Carbohydrate Response Element binding Protein (ChREBP) and its role in glucose activated transcription of Lipogenic enzymes
Excess dietary carbohydrate is converted in the liver into triglycerides. Following a carbohydrate rich diet, several enzymes catalyzing glycolysis and lipogenesis are activated including L-type Pyruvate kinase (L-PK), acetyl CoA carboxylase, fatty acid synthase and S14(Ishii et al.). The well known pathway involves insulin mediated induction of lipogenic enzyme genes via the Sterol response element binding protein (SREBP). However, now evidence exists that increased glucose uptake and metabolism in liver cells plays an equally important role in regulating this process. Glucose stimulated increase in L-PK synthesis is mediated through a DNA cis element located -183 to -96 base pairs upstream of the cap site of the L-PK gene. This cis element is termed the carbohydrate response element (ChRE). The transcription factor that binds to ChRE and activates transcription is called the Carbohydrate Response Element binding protein (ChREBP), also called WBSCR14 or MondoB. ChREBP belongs to the family of basic-helix-loop-helix leucine zipper proteins (bHLH-LZ) and binds to the ChRE as an obligatory heterodimer with another bHLH-LZ protein called Max-like protein X (Mlx). Moreover, this trans-acting factor is activated by high glucose and insulin, while low glucose, high fat diet and glucagon inhibit its activity.
Functional domains of ChREBP
ChREBP has an N-terminal 19 amino acid (HRKPEAVVLEGNYWKRRIE) nuclear localization signal (NLS) that is essential for its translocation into the nucleus. Beside this, it also has a bHLH-LZ domain, a Proline rich domain and a ZIP-like domain. It also has three consensus phosphorylation sites for cAMP activated Protein kinase A (PKA) at S196, S626 and T666. A deletion analysis of the various domains of ChREBP showed that while loss of the NLS or bHLH-LZ domains resulted in complete loss of its high-glucose mediated transcription activation activity, deletion of the Pro or ZIP-like domains had no effect. Thus, the NLS and leucine zipper domains are essential for the transcription function of ChREBP. Inhibitors of PKA (like H-19) stimulated the activity of ChREBP, while it was inhibited by using inhibitors of Protein phosphatase 2A (like cantharidin). Further, inhibitors of other protein kinases (calmodulin activated protein kinase II, MAP kinase and tyrosine kinase) had no effect on transcription activating property of ChREBP in high-glucose conditions. This suggests that PKA inhibits ChREBP by phosphorylation, and its activity is restored by protein phosphatase 2A. ChREBP was found to be localized to the cytoplasm under low glucose conditions and migrates into the nucleus when cells are transferred to high glucose containing media. In order to assess which of the three phosphorylation sites was important in nuclear translocation of ChREBP, various mutants were generated wherein these S196, S626 and T666 were replaced with similarly hydrophobic residues (Alanine) or strongly acidic residue (Glutamate) followed by culturing in low/high glucose containing media. The results of these experiments show that phosphorylation of S196 (forms a part of the consensus phosphorylation site 193RRSS196 designated as P1) is an independent regulator of nuclear translocation of ChREBP, independent of phosphorylation at the remaining two sites. Next, using a recombinant C-terminal of ChREBP containing only a single phosphorylation site (T666), it was demonstrated that PKA by phosphorylating this residue disrupts DNA binding of the protein, and DNA binding activity could be rapidly restored by PP2A. Further, using double mutants where both the S626 and T666 were mutated showed that while phosphorylation of both residues occurred during binding of ChREBP to the ChRE, phosphorylation at T666 was more important, while that on S626 may only play a supportive role.
Regulation of DNA binding of ChREBP
ChREBP like MyoD belongs to the family of bHLH-LZ transcription factors, and its basic domain binds to an E box sequence of CACGGG or GTGCCC which are separated by 5 bases. In all naturally occurring ChRE elements characterized till date, the spacing between the E boxes is always 5 bp. Mutants with 6 bp spacing show a substantially reduced glucose response, while those with 4 bp spacers are completely unresponsive to glucose. It has been suggested that ChREBP forms a dimer with Mlx and homodimerizes with another ChREBP molecule to form a heterotetrameric structure that binds to two adjacent E boxes on ChRE. It has been proposed that ChREBP has dual functions, both as an activator and repressor. A heterodimer of ChREBP with Mlx binding to a single E box element acts as a repressor of gene transcription, while a heterotetramer can activate the same gene expression. This allows great flexibility in its function. A comparison with the DNA binding domain of Myo D shows that three residues Arg, Asn, and Glu of ChREBP can potentially bind to one of the half sites on the E boxes (5′-CAC-3’/5′-GTG-3′). Phosphorylation of T666 which is the target of PKA is hypothesized to break a salt bridge between Arg and Glu and establish a new salt bridge with Arg. This is the proposed mechanism of cAMP mediated inhibition of DNA binding of ChREBP (and the resultant decrease in L-PK transcription).
Overall mechanism of Glucose mediated expression of L-PK in liver cells
There are thus two key phosphorylation sites on ChREBP, S196 (designated P1) and T666 (P2). Under low glucose conditions, ChREBP exists in an inactive state in the cytosol, being phosphorylated at P3. Upon glucose stimulation, the first step is dephosphorylation of P1 serine residue resulting in the translocation of ChREBP into the nucleus. Dephosphorylation is catalyzed by protein phosphatase 2A (PP2A) which is activated by an unknown metabolite (suggested to be Xylulose-5-phosphate). The P3 site can be in either form during nuclear translocation. Once inside the nucleus, ChREBP is activated further to bind to the ChRE of the L-PK gene by PP2a (again presumed to be activated by an unknown metabolite X). When glucose levels go down, ChREBP is transported back into the cytoplasm by sequential phosphorylation of T666 (breaks DNA binding) and S196.
Mlx: the obligate heterodimer partner of ChREBP
Mlx was identified in a yeast two-hybrid search for potential binding partners of ChREBP. It belongs to the Myc/Mad/max family of proteins. There are two isoforms of Mlx, the and the that have been found to interact with ChREBP. There is no difference in activity of the heterodimer formed with either of these isoforms. The complex of ChREBP and Mlx can specifically recognize ChRE that are glucose-responsive, and do not bind to E-boxes of ChRE that do not respond to glucose(Stoeckman, Ma, and Towle).
Recent findings suggest existence of additional mechanisms for regulation of ChREBP.Tsatsos and Towle conducted experiments using chimeras made by combining different domains of ChREBP and MondoA. MondoA is a paralogue of ChREBP found in the skeletal muscle. It lacks all of the three “critical” phosphorylation sites and so cannot rescue cells transfected with dn-Mlx, the dominant negative form of Mlx (the obligate heterodimer partner of ChREBP). They observed that chimeras in which at least two of the three phosphorylation sites were lacking were still able to rescue cultured liver cells transfected with dn-Mlx. Closer examination revealed a hitherto unexamined domain, the WBSCR14 Mlx C-tail (WMC) domain located in the C terminal region of both WT ChREBP and MondoA. They suggest that there exists yet another , novel mechanism for glucose-sensitive modulation of ChREBP which is independent of S196/S626/T666 phosphorylation. They present several experimental results to bolster their claim. A triple mutant (TM) form of ChREBP ( S196A/S626A/T666) which would be expected to be constitutively active did not show much activation in low glucose conditions, although it was able to rescue glucose response (in dn-Mlx transfected cells) in high glucose containing media. This suggests that activation of ChREBP in response to high glucose requires additional events apart from dephosphorylation of the three proposed earlier. When TM ChREBP containing cells were treated with cAMP, there was a nearly 6-fold reduction in activity compared to WT ChREBP, thereby suggesting presence of other PKA phosphorylation sites on the protein or other components of the signaling pathway. Further, there was no significant difference in cAMP levels between hepatocytes cultured in low vs those cultured in high glucose containing media. This is contrary to the earlier model proposed by Kawaguchi et al wherein increase in cAMP in low glucose conditions is proposed to activate PKA which in turn inactivates ChREBP. Tsatsos and Towle suggest this as evidence that inactivation of ChREBP in low-glucose conditions involves some mechanism other than PKA mediated phosphorylation. To further support this, they report that there is no decrease in phosphorylation of WT- ChREBP when cells are transferred from low to high-glucose ( as would be expected if dephosphorylation at high glucose were the sole mechanism of activating ChREBP at high glucose).
Based on their findings, they suggest a modification to the signaling pathway proposed by Kawaguchi et al. Accordingly, ChREBP is inhibited by cAMP under fasting conditions by PKA mediated phosphorylation. However, as the TM form of ChREBP was not constitutively active, it suggests that there are additional PKA sites on ChREBP or its interacting protein Mlx which are important in its regulation. After a high glucose diet, ChREBP is reactivated, a process requiring reversal of the inhibitory phosphorylations. The new model proposes a separate pathway for glucose mediated ChREBP activation and cAMP mediated repression.
The authors also stress the possible role played by the WMC domain. The function of this region has not been mapped in ChREBP, but has been well studied in MondoA. The WMC domain of MondoA contains a repeating pattern of hydrophobic residues that is important in its interaction with the obligatory heterodimer Mlx. Mutation of this region led to failure of MondoA to localize to the nucleus. So, the WMC domain in MondoA is important in nuclear translocation, in conjunction with signals from the N-terminal NLS domain of the protein(Tsatsos and Towle;Kawaguchi et al.).
A recent article by Ming et al has dissected the mechanism of glucose mediated regulation of ChREBP suggesting the existence of a glucose sensing module (GSM) within ChREBP and its two components, the Low glucose inhibitory domain (LID) and the glucose response activation conserved element (GRACE). By analysis of deletion mutants, they mapped the shortest fragment of ChREBP that was responsive to glucose. This mutant, comprising the amino acids 37-298 (counting from the NH2-terminal ) was named GSM. Further, they observed that a mutant lacking amino acids 37-196 was constitutively active (always active independent of glucose levels).
They reasoned that the region from 37-196 was an inhibitory domain , which they named LID. The rest of the GSM (excluding the LID domain) was named GRACE. Ming et al suggest a novel mechanism of glucose mediated regulation of ChREBP which is independent of nuclear translocation. They hypothesize that under low glucose conditions LID represses GRACE and ChREBP is predominantly localized to the cytoplasm. This repression is possibly achieved by LID hindering the recruitment of co-activators to GRACE or by the recruitment of a co-repressor to ChREBP, which is subsequently released upon stimulation with high glucose. They further suggest that phosphorylation of S196 and T666 is a relatively recently acquired method for mediating glucose response by humans. This stems from the observation that these two phosphorylation sites are absent in the other Mondo proteins (of which ChREBP is one type) that exist in many species , from nematodes to higher mammals. The modulation of nuclear localization and DNA binding of ChREBP by phosphorylation of S196 and T666 respectively help to fine tune glucose homeostasis(Li et al.).
The discovery and characterization of this transcription factor have answered many questions regarding control of key lipogenic enzymes by glucose. Yet many questions remain unanswered. For instance, the role of ChREBP in tissues other than the liver, particularly glucose sensitive tissues like the pancreas and neurons. Ultimately, an understanding of the mechanism of ChREBP is expected to shed light on the pathogenesis of diseases like obesity, diabetes and the metabolic syndrome.
Reference List
Ishii, S., et al. “Carbohydrate response element binding protein directly promotes lipogenic enzyme gene transcription.” Proc.Natl.Acad.Sci.U.S.A 101.44 (2004): 15597-602.
Kawaguchi, T., et al. “Glucose and cAMP regulate the L-type pyruvate kinase gene by phosphorylation/dephosphorylation of the carbohydrate response element binding protein.” Proc.Natl.Acad.Sci.U.S.A 98.24 (2001): 13710-15.
Li, M. V., et al. “Glucose-dependent transcriptional regulation by an evolutionarily conserved glucose-sensing module.” Diabetes 55.5 (2006): 1179-89.
Stoeckman, A. K., L. Ma, and H. C. Towle. “Mlx is the functional heteromeric partner of the carbohydrate response element-binding protein in glucose regulation of lipogenic enzyme genes.” J.Biol.Chem. 279.15 (2004): 15662-69.
Tsatsos, N. G. and H. C. Towle. “Glucose activation of ChREBP in hepatocytes occurs via a two-step mechanism.” Biochem.Biophys.Res.Commun. 340.2 (2006): 449-56.
