Phosphoenolpyruvate carboxykinase revisited: Insights into its metabolic role

Abstract

“For every complex problem there is a simple solution. And it is always wrong.” H. L. Mencken Our understanding of the major metabolic pathways is built on a solid record of research extending back to the beginnings of biochemistry. It is unlikely, for example, that a new pathway for the utilization of major metabolic fuels will be described in mammals. What is less certain, however, are the mechanisms responsible for the regulation of the complex networks of enzymatic reactions that make up a metabolic pathway. As an example, the details of the interaction between fatty acid and carbohydrate metabolism remain to be fully clarified. However, new insights have changed the way we view key steps in metabolic pathways. As we shall develop in this article, enzymes once considered important for one metabolic pathway are now being shown to be critical for another. This is primarily because of advances made in molecular genetics and in vivo tracer methods over the past three decades that have provided a new set of tools for the study of metabolism. The result has greatly extended our understanding of the role played by individual enzymes in specific metabolic pathways. In this review we will present the latest information on the biological role of the enzyme phosphoenolpyruvate carboxykinase (PEPCK)2 in intermediary metabolism. Since its discovery in the late 1950s, PEPCK has been recognized as a key step in hepatic and renal gluconeogenesis. It is widely recognized as a critical enzyme in diabetes, because an elevation in its activity is associated with the increased rate of hepatic glucose output characteristic of that disease [1]. The regulation of transcription of the gene for the cytosolic form of PEPCK (PEPCK-C) has also been intensively studied and will be detailed in a second article in this series. The interaction of transcription factors that control the expression of the gene for PEPCK-C has been something of a model for understanding the complex mechanisms responsible for regulating transcription of an acutely controlled gene that is responsive to a wide variety of metabolically generated signals. There are two isoforms of PEPCK in animals, the mitochondrial form (PEPCK-M) and the cytosolic form (PEPCK-C). PEPCK-M was first isolated by Utter and Kurahashi [2] in 1953 from chicken liver mitochondria, and PEPCK-C was first described by Nordlie and Lardy [3] in tissues of the rat. The two isoforms of PEPCK are coded for by different nuclear genes, have virtually the same molecular weight, and catalyze the same reaction with very similar kinetic properties [4–6]. It was recognized from the time of its discovery that PEPCK was a key enzyme in gluconeogenesis [7]. PEPCK catalyzes the following reaction: oxaloacetate + GTP ↔ PEP + CO2 + GDP. The oxaloacetate required for the reaction can be generated in the mitochondria from pyruvate by pyruvate carboxylase or synthesized by the cytosolic form of malate dehydrogenase as part of a shuttle that carries reducing equivalents (NADH) from the mitochondria to the cytosol. More details of the kinetic properties and the background physiology of the two forms of PEPCK are presented in detail in a review of the biology of enzyme by Hanson and Patel [8]. PEPCK-C first appears in the livers of mammals at birth [9]. Its appearance is linked to the initiation of gluconeogenesis in the liver and to the maintenance of glucose homeostasis in the perinatal period. In contrast, the gene for PEPCK-M is present in the liver before birth [10]. Transcription of the gene for PEPCK-C is initiated by the increase in hepatic cAMP and the fall in circulating insulin that occurs at birth [11]. The administration of dibutyryl cAMP to fetal rats in utero induces PEPCK-C mRNA [11] and activity [12, 13] in the liver suggesting that the gene for the enzyme can respond to the appropriate hormonal signals. Conversely, if newborn rats are administered glucose by injection, the development of PEPCK-C can be delayed significantly. This is because of the powerful negative effect of insulin on PEPCK-C gene transcription [14]. Rats born from diabetic mothers also have a delayed development of PEPCK-C in the liver because of the high levels of glucose and insulin in the fetal blood [15]. The renal form of PEPCK-C is present before birth and increases slowly after birth [16]. Animals vary in the intracellular distribution of the two forms of PEPCK. Most species studied to date (humans, dogs, cats, cows, sheep) have both forms of PEPCK at about equal activities [17] (see Table I). The rat and mouse have 95% PEPCK-C, whereas adult chickens and other birds have 100% PEPCK-M in their livers [18]. Interestingly, the liver of the developing chicken expresses PEPCK-C to make glucose from the amino acids derived from the protein of the egg [19]. At hatching, the expression of the gene for PEPCK-C is suppressed, and transcription of the gene for PEPCK-M is initiated in the liver. The mechanism(s) underlying this interesting switch in PEPCK isoforms is not understood. It has been proposed that PEPCK-M is involved in gluconeogenesis from lactate, because the generation of PEP from oxaloacetate in the mitochondria ensures that the appropriate balance of NADH is maintained in the cytosol of the liver and kidney cortex [20, 21]. A more detailed review of the role of the isoforms of PEPCK-C in different animal species can be found in Ref. 22. The tissue distribution of PEPCK-C in animal species provides insight into the physiological role of the enzyme. As mentioned above, PEPCK is generally considered the pace-setting step in gluconeogenesis and is accordingly termed a “gluconeogenic enzyme.” However, there is detectable activity of PEPCK-C in tissues that do not release glucose, such as white adipose tissue, brown adipose tissue, muscle, brain, small intestine, lung, and mammary gland [23]. Several physiological roles (other than gluconeogenesis) have been proposed for PEPCK-C (see Table II). For example, white adipose tissue synthesizes glycerol-3-phosphate from precursors other than glucose or glycerol by an abbreviated version of the gluconeogenic pathway termed glyceroneogenesis [24–26] (see Fig. 1 for the details of this pathway). This pathway is of significance during fasting, because up to 30% of the free fatty acids released by the adipose tissue during lipolysis are re-esterified to triglyceride, requiring a continued source of glycerol-3-phosphate [27]. Another potential role for PEPCK-C is cataplerosis. During amino acid catabolism there is an inflow for four and five carbon intermediates into the citric acid cycle. The citric acid cycle cannot fully oxidize four and five carbon compounds to carbon dioxide to generate energy. The carbon skeletons of the amino acids are thus removed from the cycle by cataplerosis. Cataplerosis can be linked to biosynthetic processes such as gluconeogenesis in the liver and kidney cortex or glyceroneogenesis in adipose tissue. Finally the mitochondrial isoform of PEPCK, PEPCK-M, could function as an anaplerotic enzyme. This could be important in muscle where there is a low activity of pyruvate carboxylase; PEPCK-M would generate oxaloacetate from PEP directly in the mitochondria thus replenishing the citric acid cycle. An excellent example of the complex metabolic role of PEPCK-C in tissues is illustrated by considering its function in the small intestine. PEPCK-C mRNA [28] and enzyme activity [29] are both present in the mucosa of the small intestine (highest in the proximal gut) in suckling rats and mice, where it reaches a maximum level between 3 and 9 days after birth and then slowly decreases at weaning. The metabolic role of PEPCK-C in the small intestine has been inferred from both in vitro and in vivo studies. Hahn and Wei-Ning [29] suggested that PEPCK-C in the gut may be responsible for the glucose synthesis from lactate locally for use in the small intestine. More recently, the rates of glucose synthesis by the small intestine in the adult rat have been determined in both normal and diabetic animals [30]. Although the rates are low compared with hepatic gluconeogenesis, the glucose synthesized by the small intestine may be critical for the metabolism of the mucasal cells. The small intestine may therefore be considered a gluconeogenic tissue. The small intestine also plays an important role in the catabolism of dietary amino acids. Most of the glutamine and aspartate in the diet (about 60%) are catabolized by the mucosa of the small intestine in the first pass by oxidation to carbon dioxide. These amino acids enter the citric acid cycle as four and five carbon intermediates and exit as malate (cataplerosis) that is subsequently converted to oxaloacetate in the cytosol by NAD malate dehydrogenase. Oxaloacetate is decarboxylated to PEP by PEPCK-C; the PEP generated by this route has the following potential metabolic fates: (a) conversion of PEP to pyruvate via pyruvate kinase, followed by oxidation to acetyl-CoA and subsequent entry into the citric acid cycle to generate energy; (b) transamination of the pyruvate to alanine, which is released by the small intestine; (c) conversion of the PEP to glycerol-3-phosphate via glyceroneogenesis; or (d) conversion to glucose via gluconeogenesis. A major difference between the two isoforms of PEPCK is the nature of the regulation of expression of their genes. PEPCK-C has a relatively short half-life (6 h) in the liver, and its levels are acutely regulated by hormones or diet [31, 32]. In contrast, PEPCK-M is constitutively expressed and not controlled acutely by these factors. The gene for PEPCK-C has been well characterized, and its regulation in a variety of mammalian tissues has been studied in detail. The major control of the level of PEPCK-C in all mammalian tissues is the result of alterations in the rate of transcription of its gene. However, relatively little is known of the factors that control PEPCK-C translation. In this regard, a recent report by Scheuner et al. [33] is of considerable interest. They noted that a mutation of the phosphorylation domain of the α-subunit of the eukaryotic translation factor eIF2α resulted in the complete loss of hepatic PEPCK-C gene expression in mice. This could be because of a direct effect on the translation of PEPCK-C or because of an alteration in the translation of specific transcription factor(s) involved in the control of expression of the PEPCK-C gene. More work on the translational control of PEPCK-C is clearly needed. In general, PEPCK gene expression in the liver is induced by fasting [34] or by a diet devoid of carbohydrate and inhibited by a diet high in carbohydrate [35]. The induction noted during fasting is because of increased glucagon and because of a decrease in insulin levels characteristic of the fasted state. Glucagon (acting via cAMP) and glucocorticoids stimulate PEPCK-C gene transcription [36] whereas insulin profoundly depresses this process [14]. As an example, PEPCK-C gene transcription can be induced 10-fold in 30 min by cAMP administration to carbohydrate-fed rats [36] whereas insulin causes a 50% decrease in transcription over the same time period [14]. Other regulators include thyroid hormone, retinoic acid, epinephrine (stimulate transcription), and high concentrations of glucose (decrease transcription) (see Ref. 37 for details). Although cAMP and insulin have their effects on PEPCK-C gene transcription in all tissues studied to date, there are interesting tissue-specific variations in control of transcription. For example, in the kidney, metabolic acidosis induces PEPCK-C gene transcription, whereas metabolic alkalosis represses this process [38]. Adipose tissue has an interesting variation in the control of PEPCK-C gene transcription. Glucocorticoids inhibit transcription of this gene in adipose tissue rather than induce the process as in the liver [34, 39]. A detailed discussion of the specific transcription factors involved in the complex pattern of regulation of PEPCK-C gene transcription outlined above will be presented in the second article in this series. The ability to modify gene expression in specific tissues has created a new and powerful tool for understanding metabolic processes. As an example, PEPCK-C has been studied by overexpressing its gene in the liver [1] or by deleting the gene both in the entire animal and in a tissue-specific manner in liver [40] and in adipose tissue [41]. The results of these studies have changed the way we think about the metabolic function of this enzyme. The gene for PEPCK-C has been overexpressed in the livers of mice by generating transgenic animals that contain the PEPCK-C structural gene linked to the rat PEPCK-C gene promoter [1]. The mice produced by this technique contain greatly increased levels of PEPCK-C mRNA in their livers and elevated PEPCK activity. The metabolic consequences of the overexpression of hepatic PEPCK-C are a greatly elevated level of hepatic glucose output and a concomitant increase in the concentration of blood glucose. There was no report of altered lipid metabolism in the mice. These findings suggest that PEPCK-C is a rate-controlling step in hepatic gluconeogenesis and that its level of activity in the liver can control the overall level of glucose synthesis. This is particularly relevant to diabetes where the control of hepatic glucose output is a major therapeutic goal. A global deletion of the gene for PEPCK-C in the mouse results in severe hypoglycemia and death within 2 days after birth [40]. At their death these mice have profound fatty livers; the liver is virtually white from accumulated triglyceride. Mice with a deletion of the gene for PEPCK-C only in the liver, created using Cre recombinase, survive to adulthood and are more able to maintain normal levels of blood glucose when fasted for up to 36 h [40]. These mice also have a fatty liver because of the accumulation of both fatty acids and triglyceride [40]. Despite the large quantities of hepatic fat, the mice do not have an elevation in the concentration of ketone bodies in the blood. This suggests that the fatty acids are not entering the mitochondria for oxidation to acetyl-CoA, perhaps because of regulation at the level of carnitine acyltransferase. Further study of the reason for the failure to induce ketone body synthesis in the face of high hepatic lipid content is required. These observations have two surprising aspects. First, it was not intuitive that animals would be able to maintain normal glucose homeostasis in the absence of hepatic PEPCK-C. This is especially surprising, because, as discussed above, the overexpression of PEPCK-C resulted in an elevated rate of hepatic gluconeogenesis [1]. It is probable that glucose synthesis from glycerol in the liver (glycerol bypasses PEPCK-C in the gluconeogenic pathway) and gluconeogenesis in the kidney are capable of maintaining normal glucose output during fasting in these animals. Second, the role of PEPCK-C in hepatic lipid metabolism was not predicted based on our traditional understanding of the function of this enzyme in intermediary metabolism. One clue to the important role of PEPCK-C in hepatic lipid metabolism comes from recent studies of glyceroneogenesis. It has been known for some time that ∼30% of the fatty acid that enters the liver during fasting is re-esterified to triglyceride, packaged as very low density lipoprotein and released from the liver. The source of the glycerol-3-phosphate required for the synthesis of triglyceride remains a question. It is clear that during fasting glucose utilization by the liver is negligible so that glucose cannot be the source of the glycerol-3-phosphate. It is thus possible that the glycerol moiety in triglyceride is derived from glycerol mobilized to the liver from adipose tissue during fasting or from glyceroneogenesis. Botion et al. [42] noted that hepatic glyceroneogenesis in rats fed a high protein diet (carbohydrate-free) accounted for a significant fraction of the glycerol-3-phosphate incorporated into triglyceride in the liver. Kalhan et al. [43] reported that glyceroneogenesis was responsible for up to 60% of the glycerol found in the triglyceride in very low density lipoprotein of humans fasted overnight. Surprisingly, less than 5% of the glycerol in triglyceride was derived from glycerol in these subjects; the glycerol was converted almost exclusively to glucose. It is thus probable that hepatic glyceroneogenesis is a critical pathway for the re-esterification of fatty acids to triglyceride during periods of fasting. The longer the fast the more important this process becomes. In mice with an ablation of PEPCK-C gene in the liver, it is probable that there would be a limited rate of glyceroneogenesis and as a consequence, a failure to appropriately re-esterify fatty acids. This may interfere with the normal flux of triglyceride in the liver and account, at least in part, for the observed fatty liver in mice with a liver-specific deletion in the gene for PEPCK-C. The gene for PEPCK-C has also been deleted in white adipose tissue from the mouse using an interesting strategy with wide potential application for metabolic studies. The PEPCK-C gene promoter contains a PPARγ regulatory element (PPARE) that maps at −1000 in the promoter. The PPARE is required for the expression of the PEPCK-C gene in adipose tissue [44]. A copy of the mouse PEPCK-C gene, with a deletion in the PPARE, was introduced into the mouse genome by homologous recombination in place of the endogenous PEPCK-C gene [41] (see Fig. 2 for details of the generation of these mice). Mice with this so-called tissue-specific enhancer knock-out do not express PEPCK-C in white adipose tissue. The absence of PEPCK-C in this tissue results in mice that cannot carry out glyceroneogenesis in their adipose tissue. The animals fail to deposit triglyceride and subsequently develop severe lipodsytrophy [41]. It is of interest that the mice do not develop insulin resistance or a fatty liver perhaps because of the fact that glyceroneogenesis is important for the synthesis of the glycerol-3-phosphate for triglycerides only during fasting, glucose being the major source of glycerol-3-phosphate during the fed state. Another line of mice were generated using the tissue-specific enhancer knock-out technique described in detail in the legend to Fig. 2. These animals have the neomycin resistance gene inserted in place of the PPARγ binding site in the PEPCK-C gene promoter (third line in Fig. 2). This results in mice that do not express the gene for PEPCK-C in the liver, adipose tissue, small intestine, and heart but have normal levels of expression of PEPCK-C in the kidney. They also are able to maintain normal levels of glucose after 24 h of starvation but developed a severe fatty liver. Interestingly, ∼50% of the mice are born half the size of their control littermates (Fig. 3). The reason for the small size is not clear, but it is evident that PEPCK-C is critical for metabolic processes that extend beyond its role in gluconeogenesis. The importance of PEPCK-C and glyceroneogenesis in adipose tissue is demonstrated by a recent report of mice that have a deletion in the gene for GLUT-4, which encodes a glucose transporter [45]. These animals have a normal pattern of growth and a normal adipose tissue mass despite a marked impairment of insulin-stimulated glucose uptake by isolated adipocytes. On the other hand they are insulin-resistant and develop glucose intolerance and hyperinsulinemia. The ability of these mice to maintain normal levels of triglycerides in their adipose tissue, despite the absence of glucose uptake to generate glycerol-3-phosphate for triglyceride synthesis, strongly suggests that glyceroneogenesis provides the glyceride-glycerol in this tissue. The role of PEPCK-C in both carbohydrate and lipid metabolism, as described in this review, underlines the critical importance of PEPCK-C in diabetes. Because diabetes is a disease that involves both of these metabolic processes, it is probable that PEPCK-C is a major regulatory point in coordinating both lipid and carbohydrate metabolism. Insulin suppresses PEPCK-C gene transcription; this would reduce both glucose output from the liver and triglyceride synthesis during fasting. In Type 2 diabetes, there is an elevation in the levels of both of these compounds because of insulin resistance. It is thus reasonable to view PEPCK-C as a target for the hormonal control of metabolic processes in the liver. In addition, glucocorticoids are known to be required for the development of diabetes; adrenalectomized animals do not readily develop the symptoms of diabetes [46, 47]. Glucocorticoids stimulate PEPCK-C gene transcription in the liver thus insuring continued glucose synthesis. In diabetes the stimulatory effects of glucocorticoids are not countered by insulin, and the increased rates of gluconeogenesis produce excess blood glucose. In contrast, glucocorticoids depress PEPCK-C gene transcription in adipose tissue thereby diminishing glyceroneogenesis and increasing the outflow of free fatty acids from this tissue. The overall effect is catabolic. Because PEPCK-C is the key regulatory step in both pathways, its role as a therapeutic target for anti-diabetic drugs seems appropriate. Most metabolic pathways have long ago reached the status of enshrinement in textbooks of biochemistry. Little can be added to change our perception of how these critical biochemical routes work in various tissues. However, there are exceptions. PEPCK was discovered in the 1950s, but its exact biological role remains elusive. For example, the metabolic function of the two isoforms of PEPCK is not clear; most discussions of gluconeogenesis completely ignore the mitochondrial form of PEPCK despite the fact that it represents half the total activity of the enzyme in the livers of most mammals, including humans. The PEPCK is clearly a key factor in the control of glucose synthesis in the liver and kidney. However, it is present in tissues that do not synthesize glucose. Its role in glyceroneogenesis is less well appreciated but seems, from current information, to be equally important to gluconeogenesis in insuring a balance between carbohydrate and lipid metabolism. In this review we have discussed some of the new data that are currently emerging that are changing our understanding of the role of PEPCK in metabolism. In a future article in this series we will discuss how transcription factors coordinate the response of PEPCK-C and other genes to hormones and dietary stimuli thereby insuring the appropriate level of activity of key enzymes in metabolic pathways. Pathway of glyceroneogenesis from lactate or alanine in adipose tissue. This figure illustrates the glyceroneogenic pathway featuring the role of PEPCK-C in this process. Pyruvate is carboxylated to oxaloacetate in the mitochondria via pyruvate carboxylase and oxaloacetate reduced to malate by NAD malate dehydrogenase. Malate is transported to the cytosol and reduced to oxaloacetate by the cytosolic form of NAD malate dehydrogenase. PEPCK-C then decarboxylates oxaloacetate to PEP, which is subsequently converted to dihydroxyacetone and then to glycerol-3-phosphate via an abbreviated version of the gluconeogenic pathway. This glycerol-3-phosphate is then used for the synthesis of triglyceride. Outline of the strategy used to generate mice with a tissue-specific enhancer knock-out. The first line illustrates the 1.1-kb 5′ flanking region of the mouse PEPCK gene. It is flanked on both sides by mouse genomic sequences (dashed lines, out of scale) constituting the 6-kb EcoRI (E)-SalI (Sa) fragment (from 5′ to 3′) that was used for homologous recombination. The arrows indicate some regulatory sites including the distal PPARE (open box PP), the CAAT/enhancer-binding protein sites (P3 and P4), the HNFI site (P2), and the cAMP-response element-binding protein (CREB) recognition sites. Also shown are two genomic XhoI (X) sites and a new site immediately adjacent to the loxP, which is diagnostic for the mutation before and after the excision of the loxPneo sequence. The second line illustrates the targeting vector with the PPARE site replaced by a Cre-loxPneo cassette, containing the positive neo resistance-selectable marker gene flanked by two loxP sequences indicated by shaded boxes. At the 3′ end the diphtheria toxin α-chain gene (DT) was inserted as the negative selection against non-homologous recombination. The third line of the diagram illustrates the structure of the correctly targeted chromosome in the E14.1 embryonic stem cells (ES) (after electroporation). The fourth line presents the targeted chromosome in which the neo cassette was deleted by Cre recombinase leaving behind a single loxP box (shaded) and preserving the new XhoI site. For additional details see Olswang et al. [41]. Mice with a deletion in the PPAR γ binding site in the PEPCK-C gene promoter. Mice produced by the procedure outlined in Fig. 2 are shown. The smaller animal at the right is homozygous for a deletion in the PPARγ biding domain in he PEPCK-C gene promoter. The selectable marker neo gene was not removed during the generation of these mice. The animals do not express the gene for PEPCK-C in the liver, white and brown adipose tissue, and small intestine but do express the gene in the kidney. We note that 30 to 50% of the mice homozygous for the modified PEPCK-C gene have the small phenotype.