Glycogen is a chain of glucose subunits held together by a 1g4 glycosidic bonds, very much like starch (a-amylose). In contrast to starch, which is a single linear chain of glucose, glycogen is a branched structure. At the branch points, subunits are joined by a 1g6 glycosidic bonds. Branches occur every 8-10 residues.

Glycogen is synthesized when blood glucose levels are high. This state is reflected inside liver cells by the presence of high levels of glucose-6-phosphate. G6P is first converted to G1P by phosphoglucomutase. This reaction is analogous to the reaction catalyzed by phosphoglycerate mutase in step 8 of glycolysis, and proceeds by a similar mechanism, with a bisphosphate intermediate.

Since this reaction simply involves trading one phosphoanhydride bond for another, DG is close to 0. However, the pyrophosphate is quickly hydrolyzed by inorganic pyrophosphatase into two molecules of inorganic phosphate, an essentially irreversible reaction. This coupling drives the synthesis of UDP-glucose in the forward direction.

UDP can be reconverted to UTP by nucleoside diphosphate kinase, which reversibly transfers g-phosphate. Thus elongation of glycogen by one glucose requires energy input equivalent to hydrolysis of one molecule of ATP.

UDP + ATP D UTP + ADP

Glycogen synthase can add a glucose subunit only onto a chain that is already at least 4 residues in length. De novo glycogen synthesis requires a primer, in the same way that DNA synthesis requires an RNA primer. Priming is performed by a specific protein, glycogenin, which autocatalyzes addition of an a 1g4 linked glucose oligosaccharide chain to a specific tyrosine residue, using UDP-glucose as a substrate. After the chain is more than four residues long, glycogen synthase takes over. Glycogenin remains bound to the reducing end of glycogen (the C1 hydroxyl group at the right side of the pictures). Glycogen synthase works efficiently only when it is bound to glycogenin. Thus the number of glycogen granules in a cell is determined by the number of glycogenin molecules available, and the size of the granules is limited by the need for physical association between glycogenin and glycogen synthase. When the granule grows too large, the synthase stops working.

Formation of branches is catalyzed by "branching enzyme", amylo (a-1,4ga1,6) transglycosylase. This enzyme breaks off a chain of about 5 to 8 glucose residues from the growing end of glycogen by hydrolyzing an a 1g4 glycosidic linkage, and transfers the short chain to another residue in the same glycogen molecule that is at least four residues away from the cleavage point, forming an a 1g6 linkage.

Recall that a-amylase, secreted in saliva and in the small intestine, also cleaves a 1g4 glycosidic bonds, but it does so hydrolytically (using water to cleave) rather than phosphorolytically (using phosphate to cleave). If a-amylase were used for glycogen degradation, it would generate glucose rather than G1P. The glucose would have to be rephosphorylated in an ATP-requiring step before it could be used for glycolysis.

The first activity of debranching enzyme, oligo (a-1,4ga1,4) glucantransferase, transfers three of the four residues in the branch of a limit dextrin to the C4 end of another branch, leaving a single glucose attached at the branch point by the a 1g6 linkage. This reaction requires no energy, merely substituting one a 1g4 linkage for another. The resulting longer chain can again be degraded by glycogen phosphorylase, until only four residues remain to the nearest branch point.

The single glucose left attached by the a 1g6 linkage is then hydrolyzed by the amylo-(a 1g6)-glucosidase activity, encoded by the same polypeptide chain as the glucantransferase. This reaction is a hydrolysis, not a phosphorolysis, so the product is glucose rather than G1P. Because of the need to hydrolyze the branched residues, recovery of phosphorylated glucose from glycogen is not 100% efficient.