QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by van Milgen, J. Search for Related Content PubMed PubMed Citation Articles by van Milgen, J.

---

Nutritional Models

Modeling Biochemical Aspects of Energy Metabolism in Mammals¹

INRA, Unité Mixte de Recherches sur le Veau et le Porc, 35590 Saint-Gilles, France

²To whom correspondence should be addressed. E-mail: jaap{at}st-gilles.rennes.inra.fr.

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION LITERATURE CITED

 
A framework representing the major biochemical pathways of nutrient metabolism is developed allowing quantification of the energy efficiency of different nutritional scenarios. The model is based on a number of carbon chain pivots (glucose, pyruvate, acetyl coenzyme A, -ketoglutarate, oxaloacetate and serine) and cofactors involved in metabolism. Excess pivots yield acetyl coenzyme A, which may be used for ATP or lipid synthesis. In contrast to previous work of this kind, the framework was constructed so that new insights in nutrient metabolism can be easily incorporated. Traditionally, integral values have been used to quantify mitochondrial ATP synthesis from cofactors (i.e., 3 ATP/NADH and 2 ATP/FADH₂), but current estimates are approximately 0.20 lower than previously assumed. Based on the latter, the energy expenditure for ATP synthesis from glucose was 91.0 kJ/ATP. For lipid (tripalmitin), 96.3 kJ/ATP was required whereas for amino acids energy expenditures varied between 99.2 (glutamate) and 153.2 kJ/ATP (cysteine). Energy derived from amino acid catabolism is stored and transferred either via carbon chain pivots or cofactors. It is hypothesized that this may affect the ultimate utilization of this energy (e.g., for ATP or lipid synthesis). The energy cost of nitrogen transport appeared relatively modest for most nonessential amino acids. Likewise, the net cost of using dietary glutamate and glutamine for ATP synthesis (e.g., in the viscera) and de novo synthesis of these amino acids in muscle is relatively minor and of similar magnitude as the cost of storing glucose energy as glycogen.

KEY WORDS: • energy efficiency • nutritional models • nutrient utilization • biochemistry

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION LITERATURE CITED

 
There are numerous ways to assess the energy value of feed or food. In one of its more basic forms, energy value can be represented by the chemical composition and gross energy value of the nutrients. However, not all nutrients are absorbed and, once absorbed, nutrients are used for different metabolic purposes. Consequently, the energy value is not necessarily a feed characteristic per se but may depend on the ultimate destiny of the nutrient. This concept is to some extent used in net energy (NE)³ systems, which have been proven useful tools in formulation of animal feeds (1 ). Empirical models such as NE systems have been criticized as being overly simple, little precise, not realistic etc., resulting in an urge to develop more mechanistic models of energy metabolism (2 ,3 ).

The energy expenditure or heat production of an animal is essentially a function of two processes. First, the biochemical transformation of a nutrient to an end product is often associated with a loss of energy for the animal. For example, Baldwin (4 ) calculated that only 0.84 of the glucose energy can be conserved as tripalmitin and ATP; the remainder is lost as heat. Second, energy expenditure is due to biophysical processes requiring ATP (5 ,6 ). Both the conversion of nutrients to ATP and the actual ATP utilization contribute to the heat production. Different authors have quantified the energy efficiencies of the main biochemical processes (3 ,7 –12 ). The work of Schulz (13 ) is rather advanced and includes both catalytic and anabolic processes for carbohydrates, amino acids and lipids. It assumes that excess amino acids are converted to either glucose or ATP, which can then be used for other purposes. Birkett & de Lange (3 ) and Chudy (12 ) used a similar approach, but used acetyl coenzyme A (acetylCoA) as a common intermediate.

Most of the previous work is based on the premise that the P/O ratio (i.e., the ratio of ATP synthesized to oxygen consumed) equals three for mitochondrial NADH and two for FADH₂. For NADH, this implies that 3 mol of ATP are produced per mol of NADH oxidized. Present consensus is that P/O ratios are considerably lower (6 ,14 ), which has an important impact on efficiency estimates of ATP synthesis. It is rather difficult and delicate to incorporate such changing insights in the previously mentioned approaches (7 –13 ). The objective of this paper is therefore to provide a general framework representing the main biochemical pathways of nutrient metabolism. Rather than establishing complete catabolic and anabolic pathways, nutrient metabolism is represented as a function of nutrients that play pivoting roles in quantifying nutrient metabolism. With this approach, it is relatively easy to incorporate new information without affecting the general structure of the framework. Although the paper does not generate new knowledge per se, it combines existing knowledge in such a way so that it can be exploited easily in models of nutrient metabolism or be used for educational purposes. A number of examples will be given in which the framework is used to quantify (in terms of energy metabolism) the consequences of using different hypotheses concerning the efficiency of mitochondrial ATP synthesis, nutrient transport and storage and inter-organ nutrient metabolism.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION LITERATURE CITED

 
Intermediary metabolism and ATP synthesis.

Only the quantification of energy efficiency of intermediary metabolism is considered here, and absorbed nutrients serve as a starting point. The core of the approach is the identification of a number of carbon chain nutrients and cofactors that play pivoting roles in nutrient metabolism. The choice of these pivots is not based on biochemistry per se, but serves as a means to quantify catabolic and anabolic processes. Six carbon chain pivots (glucose, pyruvate, acetylCoA, -ketoglutarate, oxaloacetate and serine) and eight cofactors [ATP, mitochondrial NADH (NADHm), cytoplasmic NADH (NADHc), NADPH, FADH₂, NH₃, O₂ and CO₂] are used. Catabolism and anabolism of amino acids, glucose and lipids (triacylglycerides) are expressed as a function of one or more of these pivots. Serine is included to quantify 1-carbon moieties (i.e., methyl groups) that result from the degradation or synthesis of certain amino acids. Catabolic and anabolic pathways are essentially those described by Salway (14 ). The pivots acetylCoA, -ketoglutarate and oxaloacetate are assumed to be present in the mitochondria. Transfer of nutrients from the mitochondrion to the cytoplasm (e.g., for gluconeogenesis or fatty acid synthesis) is accounted for in the stoichiometric balances. Guanosine triphosphate was assumed equivalent to ATP.

Figure 1 illustrates the possible conversions between carbon chain pivots, and the corresponding stoichiometry is given in Table 1 . It includes glycolysis (eq. 1), the tricarboxylic acid cycle (TCA; eqs. 2–4), oxidative phosphorylation (eqs. 10–13) and gluconeogenesis (eq. 9). Similar to the approaches of Schulz (13 ) and Livesey (11 ), cytoplasmic and mitochondrial NADH are considered separately, which allows relating mitochondrial membrane potential to ATP synthesis. For the purpose of illustrating the concept and comparison with previous work, integral values for the ATP synthesis from NADH, NADPH and FADH₂ are initially used (eqs. 10, 12 and 13) and identical ATP yields for cytoplasmic and mitochondrial NADH are assumed (eq. 11).

View larger version (22K): [in this window] [in a new window]   FIGURE 1 Possible conversions between carbon chain pivots. The stoichiometry of each reaction is specified in Table 1 (i.e., the number between parentheses corresponds to an equation number in Table 1 ).

During digestion, dietary lipids are hydrolyzed to glycerol and fatty acids. Subsequently, these can be re-esterified (i.e., before transport or in adipose tissue) or catabolized for energetic purposes. It was assumed that hydrolysis of lipids to free fatty acids and glycerol did not result in the use or synthesis of cofactors. The ß-oxidation of saturated fatty acids requires the equivalent of two ATP for acyl coenzyme A synthesis. The synthesis of NADHm, FADH₂ and acetylCoA depends on the fatty acid chain length (eqs. 14 and 15 in Table 2 ). For the mono-unsaturated and poly-unsaturated fatty acids, one additional NADPH is required for synthesis of trans--enoylCoA, whereas the FADH₂ release depends on the number of double bonds (14 ). Glycerol can be either catabolized to acetylCoA (eq. 19) or used for gluconeogenesis (eq. 20).

Amino acid catabolism results in a wide variety of carbon chains and cofactors (Table 3 ). Although amino acids can be used for synthesis of products such as bile pigments or creatine, this stoichiometry is not considered here. Degradation of histidine, methionine and tryptophan results in the release of 1-carbon units, which are accepted by tetrahydrofolate (THF) resulting in N⁵-formimino THF, N⁵-methyl THF and N¹⁰-formyl THF, respectively. These 1-carbon units attached to THF can be converted to N⁵, N¹⁰-methylene THF, which can then be used in the conversion of glycine to serine. This allows expression of the 1-carbon units as serine equivalents. Resuming:

So that

Amino acids given in excess of protein deposition will be deaminated and ammonia is considered the end product of amino acid catabolism. The amino group from certain amino acids will be accepted by -ketoglutarate to form glutamate. The NADHm that results from the conversion of glutamate back to -ketoglutarate is included in the stoichiometry of amino acid catabolism (Table 3) . Part of the NH₃ can be used in the synthesis of nonessential amino acids. The remainder is converted to urea in the urea cycle, requiring one CO₂ and four ATP per mol of urea synthesized (eq. 41 in Table 3 ).

Fatty acid and lipid synthesis.

Synthesis of unsaturated free fatty acids from acetylCoA involves the transfer of acetylCoA from the mitochondrion to the cytoplasm, which is carboxylated subsequently to malonyl coenzyme A. Synthesis of free fatty acids from acetylCoA results in a net requirement for ATP, NADHc and NADPH (eq. 42 in Table 4 ). Part of the NADPH required for fatty acid synthesis is provided by the pyruvate/malate cycle. The remaining NADPH can be supplied by the pentose phosphate pathway from glucose or other glucogenic precursors. Equation 43 represents the complete utilization of glucose to NADPH and CO₂ and can be used to balance NADPH requirements. Saturated fatty acids may be unsaturated, which requires one NADHc and one O₂ per double bond (eq. 44). Finally, synthesis of lipid involves activation of the fatty acids (requiring the equivalent of two ATP per fatty acid) and requires glycerol-3-phosphate, which can be synthesized from glucose (eq. 45).

The stoichiometry of amino acid synthesis from carbon chain pivots is given in Table 5 . Although dietary arginine may be required for young individuals, mammals have a capacity for arginine synthesis. Histidine is sometimes considered a nonessential amino acid (17 ), but its biosynthesis by mammals remains doubtful [Fuller, M. F., Lobley, G. (Rowett Research Institute) and Reeds, P. J. (University of Illinois), personal communications] and was not considered here. Likewise, postsynthetic modifications of amino acids in proteins and peptides (e.g., synthesis of methylhistidine, hydroxyproline) are not represented. In determining nutrient requirements, methionine and cysteine and phenylalanine and tyrosine are often considered together. Although the sulfur group of cysteine may be donated by methionine, its carbon chain originates from serine. Tyrosine can be synthesized from phenylalanine requiring one mol of O₂ and NADPH (14 ).

	RESULTS AND DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION LITERATURE CITED

 
Using the framework.

The stoichiometric equations listed in the tables allow for an easy quantification of most catabolic and anabolic reactions. For simple pathways, this quantification can be employed in a spreadsheet by summation of the appropriate reaction equations. For more complex situations, a compartmental representation of the system may be required whereby each pivot is represented as a state variable. The latter application is beyond the scope of this paper.

A simple application of the framework is the calculation of the ATP yield from different nutrients. For example, oxidation of acetylCoA in the TCA cycle results in the equivalent of 12 ATP (i.e., the sum of eqs. 3, 4 and 12 plus three times eq. 10). If similar logic is used for the other carbon chain pivots, it can be shown that complete oxidation of oxaloacetate, -ketoglutarate, pyruvate, glucose and serine results in 15, 24, 15, 38 and 13 ATP/mol, respectively. Syntheses of urea from ammonia (eq. 41) and serine (eq. 47) require the equivalent of 4 and 14 ATP/mol, respectively. This information, combined with that of Tables 1 , 2 and 3 , allows calculation of the ATP yield from different nutrients.

The energy expenditure per mol of ATP synthesized is conceptually equivalent to using metabolizable energy (ME) for maintenance purposes. Complete oxidation of glucose results in an energy expenditure of 2820/38 = 74.2 kJ/mol ATP (Table 6 ), whereas that for lipid is slightly greater (77.6 kJ/mol ATP for tripalmitin). The energy expenditure per mol of ATP synthesized varies considerably between amino acids and ranges from 77.4 for glutamate to 119.7 kJ ME/ATP for cysteine. These values are, in general, similar to previously reported values (7 ,8 ,13 ) although differences exist depending on the quantification of catabolic pathways. The largest differences between this and other studies occur for glycine and methionine. The difference for glycine is due to incomplete accounting of methyl groups in the glycine cleavage system by Krebs (7 ). The difference for methionine is also partly due to differences in quantifying the methyl group. Krebs (7 ) assumed that the methyl group of methionine was excreted as creatinine in the urine. Although methionine is a potent donor of methyl groups, the (dietary) origin of these groups is not necessarily methionine. Homocysteine can be remethylated by N⁵-methyl THF using a 1-carbon unit from other sources. In this study, it was assumed that the equivalent of 3 ATP was required to synthesize S-adenosylmethionine, whereas the methyl group was quantified as the equivalent of 0.5 serine (derived via N⁵-methyl THF and requiring 0.5 NADHc and yielding 1 NADHm). This results in the equivalent of 0.5 x 13 + 0.5 x 3 = 8 ATP for the methyl group from methionine. In addition, degradation of methionine requires serine whereas cysteine is synthesized. Schulz (13 ) assumed that cysteine was catabolized to pyruvate (which occurs with a lower efficiency than assumed here). This pyruvate was then carboxylated, and serine was re-synthesized via phosphoenolpyruvate. In the current approach, cysteine was also degraded to pyruvate, but the serine required for methionine catabolism may be obtained from other sources (e.g., from glucose or from excess glycine) resulting in a greater energy efficiency. The difference for lysine [37 ATP/mol vs. 35 assumed by Krebs (7 ) and Armstrong (8 )] is due to the assumption that no free glutaric acid is formed (14 ). Consequently, this saves two ATP used for the formation of glutarylCoA. Similarly, valine degradation yields two ATP more here because it is assumed that no free propionate is synthesized (14 ). The difference for histidine (24 vs. 21 ATP) is due to the quantification of the 1-carbon unit. Finally, the difference for threonine (22 vs. 21 ATP) is due to the assumption of degradation via threonine dehydrogenase (rather than threonine dehydratase) and the subsequent degradation to glycine and acetylCoA (15 ).

ATP plays an essential role in energy metabolism for both production and maintenance. In literature on animal production, there has been (and probably will continue to be) considerable debate on the appropriateness and usefulness of maintenance energy, which covers the energy expenditure of vital service functions of the organism and individual cells (5 ). Although the maintenance energy requirement is usually expressed as an ME or NE requirement, it is acknowledged that it is essentially an ATP requirement.

Traditionally, integral values P/O ratios have been used to quantify the ATP synthesis in the mitochondrial respiratory chain, which results in synthesis of 3, 3 and 2 ATP/mol NADH, NADPH and FADH₂, respectively [eqs. 10 (and 11), 13 and 12]. These values are now being challenged, and current estimates of ATP yield appear considerably lower than previously assumed. Oxidation of NADH and FADH₂ generates a proton gradient in the mitochondrial inner membrane, which can be used to synthesize and translocate ATP to the cytoplasm. Transport of NADH from the cytoplasm to the mitochondrion (via the malate/aspartate shuttle) and conversion of mitochondrial GTP to ATP result in a reduction of the proton gradient, thereby lowering the potential for ATP synthesis. To quantify the consequence of assuming lower P/O ratios, the stoichiometry of some equations in Table 1 was modified. It was assumed that oxidation of NADHm and NADPH result in the synthesis of 2.5 ATP (eqs. 10 and 13, respectively), whereas 1.5 ATP can be obtained from FADH₂ (eq. 12). Transfer of NADH from the cytoplasm to the mitochondrion (eq. 11) was associated with the use of 0.25 ATP, whereas conversion of mitochondrial GTP to ATP (eq. 4) would yield 0.75 rather than 1 ATP. Also the ATP requirement for urea synthesis from ammonia is affected by changing the assumptions concerning P/O ratios. Half of the ammonia enters the urea cycle via the transamination route in the cytoplasm and the other half via the transdeamination route in the mitochondrion (14 ). The stoichiometric balances for amino acid catabolism (Table 3) are based on the assumption that all released ammonia is mitochondrial. To use this ammonia via the transamination route, one NADHm has to be invested (to synthesize glutamate from -ketoglutarate), which will be released as NADHc (when malate is oxidized to oxaloacetate). Based on this information, the potential ATP yield from glucose, pyruvate, acetylCoA, oxaloacetate and -ketoglutarate would be 31, 12.25, 9.75, 12 and 19.25, respectively. Synthesis of serine from glucose represents the equivalent of 11.25 ATP whereas degradation of serine via pyruvate is equivalent to 10 ATP. Assuming these nonintegral P/O ratios significantly reduces the potential ATP yield from different nutrients and thus increases the energy requirement for ATP synthesis (Table 6) . For glucose, proportionally 0.23 more energy is required to synthesize ATP when using nonintegral P/O ratios. For tripalmitin, 0.24 more energy is required whereas for amino acids this averages 0.30 (ranging from 0.26 for isoleucine to 0.38 for histidine). The important sensitivity for amino acids concerning the assumptions of P/O ratios is due to a lower ATP yield (from NADH, etc.) combined with an increased ATP requirement for urea synthesis. Consequently, changing the assumption concerning P/O ratios changes the relative energy values of nutrients. When assuming integral P/O ratios and relative to glucose (74.2 kJ/ATP = 1.00), tripalmitin and (average) amino acids given in excess of protein deposition have energy values of 0.96 and 0.83 (range 0.62–0.96), respectively. For the nonintegral P/O ratios assumed above, tripalmitin and amino acids have energy values relative to glucose (91.0 kJ/ATP = 1.00) of 0.95 and 0.78 (range 0.59–0.92), respectively.

Although the assumption of P/O ratios has important consequences on the stoichiometry and energetics of ATP synthesis, it will probably not revolutionize our view on whole animal energy expenditure. The main problem is that whole animal ATP requirements cannot be measured and that both the conversion of a nutrient to ATP and the ATP utilization itself result in the same product: heat. Observed differences in energy expenditure (e.g., by measurement of heat production or tissue oxygen consumption) can therefore be interpreted as differences in the efficiency of ATP synthesis or as a different ATP requirement per se. The situation is somewhat different for processes for which ATP requirements are (supposed to be) known. For example, it is often assumed that synthesis of a peptide bond requires five ATP. If this ATP was synthesized from glucose (i.e., the most efficient way), an energy expenditure of 2820/38 x 5 = 371 kJ/peptide is implied when using integral P/O ratios. Using the amino acid composition of protein deposition in pigs, it can be calculated that the maximum energy efficiency of protein deposition is close to 0.87. This value is considerably greater than what is found experimentally (efficiency close to 0.60). Assuming nonintegral P/O ratios, 2820/31 x 5 = 455 kJ/peptide would be required as ATP, which results in a maximum efficiency of 0.85. Changing the hypothesis concerning P/O ratios therefore only marginally changes the maximum efficiency of protein synthesis. More than 18 ATP (22 when assuming integral P/O ratios) would be required to explain an energy efficiency of 0.60 for protein deposition. The actual ATP requirements for protein deposition (including the cost of protein turnover, amino acid transport and associated costs) appear more important in explaining the energy efficiency of protein deposition than the efficiency with which this ATP can be synthesized. For the remainder of this text, the nonintegral P/O ratio indicated above will be used in quantifying aspects of energy metabolism.

Energy storage and transport.

After a meal, nutrients are stored temporarily to ensure a supply between meals. The energy loss associated with the temporary storage of glucose as glycogen depends on the storage site and the branching of glycogen. To synthesize liver glycogen, two ATP are required per glucose (one for synthesis of glucose-6-phosphate and one for glycogenesis). Neither of these are recovered when glycogen is hydrolyzed to glucose implying an energy loss of 2/31 = 0.06. Energy can also be stored temporarily as lipid. This may occur during short (e.g., between meals) and long time intervals (e.g., deposition of lipid during gestation and mobilization during lactation). For the temporary storage of glucose energy as tripalmitin, the process glucose tripalmitin ATP requires 14 mol of glucose and yields 334 ATP, which represents an energy cost of 14 x 2820/334 = 118.2 kJ/ATP. Compared with the direct utilization of glucose for ATP (91.0 kJ/ATP), temporary storage of energy as lipid requires 0.30 more energy. Although the mode of expression is somewhat different, the logic is essentially the same as that of Baldwin (Ref. 4 , Table 6.7). The less efficient storage of energy as lipid compared with glycogen is apparently the price to pay to store energy in a very dense form.

Energy from nutrients is transferred via carbon chain pivots and cofactors. For example, the energy that is stored as lipid originates from acetylCoA, ATP, NADH and NADPH (eq. 42). Synthesis of palmitic acid from acetylCoA requires the equivalent of 7 ATP per acetylCoA as cofactors. As acetylCoA itself represents the equivalent of 9.75 ATP, the energy preserved as palmitic acid originates for approximately 0.58 from acetylCoA and 0.42 from cofactors. When glucose is converted to acetylCoA, 2 x 9.75/31 = 0.63 of the glucose energy is retained as acetylCoA and 0.37 as cofactors (Table 6) . However, most of the latter can be used in lipid synthesis and glucose appears reasonably well balanced in terms of supplying carbon chains and cofactors.

A similar reasoning can be applied for amino acids. For example, degradation of aspartate yields oxaloacetate, which is equivalent to 12 ATP when oxidized in the TCA cycle. Complete oxidation of aspartate results in 14.75 ATP, 2.25 of which may be required to synthesize urea from NH₃. Consequently, 12/14.75 = 0.81 of the potentially useful energy of aspartate is preserved in the carbon chain pivot and 0.19 in the cofactors (Table 6) . For the other amino acids, between proportionally 0.44 (valine) and 1.05 (histidine) of the energy is preserved in the carbon chain pivot. The low value for valine suggests it to be a suitable source of cofactors (e.g., ATP) in situations where the capacity for ATP synthesis through the TCA cycle is limited, for example during fatty acids synthesis. The value of histidine exceeds unity and indicates that an energy investment from cofactors has to be made to synthesize -ketoglutarate from histidine.

Nonessential amino acids are not only transporters of energy but also of nitrogen. By comparing pivot balances for amino acid catabolism and synthesis, the cost of nitrogen transport can be quantified. This involves synthesis of an amino acid from its nearest carbon chain pivot followed by de-amination and catabolism to the pivot. No energy input is required to synthesize and catabolize glutamate, aspartate, alanine and glycine from and to their carbon chain pivots. Between one (glutamine) and three ATP equivalents (asparagine) are required for synthesis and breakdown of the other amino acids from and to their nearest carbon chain pivots. Except for asparagine, the stoichiometric energy cost of nitrogen transport is therefore relatively modest. Young and Ajami (19 ) judged glutamine to be an energetically more efficient N-transporter than glutamate as one NADH is required to synthesize glutamate from -ketoglutarate versus an investment of one ATP to synthesize glutamine from glutamate. However, in releasing the nitrogen, the NADH investment for glutamate synthesis can be recovered whereas the ATP investment from glutamine will be lost.

The previous section assumed that carbon chain pivots are present in sufficient quantities to ensure synthesis of amino acids. At the whole animal level, the supply of carbon chain pivots from amino acid catabolism will typically exceed the requirements for amino acid synthesis. Reeds et al. (20 ) suggested that the site of entry of amino acids in intermediary metabolism has implications on the efficiency of ATP synthesis (expressed as ATP per CO₂ produced or carbon involved). Oxaloacetate and -ketoglutarate have to be converted to pyruvate (via phosphoenolpyruvate) before the energy can be used for ATP or fatty acid synthesis. This conversion requires the equivalent of 0.25 ATP that cannot be recovered. However, this loss is minor relative to the energy involved. More importantly, expressing energy efficiency as ATP yield per mol of carbon can potentially be misleading as not all carbon chain pivots are equally oxidized. Conversion of oxaloacetate to pyruvate requires the equivalent of 0.25 ATP, whereas carboxylation of pyruvate back to oxaloacetate only requires 1 ATP. Consequently, the carbon that differentiates oxaloacetate from pyruvate represents a low energy potential (i.e., 1.25 ATP) and care must be taken when expressing energy or ATP yields relative to carbon.

Inter-organ aspects of energy metabolism.

Glutamate and glutamine are thought to contribute extensively to the energy metabolism of visceral organs (21 ,22 ). Relatively little glutamate and glutamine arrive at the portal vein, implying that most of these amino acids have to be synthesized de novo to ensure protein deposition in skeletal muscle (e.g., from glucose). This scenario seems energetically less efficient than utilizing glucose for ATP in viscera and utilizing dietary glutamate for protein deposition in muscle. Stated otherwise, is the process glucose ATP energetically more efficient than glucose glutamate ATP (with glucose glutamate in muscle and glutamate ATP in viscera)? As pyruvate is a pivot for converting glucose to glutamate (using pyruvate carboxylase) and glutamate to ATP (via phosphoenolpyruvate), the process pyruvate glutamate pyruvate quantifies the additional cost of using glutamate for ATP synthesis relative to using glucose. Although the conversion of pyruvate to glutamate and back to pyruvate involves the complete oxidation of acetylCoA, the net cost is the equivalent of only 1.25 ATP [i.e., the difference between the utilization of acetylCoA in the TCA cycle and sum of eqs. 3, 4 (with 0.75 rather than 1 ATP), 7, 8, 36 and 50, hence less than the energy involved in glycogen turnover.

The example given above can also be quantified for the other nonessential amino acids. For nonessential amino acids other than serine and glycine, this involves the cost of turnover of amino acids to/from their carbon chain pivots and (except for alanine) the cost of turnover of oxaloacetate and ketoglutarate to/from pyruvate. As indicated above, the first cost is relatively modest for most nonessential amino acids, whereas the latter is equivalent to 1.25 ATP. Consequently, the cost of pyruvate amino acid pyruvate is equivalent to 0 ATP for alanine, 1.25 for glutamate and aspartate, 2.25 for glutamine, 2.5 for arginine, 3.5 for proline and 4.25 for asparagine. Serine (and glycine) is not necessarily synthesized from pyruvate because it can be synthesized more efficiently from glucose (via 3-phosphoglycerate). The process glucose serine pyruvate costs 2 ATP (1 ATP/mol of serine) more than the direct glucose pyruvate. Summarizing, the use of nonessential amino acids, rather than glucose, as energy source does not necessarily imply an extensive loss of energy and some nonessential amino acids appear suitable sources for temporarily storing and transporting energy and nitrogen. However, storage as protein rather than as free amino acid involves a considerably greater energy loss. For example, the process pyruvate glutamate as protein ATP is accompanied by a net energy loss of at least 1.25 + 5 = 6.25 ATP compared with 1.25 for glutamate as free amino acid.

Although some amino acids are efficient energy and nitrogen transporters, some important differences exist if the site of energy release is considered. For example, amino acids can be used in the liver for gluconeogenesis. A complete cycle of gluconeogenesis and glycolysis [i.e., pyruvate oxaloacetate glucose pyruvate requires the equivalent of 2.25 ATP (the sum of eqs. 7, 8 and 0.5 times eq. 1). For glutamate and alanine, the cost of this cycle is the same. However, gluconeogenesis from alanine requires three ATP and yields one NH₃ (the sum of eqs. 7, 9 and 35). As ATP is also required for ureagenesis from NH₃, gluconeogenesis from alanine requires the equivalent of 5.25 ATP in the liver. Gluconeogenesis from glutamate requires 1.25 ATP but also yields two NADHm, one FADH₂ and one NH₃ (i.e., the sum of eqs. 4, 9 and 36) and thus results in a positive balance of 3 ATP in the liver. Using both alanine and glutamate as energy transporters may therefore be beneficial in balancing the supply of carbon and cofactors at the site of gluconeogenesis.

The purpose of this paper was to lay down a quantitative framework of intermediary metabolism. With a limited number of balance equations, the quantitatively most important pathways of metabolism are represented. Relative to other studies of this kind, it is shown that expressing nutrient metabolism as a function of only ATP and/or glucose yield can be overly simplistic and inflexible. The proposed framework allows incorporation of new insights in energy metabolism (e.g., the efficiency of ATP synthesis) and quantification of different nutritional scenarios such as inter-organ nutrient transport and metabolism. To use the information of the framework in more elaborated models of nutrient metabolism, hypotheses concerning the regulation of intermediary metabolism with respect to both time and space are required.

	ACKNOWLEDGMENTS

 
The critical discussions with Nathalie Le Floc’h, Bernard Sève, Patrick Herpin and Jean Noblet are gratefully appreciated.

	FOOTNOTES

 
¹ Presented in part at the annual meeting of the American Society of Animal Science: Van Milgen, J. & Noblet, J. (2000) A biochemical model of nutrient utilization in growing pigs. J. Anim. Sci. 78 (Suppl. 1): 186.

³ Abbreviations used: acetylCoA, acetyl coenzyme A; ME, metabolizable energy; NADHc, cytoplasmic NADH; NADHm, mitochondrial NADH; NE, net energy; P/O ratio, ratio between mitochondrial ATP synthesis and oxygen consumption; TCA, tricarboxylic acid; THF, tetrahydrofolate.

Manuscript received 22 April 2002. Initial review completed 26 June 2002. Revision accepted 1 July 2002.

	LITERATURE CITED

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS AND DISCUSSION LITERATURE CITED

 

1. Noblet, J. (1996) Digestive and metabolic utilization of dietary energy in pig feeds: comparison of energy systems. Garnsworthy, P. C. Wiseman, J. Haresign, W. eds. Recent Advances in Animal Nutrition 1996:207-231 Nottingham University Press Nottingham, UK. .

2. Birkett, S. & de Lange, K. (2001) Limitations of conventional models and a conceptual framework for a nutrient flow representation of energy utilization by animals. Br. J. Nutr. 86:647-659.[Medline]

3. Birkett, S. & de Lange, K. (2001) A computational framework for a nutrient flow representation of energy utilization by growing monogastric animals. Br. J. Nutr. 86:661-674.[Medline]

4. Baldwin, R. L. (1995) Modeling ruminant digestion and metabolism 1995 Chapman & Hall London, UK. .

5. Milligan, L. P. & Summers, M. (1986) The biological basis of maintenance and its relevance to assessing responses to nutrients. Proc. Nutr. Soc. 45:185-193.[Medline]

6. Rolfe, D. F. & Brown, G. C. (1997) Cellular energy utilization and molecular origin of standard metabolic rate in mammals. Physiol. Rev. 77:731-758.[Abstract/Free Full Text]

7. Krebs, H. A. (1964) The metabolic fate of amino acids. Munro, H. N. Allison, J. B. eds. Mammalian Protein Metabolism 1964:125-176 Academic Press, Inc New York, NY. .

8. Armstrong, D. G. (1969) Cell bioenergetics and energy metabolism. Lenkeit, W. Breirem, K. Crasemann, E. eds. Handbuch der Tierernährung 1969:385-414 Parey Buchverlag Hamburg, Germany. .

9. Milligan, L. P. (1971) Energetic efficiency and metabolic transformations. Fed. Proc. 30:1454-1458.[Medline]

10. Schulz, A. R. (1975) Computer-based method for calculation of the available energy of proteins. J. Nutr. 105:200-207.

11. Livesey, G. (1984) The energy equivalents of ATP and the energy values of food proteins and fats. Br. J. Nutr. 51:15-28.[Medline]

12. Chudy, A. (2000) Model for the interpretation of energy metabolism in farm animals. McNamara, J. P. France, J. Beever, D. E. eds. Modelling Nutrient Utilization in Farm Animals 2000:329-346 C.A.B. (Commonwealth Agricultural Bureaux) International Oxon, UK. .

13. Schulz, A. R. (1978) Simulation of energy metabolism in the simple-stomached animal. Br. J. Nutr. 39:235-254.[Medline]

14. Salway, J. G. (1994) Metabolism at a glance 1994 Blackwell Publishers Oxford, UK. .

15. Ballevre, O., Cadenhead, A., Calder, A. G., Rees, W. D., Lobley, G. E., Fuller, M. F. & Garlick, P. J. (1990) Quantitative partition of threonine oxidation in pigs: effect of dietary threonine. Am. J. Physiol. 259:E483-E491.[Abstract/Free Full Text]

16. Le Floc’h, N., Obled, C. & Sève, B. (1995) In vivo threonine oxidation rate is dependent on threonine dietary supply in growing pigs fed low to adequate levels. J. Nutr. 125:2550-2562.

17. Bender, D. A. (1985) Amino Acid Metabolism 2nd ed. 1985 John Wiley & Sons Ltd Chichester, UK. .

18. Reeds, P. J., Cadenhead, A., Fuller, M. F., Lobley, G. E. & McDonald, J. D. (1980) Protein turnover in growing pigs. Effects of age and food intake. Br. J. Nutr. 43:445-455.[Medline]

19. Young, V. R. & Ajami, A. M. (2002) Glutamine: the emperor or his clothes?. J. Nutr 131:2449S-2459S.[Abstract/Free Full Text]

20. Reeds, P. J., Burrin, D. G., Davis, T. A. & Stoll, B. (1998) Amino acid metabolism and the energetics of growth. Arch. Anim. Nutr. 51:187-197.

21. Reeds, P. J., Burrin, D. G., Stoll, B. & Jahoor, F. (2000) Intestinal glutamate metabolism. J. Nutr. 130:978S-982S.

22. Nurjhan, N., Bucci, A., Perriello, G., Stumvoll, M., Dailey, G., Bier, D. M., Toft, I., Jenssen, T. G. & Gerich, J. E. (1995) Glutamine: a major gluconeogenic precursor and vehicle for interorgan carbon transport in man. J. Clin. Inves. 95:272-277.

This article has been cited by other articles:

				S. Lemosquet, E. Delamaire, H. Lapierre, J. W. Blum, and J. L. Peyraud Effects of glucose, propionic acid, and nonessential amino acids on glucose metabolism and milk yield in Holstein dairy cows J Dairy Sci, July 1, 2009; 92(7): 3244 - 3257. [Abstract] [Full Text] [PDF]

				C. M. Ross-Inta, Y.-F. Zhang, A. Almendares, and C. Giulivi Threonine-deficient diets induced changes in hepatic bioenergetics Am J Physiol Gastrointest Liver Physiol, May 1, 2009; 296(5): G1130 - G1139. [Abstract] [Full Text] [PDF]

				J. Noblet and J. van Milgen Energy value of pig feeds: Effect of pig body weight and energy evaluation system J Anim Sci, January 1, 2004; 82(13_suppl): E229 - 238. [Abstract] [Full Text] [PDF]

				J. van Milgen and J. Noblet Partitioning of energy intake to heat, protein, and fat in growing pigs J Anim Sci, February 1, 2003; 81(14_suppl_2): E86 - 93. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by van Milgen, J. Search for Related Content PubMed PubMed Citation Articles by van Milgen, J.