Ϫ mice (33), although the molecular basis for growth retardation in GSD type 1a and G6Pase Ϫ/Ϫ mice remains to be determined (34). Given the well established role for glucagon in the maintenance of blood glucose levels (35), the simplest explanation for the elevated glucagon levels seen in female UGRP Ϫ/Ϫ mice is that this increase serves to prevent hypoglycemia.
Two additional studies were performed to investigate the hypothesis that the elevated glucagon detected in female UGRP Ϫ/Ϫ mice serves to maintain blood glucose levels. Specifically, hepatic glycogen content and G6Pase gene expression were assessed in six female wild type and UGRP Ϫ/Ϫ mice following a 6-h fast. Glucagon stimulates both hepatic glycogenolysis (2,36) and G6Pase gene expression (37). For this study we purposefully chose to use six UGRP Ϫ/Ϫ mice that exhibited elevated glucagon levels at 4 months of age (Fig. 2D) so as to magnify the anticipated changes in hepatic glycogen content FIGURE 1. Generation and authentication of UGRP knock-out mice. A, strategy used to generate UGRP knock-out mice by homologous recombination in ES cells. Schematic representations of the wild type murine ugrp locus and the targeting construct are shown. A 228-bp region representing the first coding exon (E1) and 10 bp of the first intron was replaced by a LacZ/neo cassette. The 5Ј and 3Ј homologous regions of the targeting construct were 9.0 and 1.8 kbp, respectively. Details of the targeting vector construction are described under "Experimental Procedures." X and G indicate KOS-specific primer-binding sites (27). B, genotyping of UGRP knock-out mice using PCR. Primers UTT012-1 and UTT012-23 were used to amplify a 497-bp product from the wild type (WT) allele, whereas primers UTT012-23 and Neo3a were used to amplify a 376-bp product from the targeted (KO) allele. Heterozygous (H) mice therefore have both loci. L, ladder. C, analysis of mouse UGRP mRNA expression by RNA blotting. Total RNA was isolated from wild type (WT) and UGRP knock-out (KO) mouse kidneys, and blotting analysis was then performed as described under "Experimental Procedures." The representative autoradiographs show the results of RNA blotting using the mouse UGRP ORF (top panel) or mouse ␤-actin ORF (bottom panel) as the labeled probes. DECEMBER 29, 2006 • VOLUME 281 • NUMBER 52 and G6Pase gene expression. As seen in the study with a larger number of animals (Table 1), there was little difference in blood glucose levels between groups (Fig. 3A), whereas plasma glucagon was markedly elevated in the UGRP Ϫ/Ϫ mice (Fig. 3B). Surprisingly, despite this elevation in plasma glucagon, there was no difference in hepatic glycogen content (Fig. 3C) or G6Pase gene expression (Fig. 3D) between female wild type and UGRP Ϫ/Ϫ mice. Potential explanations for these paradoxical observations are presented under "Discussion."

UGRP/G6Pase-␤ Knock-out Mice
Previous studies have shown that, in mice, genetic changes that affect glucagon processing and action are associated with long term adaptive changes in pancreatic islet alpha cell number that presumably reflect an adaptation designed to prevent hypoglycemia. Thus, in prohormone convertase 2 knock-out mice glucagon processing is impaired (38), whereas in glucagon receptor knock-out mice glucagon action is impaired (39). In both cases, there is a marked proliferation of alpha cells (38,39). To investigate whether the hyperglucagonemia seen in female UGRP Ϫ/Ϫ mice is also associated with alpha cell proliferation, immunohistochemical analyses were performed. The pancreas samples used in this experiment were isolated from the same mice as used for the analysis of hepatic glycogen abundance (Fig. 3C) and G6Pase gene expression (Fig. 3D). A comparison of pancreas tissue isolated from female wild type and UGRP Ϫ/Ϫ mice revealed that alpha cell abundance was increased in the latter (Fig. 4). This observation suggests that the hyperglucagonemia seen in female UGRP Ϫ/Ϫ mice is achieved, at least in part, through alpha cell hyperplasia rather than just enhanced glucagon secretion by a wild type alpha cell population. Interestingly, there was no evidence of alpha cell invasion into the core of the islet (Fig. 4A), which contrasts with what is seen in islets from prohormone convertase 2 and glucagon receptor knock-out mice (38,39).

DISCUSSION
To assess whether G6P may be a physiologically important substrate for UGRP in vivo, we have generated and characterized the phenotype of UGRP Ϫ/Ϫ mice. G6P hydrolytic activity is decreased in homogenates of UGRP Ϫ/Ϫ mouse brain tissue relative to wild type tissue, consistent with the ability of UGRP to hydrolyze G6P. In addition, female, although not male, UGRP Ϫ/Ϫ mice, like G6Pase Ϫ/Ϫ mice and patients with GSD type 1a, exhibit growth retardation (Table 1). However, this effect (ϳ10% reduction in body weight at 4 months) is relatively minor compared with that seen in G6Pase Ϫ/Ϫ mice (ϳ50% reduction in body weight at 1 month) (33). In contrast to G6Pase Ϫ/Ϫ mice and patients with GSD type 1a, UGRP Ϫ/Ϫ mice exhibit no change in hepatic glycogen content (Fig. 2C), blood glucose, or plasma triglyceride levels ( Table 1). Moreover, plasma cholesterol levels are actually decreased in female mice (Table 1), whereas they are markedly elevated in G6Pase Ϫ/Ϫ mice and patients with GSD type 1a. Although UGRP Ϫ/Ϫ mice are not hypoglycemic, their plasma glucagon levels are significantly elevated (Table 1). We hypothesize that this hyperglucagonemia prevents hypoglycemia and that the hypocholesterolemia is secondary to the hyperglucagonemia. As such, although UGRP can hydrolyze G6P, these results indicate that it plays a minor role in endogenous glucose produc-

UGRP/G6Pase-␤ Knock-out Mice
tion and that G6Pase is the major glucose-6-phosphatase of physiological importance for glucose homeostasis in vivo. Indeed, in GSD type 1a patients, endogenous glucose production is markedly impaired (40). Moreover, if UGRP was important for endogenous glucose production, one might expect that the phenotype of G6P transporter knock-out mice would be much more severe than that of G6Pase knock-out mice because both G6Pase and UGRP action would be impaired. However, this is not the case (41).
Although UGRP clearly plays a minor role in endogenous glucose production in otherwise normal mice, Chou and colleagues (19, 24 -26) have proposed that its contribution to this process could become significant in patients with GSD type 1a. This concept is based on the observation that the increased abundance of UGRP concomitant with an accumulation of muscle mass could explain the observed improvement in endogenous glucose production as patients with GSD type 1a age.
Although we hypothesize that the hyperglucagonemia detected in female UGRP Ϫ/Ϫ mice prevents the onset of hypoglycemia, paradoxically no changes in hepatic glycogen (Fig. 3C) or G6Pase gene expression (Fig. 3D) were detected. A trivial explanation for these observations that might explain this apparent paradox is that the sample size analyzed was too low to detect significant changes. Because the mice are a mixed 129SvEvBrd ϫ C57BL/6J background, the effect of modifier genes contributes to variability in parameters (Fig. 2) and may obscure small differences. A more interesting explanation for the paradoxical absence of a change in hepatic glycogen levels and G6Pase gene expression is that it reveals an important role for UGRP function in the liver. Hepatic UGRP expression is detected at the RNA level in both humans and mice (16,17). If deletion of the UGRP gene does FIGURE 2. Phenotypic characterization of female wild type and UGRP knock-out mice. At 4 months of age female wild type (WT) and UGRP knock-out (KO) mice were fasted for 5 h and then weighed. One hour later mice were anesthetized, their lengths were measured, and blood samples were isolated. Cholesterol and glucagon levels in plasma were then quantitated as described under "Experimental Procedures." The results show the individual values obtained in wild type and knock-out female mice for weight (A), length (B), cholesterol (C), and glucagon (D). The data were derived from Table 1; the n value appears reduced because data points with identical values are overlaid. FIGURE 3. Analysis of glycogen levels and G6Pase gene expression in female wild type and UGRP knockout mice. Wild type (WT) and UGRP knock-out (KO) female mice older than 4 months of age were fasted for 6 h and anesthetized, and blood samples were isolated. The mice were then euthanized, and liver samples were immediately frozen in liquid nitrogen. Blood glucose (A), plasma glucagon (B) and hepatic glycogen levels (C), and G6Pase gene expression (D) were quantitated as described under "Experimental Procedures." The results represent the means Ϯ S.E. of values obtained from six wild type and six knock-out animals. *, p Ͻ 0.05 versus WT. DECEMBER 29, 2006 • VOLUME 281 • NUMBER 52 directly affect hepatocyte function, there are two broad possibilities that would then explain our observations. Either the generation of cAMP in hepatocytes in response to glucagon signaling may be impaired, as seen in ob/ob mice (42), or the initial events in hepatic glucagon signaling may be normal, but mobilization of glycogen stores and stimulation of G6Pase gene expression may be selectively impaired. In either case the hyperglucagonemia could act to maintain euglycemia indirectly through effects on other tissues such as heart and adipose tissue (43). Interestingly, a similar situation is seen in GLUT2 knock-out mice in which fasting induces a 10-fold increase in plasma glucagon, but there is little change in glycogen breakdown (44). Further experimentation will be required to resolve these possibilities. . Immunohistochemical staining of female wild type and UGRP knock-out mouse pancreas with antisera raised to insulin and glucagon. Pancreas tissue was isolated from the same six female wild type (WT) and six female UGRP knock-out (KO) mice as described for Fig. 3. Fixation, preparation of mouse pancreatic slices, immunohistochemical staining with antibodies raised to insulin and glucagon, and data quantitation were then performed as described under "Experimental Procedures." A shows representative results, whereas B shows mean data Ϯ S.E. of values obtained from three wild type and knock-out mice. *, p Ͻ 0.05 versus wild type.