Do rodents have a gene encoding glycogenin-2, the liver isoform of the self-glucosylating initiator of glycogen synthesis?

The discovery of a second human gene, GYG2, encoding a liver-specific isoform of glycogenin, the self-glucosylating initiator of glycogen biosynthesis, raised the possibility for differential controls of this protein in liver and muscle. The new protein, glycogenin-2, had several properties similar biochemically to the muscle isoform, glycogenin-1, but unlike glycogenin-1, stable expression in fibroblasts led to a significant overaccumulation of glycogen. Ensuing attempts to generate reagents suitable for use with rodents, to examine the physiological regulation of glycogenin-2 by nutritional and hormonal factors, have been unsuccessful. Proof of a negative is difficult but the weight of the evidence is beginning to mitigate against the existence of a second glycogenin gene in rodents leading us to hypothesize that the presence of the GYG2 gene is limited to primates.

Surprises of genetic engineering: a possible model of polyglucosan body disease.

BACKGROUND: The authors previously reported the generation of a knockout mouse model of Pompe disease caused by the inherited deficiency of lysosomal acid alpha-glucosidase (GAA). The disorder in the knockout mice (GAA-/-) resembles the human disease closely, except that the clinical symptoms develop late relative to the lifespan of the animals. In an attempt to accelerate the course of the disease in the knockouts, the authors increased the level of cytoplasmic glycogen by overexpressing glycogen synthase (GSase) or GlutI glucose transporter. METHODS: GAA-/- mice were crossed to transgenic mice overexpressing GSase or GlutI in skeletal muscle. RESULTS: Both transgenics on a GAA knockout background (GS/GAA-/- and GlutI/GAA-/-) developed a severe muscle wasting disorder with an early age at onset. This finding, however, is not the major focus of the study. Unexpectedly, the mice bearing the GSase transgene, but not those bearing the GlutI transgene, accumulated structurally abnormal polysaccharide (polyglucosan) similar to that observed in patients with Lafora disease, glycogenosis type IV, and glycogenosis type VII. Ultrastructurally, the periodic acid-Schiff (PAS)-positive polysaccharide inclusions were composed of short, amorphous, irregular branching filaments indistinguishable from classic polyglucosan bodies. The authors show here that increased level of GSase in the presence of normal glycogen branching enzyme (GBE) activity leads to polyglucosan accumulation. The authors have further shown that inactivation of lysosomal acid alpha-glucosidase in the knockout mice does not contribute to the process of polyglucosan formation. CONCLUSIONS: An imbalance between GSase and GBE activities is proposed as the mechanism involved in the production of polyglucosan bodies. The authors may have inadvertently created a "muscle polyglucosan disease" by simulating the mechanism for polyglucosan formation.

Glycogen synthase sensitivity to insulin and glucose-6-phosphate is mediated by both NH2- and COOH-terminal phosphorylation sites.

In skeletal muscle, insulin activates glycogen synthase by reducing phosphorylation at both NH2- and COOH-terminal sites of the enzyme and by elevating the levels of glucose-6-phosphate, an allosteric activator of glycogen synthase. To study the mechanism of regulation of glycogen synthase by insulin and glucose-6-phosphate, we generated stable Rat-1 fibroblast clones expressing rabbit muscle glycogen synthase with Ser-->Ala substitutions at key phosphorylation sites. We found that 1) elimination of the phosphorylation of either NH2- or COOH-terminal sites did not abolish insulin stimulation of glycogen synthase; 2) mutations at both Ser-7 and Ser-640 were necessary to bypass insulin activation; 3) mutation at Ser-7, coupled with the disruption of the motif for recognition by glycogen synthase kinase-3 (GSK-3), did not eliminate the insulin effect; and 4) mutation of either Ser-7 or Ser-640 increased the sensitivity of glycogen synthase to glucose 6-phosphate >10-fold. We conclude that Ser-7 and Ser-640 are both involved in mediating the response of glycogen synthase to insulin and activation by glucose 6-phosphate. In Rat-1 fibroblasts, GSK-3 action is not essential for glycogen synthase activation by insulin, and GSK-3-independent mechanisms also operate.

Control of glycogen synthesis is shared between glucose transport and glycogen synthase in skeletal muscle fibers.

The effects of transgenic overexpression of glycogen synthase in different types of fast-twitch muscle fibers were investigated in individual fibers from the anterior tibialis muscle. Glycogen synthase was severalfold higher in all transgenic fibers, although the extent of overexpression was twofold greater in type IIB fibers. Effects of the transgene on increasing glycogen and phosphorylase and on decreasing UDP-glucose were also more pronounced in type IIB fibers. However, in any grouping of fibers having equivalent malate dehydrogenase activity (an index of oxidative potential), glycogen was higher in the transgenic fibers. Thus increasing synthase is sufficient to enhance glycogen accumulation in all types of fast-twitch fibers. Effects on glucose transport and glycogen synthesis were investigated in experiments in which diaphragm, extensor digitorum longus (EDL), and soleus muscles were incubated in vitro. Transport was not increased by the transgene in any of the muscles. The transgene increased basal [(14)C]glucose into glycogen by 2.5-fold in the EDL, which is composed primarily of IIB fibers. The transgene also enhanced insulin-stimulated glycogen synthesis in the diaphragm and soleus muscles, which are composed of oxidative fiber types. We conclude that increasing glycogen synthase activity increases the rate of glycogen synthesis in both oxidative and glycolytic fibers, implying that the control of glycogen accumulation by insulin in skeletal muscle is distributed between the glucose transport and glycogen synthase steps.

Novel aspects of the regulation of glycogen storage.

The storage polysaccharide glycogen is widely distributed in nature, from bacteria to mammals. Study of its regulated accumulation has resulted in the discovery or elaboration of several important biochemical principles. Many aspects of the control of glycogen storage still remain poorly understood and glycogen metabolism continues to provide interesting models of more general relevance.

Glycogenin-2, a novel self-glucosylating protein involved in liver glycogen biosynthesis.

Glycogenin is a self-glucosylating protein involved in the initiation phase of glycogen biosynthesis. A single mammalian gene had been reported to account for glycogen biogenesis in liver and muscle, the two major repositories of glycogen. We describe the characterization of novel forms of glycogenin, designated glycogenin-2 (GN-2), encoded by a second gene that is expressed preferentially in certain tissues, including liver, heart, and pancreas. Cloning of cDNAs encoding glycogenin-2 indicated the existence of multiple species, including three liver forms (GN-2alpha, GN-2beta, and GN-2gamma) generated in part by alternative splicing. Overall, GN-2 has 40-45% identity to muscle glycogenin but is 72% identical over a 200-residue segment thought to contain the catalytic domain. GN-2 expressed in Escherichia coli or COS cells is active in self-glucosylation assays, and self-glucosylated GN-2 can be elongated by skeletal muscle glycogen synthase. Antibodies raised against GN-2 produced in E. coli recognized proteins of Mr approximately 66,000 present in extracts of rat liver and in cultured H4IIEC3 hepatoma cells. In H4IIEC3 cells, most of the GN-2 was present as a free protein but some was covalently associated with glycogen fractions and was only released by treatment with alpha-amylase. H4IIEC3 cells also expressed the muscle form of glycogenin (glycogenin-1), which was attached to a chromatographically separable glycogen fraction.

Glycogen biogenesis in rat 1 fibroblasts expressing rabbit muscle glycogenin.

Glycogenin, a self-glucosylating protein involved in the initiation of glycogen biosynthesis, varies in intracellular concentration from barely detectable in liver to a high level in muscle. The effect of increasing the glycogenin level on glycogen synthesis was studied in rat 1 fibroblasts stably overexpressing rabbit muscle glycogenin. In the presence of glucose, all of the expressed glycogenin was attached to polysaccharide and the free protein could only be detected by western blot analysis after incubation of cells in a glucose-depleted medium or treatment of the cell extract with alpha-amylase. In control cells, increased extracellular glucose concentrations promoted translocation of glycogen synthase from the soluble to the pellet fraction with an increase in the associated glycogen. Overexpression of glycogenin did not affect total intracellular glycogen and glycogen synthase levels at any concentration of glucose but significantly reduced glucose-induced accumulation of insoluble glycogen and translocation of glycogen synthase. Immunofluorescence analysis revealed a diffuse cytoplasmic distribution of glycogenin expressed in rat 1 cells. In rat 1 cells incubated with glucose, discrete deposits of glycogen were detected by staining with HIO4/Schiff but this was eliminated by overexpressing glycogenin. Analysis of [14C]glucose- or [35S]methionine-labeled extracts from glycogenin-expressing cells by continuous polyacrylamide gel electrophoresis and by two-dimensional gel electrophoresis revealed a continuum of glycogenin-containing species from low molecular mass to sizes significantly greater than 400 kDa. We conclude that (a) overexpression of glycogenin does not enhance glycogen synthesis but causes production of more, smaller, glycogen molecules with a concomitant change in their intracellular localization; (b) glycogenin and elevated glucose have opposing effects on the distribution of glycogenin and glycogen synthase in rat 1 cells; and (c) the biogenesis of glycogen in rat 1 cells occurs without the accumulation of any major intermediate form.

Increased glycogen accumulation in transgenic mice overexpressing glycogen synthase in skeletal muscle.

To investigate the role of glycogen synthase in controlling glycogen accumulation, we generated three lines of transgenic mice in which the enzyme was overexpressed in skeletal muscle by using promoter-enhancer elements derived from the mouse muscle creatine kinase gene. In all three lines, expression was highest in muscles composed primarily of fast-twitch fibers, such as the gastrocnemius and anterior tibialis. In these muscles, glycogen synthase activity was increased by as much as 10-fold, with concomitant increases (up to 5-fold) in the glycogen content. The uridine diphosphoglucose concentrations were markedly decreased, consistent with the increase in glycogen synthase activity. Levels of glycogen phosphorylase in these muscles increased (up to 3-fold), whereas the amount of the insulin-sensitive glucose transporter 4 either remained unchanged or decreased. The observation that increasing glycogen synthase enhances glycogen accumulation supports the conclusion that the activation of glycogen synthase, as well as glucose transport, contributes to the accumulation of glycogen in response to insulin in skeletal muscle.

Rate-determining steps in the biosynthesis of glycogen in COS cells.

Consistent with previous results, overexpression of rabbit skeletal muscle glycogen synthase in COS cells did not lead to overaccumulation of glycogen unless activating Ser-->Ala mutations were present at key regulatory phosphorylation sites 2 (Ser7) and 3a (Ser644) in the enzyme. In addition, we found that expression of glycogenin, glycogen branching enzyme, or UDP-glucose pyrophosphorylase alone in COS cells had no effect on the glycogen level. However, coexpression of the hyperactive 2,3a glycogen synthase mutant with either glycogenin or UDP-glucose pyrophosphorylase led to higher glycogen accumulation than that obtained from the expression of glycogen synthase alone. Coexpression of glycogenin with the 2,3a mutant of glycogen synthase led to the appearance of glycogenin with a lower molecular weight suggestive of reduced glucosylation. Increased glycogen synthesis may lead to competition between glycogenin and glycogen synthase for their common substrate UDP-glucose. In summary, we conclude that (i) glycogen synthase is a primary rate-limiting enzyme of glycogen biosynthesis in COS cells, (ii) that phosphorylation of glycogen synthase is regulatory for glycogen accumulation, and (iii) once glycogen synthase is activated, the reaction mediated by UDP-glucose pyrophosphorylase can become rate-determining.

Multiple mechanisms for the phosphorylation of C-terminal regulatory sites in rabbit muscle glycogen synthase expressed in COS cells.

Glycogen synthase can be inactivated by sequential phosphorylation at the C-terminal residues Ser652 (site 4), Ser648 (site 3c), Ser644 (site 3b) and Ser640 (site 3a) catalysed by glycogen synthase kinase-3. In vitro, glycogen synthase kinase-3 action requires that glycogen synthase has first been phosphorylated at Ser656 (site 5) by casein kinase II. Recently we demonstrated that inactivation is linked only to phosphorylation at site 3a and site 3b, and that, in COS cells, modification of these sites can occur by alternative mechanisms independent of any C-terminal phosphorylations [Skurat and Roach (1995) J. Biol. Chem. 270, 12491-12497]. To address these mechanisms multiple Ser-->Ala mutations were introduced in glycogen synthase such that only site 3a or site 3b remained intact. Additional mutation of Arg637-->Gln eliminated phosphorylation of site 3a, indicating that Arg637 may be important for recognition of site 3a by its corresponding protein kinase(s). Similarly, additional mutation of Pro645-->Ala eliminated phosphorylation of site 3b, indicating a possible involvement of 'proline-directed' protein kinase(s). Mutation of Arg637 alone did not activate glycogen synthase as expected from the loss of phosphorylation at site 3a. Rather, mutation of both Arg637 and the Ser-->Ala substitution at site 3b was required for substantial activation. The results suggest that sites 3a and 3b can be phosphorylated independently of one another by distinct protein kinases. However, phosphorylation of site 3b can potentiate phosphorylation of site 3a, by an enzyme such as glycogen synthase kinase-3.

Phosphorylation of sites 3a and 3b (Ser640 and Ser644) in the control of rabbit muscle glycogen synthase.

Glycogen synthase kinase-3 inactivates rabbit muscle glycogen synthase by sequential phosphorylation of four COOH-terminal residues Ser652 (site 4), Ser648 (site 3c), Ser644 (site 3b), and Ser640 (site 3a). Effective recognition of glycogen synthase by glycogen synthase kinase-3 occurs only after the phosphorylation of Ser656 (site 5) catalyzed by casein kinase II. The present study addresses specifically the role of sites 3a and 3b in the regulation of glycogen synthase expressed in COS cells. Simultaneous Ser-->Ala substitutions at sites 3 a, b and c, 4, and 5 in the same protein molecule eliminated 32P labeling in the proteolytic fragment Arg634-Lys682, which contains these sites. This mutant enzyme (which also had a Ser-->Ala substitution at site 2 in the NH2 terminus) had a -/+ glucose-6-P activity ratio of approximately 0.8, similar to that of totally dephosphorylated enzyme. Reinstating serine residues at either site 3a or site 3b restored labeling in the Arg634-Lys682 peptide and caused a decrease in the activity ratio to 0.4-0.6. When both sites 3a and 3b were reintroduced, there was complete inactivation of the enzyme. Thus, sites 3a and 3b are sufficient for the inactivation of glycogen synthase and act synergistically to control activity. This investigation demonstrates the existence of an alternate mechanism for the phosphorylation of sites 3a and 3b that does not depend on prior phosphorylation of site 5.

Rabbit skeletal muscle glycogen synthase expressed in COS cells. Identification of regulatory phosphorylation sites.

Rabbit skeletal muscle glycogen synthase contains multiple sites for phosphorylation. To investigate the relative importance of these sites, the enzyme was overexpressed in COS M9 cells, and Ser-->Ala mutations were introduced singly, or in combinations, at nine known phosphorylation sites. Overexpressed wild-type enzyme had a very low +/- glucose-6-P activity ratio of approximately 0.01, indicative that the glycogen synthase is in a highly phosphorylated state. No single Ser-->Ala mutation was able to cause a substantial increase in activity ratio; rather, simultaneous mutation at both NH2- and COOH-terminal sites was needed. The most effective combinations were mutations at site 3a (Ser-640) or site 3b (Ser-644) together with site 2 (Ser-7). The results were consistent with site 2 phosphorylation being a prerequisite for phosphorylation of site 2a (Ser-10). Mutation of site 5 (Ser-656) perturbed COOH-terminal phosphorylation but did not prevent inactivation. Expression of the most active mutants correlated with increased glycogen accumulation in the COS M9 cells. In summary, we conclude that (i) the sites most important for activating the enzyme are sites 2, 2a, 3a, and 3b; (ii) removal of phosphate from both NH2- and COOH-terminal sites is required for activation; and (iii) sites 3a and/or 3b can be phosphorylated in COS cells by mechanisms that do not depend on phosphorylation of site 5.

Initiation of glycogen synthesis. Control of glycogenin by glycogen phosphorylase.

Glycogen biosynthesis involves a specific initiation event, mediated by a specialized protein, glycogenin. Glycogenin undergoes self-glucosylation to generate an oligosaccharide primer, which, when long enough, supports the action of glycogen synthase to elongate the polysaccharide chain, leading ultimately to the formation of glycogen. We report that primed glycogenin is also a substrate for glycogen phosphorylase. Phosphorylase removed glucose from the oligosaccharide attached to glycogenin in a phosphorolysis reaction that required phosphate and produced glucose 1-phosphate. The phosphorylated form, phosphorylase a, was much more effective than the dephosphorylated phosphorylase b. However, in the presence of the allosteric effector AMP, phosphorylase b also catalyzed the phosphorolysis reaction. Glucose, an allosteric inhibitor of phosphorylase, inhibited the reaction. Glycogen, but not a short oligosaccharide (maltopentaose), also inhibited the reaction. Treatment of fully primed glycogenin with phosphorylase converted the glycogenin to a form with slightly lower apparent molecular weight, which was less effective as a substrate for glycogen synthase. These results suggest a novel role for phosphorylase in the control of glycogen biosynthesis. We propose that the glucosylation level of glycogenin would be determined by the balance between the self-glucosylation reaction and the opposing action of phosphorylase. The level of glucosylation would in turn determine whether or not glycogenin was an effective primer for glycogen synthase. In this way, several known controls of phosphorylase activity, such as epinephrine, glucagon, and insulin, could influence not only the elongation/degradation stage of glycogen metabolism but also its initiation.

Glucose control of rabbit skeletal muscle glycogenin expressed in COS cells.

Glycogenin is a self-glucosylating protein involved in the initiation of glycogen synthesis. Rabbit skeletal muscle glycogenin, transiently expressed in COS cells, was found exclusively in the low speed supernatant fraction, with M(r) 37-38,000. The protein was capable of self-glucosylation and was, if suitably primed, an effective substrate for glycogen synthase. Rabbit muscle glycogen synthase was similarly expressed, to a level 7-10-fold over the endogenous activity. Most of the expressed protein was found in the low speed pellet fraction. However, when co-expressed with glycogenin, a significant increase in the proportion of glycogen synthase in the soluble fraction was observed. Therefore, glycogenin interacts with glycogen synthase in the cell and redistributes the synthase to the soluble fraction. Co-expression of an inactive form of glycogenin did not affect glycogen synthase localization. The expressed glycogenin could be detected as two distinguishable species, differing slightly in electrophoretic mobility, depending on the glucose concentration of the cell culture medium. At 25 mM glucose, a form of M(r) 38,000 was observed; however, upon transfer to 5 mM glucose, it converted to a species of slightly lower M(r). The M(r) of the 38,000-dalton species could be also be reduced by treatment of the cell extract with alpha-amylase. It was additionally found that the 38,000-dalton glycogenin was a much more effective glycogen synthase substrate than the lower M(r) species. These results, therefore, raise the possibility of a novel mechanism for the control of glycogen metabolism in which glucose levels would regulate the glucosylation state of glycogenin, which in turn would determine glycogenin's efficacy as a substrate for elongation by glycogen synthase.

Nitroprusside stimulates the cysteine-specific mono(ADP-ribosylation) of glyceraldehyde-3-phosphate dehydrogenase from human erythrocytes.

In human erythrocyte membranes incubated with [adenylate-32P]NAD the 36 kDa protein is predominantly labeled. The labeling is greatly stimulated by nitroprusside in the presence of dithiothreitol. We have purified the 36 kDa protein and identified this modification as cysteine-specific mono(ADP-ribosylation) because: (i) labeling occurred only when [32P]NAD was replaced by adenine[U-14C]NAD, but not by [carbonyl-14C]NAD; (ii) treatment of the prelabeled protein with snake venom phosphodiesterase led to releasing 5'-[32P]AMP; (iii) the bond between the protein and the nucleotide was hydrolyzed by HgCl2, but was resistant to hydroxylamine. The 36 kDa protein reacted on Western blots with two different monoclonal antibodies (MAbs) against glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and was immunoprecipitated by both MAbs.

[Isolation of GTP-binding Gs-protein liberated from human erythrocyte membranes under the action of fluoride ions]. / Vydelenie reguliatornogo GTP-sviayvaiushchego Gs-belka, osvobozhdaiushchegosia iz membran éritrotsitov cheloveka pod deistviem ftorid-iona.

Human erythrocyte membranes were incubated in the presence of sodium fluoride. After centrifugation at 30,000 g for 30 min the supernatant was able to stimulate the catalytic subunit of adenylate cyclase. The stimulatory factor was purified from the supernatant of fluoride-treated membranes by three subsequent chromatographic steps including DEAE-Sephacel ion-exchange chromatography in the absence of detergent, gel-filtration on Ultrogel AcA 44 in the presence of 1% sodium cholate and phenyl-Sepharose CL/4B hydrophobic chromatography. The final preparation showed approximately 120-fold purification in stimulatory activity over the initial extract and contained two polypeptides (Mr 42 kDa and 36 kDa). The stimulator activity of the preparation was inhibited by 60% by beta gamma-subunits of the GTP-binding protein of bovine brain membranes, G0. The data obtained suggest that the regulatory GTP-binding stimulatory protein of adenylate cyclase, GS, dissociates from human erythrocyte membranes as a result of fluoride-ion treatment.

[Substrates of ADP-ribosyltransferase of human erythrocytes]. / Izuchenie substratov ADP-riboziltransferazy éritrotsitov cheloveka.

Incubation of human erythrocyte membranes with [32P]NAD resulted in the label incorporation into the 37 kDa and 41 kDa proteins as determined by SDS-PAAG electrophoresis with subsequent autoradiography. Treatment of labeled membranes with HgCl2 caused a significant decrease of the protein band intensity in the autoradiograms. Incubation of purified GTP-binding protein from rat brain (Go-protein) with membranes and erythrocyte cytoplasm in the presence of [32P]NAD resulted in the label incorporation into the Go-protein alpha-subunit. This incorporation was markedly decreased after treatment of radiolabeled Go-protein with HgCl2. The data obtained testify to the existence in human erythrocytes of cytoplasmic and membrane-bound forms of cysteine-specific ADP-ribosyl transferase, for which the Go-protein serves as a substrate. The 37 kDa and 41 kDa proteins are substrates for the membrane-bound form of the enzyme.

[Release of adenylate cyclase stimulating, GTP-binding protein, Gs from human erythrocyte membranes under the effect of fluoride-ion and guanine nucleotides]. / Osvobozhdenie stimuliruiushchego adenilattsikalzu GTP-sviazyvaiushchegobelka, Gs, iz membran éritrotsitov cheloveka pod deistviem ftorid- iona i guanilovykh nukleotidov.

Human erythrocyte membranes were incubated in the presence of sodium fluoride or guanylylimidodiphosphate (GppNHp), a nonhydrolysable GTP analog. After centrifugation at 100000 g the activity of the adenylate cyclase-stimulating GTP-binding protein, Gs, was detected in the supernatant fraction. The release of the Gs activity from the membranes closely resembles Gs activation by GppNHp. The Gs activity release from the GppNHp-induced membranes is characterized by a lag period. The nucleotide concentration causing a half-maximal solubilization is about 9.10(-7) M. Approximately 50% of the Gs activity released from sodium fluoride-treated human erythrocyte membranes was associated with the cytoskeletal fraction extracted by a low ionic strength solution. The data obtained suggest that Gs exists in the membrane at lease in two compartmentalized states and is solubilized from both states during its activation.