tricarboxylic acid cycle intermediate pool size

Sports Med 2007; 37 (12): 1071-1088REVIEW ARTICLE 0112-1642/07/0012-1071/$44.95/0

© 2007 Adis Data Information BV. All rights reserved.

Tricarboxylic Acid Cycle IntermediatePool SizeFunctional Importance for Oxidative Metabolism inExercising Human Skeletal Muscle

Joanna L. Bowtell,1 Simon Marwood,2 Mark Bruce,3 Dumitru Constantin-Teodosiu4

and Paul L. Greenhaff4

1 Academy of Sport, Physical Activity and Wellbeing, London South Bank University,London, UK

2 Department of Sports Studies, Liverpool Hope University, Liverpool, UK3 School of Sport and Exercise Sciences, Loughborough University, Leicestershire, UK4 School of Biomedical Sciences, Nottingham University, Nottingham, UK

ContentsAbstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10711. Oxidative Metabolism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1074

1.1 Control of Oxidative Phosphorylation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10741.2 Tricarboxylic Acid (TCA) Cycle and its Regulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1075

2. Exercise-Induced Modulation of TCA Cycle Intermediate (TCAi) Pool Size . . . . . . . . . . . . . . . . . . . . . 10753. Anaplerotic and Cataplerotic Pathways . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10764. Functional Significance of Changes in TCAi Pool Size . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1078

4.1 Effect of Exercise Training . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10794.2 Effect of Pharmacological Interventions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10814.3 Effect of Amino Acid Supplementation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10824.4 Effect of Altering Carbohydrate Availability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1083

5. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1084

The tricarboxylic acid (TCA) cycle is the major final common pathway forAbstractoxidation of carbohydrates, lipids and some amino acids, which produces reduc-ing equivalents in the form of nicotinamide adenine dinucleotide and flavinadenine dinucleotide that result in production of large amounts of adenosinetriphosphate (ATP) via oxidative phosphorylation. Although regulated primarilyby the products of ATP hydrolysis, in particular adenosine diphosphate, the rate ofdelivery of reducing equivalents to the electron transport chain is also a potentialregulatory step of oxidative phosphorylation. The TCA cycle is responsible for thegeneration of ≈67% of all reducing equivalents per molecule of glucose, hencefactors that influence TCA cycle flux will be of critical importance for oxidativephosphorylation. TCA cycle flux is dependent upon the supply of acetyl units,activation of the three non-equilibrium reactions within the TCA cycle, and it hasbeen suggested that an increase in the total concentration of the TCA cycle

1072 Bowtell et al.

intermediates (TCAi) is also necessary to augment and maintain TCA cycle fluxduring exercise.

This article reviews the evidence of the functional importance of the TCAipool size for oxidative metabolism in exercising human skeletal muscle. Inparallel with increased oxidative metabolism and TCA cycle flux during exercise,there is an exercise intensity-dependent 4- to 5-fold increase in the concentrationof the TCAi. TCAi concentration reaches a peak after 10–15 minutes of exercise,and thereafter tends to decline. This seems to support the suggestion that theconcentration of TCAi may be of functional importance for oxidative phosphoryl-ation. However, researchers have been able to induce dissociations between TCAipool size and oxidative energy provision using a variety of nutritional, pharmaco-logical and exercise interventions.

Brief periods of endurance training (5 days or 7 weeks) have been found toresult in reduced TCAi pool expansion at the start of exercise (same absolute workintensity) in parallel with either equivalent or increased oxidative energy provi-sion. Cycloserine inhibits alanine aminotransferase, which catalyses the predomi-nant anaplerotic reaction in exercising human muscle. When infused intocontracting rat hindlimb muscle, TCAi pool expansion was reduced by 25% withno significant change in oxidative energy provision or power output. Glutaminesupplementation has been shown to enhance TCAi pool expansion at the start ofexercise with no increase in oxidative energy provision. In summary, there is aconsistent dissociation between the extent of TCAi pool expansion at the onset ofexercise and oxidative energy provision.

At the other end of the spectrum, the parallel loss of TCAi, glycogen andadenine nucleotides and accumulation of inosine monophosphate during pro-longed exercise has led to the suggestion that there is a link between muscleglycogen depletion, reduced TCA cycle flux and the development of fatigue.However, analysis of serial biopsies during prolonged exercise demonstrateddissociation between muscle TCAi content and both muscle glycogen content andmuscle oxygen uptake. In addition, the delay in fatigue development achievedthrough increased carbohydrate availability does not attenuate TCAi reductionduring prolonged exercise. Therefore, TCAi concentration in whole musclehomogenate does not seem to be of functional importance. However, TCAicontent can currently only be measured in whole muscle homogenate rather thanthe mitochondrial subfraction where TCA cycle reactions occur. In addition,anaplerotic flux rather than TCAi content per se is likely to be of greaterimportance in determining TCA cycle flux, since TCAi content is probablymerely reflective of anaplerotic substrate concentration. Methodological advancesare required to allow researchers to address the questions of whether oxidativephosphorylation is limited by mitochondrial TCAi content and/or anaplerotic flux.

The tricarboxylic acid (TCA) cycle (also termed carbon dioxide, guanosine triphosphate (GTP) and‘Krebs cycle’ or ‘citric acid cycle’) is central to reducing equivalents in the form of nicotinamideoxidative metabolism. The cycle consists of eight adenine dinucleotide (NADH) and flavin adenineenzymatically catalysed reactions that oxidise ace- dinucleotide (FADH2) that subsequently feed intotyl-coenzyme A (CoA) resulting in the formation of the electron transport chain producing large

© 2007 Adis Data Information BV. All rights reserved. Sports Med 2007; 37 (12)

TCA Cycle Intermediates and Oxidative Metabolism 1073

amounts of adenosine triphosphate (ATP) via oxida- that an increase in the total concentration of the TCAtive phosphorylation. There is no net production or cycle intermediates (TCAi) is necessary to augmentconsumption of the eight TCA cycle intermediates and maintain TCA cycle flux.[3,4] However, there isduring acetyl unit oxidation; however, five of the as yet no clear consensus opinion as to the impor-intermediates are involved in ancillary reactions that tance of anaplerosis (i.e. TCA cycle expansion) incan result in their net loss (cataplerosis) or gain skeletal muscle during exercise in regulating mito-(anaplerosis) [figure 1]. Anaplerosis refers to the chondrial energy production and thereby exerciseentry of carbon into the TCA cycle, by routes other performance. There are very few published reviewsthan acetyl CoA entry via the citrate synthase reac- that focus specifically on the role of anaplerosis intion. Cataplerosis refers to the exit of carbon from energy metabolism;[5,6] and these are now somethe TCA cycle by routes other than CO2 production. years out of date with significant new knowledge

arising in the interim. Therefore, the focus of theAt rest, demand for ATP resynthesis and thuspresent article is to consider whether modulation ofskeletal muscle TCA cycle flux is low. However,the TCAi cycle intermediate concentration is able toduring moderate to intense exercise TCA cycle fluxmodify the rate of oxidative phosphorylation, suchis estimated to increase 60- to 100-fold (based onthat functional performance during exercise is af-changes in leg muscle oxygen uptake)[1,2] in thefected. Specifically, whether enhanced expansion ofexercising skeletal muscle. In addition to sufficientthe TCAi pool at the onset of exercise is able tosupply of acetyl-CoA and activation of the threefacilitate oxidative metabolism, and whether fatiguenon-equilibrium reactions (citrate synthase, isoci-during prolonged endurance exercise is associatedtrate dehydrogenase and 2-oxoglutarate dehy-

drogenase) within the cycle, it has been suggested with depletion of the TCAi pool.

Oxaloacetate

Malate

Fumarate

1.

NADH

Citrate

Cis-aconitate

2.

3.

5.

Isocitrate

CO2 + NADH

Glutamate (A,B)Pyruvate (B)

2-Oxoglutarate

Succinyl-CoA

GTP

Aspartate (C)

FADH2

Pyruvate (D)

Succinate

Phosphoenolpyruvate (F)Pyruvate (E)

CO2 + NADH

6.

7.

8.

9.

4.

Fig. 1. Overview of the tricarboxylic acid cycle. This series of reactions is catalysed by the following enzymes, as numbered in the figure:(1) citrate synthase (condensation); (2) aconitase (dehydration); (3) aconitase (hydration); (4) isocitrate dehydrogenase (oxidative carboxyl-ation); (5) α-ketoglutarate dehydrogenase complex (oxidative decarboxylation); (6) succinyl coenzyme A (CoA) synthetase (substrate-levelphosphorylation); (7) succinate dehydrogenase (oxidation); (8) fumarase (hydration); (9) malate dehydrogenase (oxidation). The mainanaplerotic pathways in skeletal muscle are also indicated: (A) glutamate dehydrogenase; (B) alanine aminotransferase; (C) purinenucleotide cycle; (D) malic enzyme; (E) pyruvate carboxylase; (F) phosphoenolpyruvate carboxykinase. FADH2 = the reduced form of flavinadenine dinucleotide; GTP = guanosine triphosphate; NADH = the reduced form of nicotinamide adenine dinucleotide.


1074 Bowtell et al.

1. Oxidative Metabolism respiration, depending on initial hexokinase or ATPconcentration in the incubation medium.[13] In bothcases, the rate of respiration was correlated with1.1 Control of Oxidative PhosphorylationADP concentration alone. Therefore, it has been

Lardy and Wellman[7] first demonstrated that the suggested that ADP concentration is the primaryrate of oxidative phosphorylation was dependent on variable regulating oxidative phosphorylation.[13-16]

the extramitochondrial concentration of adenosine Adenine nucleotides are, however, large in sizediphosphate (ADP) and inorganic phosphate (Pi). It and charge (e.g. ATP has four negative charges) andwas hypothesised that the cytosolic concentration of

therefore have low diffusibility. Fast movement ofATP hydrolysis products provide the feedback sys-

the products of oxidative phosphorylation and ATPtem between oxidative phosphorylation and adeno-

hydrolysis between sites of production and utilisa-sine triphosphatase (ATPase) activity. The basic

tion, which would be a prerequisite for communica-assumption is that the cytosolic concentrations of

tion of the energy status of the muscle via ADP, mayADP and Pi are proportional to the rate of ATP

therefore seem unlikely. Alternatively, phosphocre-hydrolysis.

atine (PCr) is smaller and less charged and thereforeSeveral hypotheses have been proposed to deter-

more mobile in cells.[17] Furthermore, creatine kin-mine the exact mechanism by which the products of

ase (CK) exists in discrete cellular locations, depen-ATP hydrolysis determine the rate of oxidative

ding on the muscle fibre type. Between 10% andphosphorylation. A detailed examination of these

30% of CK activity is on the outer side of the innertheories is beyond the scope of this article (for a

mitochondrial membrane, 3–4% is at the M-lines ofreview see Balaban[8]). Briefly, the ATP/ADP • Pi the sarcomere and the remainder is in the cytoplasm,(phosphorylation potential) hypothesis[9,10] relies on

bound to the sarcoplasmic reticulum, sarcolemma orthe assumption that the majority of the mitochondri-other sites of ATP utilisation.[17] Hence, Bessmanal respiratory chain (i.e. from NADH up to complexand Fonyo[18] introduced the concept of the creatine-II) is in near equilibrium with the cytosolic phospho-PCr shuttle for explaining the movement of cellularrylation potential. Therefore, shifts in any of thechemical energy between mitochondria and the my-equilibrium constituents (e.g. NADH and ATP/ofibril. Creatine reacts with ATP to form PCr andADP • Pi) result in alterations in the cytochrome aa3ADP, which stimulates mitochondrial respira-(complex IV) redox state, which ultimately controlstion.[13-16] PCr diffuses rapidly from the mitochon-respiration through an irreversible step in the reduc-dria to the myofibrils, where CK reverses this reac-tion of molecular oxygen. However, it has beention to form ATP (available for muscular contrac-shown that neither the adenosine translocase (in-tion) and creatine. Creatine then diffuses back to thetramitochondrial ATP/cytosolic ADP exchanger)mitochondria for re-phosphorylation. Support fornor the ATP synthesis reaction is at equilibri-the creatine-PCr shuttle model comes from the find-um.[11,12]

ing that increases in creatine[14] and PCr levels[16]The ATP/ADP model is based on the assumption

stimulate and inhibit respiration in isolated mito-that the transport of ATP out of and ADP into thechondria and skinned (human) muscle fibre, respec-mitochondrial matrix by adenosine translocase istively.rate limiting for oxidative phosphorylation. ATP

Although ADP (via the creatine-PCr shuttle) ap-and ADP compete for binding and subsequent trans-pears to be the primary controller of the rate ofport across the mitochondrial membrane; therefore,oxidative phosphorylation, the energy required toit is assumed that the influx of ADP is responsive todrive the electron transport chain is delivered via thethe ratio of ATP to ADP. There is some evidencereducing equivalents NADH and FADH2, derivedthat adenosine translocase is rate limiting.[11] How-from the catabolism of carbohydrate, fatty andever, in isolated mitochondria, a high ATP/ADPamino acids. Therefore, the rate of delivery of re-ratio can be associated with low or high rates of



ducing equivalents to the electron transport chain is a rate-limiting step of the TCA cycle;[8,26] indeeda potential regulatory step of oxidative phosphoryl- citrate synthase has been used as a marker of oxida-ation. Indeed Koretsky and Balaban[19] found that tive capacity and hence training status with endur-raising the concentration of mitochondrial NADH ance-trained subjects having a higher skeletalincreased the maximum rate of ATP synthesis of muscle citrate synthase activity than untrained sub-isolated mitochondria. Similarly, Moreno-Sanchez jects.[27-31] It should, however, be emphasised thatet al.[20] showed that respiration of incubated rat much of the understanding of TCA cycle regulationmitochondria increased linearly with an increase in has been derived from in vitro work in rat heartthe reduction of NAD+. The mechanism is unclear, muscle and regulatory mechanisms may differ inbut could involve the equilibrium of the electron human skeletal muscle.transport chain,[9,10] interaction with the mitochon-drial membrane potential[21] or via increasing the 2. Exercise-Induced Modulation of TCAsensitivity of mitochondrial respiration to ADP con- Cycle Intermediate (TCAi) Pool Sizecentration.[19] A description of the regulation of sub-

Estimated TCA cycle flux increased 70-fold fromstrate supply is beyond the scope of this review (forresting levels during exercise at 60% maximum legreviews see Stanley et al.,[22] Neely and Morgan[23]

kicking capacity. In parallel, an ≈4- to 5-fold in-and Saks et al.[24]). However as the major finalcrease in the TCA cycle intermediate (TCAi) poolcommon pathway for the oxidation of carbohy-size/concentration has consistently been observeddrates, fats and amino acids, the TCA cycle is thewithin the first 10 minutes of moderate-intensitymajor producer of reducing equivalents for the elec-exercise.[1,32-35] Whilst Gibala et al.[1] demonstratedtron transport chain.that the magnitude of TCAi pool expansion wasdependent on exercise intensity with a significant1.2 Tricarboxylic Acid (TCA) Cycle andpositive correlation between TCAi concentrationits Regulationand estimated TCA cycle flux, this does not, ofcourse, establish cause and effect.TCA cycle flux can increase 100-fold from rest

During cycle exercise at 75% maximal oxygenduring intense leg exercise,[1] presumably resultinguptake (V̇O2max), Sahlin et al.[33] showed that thein comparable increases in the rate of NADH andsum of four TCAi (malate + citrate + fumarate +FADH2 production. The TCA cycle substrates ox-oxaloacetate) increased from a rest value of 0.49 toaloacetate and acetyl-CoA and the product NADH4.41 mmol/kg dm (dry mass) after 5 minutes ofare the critical regulators of the TCA cycle.[25] Fur-exercise. This magnitude of increase in TCAi levelsthermore, the three non-reversible reactions cat-(9-fold) has not been replicated elsewhere. This mayalysed by citrate synthase, isocitrate dehydrogenasebe because, with the exception of citrate, the inter-and 2-oxoglutarate dehydrogenase are under regula-mediates measured were from the second span of thetion by many metabolites. Isocitrate dehydrogenasecycle (2-oxoglutarate – oxaloacetate), which haveis stimulated by ADP and Ca2+ and inhibited bybeen shown to account for the majority of TCAiATP and NADH.[25] The activity of 2-oxoglutaratepool expansion during exercise.[1,32,36,37]dehydrogenase is also inhibited by high ratios of

[ATP]/[ADP], [succinyl-CoA]/[CoASH] and Many studies have demonstrated that the TCAi[NADH]/[NAD+] and stimulated by Ca2+.[3] Citrate pool expands at the onset of submaximal exercisesynthase activity is responsive to changes in the reaching a peak after 10–15 minutes, followed by a[NAD+]/[NADH] ratio and regulated directly by reduction with prolonged exercise.[33,34,38-41] For ex-accessible pools of either of its two substrates ox- ample, Sahlin et al.[33] showed that following initialaloacetate and acetyl-CoA,[3] and its product cit- expansion of the TCAi pool to a peak of 4.41 ± 0.23rate.[26] Citrate synthase is also inhibited by suc- mmol/kg dm after 5 minutes of cycle exercise atcinyl-CoA.[26] Citrate synthase has been proposed as 70% V̇O2max, the TCAi content declined to a nadir


1076 Bowtell et al.

at voluntary exhaustion (3.33 ± 0.29 after 40 min- These studies demonstrated that the muscle NH3utes; 2.83 ± 0.27 after an average of 75 minutes; release[33] and the resting aspartate concentrations[35]

both mmol/kg dm). The change in the TCAi pool were not sufficient to support the magnitude of thesize during exercise is the net effect of anaplerotic net increase in the TCAi pool size during the firstand cataplerotic reactions. Clearly, during the initial minute of exercise. It seems unlikely, therefore, thatperiod of exercise, anaplerosis predominates, the PNC makes a significant contribution to anapler-whereas the depletion of TCAi pool during pro- osis in exercising human skeletal muscle.longed exercise presumably reflects the predomi- Other possible anaplerotic reactions includenance of cataplerotic reactions. those catalysed by glutamate dehydrogenase

(GluDH), alanine aminotransferase (AAT), pyr-uvate carboxylase (PC), phosphoenolpyruvate3. Anaplerotic andcarboxykinase (PEPCK) and malic enzyme (ME),Cataplerotic Pathwaysrespectively (see equation 2):

A number of biochemical pathways could con-tribute to anaplerosis and cataplerosis; however,their quantitative importance has been the subject ofdebate (for previous reviews see Graham andGibala,[5] and Gibala et al.[35]). Whilst not the mainfocus of the present article, a re-iteration of thepredominant pathways forms the basis for subse-quent discussion.

The purine nucleotide cycle (PNC) has been pro-posed to serve an important anaplerotic functionduring exercise;[42,43] and comprises a series of reac-

glutamate + NAD+ ←(GluDH)→ α-ketoglutarate + NH3 + NADH

glutamate + pyruvate ←(AAT)→ α-ketoglutarate + alanine

pyruvate + CO2 + ATP ←(PC)→ oxaloacetate + ADP + Pi

phosphoenolpyruvate + CO2 + IDP ←(PEPCK)

→ oxaloacetate + ITP

pyruvate + CO2 + NAD(P)H + H+ ←(ME)→ malate + NAD(P)+

tions involved in the regulation of adenine nucleo- (Eq. 2)tide status (see equation 1):

where IDP = inosine diphosphate; ITP = inosinetriphosphate; NAD(P)H = nicotinamide adeninedinucleotide phosphate and nicotinamide adeninedinucleotide reduced forms; and NAD(P)+ = nico-tinamide adenine dinucleotide phosphate and nico-tinamide adenine dinucleotide.

The reactions catalysed by PC, PEPCK and MEare not thought to play a significant part in the

2ADP ←(adenylate kinase)→ AMP + ATP

AMP (AMP deaminase)→ IMP + NH3

IMP + aspartate + GTP (adenylosuccinate synthase) → adenylosuccinate + GDP + Pi

adenylosuccinate (adenylosuccinate lyase)→ AMP + fumarate

process of anaplerosis[44-47] (for reviews see Graham(Eq. 1)and Gibala,[5] and Gibala et al.[6]).where AMP = adenosine monophosphate; IMP =

inosine monophosphate; and GDP = guanosine Interestingly, the activity of GluDH in humandiphosphate. quadriceps is higher than in rodent muscle[48,49] and,

The net effect of the four reactions is the deami- therefore, this enzyme could potentially be involvednation of aspartate, the consumption of GTP and the in anaplerosis. Indeed, during knee extensor activityproduction of ammonia (NH3) and fumarate, one of at 80% of maximum workload, Gibala et al.[35]

the TCAi. Aragon and Lowenstein[42] proposed that showed a decline in intramuscular glutamate ofthe PNC is the major process responsible for the net ≈50% within the first minute of exercise, which isincrease in the TCAi pool during exercise via entry consistent with the involvement of the GluDH reac-at the level of fumarate. However, human exercise tion. However, the predicted increase in NH3 pro-studies have not supported these conclusions.[33,35] duction did not occur, nor was there any change in



net glutamine production. On the other hand, intra- cise muscle glutamate concentration (induced byprior glycogen depletion) despite an anticipated re-muscular content of alanine increased markedlyduction in pyruvate availability for the AAT reac-during exercise with a concomitant increased re-tion.[52]lease. Such an increase in alanine may provide a

At the onset of exercise, the TCAi pool expan-means of removing NH3 from skeletal muscle pro-sion reflects that anaplerosis exceeds the rate ofduced via the GluDH reaction.[50] However, it iscataplerosis. However, the concentration of themore likely that the alanine was derived from theTCAi is reduced at fatigue during prolonged exer-AAT reaction. Indeed, Sahlin et al.[33] and Gibala etcise,[33,34] although the majority of this decline oc-al.[35] conclude that the increase in alanine (with thecurs in the first 20 minutes after the initial peak atconcomitant decrease in glutamate) demonstrates≈5–10 minutes.[39] This net loss of TCAi indicatesthat a rightward shift in the AAT reaction is thethat as exercise duration increases, the rate at whichpredominant anaplerotic mechanism during exer-TCAi leave the cycle exceeds the rate of TCAi entrycise. It is likely that this rightward shift is largelyinto the cycle presumably due to (i) increased cat-driven by the rate of pyruvate formation via glycoly-aplerosis; (ii) reduced anaplerosis; or (iii) a combi-sis being faster than its rate of oxidation by pyruvatenation thereof. Sahlin et al.[33] proposed that thedehydrogenase complex (PDC), due to a lag in PDCavailability of pyruvate is reduced in parallel withactivation at the onset of exercise.[51]

depletion of glycogen stores, hence reducing pyr-In support of this, infusion of pyruvate into the

uvate-dependent anaplerosis. Studies that haveperfused rat hind limb caused a significant increase

manipulated carbohydrate availability provide somein the TCAi pool size, which coincided with an

support for this hypothesis. Muscle biopsies wereincreased release of alanine.[4] Furthermore, in rest- taken at fatigue in low carbohydrate availabilitying human muscle, diversion of pyruvate away from trials and at the same timepoint in high carbohydrateAAT, via prior dichloroacetate (DCA) infusion, availability trials. TCAi content was significantlyreduces the TCAi pool size.[36] DCA is a potent higher when carbohydrate availability was elevatedactivator of PDC, which catalyses the formation of either through glucose supplementation[40] or withacetyl units from pyruvate. By infusing dichloroace- prior glycogen supercompensation.[41]

tate, pyruvate is diverted toward the acetylation of Amino acid oxidation has been proposed as thecarnitine and away from the AAT reaction. This primary cataplerotic pathway. Wagenmakers etresulted in a fall in the resting TCAi pool size from al.[53] suggest that in the initial aminotransferase1.2–1.5 mmol/kg dm without DCA to 0.75–0.78 reaction, the oxidation of branched-chain aminommol/kg dm with DCA infusion.[36] This strongly acid (BCAA) [specifically leucine] places a carbonsuggests that in normal circumstances the AAT re- ‘drain’ on the TCA cycle at the level of 2-ox-action is the predominant anaplerotic pathway. oglutarate (see equation 3):However, after 1 minute of leg-kicking exercise,TCAi pool size was not different between trialswhere DCA or saline had been infused prior toexercise.[37] This might suggest a compensatory in-volvement of the GluDH reaction for anaplerosis.However, given previous evidence,[33,35] it is per-haps more likely that provided there is sufficient

leucine + 2-oxoglutarate ←(BCAAT)→ BCOA + glutamate

glutamate + NH3 ←(glutamine synthetase)→ glutamine

glutamate + pyruvate ←(AAT)→ 2-oxoglutarate + alanine

BCOA (BCOADH)→ acetyl-CoA + acetoacetate

(Eq. 3)pyruvate, the initial muscle glutamate concentrationwhere BCAAT = branched-chain amino acid trans-may determine flux through the AAT reaction, andaminase; and BCOA = branched-chain acid.thus the TCAi pool size.[37] This is supported by

research that showed enhanced anaplerosis of the The initial step in BCAA degradation is a revers-TCA cycle during exercise with elevated pre-exer- ible aminotransferase reaction in which BCAA are


1078 Bowtell et al.

converted to their respective branched-chain oxo and proline) decreases by ≈60% within skeletalmuscle after just 15 minutes of exercise. Thisacid (BCOA) and the amino group is donated tochange alone will probably shift the equilibrium of2-oxoglutarate to form glutamate, thus drainingthese reactions to favour cataplerosis.2-oxoglutarate from the TCA cycle. The BCOA are

TCA cycle regulation is complex due to the con-further broken down by branched-chain oxo acidstant influx and efflux of intermediates and thedehydrogenase (BCOADH) in an irreversible oxida-many enzymes involved. At the onset of exercise,tive-decarboxylation reaction, which is the rate-lim-the concentration of the TCAi increases[1,33-35] dueiting step in BCAA degradation. However, oxida-mainly to flux through the AAT reaction.[33,35] How-tion of isoleucine and valine ultimately yields suc-ever, as exercise continues, the concentration of thecinyl-CoA, such that there is no net drain on theTCAi decreases.[33,34] This is most likely due to aTCAi. Indeed, isoleucine and valine oxidation willcombination of reduced flux through the AAT reac-result in net anaplerosis if the glutamate formedtion, an increase in leucine oxidation and glutamineduring transamination feeds into the AAT reactionsynthesis,[53] and also a reversal of some of theforming alanine and regenerating 2-oxoglutarateproposed anaplerotic reactions.(use of 2-oxoglutarate and formation of succinyl

CoA and 2-oxoglutarate).4. Functional Significance of Changes inCataplerosis will only result if the glutamate de-TCAi Pool Sizerived from transamination of leucine is used to syn-

thesise glutamine; if, on the other hand, glutamateThe functional significance of TCAi pool expan-

feeds into the AAT reaction, then 2-oxoglutaratesion at the onset of contraction is a matter of contro-

will be recycled, and leucine oxidation will be TCAiversy, since a dissociation between pool size and

neutral. However, as exercise continues and fatigueoxidative energy provision has been demonstrated

is approached, ammonia concentration will in-by our own group and others.[32,39,40,55-57] Longitudi-

crease,[50] thus favouring glutamine formation. Glu-nal endurance training studies have demonstrated an

tamine synthesis and efflux from skeletal muscle ≈40% reduction in the magnitude of TCAi poolincreased significantly during prolonged exercise expansion at the start of exercise after endurance(from 3 μmol/min/kg wet weight [ww] at rest to ≈70 training, in parallel with reduced substrate levelμmol/min/kg ww during exercise[34]), thus resulting phosphorylation and presumably, therefore, in-in a net loss of glutamate derived from leucine creased aerobic energy provision.[55,57] Gibala ettransamination and hence 2-oxoglutarate from the al.[52] reported lower aerobic energy provision de-muscle. Furthermore, in conditions of increased spite a greater expansion of the TCAi pool after 10BCOADH activity, the above reactions will proceed minutes of exercise with low muscle glycogen con-to a greater extent in the glutamate formation direc- tent. Our group have demonstrated that althoughtion, also enhancing cataplerosis of the TCA cycle. glutamine ingestion augmented the TCAi pool ex-BCOADH activity is augmented by prolonged exer- pansion at the start of exercise, aerobic energy pro-cise, BCAA ingestion and low pre-exercise muscle vision was not affected.[32] More recently, Dawsonglycogen concentration.[54]

et al.[56] successfully inhibited TCAi pool expansionShifts in the equilibrium of transamination reac- in rat gastrocnemius-plantaris-soleus complex

tions such as AAT are also likely to contribute to through perfusion with the AAT inhibitor cycloser-cataplerosis, due to changes in the reactant concen- ine. After 10 minutes of stimulated contractions,trations resulting in a reversal of flux from anapler- TCAi pool expansion was 25% lower after cycloser-otic to cataplerotic. The concentration of glutamate, ine rather than saline perfusion, but there was nowhich is central to amino acid metabolism and difference in substrate level phosphorylation. Thishence many anaplerotic aminotransferase reactions suggests that TCAi pool size is not of functional(arginine, aspartate, asparagine, glutamine, histidine importance for TCA cycle flux and hence mitochon-



drial respiration at the initiation of muscle contrac- was significantly elevated and muscle glycogention. breakdown and muscle lactate concentration were

lower after 5 and 45 minutes of exercise relative toSimilarly, the evidence to support the link be-the pre-training test. This predicted reduction intween muscle glycogen depletion and TCA cycleglycogenolytic flux resulted in a smaller exercise-flux, and hence muscle fatigue, is tenuous. There is a

parallel loss of TCAi, glycogen and adenine nucleo- induced expansion of the TCAi pool. When consid-tides and accumulation of IMP during prolonged ered alongside the lower muscle alanine and higherexercise.[33,40,41] However, analysis of serial biopsies muscle glutamate concentrations after 5 and 45 min-during 90 minutes of exercise demonstrated that utes of exercise relative to the pre-training test, thisthere was no relationship between muscle TCAi suggests reduced flux through the AAT reaction.content and either muscle glycogen content or Despite the attenuated TCAi pool expansion, neithermuscle oxygen uptake.[39] In addition, Baldwin et the rate of PCr breakdown, nor power output duringal.[38] manipulated endurance capacity by prior exercise were different between trials. Aerobic ener-muscle glycogen depletion and demonstrated disso- gy delivery was not compromised despite the small-ciation between the point of volitional fatigue and er TCAi pool, suggesting that TCA cycle flux is notTCAi content. In the next sections 4.1–4.4, the func- dependent upon the concentration of TCAi.tional effects of modulation of the exercise-induced Howarth et al.[57] reproduced these findings inTCAi pool expansion via prior endurance training,

male subjects after a more extended period of train-pharmacological and nutritional interventions will

ing (7 weeks, 1 hour/day, 5 days/week). The TCAibe examined in detail as summarised in table I.

pool was ≈50% smaller in the post-training com-pared with the pre-training trial, after 5 minutes of

4.1 Effect of Exercise Trainingcycling at 80% V̇O2peak (pre-training V̇O2peak).There appeared to be no functional consequencesEndurance training, even short-term (<2 weeks),after 5 minutes of cycling, since power output wasis known to profoundly alter carbohydrate metabol-similar for both trials. Consistent with findings ofism, increasing resting muscle glycogen, increasingDawson et al.,[55] the exercise-induced reduction inmaximal activity of glycolytic and oxidative path-muscle glutamate and increase in muscle alanineway enzymes, and blunting glycogenolysis and gly-were attenuated, suggesting that reduced anaplerosiscolysis during submaximal exercise.[28,60] Research-through AAT was responsible for the attenuation ofers therefore hypothesised that after even a shortTCAi pool expansion. The exercise-induced reduc-period of endurance training, there would be a bettertion in muscle PCr and increase in lactate concentra-match between the rate of pyruvate formation viations were attenuated after training, suggesting thatglycolysis and pyruvate oxidation. The availabilitydespite the reduced TCAi pool expansion, oxidativeof pyruvate for anaplerotic reactions such as thatenergy delivery was, if anything, increased. In thiscatalysed by AAT, would therefore be reduced, po-study, subjects continued to exercise until volitionaltentially attenuating the exercise-induced expansionfatigue in both trials, and endurance capacity wasof the TCAi pool.unsurprisingly ≈49 minutes longer on average afterIn order to test this hypothesis, ten recreationallytraining. However, at fatigue, the TCAi concentra-active women completed a 5-day training pro-tion was not significantly different between trials,gramme (45 minutes at 70% peak oxygen uptakewhich might suggest that TCAi declined to a mini-[V̇O2peak] per day[55]). Before and after the trainingmum fatigue-inducing level in both trials. However,period, subjects completed 45 minutes of cycling atan extra biopsy was taken in the post-training trial to70% V̇O2peak (pre-training V̇O2peak, same absolutecorrespond to the point of fatigue in the pre-trainingworkload), and biopsies were taken before and aftertrial and there was no further decline in TCAi con-5 and 45 minutes of exercise. During the post-tent after this time. This provides a clear dissociationtraining test, resting muscle glycogen concentration


1080B

owtell et al.

© 2007 A

dis D

ata

Info

rma

tion

BV. A

ll righ

ts rese

rved

.Sp

orts M

ed

2007; 37 (12)

Table I. The effects of endurance training, pharmacological and nutritional interventions on tricarboxylic acid intermediate (TCAi) concentration and oxidative energy provision

Study Subjects Intervention Exercise protocol OutcomesEndurance trainingDawson et al.[55] 8 F 5d endurance training 45 min @ 70% V̇O2peak ↓ TCAi pool expansion

↔ PCr degradation↓ Lactate production

Howarth et al.[57] 8 M 1 h/d × 5 d/wk for 7wk Cycle to exhaustion @ 80% After 5 min exercise, post- vs pre-training:V̇O2peak ↓ TCAi pool expansion

↓ PCr degradation↓ Lactate production2-fold ↑ endurance capacityAt fatigue: ↔ TCAi2-fold ↑ endurance capacity

PharmacologicalGibala and Saltin[37] 4 M, 2 F Saline vs DCA 15 min leg kicking @ 70% max At rest: [TCAi] in DCA < saline trial

knee extension capacity During ex: [TCAi] in DCA ↔ saline trialfollowed by ex to exhaustion @ Δ PCr and La in DCA ↔ saline trial100% max knee extensioncapacity

Timmons et al.[58] Canine gracilis muscle Saline vs DCA infusion 6 min tetanic stimulation ↓ PCr degradation (DCA trial) ↔ TCAiDawson et al.[56] Canine gastroc- 1h perfusion with saline vs 10 min isometric contraction [TCAi] DCA < saline ↔ PCr, La, ATP and

plantaris-soleus cycloserine infusion (ALAT (1Hz, 0.3ms, 2V) muscle tensioninhibitor)

Nutrition: amino acidsGibala et al.[59] 6 M Low glycogen ± 0.3 g/kg BCAA 15 min cycling at 70% V̇O2peak ↔ [TCAi]Bruce et al.[32] 7 M 0.125 g/kg oral GLN, 0.125 g/ 10 min cycling at 70% V̇O2peak At 10 min exercise: [TCAi] GLN > CON and

kg OKG or CON OKG ↔ PCr degradation and lactateaccumulation

Nutrition: carbohydrateSpencer et al.[40] 7 M ± CHO supplements Cycle to exhaustion @ 70% ↑ End-capacity post-ex [TCAi] CHO > placebo

V̇O2peak trialSpencer et al.[41] 7 M LG and HG trials via exercise Cycle to exhaustion @ 75% [TCAi] HG > LG; [IMP] LG > HG

and diet V̇O2peak in LG trial and forsame duration in HG trial

Gibala et al.[39] 6 M Timecourse of exercise 90 min leg kicking @ 70% max Peak [TCAi] after 10 min exknee extension capacity Steepest ↓ [TCAi] from 10–30 min ex

No association with [glycogen]No effect on oxidative energy provision

Baldwin et al.[38] 7 M Glycogen depletion ride Cycle to exhaustion @ 70% LG vs HG: ↔ [TCAi] after 15 min exfollowed by low (LG) or high V̇O2peak. An additional biopsy End capacity 103 vs 155 min(HG) CHO diet for 24h was taken in HG trials at point ↔ [TCAi] at fatigue

of fatigue in LG trials In HG trial, no further change in TCAi duringlast ≈50 min exercise

Gibala et al.[59] 6 M Low glycogen ± 0.3 g/kg BCAA 15 min cycling at 70% V̇O2peak ↔ [TCAi]ALAT = alanine aminotransferase; ATP = adenosine triphosphate; BCAA = branched-chain amino acid; CHO = carbohydrate; CON = control/placebo; DCA = dichloroacetate; ex =exercise; F = females; GLN = glutamine; HG = glycogen supercompensated; IMP = inosine monophosphate; La = lactate; LG = glycogen depleted; M = males; max = maximum;OKG = ornithine α-ketoglutarate; PCr = phosphocreatine; V̇O2peak = peak oxygen uptake; ↓ indicates decrease; ↑ indicates increase; ↔ indicates no change; Δ indicates change.


between the size of the TCAi pool, the capacity for of muscle lactate.[36] Hence, it has been suggestedoxidative energy delivery and fatigue development. that the 3- to 4-fold increase in TCAi at this time

was a consequence of mass action forcing pyruvateHowever, there is a need for caution in interpret-ing these data, not least because many factors other through anaplerotic reactions such as AAT, whichthan TCAi expansion will have been altered by appears to be the most quantitatively importanttraining. The sensitivity of mitochondrial respiration anaplerotic reaction.to creatine is increased after endurance training,[61]

PDC is present in an active dephosphorylatedand one could argue that these findings could be form and an inactive phosphorylated form, with theexplained by an increased sensitivity of mitochon- interconversion between the two being regulated bydrial respiration to changes in TCAi after endurance the relative activity of PDC phosphatase and kinase,training. In addition, for both studies,[55,57] the post- respectively. PDC is activated by pyruvate, Ca2+,training trial was undertaken at the same absolute NAD+, ADP, insulin and muscle contraction; theworkload as the pre-training trial. It is not surpris- latter most likely mediated through increased Ca2+.ing, therefore, that non-oxidative energy provision Dichloroacetate (DCA) is a potent inhibitor of PDCwas reduced at the lower relative exercise intensity kinase causing activation of PDC and hence promot-post-training. After the relatively brief training ing formation of acetyl CoA and acetylcarnitineperiod in the Dawson et al.[55] study, there was only from pyruvate. DCA infusion results in a 4- to 5-folda small reduction in relative exercise intensity (pre- increase in PDC activation status at rest comparedtraining: 73% vs post-training: 71% V̇O2max); how- with control (4.0 vs 0.9 mmol/kg ww[63]). During theever, these values were markedly different in the initial 1–3 minutes of subsequent moderate-intensityHowarth et al.[57] study (pre-training: 80% vs post- exercise, prior DCA infusion results in a sparing oftraining: 69%). If the studies had employed a design

muscle PCr and reduction in the accumulation ofwhereby exercise at the same relative intensity was

muscle lactate compared with saline infusion in bothcompleted pre- and post-training, one might predict

human[63-65] and canine models.[66-68] This suggestssimilar substrate level phosphorylation and TCAi

that the increase in oxidative metabolism at theconcentration between trials. Therefore, in either

onset of moderate-intensity exercise is acceleratedcase, there is dissociation between TCA cycle flux

by prior activation of the PDC.and TCAi concentration (i) as observed, at the same

At rest, infusion of DCA reduced the TCAi poolabsolute exercise intensity aerobic energy provisionsize presumably via a reduction in availability ofis maintained in the face of lower TCAi concentra-pyruvate for anaplerotic reactions due to increasedtion; or (ii) as predicted, at the same relative exercisepyruvate flux through PDC.[36,37] However, after 1intensity, aerobic energy provision is increased atand 15 minutes of moderate leg-kicking exercise,the same TCAi concentration. In section 4.2, thethere was no significant difference in TCAi pooleffect of pharmacological interventions that inducesize between DCA and saline, indicating that neta more isolated effect on TCAi pool expansion isanaplerosis was greater in the DCA trial.[37] Moreexamined.recently, Bangsbo et al.[69] provided better temporalresolution of the effect of DCA infusion on TCAi4.2 Effect of Pharmacological Interventionspool expansion at the start of exercise. They demon-strated TCAi pool size was lower in the DCA condi-During exercise, there is a rapid activation oftion at rest and after 5 and 15 seconds of strenuousglycogen phosphorylase by increased Ca2+, Pi andleg-kicking exercise (relative to non-infused condi-adrenaline (epinephrine), which results in high ratestion); however, this difference had disappeared afterof glycogenolysis.[62] At the start of exercise the rate3 minutes of exercise. This supports the earlierof pyruvate formation exceeds its rate of entrysuggestion (see section 2) that glutamate availabili-through pyruvate dehydrogenase complex (PDC) toty, which is not affected by DCA, is an importantform acetyl CoA, as indicated by the accumulation


1082 Bowtell et al.

determinant of flux through AAT at the start of was no difference between trials in muscle PCr,lactate or ATP concentration at any time. This fur-exercise. Or, that even when PDC is maximallyther example of dissociation between aerobic energyactivated by DCA prior to exercise, there remains aprovision and TCAi pool size suggests that themismatch between carbon flux through glycolysisTCAi pool expansion is not required to supportand PDC. In this only human study to measure TCAiincreased TCA cycle flux. One cautionary note isin parallel with PCr degradation and lactate forma-that even in the control trial, only a relatively modesttion during exercise after prior DCA infusion, thereTCAi pool expansion (≈20%) was observed after 10was a ≈27% reduction in substrate level phosphoryl-minutes of isometric contraction, which is surprisingation, but this did not achieve statistical significanceconsidering the degree of muscle fatigue (≈60%(six subjects only). More recently, Timmons etreduction in muscle force). There are species differ-al.,[58] using an intense titanic contraction protocol inences with lower TCAi pool expansion in rat thancanine skeletal muscle, found that prior DCA infu-human muscle during contraction, possibly relatedsion reduced substrate level phosphorylation whilstto differences in fibre type,[70] and one must there-TCAi pool (citrate, fumarate, malate) expansion wasfore be cautious in extrapolating data across species.reduced. Collectively, this evidence suggests that

prior DCA infusion facilitates aerobic energy provi-4.3 Effect of Amino Acid Supplementationsion at the start of moderate-intensity exercise either

via increased acetyl group availability or PDC flux,Our group demonstrated that increasing musclewhilst TCAi pool expansion is either unchanged[37]

glutamine content through oral glutamine supple-or reduced.[58] This provides another experimentalmentation prior to exercise augmented the TCAimodel in which increases in oxidative metabolismpool expansion after 10 minutes of exercise.[32] Glu-and TCAi pool expansion are dissociated, callingtamine is rapidly taken up into skeletal muscleinto question the functional importance of the latter.through system human muscle glutamine transporter

Cycloserine is a potent inhibitor of alanine (Nm),[71] and then deaminated to form glutamate,aminotransferase, which catalyses what is thought to and then 2-oxoglutarate, through the action of thebe quantitatively the most important anaplerotic re- enzymes glutaminase (EC 3.5.1.2)[72] and glutamateaction at the start of exercise. Dawson et al.[56]

dehydrogenase (EC 1.4.1.2)[29] or alanine amino-employed an isolated rat hindlimb preparation and transferase (EC 2.6.1.2);[73] or glutamine trans-infused either cycloserine or saline, in order to deter- aminase (EC 2.6.1.15)[72,74] and ω-amidase (ECmine the effects of any reduction in TCAi pool 3.5.1.3),[72,74] all of which exist in skeletal muscle. Itexpansion during 10 minutes of isometric contrac- is, therefore, entirely feasible for 2-oxoglutarate de-tion. Administration of cycloserine resulted in an rived from glutamine to enter the TCA cycle and≈80% reduction in AAT activity accompanied by a increase the total content of TCAi. Interestingly, the25% reduction in TCAi pool size. If TCAi pool decline in intramuscular glutamine (≈7 mmol/kg dryexpansion was necessary to facilitate an increase in weight [dw]) and glutamate (≈11 mmol/kg dw) con-aerobic energy provision during exercise, one might tent during the first 10 minutes of exercise was moretherefore expect either that substrate level phospho- than 4-fold greater than the increase in the measuredrylation would be increased or muscle force produc- TCAi (≈4 mmol/kg dw) over the same time period intion would be reduced after cycloserine administra- the glutamine supplementation trial. There are twotion. There was some evidence that peak tension means by which this discrepancy might be ex-development was inhibited at rest and during the plained: first, drainage of the TCAi to take part infirst 6 minutes of contraction; however, this effect the many other reactions in which they are involved,did not achieve statistical significance. Despite the for example, amino acid synthesis; and secondly,attenuation of the contraction-induced TCAi pool glutamate utilisation in reactions by which there isexpansion after cycloserine administration, there no net production of TCAi. The aspartate amino-



transferase (EC 2.6.1.1)[75] is one such reaction in thesis of Wagenmakers et al.[53] should increasecataplerosis, the extent of TCAi expansion at thewhich glutamate donates its amino group to ox-start of exercise was not affected. Gibala et al.[59]aloacetate, resulting in the production of aspartateintroduced the concept of a critical minimum con-and 2-oxoglutarate. The removal of oxaloacetate iscentration of glycogen required to ensure sufficientbalanced by the production of 2-oxoglutarate; thuspyruvate flux to drive anaplerosis through reactionsthere is no net change in TCAi pool size. As observ-such as alanine aminotransferase. Indeed, despiteed after ingestion of monosodium glutamate,[76]

the lowered muscle glycogen concentration, thereplasma aspartate concentration was elevated afterwas a significant increase in muscle alanine andglutamine consumption and may therefore accountTCAi concentration, and reduction in glutamatefor a proportion of the ‘missing’ glutamate.concentration. This is indicative of alanine forma-In terms of functional significance, despite thetion, in which case despite elevated leucine oxida-successful augmentation of exercise-induced TCAition, net drainage of TCAi would not occur. It wouldpool expansion, glutamine ingestion did not reducebe interesting to examine the effect of BCAA inges-substrate level phosphorylation. More recently, us-tion and consequent elevation in leucine oxidationing an identical protocol, we have demonstrated thaton drainage of TCAi towards fatigue when gluta-

glutamine ingestion does not alter pulmonary oxy-mine formation is favoured. However, Gibala et

gen kinetics or muscle deoxygenation kinetics (viaal.[59] calculated that an increase of 1.2 μmol/min/kg

near infrared spectroscopy).[77] This dissociation be-ww in BCOADH activation state would result in

tween TCAi pool size and aerobic energy provision,0.005 mmol/min//kg dw additional depletion of

again suggests that augmentation of the TCAi poolTCAi. Most studies have demonstrated an ≈4 mmol/

expansion at the onset of exercise is not of function-kg dw increase in TCAi within the first 5–10 min-

al importance.utes of exercise. It is, therefore, unlikely that even

Wagenmakers et al.[53] suggested that BCAA oxi- when elevated by glycogen depletion and BCAAdation places a drain on the TCAi pool due to the ingestion, leucine oxidation would place a quantita-requirement for 2-oxoglutarate for the initial BCAA tively important drain upon TCAi.transamination reaction. As described in section 2,only leucine oxidation can result in a net loss of 4.4 Effect of AlteringTCAi, since valine and isoleucine oxidation result in Carbohydrate Availabilitythe generation of succinyl CoA. In order to explorethis hypothesis further, Gibala et al.[59] measured the Fatigue during prolonged exercise is associatedTCAi pool expansion during 15 minutes of exercise with glycogen depletion, but the precise bio-after BCAA or placebo ingestion in subjects with chemical nature of this association is unclear. Sahlinlow muscle glycogen (a 2-hour cycle followed by a et al.[33] found that fatigue during prolonged exerciselow-carbohydrate diet for 20 hours prior to the was characterised by energy deficiency as indicatedstudy). Elevated BCAA concentration and glycogen by increased AMP deamination, but acetyl unitdepletion both cause activation of BCOADH[54] and availability was unaffected. The authors suggested,enhance leucine oxidation.[78] The net increase in therefore, that the observed reduction in TCAi con-TCAi content during exercise was not different be- tent impaired TCA cycle flux and hence aerobictween trials (3.97 ± 0.34 vs 3.88 ± 0.34 mmol/kg energy provision, causing fatigue. In order to ex-dw). In addition, the extent of TCAi pool expansion plore the hypothesised relationship between musclewas similar to that observed in subjects who per- glycogen, TCAi and development of fatigue, Spen-formed an identical exercise bout without undergo- cer et al.[40] manipulated carbohydrate availabilitying any modification of carbohydrate availability. either by carbohydrate supplementation during exer-This suggests that despite the predicted increase in cise or modulation of muscle glycogen levels withleucine oxidation rate, which according to the hypo- prior exercise and diet.[41] At the point of fatigue in


1084 Bowtell et al.

the low-carbohydrate availability trials, muscle oxygen uptake, which remained constant oncesteady state was attained, while both muscle PCrTCAi content was significantly lower and IMP con-degradation and lactate accumulation rates declinedtent significantly higher than in the high-carbohy-throughout exercise from their peaks at 5–10 min-drate availability trials. These findings leantutes of exercise. Once again, this demonstrates acredence to the hypothesis that glycogen depletionclear dissociation between TCAi content and rate ofcould be linked to the development of fatigueoxidative energy provision.through TCAi depletion, and a consequent mis-

At fatigue, muscle pyruvate content is either un-match between ATP usage and resynthesis rates aschanged or tends to increase relative to restingindicated by elevated IMP concentration.[33] Unfor-levels,[79,80] which may seem counter-intuitive sincetunately, subjects did not continue to volitional fa-muscle glycogen is presumably relatively depletedtigue in the high-carbohydrate availability trials and,at fatigue, and one might expect reduced glycoge-therefore, it is unclear whether any further change innolysis. However, reduced PDC activity[79,80] or re-TCAi content occurred.duced PDC flux[81] in glycogen-depleted muscleMore recently, Baldwin et al.[38] compared themay contribute to this effect. The consequent diver-response of subjects cycling to exhaustion in a gly-sion of pyruvate carbon towards anaplerosis rathercogen-depleted or -supercompensated state. Predict-than formation of acetyl CoA, probably contributedably, endurance capacity was ~30% longer in theto the higher TCAi content observed in glycogen-supercompensated trial. A similar TCAi content wasdepleted muscle after 1 and 10 minutes of exerciseobserved in both trials at fatigue, which could bein most,[39-41,52] but not all studies.[38] It is possible,interpreted to indicate that fatigue occurred once atherefore, that a reduction in PDC flux or activity as

critical TCAi content was reached. However, in themuscle glycogen stores deplete during prolonged

supercompensated trial, muscle biopsies were takenexercise will lead to a greater proportion of pyruvate

not only at the point of volitional fatigue, but also atdiverting to anaplerosis rather than oxidation, thus

the earlier timepoint when fatigue had occurred inresulting in maintained or even increased TCAi pool

the glycogen-depleted state. This revealed that aftersize.[38] Certainly, although relatively depleted at≈100 minutes (fatigue in glycogen-depleted trial),fatigue, muscle TCAi content remains significantly

muscle TCAi content was similar in both trials.greater than resting levels.[33,38] Interestingly, TCAi

Nevertheless, subjects were able to cycle for a fur-content was not different from resting levels after 90

ther ≈50 minutes in the glycogen-supercompensatedminutes of leg-kicking exercise;[39] unfortunately,

trial. Collectively, when considered alongside the biopsy data at fatigue are not available in this studydissociation between aerobic energy provision and due to the already high number of biopsies. How-TCAi content, it seems clear that fatigue during ever, it is tempting to speculate that TCAi contentendurance exercise does not occur due to the attain- may have subsequently increased at fatigue.ment of a critical level of TCAi depletion.

Gibala et al.[39] measured the timecourse of 5. Conclusionchanges in muscle glycogen and TCAi content inparallel during 90 minutes of moderate-intensity It is clear that the TCA cycle is of central impor-leg-kicking exercise. TCAi content increased to a tance not only for oxidative metabolism, but also forpeak at 10 minutes of exercise, and thereafter de- the biosynthesis of glucose, fatty acids and non-clined rapidly until 30 minutes of exercise with a essential amino acids. The functional importance ofmore gradual decline thereafter. It appears, there- anaplerosis within different tissues has been recog-fore, that significant cataplerosis (removal of TCAi) nised. For example, loss of 2-oxoglutarate fromoccurs before muscle glycogen nears any ‘critical’ neurones due to release of neurotransmitters gluta-level for muscle function. In addition, the decline in mate and γ-amino butyric acid, is compensated notTCAi occurred without any consequence for thigh only by glutamine influx, but also anaplerotic pyr-



uvate carboxylation.[82] Anaplerotic products within date. For example, currently TCAi concentrationcan only be measured within whole muscle homoge-the mitochondria and cytosol of the pancreatic β-cellnate rather than isolated to the mitochondrial poolare thought to be regulators of insulin secretion.[83]

where the reactions of the TCA cycle take place. InThe alleged link between TCAi pool size andaddition, due to the equilibrium nature of mostTCA cycle flux has been explored by manipulatinganaplerotic reactions, TCAi content probably mere-TCAi content using a variety of experimental mod-ly reflects reactant concentration rather thanels. Muscle glycogen depletion and glutamine inges-anaplerotic flux. Yet, it seems likely that anaplerotiction have both been shown to increase the TCAiflux is of prime importance for TCA cycle fluxpool expansion during the first 10 minutes of exer-rather than TCAi concentration per se. 13C-NMRcise.[32,52] However, in the face of higher TCAi(carbon nuclear magnetic resonance spectroscopy)concentration, substrate level phosphorylationtechniques to measure collective flux through all of(SLP) was either higher or not different; indicatingthe anaplerotic reactions relative to TCA cycle fluxdissociation between TCAi pool size and TCA cycle(oxidation of acetyl units) have recently been devel-flux. Similarly, endurance training has been shownoped for steady-state metabolic conditions. Propor-to attenuate the exercise-induced TCAi pool expan-tional increases in anaplerotic and TCA cycle fluxsion at the start of exercise (same absolute work-occur when comparing rested to contracted skeletalload) whilst substrate level phosphorylation is atten-muscle, which supports the hypothesis that anapler-uated.[55,57] The findings from the latter studies mustotic flux plays a facilitative role for TCA cyclebe interpreted with caution since a plethora of adap-flux.[84] At present, such measurements are not tech-tations occur in response to endurance training,nically possible within human skeletal muscle orwhich may be responsible for the observed reduc-during non-steady state exercise, and thus we con-tion in SLP. However, recently, the AAT inhibitortinue to rely upon the static muscle biopsy approachcycloserine has been shown to reduce TCAi poolto tackle a dynamic metabolic system.expansion in contracting canine muscle without ap-

The current evidence indicates that muscle TCAiparent consequence for muscle force production orcontent is not of functional importance for TCASLP.[56]

cycle flux and oxidative phosphorylation. However,Several studies in the 1990s involving manipula- it remains to be seen whether mitochondrial TCAi

tion of carbohydrate availability either through car- concentration and/or anaplerotic flux in exercisingbohydrate supplementation or depletion appeared to human skeletal muscle are determining factors forsuggest that progressive depletion of TCAi during oxidative phosphorylation.prolonged exercise may be associated with the de-velopment of fatigue.[40,41] However, more recently Acknowledgementsthe timecourse of change in TCAi concentration

No sources of funding were used to assist in the prepara-during prolonged endurance exercise has been mea-tion of this review. The authors have no conflicts of interest

sured.[39] These studies have demonstrated that al- that are directly relevant to the content of this review.though TCAi concentration tends to be lower atfatigue than at the early stages of exercise, the Referencesdevelopment of fatigue cannot be directly linked to a 1. Gibala MJ, Maclean DA, Graham TE, et al. Tricarboxylic acid

intermediate pool size and estimated cycle flux in humanloss of TCAi. Indeed, there appears to be no associa-muscle during exercise. Am J Physiol 1998; 275: E235-42

tion between TCAi content and volitional fatigue.[38]2. Blomstrand E, Radegran G, Saltin B. Maximum rate of oxygen

uptake by human skeletal muscle in relation to maximal activi-Collectively, these data imply that there is not aties of enzymes in the Krebs cycle. J Physiol 1997; 501 (2):simple cause-effect relationship between TCAi con-455-60

centration and TCA cycle flux. 3. Williamson JR, Cooper RH. Regulation of the citric acid cyclein mammalian systems. FEBS Lett 1980; 117: K73-85However, there are methodological limitations to 4. Lee S-H, Davis EJ. Carboxylation and decarboxylation reac-

the approaches brought to bear on this question to tions. J Biol Chem 1979; 254: 420-30


1086 Bowtell et al.

5. Graham TE, Gibala MJ. Anaplerosis of the tricarboxylic acid 25. Kunz W, Gellerich FN, Schild L. Contribution to control ofcycle in human skeletal muscle during exercise: magnitude, mitochondrial oxidative-phosphorylation by supplement of re-sources, and potential physiological significance. Adv Exp ducing equivalents. Biochem Med Metab Biol 1994; 52 (1):Med Biol 1998; 441: 271-86 65-75

6. Gibala MJ, Young ME, Taegtmeyer H. Anaplerosis of the citric 26. LaNoue KF, Bryla J, Williams JR. Feedback interactions inacid cycle: role in energy metabolism of heart and skeletal control of citric-acid cycle activity in rat-heart mitochondria. Jmuscle. Acta Physiol Scand 2000; 168: 657-65 Biol Chem 1972; 247 (3): 667-79

7. Lardy HA, Wellman H. Oxidative phosphorylations: role of 27. Terblanche SE, Gohil K, Packer L, et al. The effects of endur-inorganic phosphate and acceptor systems in control of meta- ance training and exhaustive exercise on mitochondrial en-bolic rates. J Biol Chem 1952; 195: 215-24 zymes in tissues of the rat (Rattus norvegicus). Compar Bio-

8. Balaban RS. Regulation of oxidative-phosphorylation in the chem Physiol 2002; 128 (4): 889-96mammalian-cell. Am J Physiol 1990; 258 (3): C377-89 28. Spina RJ, Chi MMY, Hopkins MG, et al. Mitochondrial en-

9. Wilson DF. Factors affecting the rate and energetics of mito- zymes increase in muscle in response to 7-10 days of cyclechondrial oxidative-phosphorylation. Med Sci Sports Exerc exercise. J Appl Physiol 1996; 80 (6): 2250-41994; 26 (1): 37-43

29. Starritt EC, Angus D, Hargreaves M. Effect of short-term train-10. Wilson DF, Erecinska M, Drown C, et al. Oxygen dependence ing on mitochondrial ATP production rate in human skeletal

of cellular-energy metabolism. Arch Biochem Biophys 1979; muscle. J Appl Physiol 1999; 86: 450-4195 (2): 485-93

30. Bell C, Paterson DH, Kowalchuk JM, et al. Determinants of11. Tager JM, Groen AK, Wanders RJA, et al. Control of mitochon-oxygen uptake kinetics in older humans following single-limbdrial respiration. Biochem Soc Trans 1983; 11 (1): 40-3endurance exercise training. Exp Physiol 2001; 86 (5): 659-65

12. Kingsleyhickman PB, Sako EY, Mohanakrishnan P, et al. P-3131. Leek BT, Mudaliar SRD, Henry R, et al. Effect of acute exerciseNmr-Studies of ATP synthesis and hydrolysis kinetics in the

on citrate synthase activity in untrained and trained humanintact myocardium. Biochem 1987; 26 (23): 7501-10skeletal muscle. Am J Physiol 2001; 280 (2): R441-713. Jacobus WE, Moreadith RW, Vandegaer KM. Mitochondrial

32. Bruce M, Constantin-Teodosiu D, Greenhaff PL, et al. Gluta-respiratory control: evidence against the regulation of respira-mine supplementation promotes anaplerosis but not oxidativetion by extra-mitochondrial phosphorylation potentials or byenergy delivery in human skeletal muscle. Am J Physiol 2001;[ATP] [ADP] ratios. J Biol Chem 1982; 257 (5): 2397-402280 (4): E669-7514. Jacobus WE, Diffley DM. Creatine-kinase of heart-mitochon-

33. Sahlin K, Katz A, Broberg S. Tricarboxylic acid cycle inter-dria: control of oxidative-phosphorylation by the ex-mediates in human muscle during prolonged exercise. Am Jtramitochondrial concentrations of creatine and phosphocrea-Physiol 1990; 259: C834-41tine. J Biol Chem 1986; 261 (35): 6579-83

15. Saks VA, Kongas O, Vendelin M, et al. Role of the creatine/ 34. Gibala MJ, Tarnopolsky MA, Graham TE. Tricarboxylic acidphosphocreatine system in the regulation of mitochondrial cycle intermediates in human muscle at rest and during pro-respiration. Acta Physiol Scand 2000; 168 (4): 635-41 longed cycling. Am J Physiol 1997; 272: E239-44

16. Walsh B, Tonkonogi M, Soderlund K, et al. The role of 35. Gibala MJ, Maclean DA, Graham TE, et al. Anaplerotic pro-phosphorylcreatine and creatine in the regulation of mitochon- cesses in human skeletal muscle during brief dynamic exer-drial respiration in human skeletal muscle. J Physiol 2001; 537 cise. J Physiol 1997; 502 (3): 703-13(3): 971-8 36. Constantin-Teodosiu D, Simpson E, Greenhaff PL. The impor-

17. Bhagavan NV. Medical biochemistry. 4th ed. Burlington (MA): tance of pyruvate availability to pyruvate dehydrogenase acti-Harcourt Academic Press, 2002 vation and anaplerosis in human skeletal muscle. Am J Physiol

18. Bessman SP, Fonyo A. The possible role of the mitochondrial 1999; 39: E472-8bound creatine kinase in regulation of mitochondrial respira- 37. Gibala MJ, Saltin B. PDH activation by dichloroacetate reducestion. Biochem Biophys Res Comm 1966; 22: 597-602 TCA cycle intermediates at rest but not during exercise in

19. Koretsky AP, Balaban RS. Changes in pyridine-nucleotide humans. Am J Physiol 1999; 77: E33-8levels alter oxygen-consumption and extra-mitochondrial

38. Baldwin J, Snow RJ, Gibala MJ, et al. Glycogen availabilityphosphates in isolated-mitochondria: a P-31-NMR anddoes not affect the TCA cycle or TAN pools during prolonged,NAD(P)H fluorescence study. Biochim Biophys Acta 1987;fatiguing exercise. J Appl Physiol 2003; 94 (6): 2181-7893 (3): 398-408

39. Gibala MJ, Gonzalez-Alonso J, Saltin B. Dissociation between20. Moreno-Sanchez R, Hogue BA, Hansford RG. Influence ofmuscle tricarboxylic acid cycle pool size and aerobic energyNAD-linked dehydrogenase-activity on flux through oxida-provision during prolonged exercise in humans. J Physioltive-phosphorylation. Biochem J 1990; 268 (2): 421-82002; 545 (2): 705-1321. Chance B. Pyridine-nucleotide as an indicator of oxygen re-

40. Spencer MK, Yan Z, Katz A. Carbohydrate supplementationquirements for energy-linked functions of mitochondria. Circattenuates IMP accumulation in human muscle during pro-Res 1976; 38 (5): 31-8longed exercise. Am J Physiol 1991; 261: C71-622. Stanley WC, Recchia FA, Lopaschuk GD. Myocardial substrate

41. Spencer MK, Yan Z, Katz A. Effect of low glycogen on carbo-metabolism in the normal and failing heart. Physiol Rev 2005;hydrate and energy-metabolism in human muscle during exer-85 (3): 1093-129cise. Am J Physiol 1992; 262 (4): C975-923. Neely JR, Morgan HE. Relationship between carbohydrate and

lipid-metabolism and energy-balance of heart-muscle. Ann 42. Aragon JJ, Lowenstein JM. The purine-nucleotide cycle: com-Rev Physiol 1974; 36: 413-59 parison of the levels of citric-acid cycle intermediates with the

24. Saks V, Favier R, Guzun R, et al. Molecular system bioenerget- operation of the purine nucleotide cycle in rat skeletal-muscleics: regulation of substrate supply in response to heart energy during exercise and recovery from exercise. Eur J Biochemdemands. J Physiol 2006; 577 (3): 769-77 1980; 110 (2): 371-7



43. Flanagan WF, Holmes EW, Sabina RL, et al. Importance of 62. Watt MJ, Howlett KF, Febbraio MA, et al. Adrenaline increasespurine nucleotide cycle to energy-production in skeletal- skeletal muscle glycogenolysis, pyruvate dehydrogenase acti-muscle. Am J Physiol 1986; 251 (5): C795-802 vation and carbohydrate oxidation during moderate exercise in

humans. J Physiol 2002; 534 (1): 269-7844. Davis EJ, Spydevold O, Bremer J. Pyruvate-carboxylase andpropionyl-Coa carboxylase as anaplerotic enzymes in skeletal- 63. Howlett RA, Heigenhauser GJF, Hultman E, et al. Effects ofmuscle mitochondria. Eur J Biochem 1980; 110 (1): 255-62 dichloroacetate infusion on human skeletal muscle metabolism

at the onset of exercise. Am J Physiol 1999; 277: E18-2545. Newsholme EA, Arch JRS, Brooks B, et al. The role of substratecycles in metabolic-regulation. Biochem Soc Trans 1983; 11 64. Timmons JA, Gustafsson T, Sundberg CJ, et al. Muscle acetyl(1): 52-6 group availability is a major determinant of oxygen deficit in

humans during submaximal exercise. Am J Physiol 1998; 274:46. Mole PA, Baldwin KM, Terjung RL, et al. Enzymatic pathwaysE377-80of pyruvate metabolism in skeletal-muscle: adaptations to ex-

ercise. Am J Physiol 1973; 224 (1): 50-4 65. Parolin ML, Spriet LL, Hultman E, et al. Effects of PDHactivation by dichloroacetate in human skeletal muscle during47. Lindena J, Trautschold I. Catalytic enzyme activity concentra-exercise in hypoxia. Am J Physiol 2000; 279 (4): E752-61tion in plasma of man, sheep, dog, cat, rabbit, guinea pig, rat

and mouse. J Clin Chem Clin Biochem 1986; 24: 11-8 66. Roberts PA, Loxham SG, Poucher SM, et al. The acetyl groupdeficit at the onset of contraction in ischaemic canine skeletal48. Wibom R, Hultman E. ATP production rate in mitochondriamuscle. J Physiol 2002; 544 (2): 591-602isolated from microsamples of human muscle. Am J Physiol

1990; 259: E204-9 67. Timmons JA, Poucher SM, ConstantinTeodosiu D, et al. Meta-bolic responses from rest to steady state determine contractile49. Wibom R, Hultman E, Johansson M, et al. Adaptation of mito-function in ischemic skeletal muscle. Am J Physiol 1997; 36chondrial ATP production in human skeletal muscle to endur-(2): E233-8ance training and detraining. J Appl Physiol 1992; 73: 2004-10

68. Timmons JA, Poucher SM, ConstantinTeodosiu D, et al. In-50. Graham TE, Maclean DA. Ammonia and amino acid metabol-creased acetyl group availability enhances contractile functionism in skeletal muscle: human, rodent and canine models. Medof canine skeletal muscle during ischemia. J Clin Invest 1996;Sci Sports Exerc 1998; 30: 34-4697 (3): 879-8351. Constantin-Teodosiu D, Peirce NS, Fox J, et al. Muscle pyr-

69. Bangsbo J, Gibala MJ, Howarth KR, et al. Tricarboxylic aciduvate availability can limit the flux, but not activation, of thecycle intermediates accumulate at the onset of intense exercisepyruvate dehydrogenase complex during submaximal exercisein man but are not essential for the increase in muscle oxygenin humans. J Physiol 2004; 561 (2): 647-55uptake. Eur J Physiol 2006; 452 (6): 737-4352. Gibala MJ, Peirce N, Constantin-Teodosiu D, et al. Exercise

70. Matoba H, Allen JR, Bayly WM, et al. Comparison of fiberwith low muscle glycogen augments TCA cycle anaplerosistypes in skeletal-muscles from 10 animal species based onbut impairs oxidative energy provision in humans. J Physiolsensitivity of the myofibrillar actomyosin ATPase to acid or2002; 540 (3): 1079-86copper. Histochem 1985; 82 (2): 175-8353. Wagenmakers AJM, Coakley JH, Edwards RH. Metabolism of

71. Ahmed A, Maxwell DL, Taylor PM, et al. Glutamine transportbranched chain amino acids and ammonia during exercise:in human skeletal muscle. Am J Physiol 1993; 264: E993-1000clues from McArdle’s disease. Int J Sports Med 1990; 11:

S101-13 72. Swierczynski J, Bereznowski Z, Makarewicz W. Phosphate-dependent glutaminase of rat-skeletal-muscle: some properties54. van Hall G, Maclean DA, Saltin B, et al. Mechanism of activa-and possible role in glutamine metabolism. Biochim et Bi-tion of muscle branched chain alpha-keto acid dehydrogenaseophys Acta 1993; 1157 (1): 55-62during exercise in man. J Physiol 1996; 494: 899-905

73. Lindena J, Sommerfield U, Hopfel C, et al. Catalytic enzyme55. Dawson KD, Howarth KR, Tarnopolsky MA, et al. Short-termactivity concentration in tissues of man, dog, rabbit, guineatraining attenuates muscle TCA cycle expansion during exer-pig, rat and mouse. J Clin Chem Clin Biochem 1986; 24: 35-47cise in women. J Appl Physiol 2003; 95 (3): 999-1004

74. Wu G, Thompson JR, Baracos VE. Glutamine metabolism in56. Dawson KD, Baker DJ, Greenhaff PL, et al. An acute decreaseskeletal muscles from the broiler chick and the laboratory rat.in TCA cycle intermediates does not affect aerobic energyBiochem J 1991; 274: 769-74delivery in contracting rat skeletal muscle. J Physiol 2005; 565

(2): 637-43 75. Schantz PG, Henriksson J. Enzyme levels of the NADH shuttlesystems: measurements in isolated muscle fibres from humans57. Howarth KR, LeBlanc PJ, Heigenhauser GJF, et al. Effect ofof differing physical activity. Acta Physiol Scand 1987; 129:endurance training on muscle TCA cycle metabolism during505-15exercise in humans. J Appl Physiol 2004; 97 (2): 579-84

76. Graham TE, Sgro V, Friars D, et al. Glutamate ingestion: the58. Timmons JA, Constantin-Teodosiu D, Poucher SM, et al. Acetylplasma and muscle free amino acid pools of resting humans.group availability influences phosphocreatine degradationAm J Physiol 2000; 278: E83-9even during intense muscle contraction. J Physiol 2004; 561

(3): 851-9 77. Marwood S, Bowtell JL. Effects of glutamine and hyperoxia onpulmonary oxygen uptake and muscle deoxygenation kinetics.59. Gibala MJ, Lozej M, Tarnopolsky MA, et al. Low glycogen andEur J Appl Physiol 2007; 99 (2): 149-61branched-chain amino acid ingestion do not impair anaplerosis

during exercise in humans. J Appl Physiol 1999; 87: 1662-7 78. Bowtell JL, Leese GP, Smith K, et al. Modulation of whole bodyprotein metabolism, during and after exercise, by variation of60. Putman CT, Jones NL, Hultman E, et al. Effects of short-termdietary protein. J Appl Physiol 1998; 85 (5): 1744-52submaximal training in humans on muscle metabolism in

exercise. Am J Physiol 1998; 38 (1): E132-9 79. Mourtzakis M, Saltin B, Graham T, et al. Carbohydrate meta-61. Tonkonogi M, Sahlin K. Physical exercise and mitochondrial bolism during prolonged exercise and recovery: interactions

function in human skeletal muscle. Exer Sport Sci Rev 2002; between pyruvate dehydrogenase, fatty acids, and amino acids.30 (3): 129-37 J Appl Physiol 2006; 100 (6): 1822-30


1088 Bowtell et al.

80. Watt MJ, Heigenhauser GJF, LeBlanc PJ, et al. Rapid upregula- 84. Walton ME, Ebert D, Haller RG. Relative rates of anaplerotiction of pyruvate dehydrogenase kinase activity in human skele- flux in rested and contracted rat skeletal muscle measured bytal muscle during prolonged exercise. J Appl Physiol 2004; 97

C-13 NMR spectroscopy. J Physiol 2003; 548 (2): 541-8(4): 1261-781. Coggan AR, Raguso CA, Gastaldelli A, et al. Regulation of

glucose production during exercise at 80% of V̇O2peak inuntrained humans. Am J Physiol 1997; 273: E348-54 Correspondence: Dr Joanna L. Bowtell, Academy of Sport,

82. Hassel B. Carboxylation and anaplerosis in neurons and glia. Physical Activity and Wellbeing, FESBE, London SouthMol Neurobiol 2000; 22 (1-3): 21-4083. Macdonald MJ, Fahien LA, Brown LJ, et al. Perspective: Bank University, 103 Borough Road, London, SE1 0AA,

emerging evidence for signaling roles of mitochondrial UK.anaplerotic products in insulin secretion. Am J Physiol 2005;

E-mail: [email protected] (1): E1-15


tricarboxylic acid cycle intermediate pool size

Documents