It has long been known that eukaryotic ribosomes are not homogenous, but that individual ribosomes differ in the composition and modification state of their constituent proteins and RNAs 1; 2; 3
. Based upon the hypothesis that the heterogeneity of eukaryotic ribosomes may enable cells to execute rapid changes in protein expression, it has been speculated that subpopulations of ribosomes may control translation of specific mRNAs 4; 5
. Support for this speculation has been provided from studies in S. cerevisiae
, which contains many duplicated ribosomal protein-encoding genes. For example, cells that lack one copy of particular ribosomal protein genes developed altered effects on the translation of localized ASH1
mRNA, suggesting that the “duplicated” ribosomal protein has a distinct function 6
. Non-ribosomal proteins have also been shown to contribute to ribosome heterogeneity. For example, proteins releated to the laminin binding protein precursor LBP/p40 are required for maturation of 20S ribosomal RNA and stability of the 40S subunit in S. cerevisiae 7; 8
. In Drosophila melanogaster
, Reaper protein, a regulator of apoptosis, has been shown to affect protein biosynthesis. Reaper binding to the 40S ribosomal subunit inhibited cap-dependent mRNA translation, but not internal ribosome entry-mediated mRNA translation 9
. The receptor of activated protein kinase C (RACK1) is also a 40S-binding protein and has been shown to recruit activated protein kinase C to the ribosome, thereby linking the translation machinery to signal transduction pathways 10
. Loss of RACK1 in Saccharomyces pombe
caused only a small decrease in global cellular translation, but resulted in a large reduction of ribosomal protein L25 10
. Because the abundance of rpL25 mRNA was unaffected upon depletion of RACK1, RACK1 has been suggested to control translation of the rpL25 mRNA.
To address whether distinct subsets of ribosomes in mammalian cells preferentially translate specific mRNAs, we have purified inactive and actively translating ribosomes from HeLa cells. Analysis of these ribosome populations by mass spectrometry revealed few differences in the association of non-ribosomal proteins with inactive and active ribosomes, with one striking exception, glycogen synthase 1 (GYS1), which we found to be specifically associated with polysomes.
GYS1 belongs to the family of glycosyltransferases and catalyzes the rate-limiting step during glycogen biosynthesis 11; 12
. It uses the donor molecule UDP-glucose to lengthen glycogen chains by forming alpha-1,4-glycosidic linkages 11; 12
. The activity of GYS1 is regulated by phosphorylation by glycogen synthase kinase 3 (GSK-3) and other kinases. Phosphorylation inhibits GYS1 activity, by converting GYS1 from a glucose-6-phosphate-independent (I-form) into a glucose-6-phosphate-dependent form (D-form) 12
. Humans express two GYS isoforms encoded by the GYS1 and GYS2 genes 13; 14
. The highest levels of GYS1 are in skeletal and cardiac muscle; in contrast, GYS2 is only found in the liver 15
. Interestingly, most mice lacking GYS1 are not viable and die shortly after birth due to failure of their lungs to inflate 16
. In addition, they display abnormal heart morphology, hemorrhagic livers and venous and pulmonary congestion. The phenotype of the surviving mice is similar to the glycogen storage disease type 0 in humans. This disease displays lack of muscle glycogen, an increase in oxidative muscle fibers and mitochondrial proliferation 16; 17
. Even though the disease is rare in humans, it has been speculated that GYS1 deficiency could be a common cause of sudden cardiac arrest in infants and children 18
. Here, we have identified GYS1 as a novel protein factor associated with actively translating ribosomes. Specifically, phosphorylation of serine 640 in GYS1 correlated with association of GYS1 with translating ribosomes. Depletion of GYS1 from HeLa cells resulted in a loss of polysomes, and microarray analysis revealed changes in abundances and translation of particular subsets of mRNAs.