J Cell Biochem. 2016 Nov;117(11):2597-607. doi: 10.1002/jcb.25554.

Glycogen Synthase in Sertoli Cells: More Than Glycogenesis?

PubMed link



Glycogen synthase is the only enzyme capable to biosynthesize glycogen in mammals. A complex regulatory network modulates the activity of this enzyme, and thus the cellular glycogen content, involving allosteric factors, covalent modifications, and subcellular compartmentalization. Therefore, the amounts of this polysaccharide can be directly correlated to the regulation levels that glycogen synthase is subjected. In murine testes, glycogen was described to be necessary for sex determination during organogenesis (1), and glycogen content decreases progressively in tubular cells from the immediate postnatal period to adulthood (2). Nevertheless, the impact of this intermediate metabolite and the activity of glycogen synthase on the seminiferous tubule cells are not known. Previously, our laboratory described the presence of all the machinery that controls the muscle isoform of the glycogen synthase (MGS) activity, and the polysaccharide itself, in germ and Sertoli cells (3). While the present work is focused on determining the activity levels of glycogen synthase in Sertoli cells. Our results indicated that this enzyme is expressed in these cells obtained from mature (30 days old) mice, and is present in the nuclei as well as in the cytoplasm. Strikingly, the activity assays revealed that less than the 5% of the MGS is actually active, and thus, the vast majority of the enzyme is present in an inactive form. Glycogen content measurements validated the high levels of inactivity of the MGS in Sertoli cells. Nevertheless, we were able to activate the enzyme massively by over-expressing the Protein-Targeting to Glycogen (PTG), or very weakly by treating the cells with lithium chloride. In concordance, high levels of the phospho-inactive form of the enzyme (pS640MGS) were determined by phosphatase assays. Thus after establishing that Sertoli-MGS was present mainly in its inactive form, we wanted to explore a new feature of this metabolic protein that was related to RNA molecules. We performed an MGS RNA-immunoprecipitation followed by a microarray to determine the identity of the MGS-interacting RNAs. This approach allowed us to find several mRNAs that can be associated to primary metabolic processes. Thus, this work reinforces the idea that MGS, and probably other metabolic enzymes, can be related to non-metabolic processes.



Figure 1. Sertoli MGS shows a different 2-D electrophoresis pattern compared with muscle and brain. Five hundred micrograms of muscular, brain, and Sertoli cell extracts were used to immunoprecipitate the MGS. Each pulled down extract was precipitated overnight with acetone and solubilized in 8M urea, 20 mM DTT, and 0.5% IPG buffer. Proteins were firstly separated by isoelectric focusing between pH 4 and 7, and then by molecular weight on a 10% SDS-PAGE gel. Finally the proteins were visualized by silver stain. Black squares enclose the different MGS patterns for each tissue/cell type.


Highly Phosphorylated MGS on Sertoli cells

Different kinases inactivate glycogen synthase by phosphorylation on different serine residues. Among them, Ser-640, -644, -648, and -652, are hierarchically phosphorylated by GSK3β, where the phosphorylation on positions 640 and 644 potently inactivates the enzyme (4). High levels of pSer640MGS were described in our work, corroborating high levels of inactivity. Additionally, to properly define the presence of a protein with special characteristics (highly phosphorylated), we compared immunoprecipitated MGS from mouse muscle and brain tissue with Sertoli cells by 2-D electrophoresis. In this case MGS was immunoprecipitated with an antibody that recognizes both, the phospho and unphosphorylated forms of MGS. The 2-D analysis showed that MGS from muscle tissue is mainly located at more neutral pH regions, while brain MGS is slightly more distributed to acidic pH values (Figure 1). Interestingly, a very different pattern is observed in Sertoli MGS. A clear spot accumulated at neutral pH (7.0), and several other spots located to more acidic regions, without resembling the muscle or brain results (Figure 1), indicating the presence of a Sertoli-MGS with different post-translational modifications (PTM). With these results we suggest the presence of a highly phosphorylated form of the MGS in Sertoli cells compared to the other examined mouse tissues, which is in total agreement with our phosphatase assays described in the manuscript.



Figure 2. Lithium treatments induce cytosolic MGS granules in Sertoli cells. A) and B), confocal immunofluorescence of fixed and permeabilized cells incubated with specific antibodies against total MGS (MGS) and its phosphorylated form in sertine 640 (pS640MGS). Nuclei were stained with propidium iodide (PI). Scale bars 20 µm. In A), Sertoli 42GPA9 cells were non-treated (control), or treated with 30mM and 90mM lithium chloride as indicated. In B) primary Sertoli cells obtained from adult mice treated with 6 µM Li2CO3 per day during 30 days were assessed. These cultures were maintained for two (left) and 8 days (right) before fixation. Phase contrast is shown to visualize shape and vacuolization of the cells. C) Transmission electronic microscopy of 90mM lithium chloride treated Sertoli 42GPA9 cells. Scale bar 250 nm. A zoom in of the region of interest is shown in the lower image, scale 50 nm.



Lithium effects on Sertoli-MGS

Lithium salts are widely used as therapeutic agents to control mood instabilities. One of the targets of lithium salts is to inhibit GSK3β. Therefore, lithium salts as an inhibitor of the inhibitor (GSK3β) have been described to activate the glycogen synthesis in different cell and tissues. In testicular cells, there is no data related to the MGS, but many of the reports have described that lithium salts are toxic agents, having deleterious effects on the components of the seminiferous tubule. In our work, lithium chloride treatments showed a weak activation of the Sertoli-MGS, and thus a modest 4-fold increase in the glycogen content. Nevertheless, the subcellular localization of the enzyme (that is not shown in the manuscript) suffers big changes. In general it is known that MGS is re-localized to sites where glycogen synthesis starts in response to stimuli (5). In this case we analysed the MGS and pS640MGS localization on Sertoli cells upon treatment with lithium salts by confocal immunofluorescence (Figure 2). After the treatment with (30 and 90 mM) LiCl, the total MGS (phopho and unphosphrylated forms) and the pSer640MGS (specific phosphorylated form in Ser640) were accumulated in evident cytosolic granules that were not observed in the untreated cells (Figure 2A). As a control to discard a hyperosmotic effect by increasing the salt concentration, cells were treated with NaCl and no cytosolic granules were observed (data not shown). Additionally, we treated mice with therapeutic doses of lithium carbonate (6µmoles per day) for 30 days, and then primary Sertoli cell cultures were performed to further analyse the MGS and pS640MGS localization. In this case, we recapitulated the localisation of both MGS forms in clear cytosolic granules (Figure 2B). Moreover, it is possible to observe several vacuoles, probably as sign of stress in the cytoplasm of the primary Sertoli cells (Figure 2B), suggesting some level of toxicity of the lithium salt treatments. Then, to elucidate the ultra-structural features of these cytoplasmic granules, LiCl-treated Sertoli 42GPA9 cells were subjected to transmission electron microscopy (TEM). TEM analysis shows the formation of an electrodense structure that resembles a glycogen particle (Figure 2C). Importantly, no structure like this was observed in non-treated cells. Therefore, these results validate the small activation of the MGS by the lithium treatments, suggesting a negative effect of this activation to Sertoli cells, which shows signs of stress.


All in all, the results presented in our manuscript and in this supplement indicate that MGS in Sertoli cells does not present the same features like what have been widely described for this metabolic enzyme in other tissues. We can strongly suggest, and our hypothesis is, that besides the “classic” metabolic role, this enzyme is achieving another function probably related with RNA-related processes on Sertoli cells. This new vision of metabolic enzymes related to RNAs is supported structurally and functionally. Structurally by the Rossman fold motif described for the catalytic domain of the MGS, which have been described for dehydrogenases that interact with RNA molecules (LDH for example). And functionally by the specific translation of a subset of mRNAs by elongating ribosomes containing the pSMGS640. Finally we would like to highlight that beside being a “classical” metabolic enzyme discovered more than 40 years ago, the MGS and the glycogen metabolism in general still surprise us (like the recent finding that glycogenin is not a must for glycogen synthesis (6)).



Figure 3. Members of the Molecular Metabolism Lab. Up (from left to right): Dr. Héctor Mancilla, Karina Cereceda, Dr. Ilona I. Concha, Dr. Antonia Covarrubias, Dr. Camila López, Noemí Gutierrez, Dr. Franz Villarroel-Espíndola. Down (from left to right): Karen Vander Stelt, Dr. Constanza Angulo, David León, Dr. Rodrigo Maldonado.



1. Matoba,S., Kanai,Y., Kidokoro,T., Kanai-Azuma,M., Kawakami,H., Hayashi,Y. and Kurohmaru,M. (2005) A novel Sry-downstream cellular event which preserves the readily available energy source of glycogen in mouse sex differentiation. J. Cell Sci., 118, 1449–59.
2. Gunaga,K.P., Chitra Rao,M., Sheth,A.R. and Rao,S.S. (1972) The role of glycogen during the development of the rat testis and prostate. J. Reprod. Fertil.
3. Villarroel-Espíndola,F., Maldonado,R., Mancilla,H., vander Stelt,K., Acuña,A.I., Covarrubias,A., López,C., Angulo,C., Castro,M. a, Slebe,J.C., et al. (2013) Muscle glycogen synthase isoform is responsible for testicular glycogen synthesis: glycogen overproduction induces apoptosis in male germ cells. J. Cell. Biochem., 114, 1653–64.
4. Skurat,A. V, Wang,Y. and Roach,P.J. (1994) Rabbit skeletal muscle glycogen synthase expressed in COS cells. Identification of regulatory phosphorylation sites. J. Biol. Chem., 269, 25534–42.
5. Cid,E., Cifuentes,D., Baqué,S., Ferrer,J.C. and Guinovart,J.J. (2005) Determinants of the nucleocytoplasmic shuttling of muscle glycogen synthase. FEBS J., 272, 3197–213.
6. Testoni,G., Duran,J., García-Rocha,M., Vilaplana,F., Serrano,A.L., Sebastián,D., López-Soldado,I., Sullivan,M.A., Slebe,F., Vilaseca,M., et al. (2017) Lack of Glycogenin Causes Glycogen Accumulation and Muscle Function Impairment. Cell Metab., 26, 256–266.e4.