The American Journal of Clinical Nutrition 43: MAY 1986, pp 78 1-786. Printed in USA© 1986 American Society for Clinical Nutrition

Valine kinetics at graded valine intakesin young men1’2

Michael M Meguid, MD, PhD, Dwight E Matthews, PhD, Dennis M Bier, MD,

Carol N Meredith, PhD, and Vernon R Young, PhD, DSc

ABSTRACT Twelve young men, six subjects in each group studied in two phases, participated

in an experiment to explore the relationships between valine intake, plasma valine concentrations,

and valine kinetics, using l-[’3C]valine as a tracer. Below a valine intake of about 20 mg� kg�1 . day�’plasma valine concentrations reached a low and relatively constant level. The rate of valine oxidationfell with the decline in the intake of amino acid. Below valine intakes of 16 mg. kg�’ . day_i, themean daily rate of oxidation was estimated to be generally higher than the intake level, implying anegative valine balance during the 24 h day. These findings indicate that an intake of 10 mg valinekg’ . day’ would not be adequate to maintain protein nutritional status. Our results are discussedin relation to the currently accepted 1973 FAO/WHO value of 10mg. kg� . day�’ as being the upperrange of the valine requirement in healthy adult humans. Am J Clin Nuir 1986;43:78l-786.

KEY WORDS Yaline, amino acid requirements, young men, valine kinetics, [l-’3C]valine, plasmavaline, oxidation

Infroduction

A quantitatively significant function of thenutritionally indispensable (essential) aminoacids is to support an adequate rate of proteinsynthesis for maintenance of cell and organprotein content. The major fate of these aminoacids, following their absorption from the in-testinal tract, is their incorporation into andsubsequent liberation from proteins via pro-tein synthesis and breakdown, followed eitherby their reutilization or irreversible oxidation.Thus, it follows that a study of these dynamicprocesses, and a knowledge of how they changeunder various intakes of valine, should offerinformation that is useful in better definingthe metabolic basis for the valine requirement.In support of this, experiments in growing andadult rats have shown rates of oxidation oflysine, histidine, and threonine (1-5) are lowand relatively constant when intakes of theseamino acids are below physiological require-ments. When the intakes of each of theseamino acids are increased above about theminimum level required for maximal growth,the oxidation rates rise linearly. In addition,our previous observations on the fluctuation

of tryptophan oxygenase activity in livers ofrats receiving graded tryptophan intakes (6)and recent studies of whole body leucine ki-netics in young men (7, 8) indicate that thebiochemical mechanisms responsible formaintaining body protein and amino acid ho-meostasis are probably controlled in referenceto the individual’s physiological requirementsfor these nutrients.

In an initial series of experiments, we ex-plored the relationships between the intake ofleucine and leucine kinetics in young men (9).The data from that study suggested that mea-surement of whole body leucine dynamics of-fers another basis for increasing our under-standing of the physiological needs for this in-dispensable amino acid. We have now

From the Laboratory of Human Nutrition, Depart-

ment of Applied Biological Sciences and Clinical ResearchCenter, Massachusetts Institute of Technology, Cambridge,MA.

2Addrms reprint requests to: Dr Yernon R Young, De-partment of Applied Biological Sciences, 56-333, Massa-chusetts Institute of Technology, 77 Massachusetts Av-enue, Cambridge, MA 02139.

Received January 2, 1985.Accepted for publication October 29, 1985.

781

by guest on October 29, 2015

ajcn.nutrition.orgD

ownloaded from

782

extended this approach through studies on therelationships between valine metabolism andthe intake of valine, with total nitrogen (orprotein) remaining constant.

Methods and materials

Subjects

Twelve young male volunteers participated in the study.They were studied as outpatients in the Clinical ResearchCenter ofthe Massachusetts Institute ofTechnology. Thesubjects were assigned to one oftwo groups and the overallstudy was conducted in two phases, with six men partic-ipating in each phase. Their characteristics are given inTable 1. All were in good health asdetermined by medicalhistory, physical examination (including chest x ray andelectrocardiogram), and routine biochemical screening ofblood and urine samples. Their daily dietary intake wasdesigned to approximate their usual caloric intake. Thesubjects were encouraged to maintain customary levels ofphysical activity, but they were not allowed to participatein competitive sports.

The purpose of the study and the potential risks in-volved were fully explained to each subject. Written con-sent was obtained in accordance with the protocol ap-proved by the MIT Committee on the Use of Humans asExperimental Subjects and the Executive and PolicyCommittees of the MIT Clinical Research Center. Thesubjects were paid for their participation in the experimentand they remained healthy throughout.

Diet and experimental design

Each subject received a series of isocaloric, isomtrog-enous diets based on a crystalline amino acid mixture fur-nishing the equivalent (N X 6.25) of 0.8 g egg protein

TABLE 1

MEGUID ET AL

kg�’ . day’. The composition ofthe amino acid mixture,as well as the remainder of the diet, was as in the earlierstudy(9). Fourequal mealswere consumed at 8 AM, noon,5 PM, and 8 PM. Energy intake varied among individualsbut was held constant for each individual throughout theexperiment Mineral and vitamin supplements were given,also as previously described (9). Each diet period lasted 6days before the conduct ofa stable isotope infusion study,performed on the morning of the 7th day, as describedbelow.

The first six subjects (Group 1) began with the dietsupplying68 mgvaline kg� ‘ day1. Thisintakelevel wouldbe supplied by about 1 g milk or egg proteins per kg perday. The content ofvaline in the diet was then decreasedduring subsequent, consecutive diet periods to levels of40, 20, and 16 mg.kg’ ‘day’. The second group of sixmen (Group 2) began with a diet supplying 16 mg valinekg�1 . day’ and the content of the amino acid was thendecreased during consecutive diet periods to levels of 12,8, and 4 mg- kg�1 - day’. The dietary valine intake in thissecond group was chosen to cover the range of reporteddaily valine requirements (10, 1 1). Nitrogen intake waskept constant by substitution ofL-proline for L-valine inthe amino acid mixture.

Isotope tracer study

L�I1�UC]valine (92% UC) and sodium ‘�C bicarbonate(90% UC)were obtained from KOR Isotopes (Cambridge,MA). Optical and isotopic purity were confirmed as pre-viously described (9). Stock solutions of L�[l�UCJvalineand NaH’3C03 were prepared with aseptic technique usingsterile, pyrogen-free 0.9 NaQ solution. The stock solutionswere tested for pyrogens and sterility before aliquots wereremoved aseptically, and then diluted with sterile salinesolution for administration to the subjects as priming andconstant infusion doses. The exact concentration of thesolution infused was verified by analyzing an aliquot with

Characteristics of young men studied for whole body valine kinetics in relation to valine intake

Subject code Age

Weight

HeightDaily energy

intakeInitial Final

r kg kg on kcal

Group 101 25 65.5 64.7 175.0 300002 21 77.7 77.0 191.0 350003 24 72.4 69.9 181.0 347504 20 57.5 56.1 176.5 300005 24 66.4 68.8 169.0 294006 22 71.1 71.0 180.0 3300

Mean ± SE 22.7 ± 0.8 68.4 ± 2.8 67.9 ± 2.9 178.8 ± 7.0 3212 ± 104

Group 211 26 78.8 77.2 177.0 312012 24 71.9 71.1 176.0 318513 19 66.4 65.1 178.0 297514 18 55.5 54.4 175.0 280015 21 73.6 71.1 178.0 318516 28 72.5 70.9 179.0 3000

Mean ± SE 22.7 ± 1.6 69.8 ± 7.7 68.3 ± 3.2 177.1 ± 0.6 3045 ± 61

by guest on October 29, 2015

ajcn.nutrition.orgD

ownloaded from

YALINE METABOLISM 783

the aid ofan automatic amino acid analyzer. In all studiesL-[l)3Cjvaline was administered, at a constant knownrate ofabout 7. 1 �&mol - kg� . h� via an intravenous cath-eter placed in one arm. Blood samples were taken fromthe opposite arm after insertion into the antecubital veinof an indwelling 19 gauge, #{190}-inchintravenous samplingcatheter. The patency ofboth infusion and sampling cath-eters was maintained by slow infusions of normal saline.Starting at 8:55 AM, three expired air samples, each of 6-mm duration, were collected into Douglas bags to deter-mine the rate of expired CO2 production. Yenous bloodand expired air samples were obtained at 9:30, 9:45, and10:00 AM to determine background ‘3C enrichment inplasma valine and expired COT, respectively. The meanenrichments ofthree plasma and expired air samples wereused for determination ofpreinfusion values.

Immediately following these sample collections, a doseof NaH’3CO3 (1.3 �imol.kg’) was injected as a bolus,into the infusion catheter during a 1-mm period, in orderto prime the bicarbonate space (12). This was immediatelyfollowed by a primingdose of(l�UC]valine(0.84 mg#{149}kg�)and then the continuous infusion ofthe isotopically labeledvaline started.

As in the earlier study with leucine (9), the breakfastmeal was consumed at 8:00 AM. Then the usual 12 noonmeal was divided into three equal portions and one portionwas ingested at 10 AM prior to starting the continuous[l-�C1valine infusion. Anotherportion was ingested duringthe infusion period at 11 AM and the final portion eatenat 12 noon.

All the measurements of the isotope enrichment weremade during quasi-steady�state of plasma ‘3C valine andexpired air ‘3C02 enrichment, following 90 mm of infu-sion (9).

Sample collection, analyses, and calculations

On the infusion day, blood samples were obtained at8:00 AM, following the overnight fast, for determinationof plasma free valine and leucine concentrations. Subse-quent sampling of blood and expired air for measurementof plasma amino acid levels, [1�UC]valine enrichment, totalCO2 production, and I3(� enrichment of expired air wereas described previously for studies of [l)3Cjleucine hi-netics (9).

The mass spectrometric methods and procedures forcalculation of valine flux and oxidation were also the sameas those used in the previous [l-’3Clleucine studies (9, 13).We have discussed the precision of the methods in theinitial paper in this series (9).

Estimation of valine balance

The estimations of valine balance for the 12 h fed periodand the 24 h day were made using the procedure describedin the initial paper concerned with leucine kinetics andleucine intake (9). Thus, the intake of valine and its oxi-dation during the 12 h feeding period and the oxidationof valine during the 12 h period of fast were calculated asdescribed (9), with the latter value taken to be 48Mmol’ kg� 12 h’ based on the lowest rate of valine oxi-dation observed, which was for the 4 mg valine. lc�. 12h� test intake level. For the various reasons discussed ear-lier (9) our estimates of valine balance are probably higherthan the actual balance for each test valine intake. Thus,our estimates of intake levels of valine required to achieve

valine balance are likely to be underestimates. This beingthe case our conclusions, discussed later, are strengthened.

Statistical methods

Analysis of variance and covariance (14) were used toevaluate the relationship between the dynamic aspects ofvalinemetabolismandvalineintakeforeachsubjectgroup.We have assumed that the two groups of subjects werederived from the same population.

Results

Plasma amino acid levels

Plasma valine concentrations, in blooddrawn after an overnight fast and during theperiod when subjects were consuming meals,showed a progressive decline (p � 0.01) withthe decrease in dietary valine intake (Table 2).The fall in plasma valine with reduced dietaryvaline appeared to level off below 20mg � kg� . day� valine intakes. There was nochange in the plasma leucine concentrationwith altered valine intake. The changes inplasma valine in response to decreasing dietaryvaline were statistically significant for bothfasted and fed states, but were less dramaticthan the alterations in plasma valine pool sizeinduced by changes in dietary leucine intake(9, 15).

Valine kinetics

Plasma valine flux and oxidation rates insubjects studied while they were in the fed stateare summarized in Table 3. Valine flux de-creased linearly with a restriction in dietaryvaline content. No significant difference (p� 0.05) was found between the means of theslopes of the regression of flux on valine intakefor individuals between the two diet groups.

The relationship between the dietary valinelevel and valine oxidation, for both groups, isalso summarized in Table 3. No significantdifference was observed between the twogroups for the 16 mg’kg’ �day’ diet level.Valine oxidation decreased with a restric-tion in valine intake between the 68 and20 mg#{149}kg� day� levels but was rela-tively constant for intakes of between 20mg#{149}kg’ day� and below.

Making the same assumptions as in the caseof our leucine studies (9), estimates of valinebalance during the 12 h fed period and for the24 h day can be made and these values are

by guest on October 29, 2015

ajcn.nutrition.orgD

ownloaded from

784 MEGUID ET AL

TABLE 2Plasma valine and leucine concentrations in young men, during postabsorptive (fasted) and fed states, for varyingdietary valine intakes

Amino acid andcondition

valine in take (mg .kg’ .day*)

Group I Grou p2

68 40 20 16 16 12 8 4

�iM

ValinePostabsorptive *

Mean 294 175 106 95(4) 126(4) 110 121 120(3)SEM 16 10 8 10 3 7 6 7

FedtMean 349 201 67 56(4) 74(4) 59 53 51(3)SEM 7

‘

31 8 2 3 6 3 3Leucine

Postabsorptive *

Mean 135 129 115 120(4) 127(4) 130 138 127(3)SEM 8 8 5 5 4 6 7 9

FedtMean 126 132 138 136(4) 138(4) 152 144 171(3)SEM 3 5 10 10 5 5 4 10

* Blood sample taken prior to the breakfast meal.

t Mean value of samples taken from each subject at 1.5 and 2.0 h into the isotope infusion period. Based on six

subjects per group, except as noted in parentheses.

recorded in Table 3. A distinctly positive fed- line intakes studied. Assuming a low rate ofstate valine balance occurred at the higher in- valine oxidation of 4 �mol #{149}kg� #{149}h� duringtakes ofthe amino acid, but mean balance also the postabsorptive period valine balanceremained positive throughout the range ofva- throughout the 24 h day was negative at in-

TABLE 3Plasma valine flux and valine oxidation in young men receiving meals supplying graded intakes of valine

Dietary valine in

Group I

Indezofvahnemetaboh*m 68 40 20 16

t�k�(�g.kg”dey’)

Group 2

16 12 8 4

Flux (j�mol kg�’ . h�)Mean* 1 10 89 50 40 (4) 48 (3) 51 (4) 51 38 (2)SEM 81 5 3 5 2 2 2 2

Oxidation (�smol kg’ . h�)Mean 26 17 8 6(4) 8(3) 8(4) 6 4(2)SEM 2 1 1 0.4 1 1 0.4 0.4

Mean balanceFed-statevalineintaket 602 363 192 157 157 124 89 55Fed-state valine

oxidationt 315 207 98 74 98 98 74 50Fed-state valine

balancet +287 +156 +94 +83 +9 +6 +15 +5Postabsorptive valine

balancett -48 -48 -48 -48 -48 -48 -48 -48Dailyvalinebalance� +239 +108 +46 +35 +11 -22 -33 -43Daily valine balanc&’ +28 +13 +5.4 +4 +1.3 -2.6 -3.9 -5.0* All means based on six subjects except as noted in parentheses.

t Yalues expressed as �imol . k�’ - 12 h’.t Equivalent to postabsorptive valine oxidation.§ Yalues expressed as Mmol#{149}k�’ 24 h’.II Yalues expressed as mg kg� day’.

by guest on October 29, 2015

ajcn.nutrition.orgD

ownloaded from

VALINE METABOLISM 785

takes below about 16 mg . kg I day(Table 3).

Discussion

In the preceding paper (9) we showed a re-lationship between leucine intake and leucineoxidation in healthy young men and at intakesbelow the equivalent of about 20 mg . leucinekg . � . day ‘ oxidation rates were estimatedto exceed the 24 h intake level. Similarly, atentative picture ofthe quantitative aspects ofbody valine metabolism in relation to valineintake can be developed from the presentfindings. The design of the overall study, in-volving two separate groups of young adultmen, was intended to permit examination ofthe relationships between parameters of valinekinetics and alterations in dietary valine intakelevel within both the submaintenance and su-pramaintenance ranges of valine intake. Thequalitative nature ofthe relationships betweenthe dynamic variables of body valine kineticsand valine intakes was, in part, similar to thatfor leucine (9). Thus, as for leucine, the oxi-dation of valine represents an irreversible lossto the body of this indispensable amino acid.In order to maintain body valine balance andan adequate protein nutritional status it wouldbe necessary to match this metabolic loss ofvaline by an adequate dietary valine supply.We can therefore, estimate a dietary intakethat would be sufficient to meet the losses ofvaline via oxidative catabolism.

From the data shown in Table 3, the oxi-dation of valine proceeded at a rate equivalentto 8 �tmol kg� h� at valine intakes rangingfrom 12 to 20 mg#{149}kg’ per day. Thus, for thisintake range, during the 12 h fed state 96�tmol/kg (or 11.2 mg/kg) of valine were irre-versibly lost via oxidation. Because the oxi-dation of indispensable amino acids proceedsalso during the postabsorptive phase of aminoacid metabolism, the present estimates mustbe raised to provide a reasonable determina-tion of the daily rate of valine oxidation.

In earlier studies (7) we found that the rateof leucine oxidation during the postabsorptiveperiod in subjects consuming an essentiallyprotein-free diet approximated that found forintakes of protein below requirement levels.Making the reasonable assumption that thiswould also be the case for valine, another in-

dispensable branched-chain amino acid, wecan estimate that postabsorption valine oxi-dation under those circumstances would ap-proximate no less than about 4 smol kg’ . h’(Table 3) or the equivalent of48 �tmol/kg (5.6mg/kg) over the entire 12 h postabsorptive pe-nod. Thus, it can be calculated that total dailyvaline oxidation would exceed an intake levelof 12 mg . kg ‘ #{149}day’ (Table 3). Valine bal-ance, on this basis, would be barely achievedat an intake of 16 mg � kg ‘ day’ . Hence, itmust be concluded that the upper requirementfor valine cannot be as low as the current in-ternational estimate of 10 mg � kg ‘ . day’(1 1), based on the data of Rose (16), Levertonet al (17), and Linkswiler et al (18). Perhapsthis discrepancy between the higher require-ment estimate obtained by the present oxi-dation values and the lower requirements de-rived from N-balance studies is not surprisingsince, it should be noted, that valine intakesof 10-14 mg per kg per day would not havebeen sufficient to support N-balance ifthe N-balances calculated by these earlier investi-gators (16-18) had also included an allowancefor unmeasured nitrogen losses. It is also rel-evant to recall that Weller et al (19) was unableto confirm the adequacy of the amino acidrequirements of Rose (16). Whether this ap-plies specifically to valine or other indispens-able amino acids was not explored by Welleret al (19).

This interpretation ofour valine kinetic datais further supported by reference to the plasmavaline concentration changes. Thus, there wasa significant (p � 0.01) decline in the plasmavaline concentration when valine intakes werereduced to the 20 mg. kg’ day’ level (Table2). Below this intake level plasma valine con-centrations were relatively constant. Further-more, when the concentration of an indis-pensable amino acid declines with the inges-tion of a protein-containing meal this indicatesthat the amino acid content of the meal is lim-iting or inadequate (20). This occurred in thepresent study at valine intakes below 40mg#{149}kg’ day’. With reference to compara-ble plasma amino acid responses observed ingrowing animals (21, 22) these plasma valinedata also suggest that the valine require-ment might approximate a mean of 20mg� kg’ . day’. This intake level would besupplied by about 0.3 g egg protein per day

by guest on October 29, 2015

ajcn.nutrition.orgD

ownloaded from

786 MEGUID ET AL

and thus, at mean requirement intakes of pro-tein of about 0.6 g per kg per day, the valineintake would still be in excess of the minimalphysiologic need for the amino acid.

In summary, and in agreement with ourstudy ofleucine kinetics (9), the present studyraises questions about the adequacy of thecurrent estimates of the valine requirementsin healthy adult men, as judged from earliernitrogen balance studies. We accept that ourexperiment involves a limited examination ofthis problem, a feature characteristic of alladult human studies concerned with deter-mination of minimum physiologic require-ments for indispensable amino acids. The ca-yeats concerning a more extensive interpre-tation of these data in relation to issues ofamino acid requirements are presented in thepreceding paper dealing with leucine (9).Clearly, additional studies, particularly longer-term metabolic experiments, will be necessarybefore the precise value for the minimumphysiological requirement for valine in healthyadults can be stated with some confidence. 0

These studies received financial support from USDAgrant 5901-0410-006 and 59-2253-1-1-7700 and NIHgrants HD10667, RR88, and RR984. We thank the staffof the MIT CRC and Christine Bilmazes for their experthelp and the volunteera for making these studies possible.The encouragement and support given by Professor NevinS Scrimshaw is appreciated greatly. The L-amino acidsused for preparation of the dietswere generously donatedby Ajinomoto, USA, Inc.

References

1. Brookes JM, Owens FN, Garrigus US. Influence ofamino acid level in the diet upon amino acid oxidationby the rat. J Nutr 1972;l02:27-36.

2. Bergner H, Simon 0, Adam K. Estimation of lysinerequirement in growing rats from rate of catabolismof �C- and ‘5N-labeled lysine. Arch Tierernabrung197828:2l-9.

3. Kang-Lee TA, Harper AE. Effect of histidineintakeand hepatic histidase activityon the metabolism ofhistidinein vivo.J Nutr 1977;l07:l427-43.

4. Kang-Lee TA, Harper AE. Threonine metabolism invivo:effect of threonine intake and prior induction ofthreoninedehydratase in rats. J Nutr l978;l08:163-75.

5. Simon 0, Adam K, Bergner H. StoffwechselorientierteLysinbedarfsbestimmung bei ausgewachsenen Rattenanhand der katabolisierungsrate von �C- und ‘3N-markiertem Lysin. Arch Tierernahrung l97828:609-17.

6. Young YR., Munro HN. Plasma and tissue tryptophanlevels in relation to tryptophan requirements ofweanling and adult rats. J Nutr l973;l03:1756-63.

7. Motil KJ, Matthews DE, Bier DM, Burke JF, MunroHN, Young YR. Whole body leucine and lysine me-tabolism: response to dietary protein intake in youngmen. Am J Physiol l981240:712-2l.

8. Young YR, Robert JJ, Motil KJ, Matthews DE, BierDM. Protein and energy intake in relation to proteinturnovermn man. In: WaterlowJC, Stephen JML, eds.Nitrogen metabolism in man. London, England andEnglewood, New Jersey. Applied Science Publishers,1981:419-47.

9. Meguid MM, Matthews DE, Bier DM, Meredith CN,Soeldner iS, Young YR. Leucine kinetics at gradedleucine intakes in young men. Am i Clin Nutrl986;43:770-80.

10. Williams HH, Harper AE, Hegsted DM, Arroyave G,Holt LE Jr. Nitrogen and amino acid requirements.In:(Committee on Amino Adds, Food and NutritionBoard)Improvement ofProtein Nutriture. Washing-ton, DC: National Academy ofSciences, l97423-63.

I 1. FAO/WHO. Energy and protein requirements. WHOTech Rept 5cr No 522. Geneva, Switzerland: WorldHealth Organization, 1973.

12. Ailsop JR, Wolfe RR, Burke FJ. Tracer priming thebicarbonate pool. J Appl Physiol 1978;45: 137-9.

13. Matthews DE, Ben Galim E, Bier DM. Determination� of stable isotopic enrichment in individUal plasma

amino acids by chemical ionization mass spectrom-etry. Anal Chem 1979;5l:80-4.

14. Dixon WJ, Massey FJ. Introduction to Statisticalanalyses. 3rd ad. New Yodc McGraw-Hill, 1969:150-87 and 222-34.

15. Hambracus L, Bilmazes C, Dippel C, Scrimshaw NS,Young YR. Regulatory role of dietary leucine onplasma branched-chain amino acid levels in youngmen. J Nutr l976;l06:230-40.

16. Rose WC. The amino acid requirements ofadult man.Nutr Abstr Rev 195727:63l-47.

17. Leverton RM, Gram MR. Bmdovsky E, CbaloupkaM, Mitchell A, Johnson N. The quantitative aminoacid requirements ofyoungwomen. II. Valine. J Nutr1956;58:83-93.

18. Linkswiler H, Fox HM, Gesclwender D, Fry PS.Availability to man of amino acids from foods. II.Valine from corn. J Nutr 1958;65:455-68.

19. Weller LA, Calloway DH, Margen S. Nitrogen bal-ances of men fed amino acid mixtures based on Rose’srequirements, egg white protein, and serum free aminoacid patterns. J Nutr 197 l;lOl:l499-508

20. Langenecker JB, Hause NL Relationship betweenplasma amino acids and composition of ingested pro-tein. Arch Biochem Biophys 1959;84:46-59.

21. McLaughlan iM, lilman WI. Use of free plasma aminoacid levels for estimating amino requirements of thegrowing rat. J Nutr l967;93:2l-4.

22. Zimmerman PA, Scott HM. Interrelationship ofplasma amino acid levels and weight gain in the chickas influenced by suboptimal and supraoptimal dietaryconcentrations of a single amino acids. J Nutr l965;87:13-20.

by guest on October 29, 2015

ajcn.nutrition.orgD

ownloaded from