The Journal of Nutrition Vol. 128 No. 4 April 1998, pp. 733-739

In Vivo Glucose Contribution to Glutamate Synthesis Is Maintained While Its Contribution to Acetyl CoA Is Lowered in Adult Mice Fed a Diet with a High Fat:Carbohydrate Ratio^1,2,3

Monica Pascual, Farook Jahoor, and Peter J. Reeds⁴

U.S. Department of Agriculture/ARS Children's Nutrition Research Center, Department of Pediatrics, Baylor College of Medicine, Houston, TX 77030

	ABSTRACT

Abstract Introduction Methods Results Discussion References

To investigate the contribution of dietary carbohydrate to glutamate and acetyl CoA synthesis, two groups of adult mice were fed a high- (HCD) or a low-carbohydrate diet (LCD) in which 5% of the carbohydrate was [U-¹³C]-glucose. Four animals from each dietary group were killed after 1, 2 and 5 d. The tracer:tracee ratios of [¹³C₃] and [¹³C₆]blood glucose and of the [¹³C₂] and [¹³C₃] isotopomers of blood, mucosal, hepatic and muscle alanine and glutamate were used to calculate the fractional contribution of glucose to the 3-carbon, acetyl CoA and oxaloacetate pools of each tissue. In the HCD mice, glucose contributed 66, 33 and 31% of the acetyl CoA pool of muscle, liver and mucosa, respectively. The contribution of glucose to acetyl CoA was lowered by 33% (P < 0.05) and 55% (P < 0.01) in the liver and muscle of the LCD group, respectively, but was unaltered in the mucosa. Glucose made a minor contribution to glutamate synthesis via oxaloacetate in the liver (23%) and muscle (10%) of the HCD group. The fraction of hepatic and muscle glutamate synthesis derived from glucose was not affected by the diet. We conclude that glucose oxidation in liver and muscle parallels the contribution of carbohydrate to dietary energy and that glucose is not a major carbon precursor for muscle glutamate synthesis. Net glutamate synthesis in extraintestinal tissues is preserved when dietary carbohydrate is restricted.

	INTRODUCTION

Abstract Introduction Methods Results Discussion References

Glutamic acid is the only amino acid that can be synthesized by reductive amination of a keto acid (-ketoglutarate), so that it is the ultimate nitrogen donor for the net synthesis of other nonessential amino acids. It is also the sole carbon precursor of glutamine, an amino acid that plays a pivotal role in interorgan nitrogen transport, immune function, acid base balance and purine and pyrimidine base synthesis (summarized by Bulus et al. 1989, Lacey and Wilmore 1990, Perriello et al. 1995). Recent reports have confirmed earlier work (Windmueller and Spaeth 1975 and 1980) and have shown that the tissues of the splanchnic bed catabolize >90% of dietary glutamate and glutamine (Battezzati et al. 1995, Matthews et al. 1993) and that much of this catabolism occurs in the intestinal mucosa (Reeds et al. 1996). It follows from these observations that de novo synthesis must contribute the large majority of glutamate and glutamine elsewhere in the body. The carbon source of the high flux of glutamate and glutamine in vivo is not known with certainty, although both glucose (Lapidot and Gopher 1994, Perriello et al. 1995) and branched-chain amino acids (Chang and Goldberg 1978) are known contributors.

The synthesis of -ketoglutarate involves the condensation of oxaloacetate and acetyl CoA. Because a single molecule of glucose can lead to the synthesis of 1 mol of both acetyl CoA and oxaloacetate, glucose is an attractive candidate as a glutamate precursor although studies in fasted human beings (Perriello et al. 1995) have shown that plasma glucose is a minor contributor to forearm glutamine production. However, in the fasted state, glucose uptake and metabolism via the Krebs cycle in skeletal muscle is suppressed, and it is possible that glucose is an important precursor for glutamate synthesis in the fed state. If this were so, then dietary carbohydrate restriction could compromise the endogenous synthesis of glutamate and glutamine and have effects unrelated directly to the role of glucose as an energy substrate.

Although the complex metabolic interactions between carbohydrate and amino acids play a key role in metabolic homeostasis, their complexity poses difficulties for quantifying the relative rates of the pathways involved. This paper reports the second portion of a study in which we used oral uniformly ¹³C- labeled glucose ([U-¹³C] glucose) as a tracer to follow the metabolic fate of dietary glucose. The first paper (Pascual et al. 1997) reported data on the initial metabolic recycling of glucose. The aims of this part of the study were to quantify the partitioning of glucose to acetyl CoA and glutamate synthesis and to examine the effect of habitual carbohydrate intake on these pathways. We hypothesized that dietary carbohydrate would make a substantial contribution to the synthesis of glutamate and that this pathway would be preserved when dietary carbohydrate was in restricted supply. Conversely, we hypothesized that the contribution of glucose to acetyl CoA synthesis would vary in proportion to the contribution of glucose to the overall energy supply of the animal.

	MATERIALS AND METHODS

Abstract Introduction Methods Results Discussion References

This study received prior approval from the Animal Protocol Review Committee of Baylor College of Medicine. Care of the animals conformed to U.S. Department of Agriculture (USDA) regulations.

Isotopic tracer.  [U-¹³C] glucose (isotopic composition 91.7% [¹³C₆]glucose, 7% [¹³C₅]glucose) was purchased from Cambridge Isotopes (Andover, MD).

Animals and diets.  The results presented in this paper are derived from an experiment detailed in a previous paper (Pascual et al. 1997). In brief, the experiment involved adult female mice (ICR strain; mean body weight 25 g) purchased from Harlan Sprague Dawley (Indianapolis, IN). They were assigned randomly to two groups that received a high- (HCD) or low-carbohydrate diet (LCD; Table 1). The diets were offered in amounts that supplied similar amounts of dietary energy [HCD, 65.1 kJ/(d·mouse); LCD, 64.8 kJ/(d·mouse)] and protein [HCD, 700 mg/(d·mouse); LCD 672 mg/(d·mouse)]. The mice were given 5 d to become accustomed to the diets, after which time they were offered diets of the same overall composition but containing [U-¹³C]-glucose. After 1, 2 and 5 d, groups of four mice were anesthetized deeply with isoflurane and exsanguinated via cardiac puncture. Samples of liver, intestine and hindlimb muscles were obtained rapidly and frozen in liquid nitrogen.

Sample preparation.  Blood samples were processed as described by Pascual et al. (1997). Tissue samples were homogenized in 1 mol/L perchloric acid (10 mL/g tissue) and centrifuged. The supernatant was brought to pH 4-6 with KOH (4 mol/L) and centrifuged, and the supernatant was loaded onto a 1-mL bed volume Dowex AG-50 (H⁺ form) cation exchange column. The sample front and a water eluant (2 mL) were collected and taken for isotopic analysis of glucose (Keshen et al. 1997). The glucose isotopic labeling was determined by selected ion monitoring of the methane positive chemical ionization spectrum (mass-to-charge 331-337).

Alanine and glutamate then were eluted from the resin with 2 ml of 2 mol/L NH₄OH. One fraction of the eluant was dried and used to determine alanine after conversion to the n-propyl ester of the heptafluorobutyramide derivative (Jahoor et al. 1994). The other fraction was dried in a centrifugal evaporator and redissolved in water, and glutamate was purified through an anion exchange column (Dowex 1 acetate form) after elution with acetic acid. The eluant also was dried, esterified and derivatized as for alanine. The labeling of the amino acids was determined by selected ion monitoring of the methane negative chemical ionization spectrum. We monitored the mass-to-charge ratio 307-310 for alanine and 407-410 for glutamate. All mass spectrographic measurements were made on a Hewlett Packard 9890 gas chromatograph quadrupole mass spectrometer (Hewlett Packard, Palo Alto, CA).

Calculations.  The labeling of glucose, alanine and glutamate is expressed as tracer:tracee ratios (mol ¹³C-isotopomer/100 mol ¹²C-analyte) after correction for the isotopomer distribution in unlabeled glucose and amino acids isolated from animals that had received no isotopic tracers. We used the matrix approach of Brauman (1966) as described previously (Berthold et al. 1995). Throughout the following section we designate compounds containing 1, 2 . . . X ¹³C atoms as [M + 1], [M + 2] . . . [M + X].

The calculations involved two main assumptions: that the labeling of intracellular alanine and glutamate reflects intracellular synthesis from extracellular (plasma) glucose rather than uptake of alanine and glutamate from the circulation, that tissue alanine is in isotopic equilibrium with tissue pyruvate (Jahoor et al. 1994) and that tissue glutamate + glutamine are in isotopic equilibrium with tissue -ketoglutarate (Magnussen et al. 1991).

The calculations were based on the following general reasoning. The principal metabolic fate of glucose is metabolism via glycolysis and the pyruvate so-generated can undergo four fates: reduction to lactate, transamination to alanine, decarboxylation to acetyl CoA and carboxylation to oxaloacetate. The first two reactions lead to the synthesis of compounds, which, by being present at higher concentrations than pyruvate, are analyzed more readily isotopically. The second two reactions lead to, respectively, the synthesis of acetyl CoA (the oxidative substrate for the Krebs cycle) and oxaloacetate (a key substrate for glutamate synthesis). Neither of these can be sampled easily, but under the conditions of this experiment, their contributions to -ketoglutarate (and hence glutamate) synthesis can be estimated by suitable analysis of the ¹³C-isotopomer distribution in glutamate.

The metabolism of [U-¹³C]glucose (i.e., [M + 6]glucose) leads to the synthesis of [M + 3]pyruvate. This isotopomer can arise only from the glycolytic metabolism of glucose so that the fractional glucose contribution to the 3-carbon pool can be calculated from the relative isotopic enrichments of [M + 3]alanine (sampling the pyruvate pool) and its precursors [M + 6] and [M + 3]glucose, the latter glucose isotopomer resulting from the operation of gluconeogenesis.

Thus the fractional contribution of glucose to the 3-carbon pool equals The critical observation that underlies the subsequent calculations is that metabolism of [M + 3]pyruvate via pyruvate carboxylase introduces [M + 3]oxaloacetate into the Krebs cycle, whereas the decarboxylation of [M + 3]pyruvate introduces [M + 2]acetyl CoA. Thus [M + 3]glutamate can only derive from the operation of the anaplerotic pathway: Pyruvate + CO₂ oxaloacetate  -ketoglutarate  glutamate.

Thus the fractional contribution of three-carbon precursors to glutamate synthesis via oxaloacetate is

(2)

Equation 2 underestimates the true contribution of pyruvate carboxylation to glutamate synthesis because there is an active metabolic cycle between fumarate and oxaloacetate (Fernandez and Des Rosiers 1995) that leads to the production of an equilibrium mixture of two positional isomers of [M + 3]oxaloacetateone labeled in carbons 1-3 (arising directly from carboxylation of [M + 3]pyruvate)and one labeled in carbons 2-4 (resulting from the recycling of [M + 3]fumarate). However, the decarboxylation step between citrate and -ketoglutarate leads to the loss of the carbon introduced from carbon 1 of oxaloacetate. Thus [M + 3]-ketoglutarate (and hence [M + 3]glutamate) is derived only from [M + 3]oxaloacetate labeled in carbons 2-4 and the tracer:tracee ratio of [M + 3]glutamate underestimates the true contribution of [M + 3]oxaloacetate by a factor of 2.

Thus the true fractional contribution of three-carbon precursors to glutamate via oxaloacetate is

(3)

In contrast to the single pathway that leads to the synthesis of [M + 3]glutamate, [M + 2]glutamate can derive from three sources[M + 3]_C1-C3 oxaloacetate, [M + 2]oxaloacetate derived from recycling of [M + 3]-ketoglutarate in the Krebs cycle and [M + 2]acetyl CoA.

However, the contributions of these three sources can be calculated. We know that at isotopic equilibrium the tracer:tracee ratio of [M + 3]_C1-C3 oxaloacetate equals that of [M + 3]_C2-C4 oxaloacetate and that this is defined by the tracer:tracee ratio of [M + 3]glutamate. Furthermore, the tracer:tracee ratio of [M + 2]-ketoglutarate arising from the recycling of [M + 3]-ketoglutarate also is equal to that of [M + 3]glutamate.

Thus the contribution of oxaloacetate to the [M + 2]glutamate signal equals

(4)

The difference between the value calculated with eq. 4 and the measured value for the tracer:tracee ratio of [M + 2]glutamate measures the contribution of [M + 2]acetyl CoA to the isotopic signal. Thus

(5)

Finally, under the conditions of this experiment, there is production of [M + 2]pyruvate via tracer recycling and [M + 2]acetyl CoA can derive from both [M + 3]- and [M + 2]pyruvate. Because [M + 2]pyruvate is also an equilibrium mixture of [M + 2]_C1-C2 and [M + 2]_C2-C3 pyruvate, and only [M + 2]_C2-C3 pyruvate, yields [M + 2]acetyl CoA, the fraction of [M + 2]acetyl CoA derived from the three-carbon pool is

(6)

Statistics.  All data are expressed as means ± 1 SD. The labeling data were subjected initially to two-way analysis of variance with time of feeding [U-¹³C]glucose and diet as independent variables. This analysis revealed no time trend and no time × diet interactions so the results are shown as the mean for all three time points. Differences between tissues, for the same metabolite isotopomer and diet group were assessed by paired t tests, with the appropriate Bonferroni correction for multiple comparisons. Differences within tissues between diet groups were assessed by grouped t tests. A P-value (2-tailed) of <0.05 was taken as statistically significant.

	RESULTS

Abstract Introduction Methods Results Discussion References

After 24 h of administration of [U-¹³C] glucose, equilibrium was attained in the key isotopomers of whole blood glucose, alanine and glutamate (Fig. 1). Table 2 shows the tracer:tracee ratios of the [¹³C₂] and [¹³C₃] isotopomers of alanine and glutamate in blood, liver, gut mucosa and muscle. Irrespective of the diet, liver and muscle, [¹³C₃]alanine had a higher (P < 0.05) tracer:tracee ratio than blood alanine. The tracer:tracee ratio of mucosal [¹³C₃]alanine was significantly (P < 0.01) lower than that of circulating alanine.

View larger version (19K): [in this window] [in a new window]   Fig 1. The tracer:tracee ratios (mol isotopomer/100 mol tracee) of whole blood [13C3]alanine and [13C3]glutamate + glutamine together with the labeling of their precursor ([13C6]glucose from mice fed, for 1-5 d, a high carbohydrate [650 g/kg] diet containing [U-13C]glucose [32.5 g/kg diet]. Data are means ± 1 SD for four mice killed at each time point. The mean values for alanine and glucose are not significantly different.

The ratio of [¹³C₃]alanine:[¹³C₂]alanine is a semiquantitative measure of the respective contributions of glycolysis and the decarboxylation of oxaloacetate to the three-carbon pool. The ratios of [¹³C₃]alanine:[¹³C₂]alanine in blood (3.2 ± 0.6), liver (3.5 ± 0.6) and mucosa (3.8 ± 1) did not differ. However, the ratio in skeletal muscle (5.6 ± 1.1 HCD and 6.2 ± 1.3 LCD) was significantly (P < 0.05) higher than that in the other sampled alanine pools. The ratio [¹³C₃]alanine:[¹³C₂]alanine was unaffected by the diet.

In the HCD group, the tracer tracee ratio of [¹³C₃]glutamate (the isotopomer derived from [¹³C₃]_C2-C4 oxaloacetate) was similar in blood and liver. The value was slightly lower (P < 0.05) in intestinal mucosa and markedly (P < 0.01) lower in skeletal muscle. Consumption of the LCD was associated with a 48% (P < 0.01) lower value for the tracer:tracee ratio of [¹³C₃]glutamate in the liver and a 72% (P < 0.001) lower value in the mucosa. Diet apparently did not alter the absolute tracer:tracee ratio of [¹³C₃]glutamate in skeletal muscle.

The ratio [¹³C₂]glutamate:[¹³C₃]glutamate is a crude measure of entry of [¹³C₂]acetyl CoA into the Krebs cycle and has a limit of 2 if no labeled carbon enters the Krebs cycle as [¹³C₂]acetyl CoA. The ratio [¹³C₂]glutamate:[¹³C₃]glutamate was significantly (P < 0.01) greater than 2 in all three tissues and was particularly high in the skeletal muscle of the HCD group (14.7). Consumption of the LCD was associated with a 53% (P < 0.001) lower ratio in muscle (14.7, HCD; 7.0, LCD) but an 86% (P < 0.001) higher value in the mucosa (5.1, HCD; 9.7, LCD). In skeletal muscle, the lower ratio of [¹³C₂]glutamate:[¹³C₃]glutamate in the LCD group reflected the 62% (P < 0.001) lower tracer:tracee ratio of [¹³C₂]glutamate, whereas the higher value in the mucosa of the LCD mice was largely due to a 73% fall in the tracer:tracee ratio of [¹³C₃]glutamate.

The calculated contribution of glucose to alanine synthesis is shown in Table 3. Despite the fact that the intestinal mucosa had been directly exposed to dietary [U-¹³C]glucose, glucose made a smaller contribution to mucosal alanine than it did to the alanine in the liver and muscle. The calculation shown in Table 3 was based on the use of the tracer:tracee ratios of plasma glucose as the denominator, and if we use the tracer:tracee ratio of the dietary [U-¹³C]glucose (4.8 mol/100 mol) in the calculation, only 10% (HCD) and 6% (LCD) of the mucosal [M + 3]alanine could have arisen from the dietary glucose tracer. In the liver, glucose and alanine were apparently in complete isotopic equilibrium. Paradoxically, in skeletal muscle, the ratio [¹³C₃]alanine:[¹³C₆]glucose +(0.5 × [¹³C₃]glucose)] was significantly greater than unity (HCD 1.44 ± 0.28; LCD 1.52 ± 0.21). There were no diet related effects on the relationship between tissue alanine and plasma glucose labeling.

The calculated contribution of the three-carbon pool to glutamate synthesis via oxaloacetate also is shown in Table 3. In the HCD group, 39 and 46% of the blood and mucosal glutamate, respectively, had been synthesized from oxaloacetate derived from three-carbon metabolism. These values were 61 and 58% lower (diet effect P < 0.001), respectively, in the blood and mucosa of the LCD group. The proportion of glutamate derived from three-carbon precursors in the liver (23%) and muscle (10%) of the HCD group were both significantly (P < 0.001) lower than in the mucosa. However, in contrast to the mucosa, the fractional contribution of pyruvate carboxylation to liver and muscle glutamate synthesis was unaffected by the consumption of the LCD.

Calculated estimates of the contribution of three-carbon precursors and glucose to tissue acetyl CoA are presented in Table 4. In the mucosa, diet appeared to have little effect on the contribution of three-carbon precursors to acetyl CoA (HCD, 66%; LCD, 61%). The contribution of three-carbon precursors to hepatic acetyl CoA (HCD, 33%; LCD, 23%) was lower than their contribution to mucosal acetyl CoA, and there was a significant diet effect (P < 0.01). In the skeletal muscle of the HCD group, three-carbon precursors contributed 66% of the acetyl CoA, but their calculated contribution (26%) was substantially (P < 0.001) lower in the LCD group. Combining the contribution of glucose to the three-carbon pool (Table 3) and of the three-carbon pool to acetyl CoA (Table 4) showed that at a maximum, 32% of the mucosal acetyl CoA derived from glucose and that this contribution was unaffected by the prior carbohydrate intake. Once again, if enteral glucose was taken as the precursor for mucosal metabolism, only 6.6% of the mucosal acetyl CoA of the HCD group and 3.7% of that of the LCD group could have derived directly from dietary carbohydrate. In the liver of the HCD group, apparently 33% of the acetyl CoA had derived from glucose and consumption of the LCD lowered the contribution by 42-19% of total.

In muscle, the calculation of the contribution of glucose to acetyl CoA synthesis is made difficult by the paradoxically high tracer:tracee ratio of [¹³C₃]alanine. In the calculations summarized in Table 4, we have constrained the contribution of glucose to muscle alanine to 1. With this constraint, it appeared that 66% of muscle acetyl CoA derived from glucose in the HCD group and consumption of the LCD lowered the contribution of glucose to acetyl CoA by 60% (P < 0.001) to 26% of the total.

	DISCUSSION

Abstract Introduction Methods Results Discussion References

Glucose as a glutamate carbon precursor.  Very little dietary glutamate and glutamine escapes degradation by the tissues of the splanchnic bed (Batezzatti et al. 1995, Matthews et al. 1993, Reeds et al. 1996, Windmueller and Spaeth 1980). This implies that the large majority of extrahepatic glutamate and glutamine must derive from synthesis de novo. This has important nutritional implications because the synthesis of glutamate and glutamine must require the continued provision of their carbon precursors. The first objective of this study was to quantify the role of dietary glucose in this respect.

The net synthesis of -ketoglutarate, and hence of glutamate and glutamine, requires the entry of both oxaloacetate and acetyl CoA into the Krebs cycle. Glucose is a particularly attractive candidate as a major glutamate biosynthetic precursor because it is capable of leading to the synthesis of both oxaloacetate and acetyl CoA and thereby can donate both portions of the glutamate molecule.

Measurements of this pathway based solely on the incorporation of labeled glucose-carbon into glutamate can give a misleading impression of the true contribution of glucose to glutamate synthesis. This is because it is possible that the label has entered the pathway only as acetyl CoA. For example, in the present study, the acetyl CoA portion of muscle glutamate was highly enriched with ¹³C, and if we had simply used the relative ¹³C-isotopic enrichments of glucose and glutamate (as would have been the case in a ¹⁴C-based study) as the basis of the calculation, we would have overestimated the contribution of glucose to muscle glutamate/glutamine synthesis by between 3.5- and 5-fold.

The combination of a [U-¹³C]tracer (glucose in the present case) and selected ion monitoring gas chromatography-mass spectrometry, however, allows the labeling of products of tracer metabolism to be dissected into isotopomers bearing different numbers of ¹³C atoms. Such data allow the identification of the pathways that led to the production of the labeled product from the [U-¹³C]precursor. In the present study, the key measurement was the labeling of [¹³C₃]glutamate, a labeled form that can only arise from the pathway [¹³C₆]glucose [¹³C₃]pyuvate + CO₂  [¹³C₃]oxaloacetate  [¹³C₃]-ketoglutarate  [¹³C₃]glutamate.

On the basis of the relative tracer:tracee ratios of [¹³C₃]alanine and glutamate, the contribution of three-carbon precursors (pyruvate, lactate and alanine) to the oxaloacetate portion of the glutamate molecule varied among the tissues. The highest proportional contribution was found in the mucosa of the HCD group, in which 46% of the glutamate flux apparently was derived from a pool that was in isotopic equilibrium with alanine. The contribution was lower in liver and was particularly low in skeletal muscle. This last observation is of critical importance because skeletal muscle is both a major contributor to whole body glutamine synthesis and a substantial site of glucose metabolism. The result is, nevertheless, compatible with in vitro data (Chang and Goldberg 1978).

There were also differences among the tissues in the metabolic response to a period of lower carbohydrate (and concomitantly higher triglyceride) intake. In the mucosa, the consumption of the LCD was associated with a markedly lower fractional contribution of three-carbon precursors to the tissue glutamate flux, whereas in the liver and muscle, their fractional contribution was not changed significantly. Given that this pathway involves pyruvate carboxylation as a key step, the results imply that there is tissue-specific regulation of pyruvate carboxylase in relation either to absolute carbohydrate intake or to the carbohydrate:lipid ratio of the diet.

In discussing this idea, it is important to recognize that pyruvate carboxylase may serve different functions in the three tissues. In the liver, the carboxylation of pyruvate is clearly a key factor in glucose synthesis. In this respect, the hepatic enzyme is activated by the presence of end products of the metabolism of alternative energy substrates, notably acetyl CoA generated from fatty acid oxidation (Agius and Alberti 1985). As in the present study, triglyceride was substituted for carbohydrate in the LCD diet, the input of lipid fuels was increased. Thus a potential restriction of the synthesis of pyruvate from carbohydrate in the liver was, conceivably, balanced by the upregulation of pyruvate carboxylase by the higher contribution of triglyceride to the diet.

In skeletal muscle and the intestinal mucosa, the function of pyruvate carboxylase is less clear. Neither tissue is a direct contributor to circulating glucose so pyruvate carboxylase is not serving a primary gluconeogenic role. Nevertheless, the present data show that mucosal pyruvate carboxylase was active in the HCD group but that its activity, at least as measured from the contribution of three-carbon precursors to glutamate synthesis, was suppressed by the consumption of the LCD. In muscle, on the other hand, pyruvate carboxylase was low in activity and unresponsive to the diet.

We would argue that in both the small intestinal mucosa and skeletal muscle, the activity of pyruvate carboxylase is related to amino acid metabolism. In the mucosa, there is good evidence that glutamate and glutamine are major oxidative substrates (Windmueller and Spaeth 1975). It is important to recognize, however, that, as pointed out by Chang and Goldberg (1978), the complete oxidation of amino acids (e.g., glutamate) that enter the Krebs cycle at or beyond -ketoglutarate requires their metabolism via acetyl CoA and hence necessitates their conversion to phosphoenolpyruvate by phosphoenolpyruvate carboxykinase. However, intestinal mucosal cells carry out substantial net synthesis of alanine (Ebner et al. 1994, Reeds et al. 1996, Watford 1994, Windmueller and Spaeth 1975) and both glutamine and glutamate (Watford 1994) are important carbon sources for this alanine synthesis. Thus in vivo, alanine loss in the portal blood potentially depletes the mucosal oxaloacetate pool, and we would argue that the carbon used to replenish the oxaloacetate pool is derived from glucose. In other words, although the mucosal cells are not a site of net glutamate/glutamine synthesis, there is substantial metabolic, and hence label, cycling within the mucosal glutamate pool, and this cycle is maintained by the entry of glucose carbon into oxaloacetate. However, when carbohydrate is in restricted supply, the cycling is lowered substantially.

Skeletal muscle, on the other hand, is a site of net glutamine (and hence glutamate) synthesis, but the present results suggest that glucose is a minor glutamate carbon precursor. Because fatty acid carbon cannot supply the necessary oxaloacetate for net glutamate synthesis, it would seem that much of the glutamate/glutamine carbon must derive from amino acid, probably branched-chain, catabolism (see Chang and Goldberg 1978). However, a small portion (~10%) (Chang and Goldberg 1978) of valine and isoleucine are oxidized via the three-carbon pool, and we would argue that this carbon also is replenished via carboxylation of pyruvate derived from glucose. However, unlike the mucosa, under fed conditions, the overall rate of muscle glucose metabolism is high, and the synthesis of oxaloacetate represents a low proportion of total glucose metabolism in skeletal muscle. Maintenance of this biosynthetic route of glucose metabolism can be sustained without compromising overall carbohydrate metabolic status even when glucose supply is lowered.

Glucose contribution to tissue acetyl CoA.  There is substantial literature on measurements of carbohydrate oxidation by indirect calorimetry (see for example, Flatt 1993) or by labeled CO₂ production from tracer glucose (Glamour et al. 1995). Although such measurements give important information on the contribution of carbohydrate to whole body energy expenditure and carbon balance, they give little insight into the role of glucose as an oxidative substrate in different tissues. Furthermore, the production of labeled CO₂ from ¹⁴C- or ¹³C-glucose does not necessarily indicate true oxidation because recycling of non-acetyl CoA carbon in the Krebs cycle leads to labeled CO₂ production without there necessarily being net substrate oxidation (Chang and Goldberg 1978).

It follows that to measure the true contribution of a given substrate to tissue energy generation, it is necessary to measure its contribution to acetyl CoA synthesis. The second objective of the study was to combine the use of a [U-¹³C]glucose tracer with measurements of the relative degrees of labeling of [¹³C₃]alanine, [¹³C₃]glutamate and [¹³C₂]glutamate to measure the contribution of three-carbon precursors, and hence glucose, to the acetyl CoA pool.

The results suggested that glucose made quite different proportional contributions to the energy supply of the three tissues that we studied. In the mucosa of the HCD group, no more than 30% of the acetyl CoA had derived from plasma glucose and, as emphasized above, if the tracer:tracee ratio of dietary glucose was used in the calculation, glucose was a very minor (<10%) contributor to mucosal energy. Glucose also made a relatively small contribution to hepatic acetyl CoA, and it was only in the skeletal muscle of the HCD mice that glucose made a major contribution to oxidative activity. Even so, this conclusion has to be drawn with caution because of the anomalous labeling of muscle [¹³C₃]alanine that we observed. We have no explanation for the data on alanine labeling but suspect that it relates to differences in the rate constants of turnover of the plasma glucose and muscle alanine pools. It should be noted therefore that the constraint we imposed on the calculations of muscle metabolism would have maximized the estimate of the contribution of glucose to muscle acetyl CoA.

It appeared that the low contribution of glucose to mucosal and hepatic acetyl CoA synthesis in the HCD group reflected differences in the organization of intermediary metabolism in the two tissues. In the mucosa, the activity of pyruvate dehydrogenase was high (i.e., 3-carbon precursors made a high contribution to acetyl CoA), but it appeared that glucose contributed a minor portion of the mucosal three-carbon intermediate flux. On the other hand, in the liver, glucose was the dominating source of alanine (and hence pyruvate) synthesis, but the three-carbon pool made a relatively low contribution to acetyl CoA synthesis. On the basis of these results, we conclude that in the mucosa there was an alternative source of three-carbon units, presumably amino acids, whereas in the liver, there was an alternative source of acetyl CoA, presumably fatty acids. The results from the mice that received the LCD showed also that the regulation pyruvate dehydrogenase activity was tissue-specific. The consumption of the LCD had no effect on the contribution of three-carbon precursors to mucosal acetyl CoA, but lowered substantially their contribution to acetyl CoA in both the liver and skeletal muscle. We believe that the lack of change in the contribution of three-carbon precursors to mucosal acetyl CoA reflects the continuing availability of amino acids as a mucosal energy source, whereas the change in the liver and skeletal muscle primarily reflected a switch toward the use of lipids, supported perhaps by a lower insulin level in the LCD mice (see for example, Denton et al. 1996, Sugden et al. 1996).

In summary, we believe that this study illustrates the potential usefulness of [U-¹³C]glucose as a tracer to investigate dietary carbohydrate metabolism in vivo. One particular strength of the approach is, we believe, its ability to yield information on the contribution of pyruvate to tissue acetyl CoA synthesis. When this measurement is combined with information on the contribution of glucose to the pyruvate pool, inferences can be drawn with regard to the relative contributions of glucose, amino acid and fatty acids to energy production in different tissues. On the basis of the present results, we conclude that the mucosal cells derive much of their acetyl CoA from pyruvate but that glucose is a relatively minor source of this pyruvate. This we believe further supports our recent data (Stoll et al. 1997) in piglets that suggest that amino acids (probably of dietary origin) are key energy substrates for the intestinal mucosa. On the other hand, the present results suggest that even in animals fed high carbohydrate diets, pyruvate makes a minor contribution to hepatic acetyl CoA and that lipids may be the most important hepatic energy source. Furthermore, the results suggest that even in the fed state glucose plays a minor role in mucosal glutamate/glutamine synthesis and that dietary carbohydrate intake affects the metabolic use of dietary glucose in a tissue-specific manner.

	ACKNOWLEDGEMENT

We are grateful to L. Loddeke for sound editorial advice.

	FOOTNOTES

Manuscript received 6 August 1997. Initial reviews completed 12 September 1997. Revision accepted 1 December 1997.

	LITERATURE CITED

Abstract Introduction Methods Results Discussion References

• Agius L., Alberti K. G. Regulation of flux through pyruvate dehydrogenase and pyruvate carboxylase in rat hepatocytes. Effects of fatty acids and glucagon. Eur. J. Biochem. 1985; 152:699-707 [Medline]
• Battezzati A., Brillon D. J., Matthews D. E. Oxidation of glutamic acid by the splanchnic bed in humans. Am. J. Physiol. 1995; 269:E269-E276 [Abstract/Free Full Text]
• Berthold H. K., Reeds P. J., Klein P. D. Isotopic evidence for the differential regulation of arginine and proline synthesis in man. Metabolism 1995; 44:466-473 [Medline]
• Brauman J. I. Least squares analysis and simplification of multi-isotope mass spectra. Anal. Chem. 1966; 38:607-610
• Bulus, N., Cerosimo, E., Ghishan, F. & Abumrad, N. (1989) Physiologic importance of glutamine. Metabolism 38 (suppl. 1): 1-5.
• Chang T. S., Goldberg A. L. The metabolic fates of amino acids and the formation of glutamine in skeletal muscle. J. Biol. Chem. 1978; 253:3685-3696
• [Free Full Text] Denton R. M., McCormack J. G., Rutter G. A., Burnett P., Edgell N. J., Moule S. K., Diggle T. A. Hormonal regulation of pyruvate dehydrogenase complex. Adv. Enzyme Regul. 1996; 36:183-198 [Medline]
• Ebner S., Schoknecht P., Reeds P. J., Burrin D. G. Growth and metabolism of gastrointestinal and skeletal muscle tissues in protein malnourished neonatal pigs. Am. J. Physiol. 1994; 266:R1736-R1743 [Abstract/Free Full Text]
• Fernandez C. A., Des Rosiers C. Modeling of liver citric acid cycle and gluconeogenesis based on ¹³C mass isotopomer distribution analysis of intermediates. J. Biol. Chem. 1995; 270:10037-10042 [Abstract/Free Full Text]
• Flatt J. P. Dietary fat, carbohydrate balance, and weight maintenance. Ann. N. Y. Acad. Sci. 1993; 683:122-140 [Medline]
• Glamour T. S., McCullough A. J., Sauer P.J.J., Kalhan S. C. Quantification of carbohydrate oxidation by respiratory gas exchange and isotopic tracers. Am. J. Physiol. 1995; 268:E789-E796 [Abstract/Free Full Text]
• Jahoor F., Burrin D. G., Reeds P. J., Frazer M. Measurement of plasma protein synthesis rate in the infant pig: an investigation of alternative tracer approaches. Am. J. Physiol. 1994; 267:R221-R227 [Abstract/Free Full Text]
• Keshen T., Miller R., Jahoor F., Jaksic T., Reeds P. Glucose production and gluconeogenesis are negatively related to body weight in mechanically ventilated, very low birth weight neonates. Pediatr. Res. 1997; 41:132-138 [Medline]
• Lacey J. M., Wilmore D. W. Is glutamine a conditionally essential amino acid? Nutr. Rev. 1990; 48:297-309 [Medline]
• Lapidot A., Gopher A. Cerebral metabolic compartmentation. Estimation of glucose flux via pyruvate carboxylase/pyruvate dehydrogenase by ¹³C NMR isotopomer analysis of D-[U-¹³C]glucose metabolites. J. Biol. Chem. 1994; 269:27198-27208 [Abstract/Free Full Text]
• Magnusson I., Schumann W. C., Bartsch G. E., Chandramouli V., Kumaran K., Wahren J., Landau B. R. Noninvasive tracing of Krebs cycle metabolism in liver. J. Biol. Chem. 1991; 266:6975-6984 [Abstract/Free Full Text]
• Matthews D. E., Marano M. A., Campbell R. G. Splanchnic bed utilization of glutamine and glutamic acid in humans. Am. J. Physiol. 1993; 264:E848-E854 [Abstract/Free Full Text]
• Pascual M., Jahoor F., Reeds P. J. Dietary glucose is extensively recycled in the splanchnic bed of fed adult mice. J. Nutr. 1997; 127:1480-1489 [Abstract/Free Full Text]
• Perriello G., Jorde R., Nurjhan N., Stumvoll M., Dailey G., Jenssen T. G., Bier D. M., Gerich J. E. Estimation of glucose-alanine-lactate-glutamine cycles in postabsorptive humans: role of skeletal muscle. Am. J. Physiol. 1995; 269:E443-E450 [Abstract/Free Full Text]
• Reeds P. J., Burrin D. G., Jahoor F., Wykes L., Henry J., Frazer E. M. Enteral glutamate is almost completely metabolized in first pass by the gastrointestinal tract of infant pigs. Am. J. Physiol. 1996; 270:E413-E418 [Abstract/Free Full Text]
• Sugden M. C., Fryer L. G., Holness M. J. Regulation of hepatic pyruvate dehydrogenase kinase by insulin and dietary manipulation in vivo. Studies with the euglycaemic-hyperinsulinaemic clamp. Biochim. Biophys. Acta 1996; 1316:114-120 [Medline]
• Stoll B., Henry J., Reeds P. J., Yu H., Jahoor F., Burrin D. G. Catabolism dominates the first-pass intestinal metabolism of dietary essential amino acids in milk-replacer fed piglets J. Nutr. 1997; 128:606-614
• Watford M. Glutamine metabolism in rat small intestine: synthesis of three-carbon products in isolated enterocytes. Biochim. Biophys. Acta 1994; 1200:73-78 [Medline]
• Windmueller H. G., Spaeth A. E. Intestinal metabolism of glutamine and glutamate from the lumen as compared to glutamine from blood. Arch. Biochem. Biophys. 1975; 171:662-672 [Medline]
• Windmueller H. G., Spaeth A. E. Respiratory fuels and nitrogen metabolism in vivo in small intestine of fed rats. J. Biol. Chem. 1980; 22:107-112

0022-3166/98 $3.00 ©1998 American Society for Nutritional Sciences