Transcriptomic and network component analysis of glycerol kinase in skeletal muscle using a mouse model of glycerol kinase deficiency

Abstract

Glycerol kinase (GK) is at the interface of fat and carbohydrate metabolism and has been linked to obesity and type 2 diabetes mellitus (T2DM). The purpose of this study was to investigate the role of GK in fat metabolism and insulin signaling in skeletal muscle (an important end organ tissue in T2DM). Microarray analysis determined that there were 525 genes that were differentially expressed (1.2 fold, p-value <0.05) between knockout (KO) and wild-type (WT) mice. Quantitative PCR (qPCR) confirmed the differential expression of genes including glycerol kinase (Gyk), phosphatidylinositol 3-kinase regulatory subunit, polypeptide 1 (p85 alpha) (Pik3r1), insulin-like growth factor 1 (Igf1), and growth factor receptor bound protein 2-associated protein 1 (Gab1). Network component analysis demonstrated that transcription factor activities of myogenic differentiation 1 (MYOD), myogenic regulatory factor 5 (MYF5), myogenin (MYOG), nuclear receptor subfamily 4, group A, member 1 (NUR77) are decreased in the Gyk KO whereas the activity of paired box 3 (PAX3) is increased. The activity of MYOD was confirmed using a DNA binding assay. In addition, myoblasts from Gyk KO had less ability to differentiate into myotubes compared to WT myoblasts. These findings support our previous studies in brown adipose tissue and demonstrate that the role of Gyk in muscle is due in part to its non-metabolic (moonlighting) activities.

Introduction

Glycerol kinase (GK) (EC 2.7.1.30) catalyzes the phosphorylation of glycerol to glycerol 3-phosphate (G3P) and is at the interface of fat and carbohydrate metabolism [1]. G3P is important in the formation of triacylglycerol, which is the foundation for fat storage (Figure 1). When triacylglycerol is hydrolyzed to form glycerol and fatty acids, glycerol is transported to the liver where it is converted by GK to G3P, an important precursor for lipid synthesis and gluconeogenesis. GK is predominantly found in liver and kidney, but also has been found in brain, adipocytes, testis, and cardiac and skeletal muscle [1].

Figure 1.

Glycerol kinase at the interface of carbohydrate and lipid metabolism. GK converts glycerol to G3P which then can be used downstream in gluconeogenesis and lipid synthesis.

Glycerol kinase deficiency (GKD) is an X-linked inborn error of metabolism that is characterized biochemically by hyperglycerolemia and glyceroluria and is due to mutations within or deletions of the GK gene Xp21 [1]. GKD can be part of a contiguous gene syndrome (the complex form) involving other genes in this area [nuclear receptor subfamily 0, group B, member 1 (NROB1), and Duchenne muscular dystrophy (DMD)] or due to mutations or deletions within the GK gene (isolated GKD). The isolated form of GKD (iGKD) can be symptomatic (juvenile form) or asymptomatic (benign adult form) [1-3]. We and others have previously shown that there is a lack of genotype-phenotype correlation in iGKD [2-4]. This includes identical DNA mutations with different phenotypes within a family [2, 3] and similar GK enzymatic activity in both symptomatic and asymptomatic individuals with iGKD [4]. This suggests that other environmental and biological factors (such as metabolic flux through related pathways or the role of modifier genes) are important in the phenotype of GKD [5]. Another possibility for this complexity is that the phenotype relates to the other nonenzymatic (moonlighting) functions of the GK protein [6, 7].

Most metabolic enzymes exhibit moonlighting functions; seven out of the ten glycolytic enzymes and seven out of the eight enzymes in the TCA cycle exhibit moonlighting functions [6]. In addition to its enzymatic role, glycerol kinase is also the ATP stimulated translocation protein (ASTP) which enhances the nuclear binding of the glucocorticoid – receptor complex [8, 9]. GK also has a role in apoptosis [10], and binds to histones and interacts with the voltage dependent anion channel in the outer mitochondrial membrane [11, 12].

There is an emerging role of GK in type 2 diabetes mellitus (T2DM). Patients with a GK missense mutation, N288D, have the asymptomatic form of GKD, increased risk for obesity, insulin resistance, and T2DM [13]. Thiazolidinediones (TZDs) are effective drugs in the treatment of type 2 diabetes [14] and have been shown to induce GK expression in adipocytes [15, 16]. We have previously shown altered gene expression in liver and brown adipose tissue of the mouse ortholog of GK, Gyk, KO mice of genes involved in lipid and carbohydrate metabolism, insulin signaling and T2DM [17, 18]. Skeletal muscle is another important end organ for insulin resistance and it has been shown that GK is functional in muscle cells and contributes to intramuscular triglyceride (TG) synthesis [19, 20].

Satellite cells are found in skeletal muscle tissue and can differentiate into myotubes to replace damaged myofibers [21]. In the past decade, many studies have used primary myoblast cultures derived from satellite cells to study muscle development [22-26]. The myogenic regulatory factors (MRFs) which are basic helix-loop-helix (bHLH) transcription factors and include MYOD, MYF5, MRF4, and myogenin (MYOG) are crucial for the determination and differentiation of skeletal muscle (for reviews, see [27, 28]). The Myod1 and Myf5 genes are expressed early in muscle cell differentiation and are required for commitment to differentiation, whereas Myog is expressed later in differentiated myotubes. MRFs and other transcription factors, such as the MEF2 family and the paired box family, control gene expression in a distinct pattern that leads to myotube formation. In this study, we tested the hypotheses that the role of GK in skeletal muscle includes its role in central carbon metabolism as well as its moonlighting activities, which include a new function in skeletal muscle differentiation.

Materials and Methods

Animal care

Animal Care: Gyk deficient mice were courtesy of W.J. Craigen (Baylor College of Medicine) [29]. The heterozygous female strain (129SvJ-C57BL/6J) was bred with WT males from the same breeding colony. The mice were kept on a normal diet (Harlan Tekland) and all procedures and experiments were per a protocol approved by the UCLA Chancellor’s Animal Research Committee.

Gene expression analysis

Day of life one (dol 1) WT and Gyk KO mice were sacrificed and skeletal muscle from hind limb was extracted. RNA isolation, microarray analysis, and quantitative PCR (qPCR) was carried out as described previously [17]. The microarray data was analyzed using the DNA-Chip analyzer (dChip) software package [30]. Differentially expressed genes were filtered with absolute fold change > 1.2 between WT and KO, absolute difference in the expression level between WT and KO > 100, student t-test p value < 0.05, and percent present call of ≥ 20%. COMBAT, an empirical Bayes method was used to adjust for batch effects [31].

Network component analysis (NCA)

NCA was carried out as previously described [17, 32, 33]. Connectivity matrices between transcription factors (TFs) important in skeletal muscle tissue and genes differentially expressed in the microarray analysis was manually constructed using PUBMED. Five TFs [myogenic differentiation 1 (MYOD), myogenic regulatory factor 5 (MYF5), myogenin (MYOG), nuclear receptor subfamily 4, group A, member 1 (NUR77), and paired box 3 (PAX3)] were used to construct the final connectivity matrix. The data matrices were decomposed and the control strengths (CS), and transcription factor activity (TFA) matrices were obtained using the NCA toolbox (http://www.seas.ucla.edu/~liaoj/downloads/htm).

Primary myocytes

Muscle tissue was isolated from hind limb of day of life three (dol 3) mice and myoblasts were cultured in DMEM/F10 media containing fetal calf serum, basic fibroblast growth factor, and penicillin/streptomycin as described [34]. 500,000 cells were seeded and after three days, the cells were counted and 200,000 cells were seeded in fusion media (DMEM, horse serum, and penicillin/streptomycin) in duplicate. Day 0 is defined as the day fusion media was added. Cells were counted again on day 5 and fusion index analysis was performed.

Fusion indices analysis

Fusion indices were measured on day 3, 4, and 5 as described previously [22]. Ten random fields were chosen per plate, photographed, and counted manually. The fusion index was calculated as the ratio of the number of nuclei in myotubes to the total number of nuclei [22]. Myotubes were defined as cells containing three or more nuclei.

MYOD activity

Skeletal muscle was isolated from dol 3 Gyk KO and WT mice and nuclear extracts were prepared using a nuclear extraction kit (Active Motif, Carlsbad, CA) per manufacturer’s instructions. MYOD activity was measured using the TransAM DNA binding assay for MYOD (Active Motif) as instructed by the manufacturer.

Results

Gene expression is altered in skeletal muscle of the Gyk KO mice

Hierarchical clustering of microarray analysis using the 4419 most varying probesets revealed that the Gyk KO samples and WT samples cluster separately (Figure 2). This unsupervised learning analysis demonstrates that Gyk KO mice have a distinct global gene expression profile in muscle compared to WT mice. Differential gene expression analysis (gene filtering) revealed 546 probesets (representing 525 genes) significantly differentially expressed between KO and WT mice. Of the 546 probesets, 323 genes were down-regulated and 223 were up-regulated.

Figure 2.

Unsupervised learning analysis of Gyk KO and WT mice samples using the 4,419 most varying genes. Average linkage hierarchical clustering tree of Gyk KO and WT samples. mmusko represents skeletal muscle from Gyk KO mice and mmuswt represents skeletal muscle from male WT mice.

Enriched biological themes in the differentially expressed genes were determined using the gene ontology analysis Expression Analysis Systematic Explorer (EASE) [35] (Table 1). Significantly (p value < 0.05) enriched biological processes and molecular functions include: ion homeostasis, cell matrix adhesion, IGF receptor signaling pathway, regulation of cell growth, lipid biosynthesis, regulation of glucose import, glucose import, protein binding, insulin-like growth factor binding, growth factor binding, nucleic acid binding, and lipoprotein kinase activity.

Table 1.

EASE analysis grouping of biological processes and molecular functions using the filtered 525 genes (546 probesets).

	Gene Category	List Hits	List Total	Population Hits	Population Total	Probability
GO Biological Process	ion homeostasis	9	348	64	10437	0
	cell-matrix adhesion	7	348	57	10437	0.003
	IGF receptor signaling pathway	2	348	4	10437	0.006
	regulation of cell growth	6	348	56	10437	0.01
	lipid biosynthesis	10	348	153	10437	0.032
	regulation of glucose import	1	348	1	10437	0.033
	glucose import	1	348	1	10437	0.033
GO Molecular Function	protein binding	103	358	1762	10943	0
	insulin-like growth factor binding	4	358	17	10943	0.002
	growth factor binding	5	358	29	10943	0.002
	nucleic acid binding	96	358	2282	10943	0.004
	lipoprotein lipase activity	1	358	1	10943	0.033

Twenty genes (including Gyk) involved in lipid metabolism (Table 2A) were differentially expressed in the Gyk KO mice compared to the WT (fold change >1.2). Twelve of those genes were down-regulated including: adipocyte complement related protein (Acrp30), Gyk, protein kinase AMP-activated beta 2 non-catalytic subunit (Prkab2), glycerol-3-phosphate dehydrogenase 1 (Gpd1), lipoprotein lipase (Lpl), glyceronephosphate O-acyltransferase (Gnpat), testis expressed gene 261 (Tex261), monoglyceride lipase (Mgll), phosphate cytidyltransferase 1 choline alpha isoform (Pcyt1a), adiponectin receptor 1 (Adipor1), and RIKEN c DNA 4930570C03 gene (Table 2A). Eight genes were up-regulated including: 3-hydroxy-3-methylglutaryl-Coenzyme A synthase 1 (Hmgcs1), cardiolipin synthase1 (Crls1), NADH dehydrogenase (ubiquinone) 1 alpha/beta subcomplex (Ndufab1), fatty acid-Coenzyme A ligase log chain 4 (Facl4), cytochrome P450, 51 (Cyp51), phospholipase D1 (Pld1), sterol O-actyltransferase 1 (Soat1), and SA rat hypertension-associated homolog (Sah) (Table 2A). In addition, ten genes (including Gyk) involved in carbohydrate metabolism were differentially expressed in the Gyk KO mice compared to the WT (Table 2B). Eight of these were down-regulated including Gyk, hexokinase 1 (Hk1), glycogen synthase 3 (Gys3), carbohydrate sulfotransferase 14 (Chst14), Gpd1, sulfatase 2 (Sulf2), glucosamine-6-sulfatase (Gns), and aconitase 1 (Aco1). Two genes involved in carbohydrate metabolism, activating transcription factor 4 (Atf4), and glycosyltransferase 8 domain containing 1 (Glt8d1), were up-regulated (Table 2B). A total of ten genes that relate to insulin signaling, insulin resistance or diabetes were differentially expressed in muscle between Gyk KO and WT mice (Table 3). Of these, eight genes were down-regulated including: insulin-like growth factor binding protein 5 (Igfbp5), growth factor receptor bound protein 2-associated protein 1 (Gab1), glycogen synthase 1, muscle (Gys1), insulin-like growth factor 1 (Igf1), growth factor receptor bound protein 10 (Grb10), protein tyrosine phosphatase, receptor type, F (Lar), Gpd1, eukaryotic translation initiation factor 2B, subunit 1 alpha (Eif2b1). Pleckstrin homology domain interacting protein (Phip), and phosphatidylinositol 3-kinase, regulatory subunit, polypeptide 1, p85 alpha (Pik3r1) were up-regulated (Table 3).

Table 2.

A: List of differentially expressed genes involved in lipid metabolism altered in the KO male vs. the WT male mice. * Student t test p value < 0.05.
LocusLink Numbers	Gene Name	Gene Symbol	Fold Change	Lowerbound	Upperbound
11450	adipocyte complement related protein	Acrp30	-3.41	-1.93	-6.9
14933	glycerol kinase	Gyk	-3.09	-2.46	-3.98
108097	protein kinase, AMP-activated, beta 2 non-catalytic subunit	Prkab2	-1.54	-1.31	-1.79
14555	glycerol-3-phosphate dehydrogenase 1 (soluble)	Gpd1	-1.5	-1.17	-1.92
16956	lipoprotein lipase	Lpl	-1.44	-1.25	-1.69
14712	glyceronephosphate O-acyltransferase	Gnpat	-1.33	-1.2	-1.48
21766	testis expressed gene 261	Tex261	-1.33	-1.26	-1.4
23945	monoglyceride lipase	Mgll	-1.31	-1.15	-1.5
68832	RIKEN cDNA 1110057K04 gene	1110057K04Rik	-1.31	-1.25	-1.38
13026	phosphate cytidylyltransferase 1, choline, alpha isoform	Pcyt1a	-1.26	-1.15	-1.39
72674	adiponectin receptor 1	Adipor1	-1.23	-1.12	-1.35
67739	RIKEN cDNA 4930570C03 gene	4930570C03Rik	-1.23	-1.15	-1.315
208715	3-hydroxy-3-methylglutaryl-Coenzyme A synthase 1	Hmgcs1	1.22	1.11	1.34
66586	cardiolipin synthase 1	Crls1	1.22	1.12	1.33
70316	NADH dehydrogenase (ubiquinone) 1, alpha/beta subcomplex, 1	Ndufab1	1.24	1.15	1.35
50790	fatty acid-Coenzyme A ligase, long chain 4	Facl4	1.26	1.14	1.38
13121	cytochrome P450, 51	Cyp51	1.31	1.14	1.52
18805	phospholipase D1	Pld1	1.32	1.18	1.46
20652	sterol O-acyltransferase 1	Soat1	1.5	1.35	1.67
20216	SA rat hypertension-associated homolog	Sah	1.61	1.36	1.88
B: List of differentially expressed genes involved in carbohydrate metabolism altered in the KO male vs. the WT male mice. * Student t test p value < 0.05.
LocusLink numbers	Gene Name	Gene Symbol	Fold Change	Lower bound	upper bound
14933	glycerol kinase	Gyk	-3.09	-2.46	-3.98
15275	hexokinase 1	Hk1	-1.74	-1.38	-2.17
14937	glycogen synthase 3, brain	Gys3	-1.68	-1.51	-1.87
72136	carbohydrate (N-acetylgalactosamine 4-0) sulfotransferase 14	Chst14	-1.53	-1.36	-1.7
14555	glycerol-3-phosphate dehydrogenase 1 (soluble)	Gpd1	-1.5	-1.17	-1.92
72043	sulfatase 2	Sulf2	-1.31	-1.24	-1.39
75612	glucosamine (N-acetyl)-6-sulfatase	Gns	-1.25	-1.16	-1.35
11428	aconitase 1	Aco1	-1.24	-1.13	-1.35
11911	activating transcription factor 4	Atf4	1.46	1.21	1.72
76485	glycosyltransferase 8 domain containing 1	Glt8d1	1.51	1.3	1.75

Table 3.

Genes involved in insulin signaling and insulin resistance differentially expressed in the Gyk KO male compared to WT male mice. Student t test p value < 0.05

Gene	Symbol	Fold Change
insulin-like growth factor binding protein 5	Igfbp5	-2.22
growth factor receptor bound protein 2-associated protein 1	Gab1	-1.97
glycogen synthase 1, muscle	Gys1	-1.68
insulin-like growth factor 1	Igf1	-1.58
growth factor receptor bound protein 10	GRb10	-1.54
protein tyrosine phosphatase, receptor type, F	Lar	-1.51
glycerol-3-phosphate dehydrogenase 1 (soluble)	Gpd1	-1.5
eukaryotic translation initiation factor 2B, subunit 1 (alpha)	Eif2b1	-1.24
pleckstrin homology domain interacting protein	Phip	1.27
phosphatidylinositol 3-kinase, regulatory subunit, polypeptide 1 (p85 alpha)	Pik3r1	1.37

In order to confirm the microarray data, we chose to analyze the following genes using qPCR analysis; Gyk was chosen to confirm our KO status, Myod1 was chosen to confirm the myocyte differentiation role that GK has, and the other four genes (Gab1, Igf1, Pik3r1, and Eif2b1) are involved in insulin signaling. qPCR confirmed the differential expression of five out of six genes (Figure 3 and Table 4). The down-regulation of Gyk (5.5 fold), growth factor receptor bound protein Gab1 (2.0 fold), insulin-like growth factor 1 (Igf1) (1.1 fold), and Myod1 (3.87 fold), and the up-regulation of Pik3r1 (1.8 fold) were confirmed. qPCR data for (Eif2b1) did not correlate with microarray data (data not shown). Fold differences for each of the genes were calculated using the ΔΔCT method [36].

Figure 3.

qPCR data (mean ± SEM) expressed as relative mRNA levels of Gyk KO compared to WT mice. Phosphatidylinositol 3-kinase, regulatory subunit, polypeptide 1 (Pik3r1), glycerol kinase (Gyk), growth factor receptor bound protein 2-associated protein 1 (Gab1), insulin-like growth factor 1 (Igf1), myogenic differentiation factor 1 (Myod1) *p< 0.05, student t test. White bars represent Gyk KO mice and black bars represent WT mice.

Table 4.

qPCR and Microarray expression data

Gene	RT- PCR	Microarray
Pik3r1	1.78001	1.37
Gyk	5.469096	-3.09
Gab1	-2.045313	-1.97
Igf1	-1.145992	-1.58
Myod1	-3.872351	-1.33

To further understand the metabolic role of GK in muscle, metabolic flux analysis was performed on primary myocytes from Gyk KO and WT cells. Isotopomer abundances and external fluxes of Gyk KO and WT cells were not statistically significant (Supplemental Material 1). In addition, selected internal fluxes were not statistically significant between Gyk KO and WT cells.

NCA revealed altered TF activity in skeletal muscle of the Gyk KO mice

To examine the role of Gyk in muscle cell gene expression, NCA of the microarray data was performed to identify TFAs of TFs important in skeletal muscle in the absence of Gyk. Forty-six genes and five TFs resulted from the analysis. The transcription factor activities of myogenic differentiation 1 (MYOD), myogenic regulatory factor 5 (MYF5), myogenin (MYOG), nuclear receptor subfamily 4, group A, member 1 (NUR77) were decreased in the Gyk KO; whereas the activity of paired box 3 (PAX3) is increased (Figure 4). The expression, connectivity, and CS matrices, and references used to deduce TFA and control strengths (CS) are provided as supplemental material 2.

Figure 4.

Network component analysis: Transcription factor activities of selected TFs important in muscle tissue. The transcriptional network contains 5 TF and 46 genes found in the muscle resulting from microarray analysis at ≥1.2 absolute fold change. Myogenic differentiation 1 (MYOD), myogenic regulatory factor 5 (MYF5), myogenin (MYOG), paired box 3 (PAX3), nuclear receptor subfamily 4, group A, member 1 (NUR77).

Gyk KO cells have a significantly lower fusion index and MYOD activity

NCA analysis revealed that GK affects transcription factors important in muscle differentiation, therefore, we examined the role of GK in the differentiation of the myoblast cells into myotubes. Fusion indices were calculated for primary myoblasts after induction to fuse into mature myotubes as described in Materials and Methods. Gyk KO cells had a fusion index significantly lower (0.75) than the WT cells (0.82) (p value <0.05) (Figures 5A and B). Consistent with this, MYOD activity in the skeletal muscle of Gyk KO mice is significantly decreased (p < 0.05) compared to that of WT (Figure 6).

Figure 5.

Myotube fusion index of Gyk KO and WT primary muscle cells: A. fusion index measured as the ratio of the number of nuclei in myotubes (3 or more nuclei) to the total number of nuclei counted of day 6 after the addition of fusion media as described in Materials and Methods. *p value <0.05, student t test. Black bar represents WT muscle cells and white bar represents Gyk KO muscle cells. B. Photograph of WT and Gyk KO muscle cells on day 5 after supplemented with fusion media.

Figure 6.

MYOD activity in skeletal muscle. MYOD activity in the nuclear extracts of skeletal muscle from hind limb of dol 3 Gyk KO (n=3) and WT (n=10) mice was measured as described in Materials and Methods. Black bar represents the WT animals and the white bar represents the Gyk KO mice. *p value <0.05, student t test.

Discussion

In this study, we demonstrate that skeletal muscle from Gyk KO mice have a distinct global gene expression pattern compared to WT. Many enriched biological groups were significantly altered including metabolism (lipid biosynthesis, glucose import, and regulation of glucose transport) and other cellular functions (ion homeostasis and cell matrix adhesion). This confirms the known metabolic role of GK and suggests that GK has other non-enzymatic (moonlighting) functions in muscle cells as it does in the liver [9].

Specifially using microarray and qPCR, we determined that Gpd1 is down-regulated. This encodes GPD1, the enzyme adjacent to GK in the enzymatic pathway (Figure 1) which is expected to be down regulated in the absence of GK. Adiponectin is down-regulated, involved in metabolism (glucose and lipid), and is altered in obese and diabetic mice and humans [37-39]. Adiponectin has insulin sensitizing activity by reducing TG content, increasing fatty acid oxidation by inducing PPARα transcriptional activity, and up-regulating insulin signaling in liver and skeletal muscle [37, 40]. Adipor1, encodes an adiponectin receptor and is down-regulated which is in contrast to our previous study in brown fat tissue of Gyk KO mice [17]. This is expected as adiponectin is secreted by adipose tissue and its role in insulin sensitization has only been reported in liver and skeletal muscle. Perhaps the Gyk KO mouse is trying to make up for the decrease of adiponectin in muscle cells by releasing more from fat. In addition, Adipor1 is regulated negatively by insulin and is down-regulated in obese mice [41, 42].

Further evaluation of genes involved in the insulin signaling pathway and insulin resistance, revealed that five docking proteins (Gab1, Grb10, Lar, Phip, and Pik3r1) are differentially expressed. Gab1 is decreased in the Gyk KO mice and it encodes a docking protein that is phosphorylated by the insulin receptor (IR) and associates with PI3-K and protein tyrosine phosphatase, non-receptor type 11 (SHP-2) in the insulin signaling pathway [43], which suggest a decrease in insulin signaling in the Gyk KO mice. Grb10 is associated with the insulin receptor and is a negative regulator of insulin signaling [44]. Deletions of Grb10 in mice improve insulin sensitivity and glucose tolerance [44, 45]. The decreased Lar expression in the Gyk KO mice is intriguing because muscle specific transgenic over-expression of Lar causes insulin resistance [46]. Both the decrease in expression of Grb10 and Lar point to insulin sensitivity, which is contrary to our hypothesis that Gyk deletions cause insulin resistance; perhaps the down regulation of these genes is a mechanism to improve insulin sensitivity in the Gyk KO mice or the insulin resistance is limited to fat and liver.

The lack of significant difference in metabolic flux analysis may be due to limitations of the methods used. First, proteinogenic amino acids are used as a surrogate for isotopomeric compositions of central carbon metabolites because amino acids are more abundant. This involves the assumption that cell metabolism is homogenous and that cells producing large amounts of protein reflect the metabolism of the entire cell population [47]. The metabolic flux presented was performed on differentiated primary cells, which may be a heterogeneous cell population making this assumption less valid which is supported by the variation in the data. Second, any true differences in metabolic flux may have been masked by the large isotopomer concentration standard deviation [9]. A third possibility is that there is no true difference in central carbon metabolism fluxes in the myoctes. In which case, studies investigating the flux in other metabolic networks, such as lipid metabolism may be needed to document changes in the metabolic flux.

NCA revealed that the TFA several muscle specific TF were altered (MYOD, MYF5, MYOG, and NUR77). Myod1 gene expression was decreased in the Gyk KO muscle tissue. However, the other TF identified with altered TFA by NCA did not have alterations in gene expression level, highlighting the role of NCA in understanding the biological significance of microarray data [17, 18, 32, 33]. The decreased Myod1 gene expression and TFA suggested that Gyk has a role in regulating transcription in muscle cells. This was validated by the lower fusion index and MYOD activity in the Gyk KO cells. MYOD is an important factor in determining skeletal muscle differentiation [48]. Since glucocorticoids decrease MYOD protein levels through the glucocorticoid receptor (GR) [49], the effect of GK on MYOD may be due in part to the GK ASTP activity. These data suggests that GK has a role in muscle specific transcriptional control and myocyte differentiation.

In this study, we examine the role (metabolic and non metabolic) of GK in skeletal muscle using transcriptomic and network component analysis. We show that Gyk deletion alters the level of expression of genes involved in metabolism and insulin signaling as has been seen in other tissues [17, 18]. In addition, GK affects gene expression in other cellular processes. In particular, we report that GK affects the activity of muscle specific transcription factors and appears to play a role in myotube differentiation. This corroborates our previous findings that GK has nonenzymatic (moonlighting) functions that are important in non metabolic processes and shows that these roles are tissue-specific. Understanding these moonlighting functions and their tissue specificity, will be important to understand the pathogenesis of GKD as well as a model system for understanding the complexity of single gene disorders.

List of Abbreviations

Aco1

aconitase 1

Acrp30

adipocyte complement related protein

Adipor1

adiponectin receptor 1

ASTP

ATP stimulated translocation protein

Atf4

activating transcription factor 4

ATP

adenosine-5’-triphosphate

BHLH

basic helix-loop-helix

Chst14

carbohydrate sulfotransferase 14

Crls1

cardiolipin synthase1

CS

control strength

Cyp51

cytochrome P450, 51

dchip

DNA-Chip analyzer

DHAP

dihydroxyacetone phosphate

DMD

Duchenne Muscular Dystrophy

DMEM

Dulbecco’s Modified Eagle Medium

dol

day of life

EASE

Expression Analysis Systematic Explorer

Eif2b1

eukaryotic translation initiation factor 2B, subunit 1 alphal

Facl4

fatty acid-Coenzyme A ligase log chain 4

G3P

glycerol 3-phosphate

Gab1

growth factor receptor bound protein 2-associated protein 1

GK

glycerol kinase

GKD

glycerol kinase deficiency

Glt8d1

glycosyltransferase 8 domain containing 1

Gnpat

glyceronephosphate O-acyltransferase

Gns

glucosamine-6-sulfatase

Gpd1

glycerol-3-phosphate dehydrogenase 1

Grb10

growth factor receptor bound protein 10

Gyk

mouse ortholog of glycerol kinase

Gys1

glycogen synthase 1, muscle

Gys3

glycogen synthase 3

Hk

hexokinase 1

Hmgcs1

3-hydroxy-3-methylglutaryl-Coenzyme A synthase 1

Igf1

insulin-like growth factor 1

Igfbp5

insulin-like growth factor binding protein 5

iGKD

isolated GKD

KO

knockout

Lar

protein tyrosine phosphatase, receptor type, F

Lpl

lipoprotein lipase

MEF2

myocyte enhancing factor 2

Mgll

monoglyceride lipase

MRF

myogenic regulatory factor

Myf5

gene coding for MYF5

MYF5

myogenic regulatory factor 5

Myod

gene coding for MYOD

MYOD

myogenic differentiation 1

Myog

gene coding for MYOG

MYOG

myogenin

NCA

network component analysis

Ndufab1

NADH dehydrogenase (ubiquinone) 1 alpha/beta subcomplex

NROB1

nuclear receptor subfamily 0, group B, member 1

NUR77

nuclear receptor subfamily 4, group A, member 1

PAX3

paired box 3

Pcyt1a

phosphate cytidyltransferase 1 choline alpha isoform

Phip

Pleckstrin homology domain interacting protein

Pik3r1

phosphatidylinositol 3-kinase regulatory subunit, polypeptide 1 (p85 alpha)

Pld1

phospholipase D1

PPAR

peroxisome proliferator-activated receptor

Prkab2

protein kinase AMP-activated beta 2 non-catalytic subunit

qPCR

quantitative PCR

Sah

SA rat hypertension-associated homolog

SHP-2

protein tyrosine phosphatase, non-receptor type 11

Soat1

sterol O-actyltransferase 1

Sulf2

sulfatase 2

T2DM

type 2 diabetes mellitus

TCA

tricarboxylic acid cycle

Tex261

testis expressed gene 261

TFA

transcription factor activity

TFs

transcription factors

TG

triglyceride

TZD

thiazolidinediones

WT

wild-type