The Journal of Nutrition Vol. 128 No. 12 December 1998, pp. 2563S-2576S

Metabolic Demands for Amino Acids and the Human Dietary Requirement: Millward and Rivers (1988) Revisited^1,2

D. Joe Millward

Centre for Nutrition and Food Safety, School of Biological Sciences, University of Surrey, Guildford, Surrey GU2 5XH, UK

	ABSTRACT

Abstract References

In 1988, Millward and Rivers reappraised existing metabolic models for amino acid requirements. The metabolic demand for amino acids was reviewed in relation to both obligatory metabolic consumption and adaptive pathways of amino acid oxidation. The obligatory demand pattern was deemed unknowable from first principles except that the level of one amino acid would be similar to its concentration in an amount of tissue protein equivalent to the obligatory nitrogen loss. The adaptive demand pattern was predicted to vary in relation to the amount and the periodicity of food protein intake that influenced the amplitude of the diurnal cycle of gains and losses. A regulatory influence of protein intake on anabolism, the anabolic drive, was identified in animal studies; benefit appeared to derive from intakes in excess of the minimum for balance, which could facilitate definition of an optimal requirement. The inherent and design-related limitations of both nitrogen and stable isotope balance studies of requirement were recognized as a major problem in identifying secure values for indispensable amino acid requirements. A decade of research of increasing methodological sophistication has generated much new information, confirming the adaptive diurnal model of balance regulation and allowing development of the anabolic drive into a general protein-stat theory for coordinated control of growth and maintenance of the lean body mass. However, notwithstanding several new estimates of amino acid requirement values, definition of a widely accepted human amino acid requirement pattern remains unresolved. Although a case can be made for an adjusted 1985 FAO adult requirement pattern being a reasonable estimate of the obligatory indispensable amino acid requirements for human maintenance, the problems posed by adaptation, methodological inadequacies and lack of independent measures of adequacy mean that assessment of the adequacy of the human diet to satisfy amino acid needs remains inherently difficult.

	ARTICLE

Abstract References

	BACKGROUND

John Rivers' impact on the science of nutrition, disaster relief and his contribution to the understanding of the essential fatty acid requirements and metabolism in the cat were described in the first memorial lecture (Sinclair 1994). His interest in the protein and amino acid requirement debate in the 1980s derived from his belief in learning the lessons of history, in this case the arguments that raged in the latter part of the 19th century in relation to the relative merits of high or low protein intakes. Those fortunate enough to been taught by him at the London School of Hygiene will remember his detailed analysis of the development of ideas about the nature and magnitude of the need for protein in the human diet: from Playfair's identification of the very high dietary protein intakes of laborers in Edinburgh, suggesting needs of ~180 g/d, through to Russel Chittendon`s studies with low protein diets, resulting in his advocacy of ~50 g/d as the optimal intake. The latter value is remarkably similar to the currently accepted allowance as identified by FAO/WHO in 1985 (FAO/WHO/UNU 1985) after a series of committees of FAO and WHO, at a stroke of their joint committee's pen, had either lowered or increased the protein requirements and, consequently, both the world prevalence of protein deficiency and the economic value of the West's agricultural surpluses.

Against this background, John Rivers and I began to examine the consequences of the 1985 FAO report from the perspective of indispensable amino acid (IAA)³ requirements and protein quality, especially the age-specific amino acid requirement and scoring patterns to be used in evaluating protein quality. The key feature was the marked decline in the IAA requirement values with age, from >50% of total protein requirement in infants to only 16% in adults. This low requirement level of IAA for adults meant that their supply from all natural diets and food proteins would be adequate, and apart from digestibility, protein quality ceased to be an issue in the nutrition of adults.

View larger version (59K): [in this window] [in a new window]   Fig 1. Metabolic fate of amino acids in relation to the metabolic demand and protein requirement. The metabolic demand comprises both obligatory and adaptive metabolic consumption, which occurs continuously throughout the postprandial and postabsorptive phase of the daily cycle. During feeding, the dietary intake provides for the daily metabolic demand in terms of fed state obligatory and adaptive metabolic losses as well as repletion of tissue protein lost in providing for the metabolic demand during the postabsorptive phase of the diurnal cycle. Any additional fed-state losses reflect the degree of inefficiency of utilization, which is also part of the requirement. Modified from Millward and Rivers (1988).

View larger version (27K): [in this window] [in a new window]   Fig 2. Diurnal changes in nitrogen balance in subjects habituated to increasing protein intakes. Postabsorptive and postprandial N excretion is shown partitioned into obligatory and adaptive metabolic losses and protein gain; the values represent rates per 12 h. The obligatory nitrogen loss is assumed to be 27 mg/(kg · 12 h) during postabsorptive and postprandial periods. Additional postabsorptive loss is identified as the adaptive metabolic demand with a similar amount assumed to occur in the postprandial state. Thus increased excretion in the postprandial state represents a metabolic inefficiency. Postprandial gains are measured by 12-h nitrogen balances adjusted for changes in the urea pool and validated with measurements of 13C leucine oxidation and balance in subjects habituated to the intakes over 2 wk. Overall daily balance is the difference between total postabsorptive losses and postprandial deposition and was negative at the two lowest intakes and positive at the two highest intakes (data from Price et al. 1994).

View larger version (16K): [in this window] [in a new window]   Fig 3. Mean daily rate of protein turnover with increasing protein intakes. Values calculated from 13C leucine kinetics measured in subjects fed the indicated protein intakes for 2 wk (data from Pacy et al. 1994). Hourly rates measured in postabsorptive and postprandial states were scaled to 24 h assuming a 12 + 12 h daily cycle.

View larger version (17K): [in this window] [in a new window]   Fig 4. Rate of change in the adaptive metabolic demand during transition from a high to lower protein diet. Data are obligatory and adaptive metabolic demands calculated from postabsorptive leucine oxidation converted to N excretion per 24 h. Similar changes were observed in actual N excretion; because of the slow change in the adaptive metabolic demand after the reduction in protein intake, there was a negative nitrogen balance throughout the experimental period after the diet change (data from Quevedo et al. 1994).

View larger version (29K): [in this window] [in a new window]   Fig 5. Fed-state balance during the initial feeding of a reduced protein intake. Twelve-hour N balance data measured in the postprandial state. The lack of change of the adaptive oxidation component results in none of the intake being deposited (data from Quevedo et al. 1994).

Disquiet about the 1985 report was initiated by Young (1986), who argued that the adult IAA requirement values were seriously flawed because of the way Rose conducted his N balance studies (mainly excess energy and no account for miscellaneous N losses). This was followed with a series of stable isotope tracer balance studies suggesting much higher requirements for leucine (Meguid et al. 1986a), valine (Meguid et al. 1986b), threonine (Zhao et al. 1986) and lysine (Meredith et al. 1986). In 1988, Millward and Rivers published a review that reappraised existing metabolic models for protein and amino acid requirements with a particular focus on the adaptive behavior of pathways of amino acid oxidation in relation to intake, how this would influence the apparent requirement and make the apparent requirement vary according to experimental design. From this perspective, the low adult requirement values reflected experimental designs that would have minimized oxidative losses of amino acids. A new model was proposed that took account of adaptive losses of amino acids within the requirement and also included the idea of metabolic responses to protein independent of balance, the "anabolic drive" (Millward and Rivers 1988). Thus, a minimum requirement value could be defined only under specific artificial conditions that minimized the adaptive component; it was concluded that "current estimates of adult requirements may be close to this level. A requirement of practical value will only be definable when the anabolic drive can be evaluated."

View larger version (21K): [in this window] [in a new window]   Fig 6. Postprandial protein utilization of milk and wheat gluten. Cumulative 13CO2 excretion after a single meal of either milk protein (solid box) or wheat gluten (open box) during a 13C-1 leucine infusion (data from Fereday et al. 1996). PPU, the efficiency of postprandial protein utilization calculated from the cumulative increase in leucine oxidation was 0.71 for wheat and 0.77 for milk protein.

View larger version (14K): [in this window] [in a new window]   Fig 7. Nitrogen balances observed in young adults fed diets containing the FAO, MIT and an egg pattern. Diets contained 9.8 g N, 0.9 g protein/kg and were fed for 3 wk (data from Marchini et al. 1993). There were no significant differences among the balances with consumption of any of the diets.

View larger version (22K): [in this window] [in a new window]   Fig 8. The anabolic drive within a protein-stat mechanism of growth control. During growth, whole-body protein content is controlled by protein intake, itself regulated through an aminostatic appetite mechanism, controlling primarily skeletal muscle mass at a level set by the linear dimensions of the organism. Bone lengthening occurs at rates determined by genetic programming (canalization) and an appropriate anabolic drive, exerted by dietary protein. Muscle protein deposition requires activation of connective tissue synthesis and myofibrillar protein deposition by passive stretching of skeletal muscle by bone length growth. In muscle, connective tissue is proposed to act like a bag that limits myofiber volume so that during growth, connective tissue remodeling, bag enlargement, must occur to allow increased myofiber volume. With connective tissue growth, i.e., bag size, limited by passive stretch, consequent muscle size will be regulated at a phenotypic muscle weight:bone length ratio. Provision of amino acids to allow muscle to increase to its phenotypic size, bag filling, is regulated through appetite stimulation in response to net amino acid flow into muscle at a rate greater than the dietary intake. Maximal phenotypic muscle size also requires some physical activity. Muscle growth, i.e., bag enlargement, ceases when bone length growth ceases. The growth of most other organs is secondary to this main interaction, determined primarily by the level of protein intake and the consequent metabolic work and functional demand for the organ and is not specifically limited in size. Operation of the protein stat occurs only at dietary protein levels sufficient to provide both substrate and an appropriate anabolic drive acting on protein deposition in bone, muscle myofibers (but not connective tissue synthesis) and other organs (modified from Millward 1995).

View larger version (39K): [in this window] [in a new window]   Fig 9. Amino acid requirement levels and dietary adequacy. Requirements can be defined for rapid growth, in this case, rapid growth at 1 mo of age on the basis of the factorial method proposed by Dewey et al. (1996) for growth and maintenance, using the revised lysine value showed in Table 4. The minimum obligatory metabolic demand is also represented by the adjusted values in Table 4 in relation to a maintenance requirement of 0.6 g protein/(kg · d). Assessment of the nutritional value of intakes in the adaptive range between these two extremes is problematic.

In practical terms, probably the most important aspect of the review was the attempt to identify which IAA might be rate limiting for the obligatory nitrogen losses (ONL). This involved calculation of the pattern of IAA losses from tissue protein mobilized to provide for the ONL, defined as the obligatory oxidative amino acid losses (OOL). These values for the OOL subsequently became the basis for the MIT scoring pattern for protein quality evaluation in adults and children published by Young et al. (1989) after some small adjustments of lysine, threonine and valine were made on the basis of their stable isotope studies. When the MIT scoring pattern was used to assess the adequacy of protein intakes in developing countries, a lysine deficiency was identified; this was rectifiable by substitution of cereal-based diets with one third animal protein (Young and Pellet 1990). Subsequently a report on protein quality evaluation (FAO/WHO 1991) was unable to identify any valid adult scoring pattern and proposed the use of a preschool child scoring pattern with values close to or somewhat higher than the MIT pattern for older children and adults as a strict interim measure. In my view, this decision was flawed because neither the preschool child nor the MIT pattern was suitable for protein quality evaluation (see Millward 1990, 1992, 1993 and 1994, Millward et al. 1989 and1990, Millward and Pacy 1995). The issue was considered further by an international expert group in London in 1994 and, contrary to what was published (Clugston et al. 1996), the MIT pattern was not endorsed (see Millward and Waterlow 1996).

The purpose here is to revisit the main features of the Millward and Rivers review in the light of current knowledge.

	METABOLIC DEMANDS AND PROTEIN REQUIREMENTS

Protein and amino acid requirements are a function of the metabolic demands (MD) of the organism and the efficiency with which diet can be utilized to meet the MD. The key features of the Millward and Rivers 1988 review were the separation of the MD into two components, obligatory (or fixed) and adaptive (or variable), and recognition of the need to take account of the diurnal nature of overall daily balance in which fed-state gains must balance fasting losses (see Fig. 1). It was further suggested that because the obligatory metabolic demand is small (much less than normal intake), most individuals have requirements that include a substantial adaptive component and are therefore greater than the minimum requirement. This not only makes the understanding of protein requirements difficult but also means that measurement is also difficult because of the need to account for adaptation. From this perspective, the main issues to be revisited here are the nature and magnitude of the obligatory metabolic demand for amino acids and the adaptive metabolic demands, especially in relation to diurnal cycling. Also reviewed here will be the all important issue of methodology, the main stumbling block to developments in this area. Finally the anabolic drive will be briefly revisited

The obligatory metabolic demand.  The obligatory MD is best understood as a demand for amino acid precursors for two components as follows: 1) net protein synthesis of tissue proteins and 2) a range of non-protein products such as nucleic acids, diverse smaller molecules such as creatine, taurine, glutathione, catecholamines, thyroxine, serotonin, dopamine and nitric oxide, and irreducible catabolism, including that of carbon skeletons of amino acids that pass into the large bowel and are lost during bacterial fermentation with amino nitrogen reabsorbed as ammonia. The importance of this division is that the qualitative nature of the first component is straightforward and determined by the amino acid pattern of tissue protein deposited. However, the qualitative nature of the second component is complex and unknowable from first principles, reflecting the sum of all the individual amino acid transformations involved.

Protein turnover is often invoked as part of the MD; in fact, little metabolic demand for amino acids is generated in this way because of amino acid recycling, which can occur indefinitely. Thus "wear and tear" as a driver of MD is not an appropriate biological analogy, and only net protein synthesis can contribute to the MD. The qualitative nature of this is usually assumed to be influenced only by the changes in body composition occurring during growth and is generally assumed not to vary with the diet within cells and tissues; however, as discussed by Fuller and Garlick (1994), some evidence does exist for changes in amino acid content of tissues during growth when amino acid-limiting diets are consumed.

In human nutrition, with growth occurring very slowly after the first year of life, net protein synthesis contributes a small and decreasing component of the MD. In the adult, it comprises only that associated with continuing growth of skin and hair, various secretions and the synthesis of those gastric secretions (e.g., threonine-rich mucus glycoproteins) that pass into the colon to be utilized for bacterial metabolism. As discussed below, this amounts to ~22 mg/kg of essential amino acids, 3 mg/kg N. Thus for the adult organism, normally at nitrogen or amino acid equilibrium, this net protein synthesis may represent <10% of the total MD, mainly non-protein pathways of amino acid metabolism and catabolism associated with maintenance of normal function and composition.

Obligatory metabolic demands: comparative information.  Millward and Rivers (1988) argued that the composition of the obligatory MD could not be predicted from first principles. Because there appear to be few major differences among mammalian species as far as the fundamentals of amino acid and protein metabolism, with the obvious exceptions, (e.g., arginine requirements for growing cats and growing and adult dogs, the taurine requirement kittens and a high maintenance amino acid requirement in avian species for feather growth), then interspecies comparisons of robust animal data should provide general principles about the nature of human needs.

Two sorts of study are pertinent. The first involves deletion studies in which individual amino acids are removed from the diet and the extent of the negative balance is monitored. If the maintenance requirement patterns corresponded exactly to the patterns of tissue protein, then there should be a similar negative balance on removal of each IAA. If not, then negative balance will occur in proportion to the ratio of obligatory MD to tissue content of each amino acid. Only one report exists for the adult rat (Said and Hegsted 1970), which is an excellent study based on measured changes in body water. Gahl et al. (1991) reported data for young rats, and Fuller et al. (1989) studied 41-kg pigs with N balances.

Table 1 shows the relative losses normalized for the response to a protein-free diet. The first, second and third limiting amino acids are threonine, total sulfur amino acids (TSA) and isoleucine for the growing rat, threonine, isoleucine and tryptophan for the adult rat, and TSA, threonine and tryptophan for the pig. The most highly conserved, least limiting amino acids are lysine and leucine in the rat and all three branched-chain amino acids (BCA) and lysine in the pig.

Supplementation with each limiting amino acid allows the slope of the balance curve and the consequent requirement values to be established; these are shown in Table 2 compared with carcass protein content. These patterns have been normalized for threonine for ease of comparison. Leucine and lysine are the two most abundant amino acids in carcass proteins and in the growth requirement patterns for both rat and pig, whereas in the maintenance requirement patterns, the most abundant are threonine and TSA in the pig, and threonine, isoleucine, valine and TSA in both adult and growing rats.

The main implication of these animal data is that there are marked differences between the MD for maintenance and for growth. It is clear that in growing rats and pigs and adult rats, leucine and lysine exhibit the largest difference between growth and maintenance patterns, i.e., they are most abundant for growth and among the least abundant for maintenance. The practical consequence of this as pointed out by Hegsted (1973) is that the balance-intake curve is extremely shallow for leucine and lysine in both the submaintenance and growth ranges; this means that measurement of a requirement value for maintenance is very difficult because it is dependent on the exact criterion for adequacy. Although several early reports of rats maintaining body weight for 6-mo periods while consuming very low lysine diets [e.g., zein, (Osborne and Mendel 1916)] or even lysine-free diets (Bender 1961) are probably explained by coprophagy, given the clear evidence of a metabolic need for lysine in terms of the rapid onset of symptoms on a lysine-free diet in humans (Rose 1957), no evidence exists for anything other than a low metabolic requirement for lysine.

Yoshida (1983) has done the most to explore the concept that rate-limiting amino acids at maintenance differ from those that rate-limit growth. Having shown that the most rate-limiting amino acids were threonine and TSA in adult rats fed a protein-free diet, he then showed that in adult rats fed limiting amounts of rice or wheat diets, the limiting amino acids were threonine and the sulfur amino acids which, when added to the cereal diets, restored nitrogen balance and transformed body weight loss to growth. This may explain why attempts to show that lysine is the limiting amino acid in wheat in supplementation trials in human adults were so disappointing (Scrimshaw et al. 1973).

Fuller's work has pointed to ileal amino acid losses as a partial explanation of the relative metabolic need for individual amino acids. Table 3 compares ileal losses of the pig (Wang and Fuller 1994) and human (Fuller et al. 1994). These account for ~60% of pig amino acid maintenance requirements, and it is clear that in each case threonine is the largest component. Although the patterns differ to a certain extent, most importantly, the absolute values are much lower in humans than in pigs. Thus whatever is said about the pig as an inappropriate model for humans, to the extent that ileal losses comprise a component of obligatory MD, these data point to a lower MD in humans than in pigs.

In summary then, as far as the obligatory MD is concerned, there is a consistent and extensive body of animal data showing that the maintenance pattern differs from that for growth with lower levels of lysine and leucine in the maintenance pattern. Furthermore, the effect is that the rate-limiting amino acids in dietary proteins can differ between maintenance and growth using the example of lysine, which limits wheat for growth but may not limit maintenance.

Obligatory metabolic demands and obligatory oxidative losses.  A second approach to identifying the obligatory MD for at least one amino acid is through calculation of the OOL. When a protein-free diet is consumed, the obligatory MD is fueled by tissue protein, giving rise to the ONL, ~54 mg N or 0.34 g protein/(kg · d). The OOL is the amino acid pattern of the ONL, assumed to be muscle protein as a first approximation. Thus, of the amino acids that comprise the OOL, one will be rate limiting, i.e., having the highest ratio of obligatory MD to OOL. All other amino acids with a lower ratio would be present in excess in the OOL but would be oxidized nevertheless because they could not be returned on their own to the tissue protein pool. Thus if protein turnover is tightly regulated, allowing just sufficient net proteolysis to provide the MD of the rate-limiting amino acid, the OOL of this amino acid should be a reasonable guide to its requirement; for all others, the values for the OOL should be greater than the maintenance requirement.

The identification of this rate-limiting amino acid cannot be done from first principles, only by reference to the actual pattern of the obligatory MD, which should comprise values equal to or less than the OOL pattern. In fact, comparison of the OOL pattern with the FAO pattern showed that the OOL of the TSA were quite close to the FAO requirement values, with the OOL of all other amino acids up to 2.5 times higher. This would mean that the TSA are rate determining for the mobilization of tissue protein to provide for obligatory MD if the TSA requirement value in the FAO pattern is correct. The possibility that the TSA are the rate-limiting amino acids that drive the ONL has been suggested by studies in both dogs (Allison et al. 1947) and rats (Yoshida and Moritoki 1974) in which supplementation by S amino acids reduces N excretion when a protein-free diet is consumed. However, apart from showing that the FAO pattern was feasible, in that no values were greater than the OOL and the TSA appeared to be rate limiting, no more information could be gained from this analysis about the composition of the metabolic demand.

Use of the OOL pattern to predict a requirement pattern.  Young and colleagues have taken a different view of the significance of the OOL. Although animal data clearly indicate that the amino acid pattern of the obligatory maintenance MD differs from that of tissue protein, the OOL pattern was used as the basis for the MIT pattern (Young et al. 1989), and the assumption that maintenance requirements are broadly similar to tissue protein has been subsequently justified (Young and El-Khoury 1995). The view has not been previously put forward and is certainly contrary to the comparative animal data discussed above. A critical analysis questioning the validity of the MIT scoring pattern is presented elsewhere (Millward 1990).

The adaptive component of the metabolic demand.  Although the obligatory MD is straightforward theoretically, the adaptive component of the MD shown in Figure 1 brings considerable complexity to understanding and measuring the maintenance MD.

The ONL at 54 mg N or 0.34 g protein/(kg · d) (FAO/WHO 1973) is only 50% of current estimates of the protein requirement [0.6g/(kg · d)], and the nature of this difference has proved difficult to account for apart from an inefficiency of utilization. However, why proteins such as milk, egg or meat were not utilized more efficiently was always puzzling. This is much more easily understood as representing an adaptive component of MD.

When subjects are fed a protein-free diet, their urinary nitrogen losses fall from an initial level set by their normal dietary protein intake to a low stable output level after 7-14 d (FAO/WHO 1973). This means that there is an additional loss of body nitrogen on a daily basis for some time before equilibrium is reached at the lower level; this additional loss reflects the adaptive component of the MD. Traditionally, this has been defined as "the labile protein reserves," implying that it represented some pool of protein that varied in size with the dietary protein intake. Although liver and visceral protein content does vary directly with dietary protein intake in the rat (Munro 1964), in support of the labile protein reserve concept, it is not known to what extent this occurs in humans and from which tissues the losses occur when a protein-free diet is fed.

Millward and Rivers (1988) were dissatisfied with the labile protein reserve concept but were unable to explain the adaptive losses completely. They developed complex kinetic arguments relating to diurnal cycling, whose overall amplitude they predicted would increase with protein intake, thus implying a postabsorptive negative protein balance that would increase with increased protein turnover induced by increasing protein intakes. Thus these oxidative losses were viewed as occurring mainly in the postabsorptive state and were a function of an inability to maintain protein balance. Subsequent work indicates that this explanation was overcomplicated.

Price et al. (1994) reported the distribution of amino acid oxidation measured as urinary N, ¹³C leucine oxidation and ²H phenylalanine hydroxylation between the postabsorptive and postprandial periods in subjects adapted to a wide range of protein intakes (0.36-2.3g/kg) for 2 wk. This showed that both postabsorptive and postprandial losses of nitrogen increased with increasing intakes. Comparisons of the rate of ¹³C leucine oxidation and total N excretion (corrected for acute changes in the body urea pool) indicated that the two measures were proportional over the entire range of intakes, although leucine oxidation underestimated N excretion by ~25% (see below). These studies confirmed and quantified the extent of the components of the metabolic model describing the fate of dietary amino acids in relation to the metabolic demand shown in Figure 1. Thus the MD at maintenance includes irreversible amino acid consumption in obligatory metabolism for non-protein products as well as adaptive habitual diet-related oxidative amino acid catabolism. Furthermore, each of these two components can be shown to occur at equal rates in the postprandial and postabsorptive states. Any additional amino acid catabolism during feeding reflects an inefficiency of protein utilization. The diet supplies the MD directly in terms of postprandial losses and indirectly for postabsorptive losses by net protein synthesis for the repletion of postabsorptive losses. During growth, the MD will also include additional net protein synthesis for growth. The actual magnitude of the components shown in Figure 1, in terms of the various components of the postabsorptive losses and the metabolic fate of the dietary intake (Price et al. 1994), are shown in Figure 2. At the lowest intake, total losses were in excess of the obligatory MD, indicating that subjects had not fully adapted to this low intake in the 2 wk of the study. At the highest intake, the adaptive component of the MD had increased to more than four times the obligatory MD. The leucine oxidation data indicated a similar pattern except that at all intakes, overall losses were lower (79%) when calculated as equivalent nitrogen.

Protein turnover rates were measured in these studies partly to identify mechanisms of the diurnal changes in balance but also because changes in losses with intake were thought to be caused by increasing turnover with intake. In fact the turnover data were surprising. Thus although the daily gains and losses were mediated mainly by changes in whole-body proteolysis of increased magnitude with protein intake, with a slightly increasing postabsorptive rate of proteolysis with intake, postabsorptive protein synthesis did not change much nor did mean daily rates of protein synthesis or proteolysis (see Fig. 3). Thus the simplest explanation for the adaptive component of the MD is that increasing activities of the pathways of oxidation of amino acids with increasing protein intakes during feeding and fasting represent an actual metabolic demand generated as a response to protein intakes in excess of minimal metabolic needs.

The persistence of this adaptive component of the metabolic demand throughout the daily cycle encouraged us to examine the time course of its adaptation to a lower protein intake (Quevedo et al. 1994). We showed with both ¹³C leucine and N balance studies over periods of up to 14 d that the negative balance induced by the reduced protein intake involved lags in reduction of nitrogen losses in both postabsorptive and postprandial states. Figure 4 shows the changes with time in postabsorptive losses, whereas Figure 5 shows the postprandial response before and immediately after the change in intake. As a result of these slow responses in reducing the adaptive MD, there was an overall marked negative balance; even at the end of 2 wk, equilibrium had not been reached. This implies that the transitional losses during adaptation to a lower intake reflect the time taken for catabolic pathways of amino acid metabolism to adapt from a level set to deal with one level of protein intake to that required for a lower level. If the mechanism was based on labile protein reserves in which the key regulator was protein synthesis or proteolysis rates regulating the size of these reserves so that the fed state losses were due entirely to an overspill once the reserves were repleted, then the expected response to a reduced intake would be an immediate marked reduction in fed state losses because most of the intake was used to replenish the labile reserves as usual with much less overspill. This was not observed because, as shown in Figure 5, on the first day of the lower intake, there was no change in the postprandial losses and a complete abolition of any gain, i.e., there was a small negative balance. This is strong evidence for amino acid oxidation rates acting as the main regulator rather than maintenance of protein stores.

It would appear that during slow growth or at weight maintenance, pathways of oxidative amino acid catabolism adapt to operate at the appropriate rate set by habitual protein intakes so as to be able to rapidly dispose of dietary protein in excess of minimal needs and maintain the very low tissue concentrations of those amino acids such as the branched-chain, aromatic and sulfur amino acids, which may be toxic at higher concentrations. This rate continues regardless of the actual acute intake, utilizing tissue protein if the dietary level falls or during the postabsorptive state, for as long as it takes to adapt to the lower level of intake. This becomes part of the metabolic demand.

Diurnal cycling: a qualitative influence on the metabolic demand?  There is an important but complex implication of diurnal feeding and fasting for the IAA pattern of the MD. Clearly, from the above discussion, with increasing dietary protein intake there will be an increasing MD generated by the adaptive oxidative losses. With both adaptive and obligatory metabolic demand occurring continuously, overall daily balance requires postprandial repletion of tissue proteins mobilized to provide for the postabsorptive demand. Furthermore, the amplitude of this cycle increases with the increasing habitual level of protein intake (Price et al. 1994). The key question is to what extent does this diurnal cycle of body protein influence the IAA composition of the adaptive MD? Young and El-Khoury (1995) have assumed that regardless of the amino acid composition of the intake that induces the adaptive MD, the adaptive MD will have a tissue protein pattern of amino acids to enable postprandial protein deposition. In fact, the actual amount of amino acids needed in the diet to provide for this adaptive requirement cannot be predicted for three reasons.

First, the amount of postabsorptive tissue protein loss will depend on the extent of the true postabsorptive state during a day. Price et al. (1994) utilized 12-h diurnal cycles in their studies, but the actual amount will vary with the pattern of meal-feeding, and individuals consuming both an early breakfast and late supper may spend <12 h in a true postabsorptive state and mobilize less tissue protein. In subjects in the postprandial state, the diet will provide directly for the losses without inducing tissue protein loss.

Second, when true postabsorptive losses of tissue protein do occur, it does not follow that all IAA liberated from the net tissue proteolysis are oxidized. Although increases in the concentrations of the BCA, aromatic and sulfur amino acids liberated from tissue protein will be associated with increased oxidation to minimize any increase in concentration, this is less likely to be the case for lysine and threonine. These two amino acids differ from other IAA with higher K_m values for their main oxidative catabolic enzymes (18 and 52 mmol/L, respectively), larger pool sizes (~1 mmol/L) and with less fine control of their catabolic pathways. The difference in the handling of lysine and threonine may well be important in allowing conservation of these amino acids and recycling from postabsorptive losses for postprandial gains.

Evidence for conservation of lysine and threonine comes from muscle biopsy studies after feeding protein (albumin) or a protein-free diet (Bergstrom et al. 1990). These investigators biopsied human muscle at 3 and 7 h after a meal of 50 g of albumin and showed that for leucine and lysine, although the intakes were the same and their removal into protein will be at the same rate (because their concentrations in protein are similar), the increase in the concentration of lysine was twice that of leucine and the same was true for threonine in comparison with valine. By 7 h, the concentration of all of the BCA, methionine and the aromatic amino acids had fallen below the base line, but for threonine and lysine, there was still an excess of amino acid over the base line value.

Some indication of how much of these free pools of lysine and threonine might be available to supplement dietary protein that was inadequate in lysine and threonine was indicated after feeding a protein-free meal, after which protein deposition would have been entirely dependent on the free amino acid pools (Bergstrom et al. 1990). The fall in the free lysine and threonine pools in muscle was sufficient to have enabled ~250 mg protein deposition/kg body weight, a substantial component of the observed deposition in our studies in subjects consuming 0.75g/(kg · d). Thus the free pool of lysine and threonine could in theory contribute most of their needs if the dietary source was inadequate. If this is the case, then the organism in overall balance may be less sensitive than would be expected to poor dietary quality in terms of lysine and threonine needs for transient protein deposition.

Evidence for this comes from the very small differences in the efficiency of postprandial protein utilization of wheat protein compared with milk in normal adults (Millward et al. 1997). Studies with either frequent small meals (Fereday et al. 1994) or a larger single meal (Fereday et al. 1997) indicate that ~80% of dietary wheat protein is deposited in the tissues even though the lysine and threonine content of the wheat is insufficient to allow this. The cumulative ¹³CO₂ excretion after a meal of either milk protein or wheat gluten is shown in Figure 6. Recycling of the free lysine and threonine liberated in the postabsorptive state allows efficient postprandial protein utilization.

Third, there is evidence that the amplitude of diurnal cycling and the consequent need for postprandial protein deposition are adaptive, according to the IAA composition of the diet. Thus Marchini et al. (1993) showed that the postabsorptive leucine balance was 7.8, 9.3 and 12.8 µmol/(kg · h) when isonitrogenous diets were fed for 3 wk with FAO (very low IAA), MIT (intermediate IAA) or egg (high IAA) patterns of amino acids, i.e., postabsorptive losses with the FAO pattern fell to <60% of those induced by the egg diet even though the overall level of total amino acid intake was the same. This is an important adaptive response. More recent 24-h ¹³C leucine balances (El Khoury et al. 1994a and 1994b) in subjects fed purified amino acid diets that varied only in terms of leucine content [14, 38 and 89 mg/(kg · d)] also indicated adaptive reduction in postabsorptive losses, with postabsorptive leucine oxidation rates equivalent to 15, 20.9 and 34.7 mg leucine/(kg · 12 h). Thus these results point to substantial adaptive reductions in the amplitude of diurnal cycling and in the consequent metabolic demand for IAA in response to reductions in dietary protein quality. This means that the qualitative influence of diurnal cycling on the metabolic demand cannot be predicted.

Summary of the metabolic demand for indispensable amino acids.  From the above considerations, it can be concluded that the metabolic demand for IAA in normal adults includes the following two components: 1) an obligatory component that in most cases is less than that contained in 0.33 g tissue protein and has a pattern that cannot be predicted from first principles but that, on the basis of animal data, is likely to contain considerably less lysine and leucine than in the tissue protein pattern; and 2) an adaptive component whose amount and composition are variable according to the nature and feeding pattern of the habitual dietary protein intake, with a particular adaptive mechanism allowing conservation of lysine and threonine when their intakes are low.

Against this metabolic background, the FAO values for IAA requirements are certainly feasible; given the design of the studies in which they were measured, low levels of IAA and excess nonspecific nitrogen, Millward and Rivers (1988) argued that the adaptive component of the MD would have been minimized. What was measured may therefore have been close to the minimum MD. However, given that the main criticism of the FAO values is leveled at the N balance method itself, it is worthwhile to reexamine briefly some of the main potential problems.

The FAO requirement pattern: limitations of N balance studies.  Some of the difficulties reflect the practical problems that arise in performing N balance studies, i.e., lack of precision of balance estimate as the small difference between larger values of N intake and N excretion (Forbes 1973, Wallace 1959) together with overestimates of intake and underestimates of loss (Hegsted 1976). Underestimated losses include loss of N gas after denitrification by colonic bacteria (Costa et al. 1968), urea losses from skin and expired ammonia (Calloway et al. 1971), and nitrate (food and urine) not measured by the Kjeldahl technique (Kurzer and Calloway 1981) and which may arise in urine from oxidation of endogenous NO production (Anggard 1994). These factors result in typical errors of -11 to +11 mg/(kg · d) (Millward and Roberts 1996), more than the entire miscellaneous losses (8 mg N/kg). In very short-term studies, failure to account for the changing size of the body urea pool can result in errors of up to 66% of the observed N balance (Price et al. 1994).

The nonlinearity of the balance curve means that studies conducted at low intakes underestimate requirements, whereas studies conducted with supramaintenance intakes overestimate requirements (see Millward and Roberts 1996). The logical outcome of this is that reliable balance studies are those that are conducted with intakes very close to the actual requirement, and studies with intakes based on preconceived requirement values will tend to confirm such preconceptions.

A conceptual limitation raised by Young (1986) relates to whether attainment of balance involves adaptation or accommodation, with the latter distinguished by an adverse response such as loss of body tissue or reduction of protein turnover rates (Young et al. 1987). The most obvious solution is to define adequacy as attainment of balance at an appropriate body composition. This raises a major design difficulty in balance studies. Losses of body N and achievement of balance at a lower body N level must be differentiated from any transient losses during an adaptation period that are gradually replaced, thus restoring appropriate body protein stores at the lower intake level. This means that studies have to be long enough not only for balance to be achieved at the lowered level, but also for repletion of losses induced during the adaptation to occur. Balance studies with purified amino acid mixtures raise the problem of utilization efficiency. Collin-Vidal et al. (1994) reported a ¹³C leucine balance study showing a higher utilization of casein compared with a casein hydrolysate, through higher plasma leucine and other amino acid levels and lower inhibition of proteolysis with the hydrolysate. These data imply that requirements for balance with purified amino acids could be at least 33% higher than with whole protein.

Finally, Fuller and Garlick (1994) have pointed to animal data indicating that depletion of the tissue peptides, carnosine (-alanyl histidine) and anserine (-alanyl-L-methyl histidine), and the depletion of hemoglobin, a protein very rich in histidine, with histidine-free diets as fed in the original balance studies of Rose and others, suggest that nitrogen equilibrium may not necessarily mean amino acid equilibrium. They also point to animal experiments showing changes in the amino acid composition of the body with amino acid-deficient diets.

Clearly, on the basis of the above problems, the question can be posed whether any credibility can be given to any N balance data. Young and colleagues have argued that because of "conceptual limitations" and poor reproducibility, N balance studies in adults in relation to assessment of amino acid requirements are of questionable value even when properly conducted (Marchini et al. 1993, Young 1986). Others take what is in my view a more balanced approach, namely, "The potential for error and bias has led some to reject nitrogen balance data in their entirety. It is foolish to accept poor data or to fail to recognize the limitations of good data, but it is equally inappropriate to reject reliable data as long as limitations are recognized" (Scrimshaw 1996). On this basis, there is value in assessing the reliable N balance data, i.e., data for which experimental design cannot be criticized. For example, a recent well-conducted N balance trial of the FAO requirement pattern was reported by Young and colleagues (Marchini et al. 1994). They tested the 1985 FAO pattern fed for 3 wk at a higher total N level (9.8 g N, 0.9 g protein/kg) against the MIT and an egg pattern in balance studies that cannot be faulted. They showed that all three diets maintained N balance at the same level as shown in Figure 7. Young rejects these data as evidence for the validity of the FAO requirement values, preferring ¹³C leucine balance data that show that the FAO pattern fails to maintain balance. On the basis of Scrimshaw`s comments above, this rejection is inappropriate.

Adjustment of the FAO requirement values.  The FAO requirement values involved balance in which no account was taken of unmeasured losses (although some did use positive N balance as their criteria of adequacy). However, this can be remedied by adding in unaccounted losses as suggested by Hegsted (1963). Hegsted performed regression analysis on published data to calculate the requirement for zero balance and reported estimates of requirements for a positive balance of 0.5 g N/d, i.e., accounting for unmeasured losses. As discussed by Fuller and Garlick (1994), this value may be excessive. Such losses are variable with protein intake and exercise; Calloway et al. (1971) proposed a lower value for sedentary subjects receiving moderate nitrogen intakes of 0.30 g N/d/, i.e., 4-5 mg N/kg.

The effect of this addition is shown in Table 4. The values for isoleucine, leucine, valine, threonine and tryptophan were recalculated with the adjustments from Hegsted`s regressions. For lysine, the original lysine balance studies of Jones et al. (1958) have been completely recalculated (see Millward 1999). This study involved 14 women studied at various levels in 50 different balances periods both in the sub- and supramaintenance range with adequate but not excessive energy. Because the sulfur amino acid and aromatic amino acid data available at the time of Hegsted`s analysis were inadequate, and because the OOL of the sulfur amino acids appear to be similar to the minimum MD as discussed above, a value calculated from the OOL by Young et al. (1989) has been used. Clearly, the total requirement values in most cases were doubled from 10% of the protein requirement to 20%. Nevertheless, compared with the OOL pattern (48% total), they still represent a lower fraction, with lysine in particular at half of the concentration in the OOL pattern. None of these studies involved excessive energy intakes (the Rose values were not included in Hegsted`s analysis); thus the recalculated values in Table 4 represent reasonable estimates of the minimum requirement for indispensable amino acids available from balance studies. Furthermore, they can be validated to some extent. As discussed elsewhere (Millward 1999), these estimates of the maintenance requirement can be used together with a growth requirement pattern, based on the human tissue amino acid content, to construct an age-related requirement and scoring pattern for infants and children. When various plant proteins are scored for infants and the values compared with the observed biological value obtained in infants, there is very good agreement between the predicted and observed values. Although those that reject N balance data will describe this as a circular argument, data that validate the lysine value exist independent of N balance. Thus adults fed diets providing the adjusted lysine level for several weeks maintain body weight, lean body mass and fitness as well as N balance (Bolourchi et al. 1968, Edwards et al. 1971). Such data cannot simply be ignored.

Stable isotope estimates of requirement values.  Isotopic studies of amino acid oxidation, in particular those based on ¹³C, were proposed as an alternative to N balance by Young (1986). Daily balances were calculated from measured rates of amino acid oxidation over a range of intakes and requirements reported for leucine (Meguid et al. 1986a) lysine, (Meredith et al. 1986), threonine (Zhao et al. 1986) and valine (Meguid et al. 1986b). Millward and Rivers (1988) discussed the limitations of these studies. Subsequently, a large number of additional ¹³C leucine and some valine balances have been reported in an attempt to improve the database, culminating with the most recent reports in which leucine oxidation and balance were monitored for entire 24-h periods (El Khoury et al. 1994a,b and 1995). Taken together, this represents an impressive and very large and complex body of experimental work. Reviewing the work is difficult because the experimental design and calculation methods have changed over time as potential problems have emerged. For example, the way in which leucine oxidation has been calculated has changed from an approach that, by attempting to compensate for the tracer, would have underestimated the true oxidation by 5-15% (Millward 1992) to a more appropriate calculation of the rate in current reports (El Khoury et al. 1994b). The current 24-h studies represent state of the art although, as discussed below, some problems persist. However, as discussed by Millward and Rivers (1988), notwithstanding the gargantuan effort that went into them, the initial leucine, lysine, valine and threonine studies used as experimental support for the MIT scoring pattern are deeply flawed.

In essence, ¹³C tracer balances can be applied in one of two ways, i.e., as measures of the intake-balance relationship for the particular tracer under examination as with the leucine, lysine, valine and threonine studies, or, alternatively, the balance of one amino acid such as leucine can be used as an alternative to N balance. This latter approach was used by Marchini et al. (1993) to compare the FAO, MIT and egg IAA patterns.

The kinetic assumptions and isotope-related problems associated with this approach are listed in Table 5. This list of problems is by no means exhaustive, but has been separated into those problems that are ultimately solvable and those that are currently less tractable. Problem 1 is unavoidable with current methodology. Although the 24-h studies (El Khoury et al. 1994a,b and 1995) represent heroic experiments, they nevertheless represent a very short period in the context of balance studies in general, with subjects necessarily restricted with unknown consequences for protein balance. Day-to-day variation is large for N balance studies, some of which may reflect the operation of the regulator of body composition. The extent of day-to-day variation in leucine balance is currently unknown, although the balance data of Young`s group (El-Khoury et al. 1994a, 1994b and 1995) indicate that it may be less variable than N balance.

Problems 2-4 are solvable by appropriate methodologies. Problem 5 can probably be ignored for leucine and most carboxyl-labeled amino acids but possibly not for threonine. Threonine exhibits unrealistic positive balances (Zhao et al. 1986), most likely because of substantial fixation as glycine (Ballevre et al. 1990). Problem 6 is important when the oxidation rates are measured as a tracer for overall amino acid oxidation and N excretion. Thus, during the fed state with milk-based food, for example, the higher leucine content in milk (10%) compared with tissue protein (8%) means that excess leucine oxidation during feeding will occur, which, unless corrected for, will result in an underestimate of overall amino acid balance during feeding (Price et al. 1994). Problem 7 is important but is usually ignored. Any route of leucine loss as the entire amino acid (free or protein-bound in skin, hair, urinary and fecal peptides or fecal bacterial protein) will not be measured. Although the magnitude of this is unknown, it could represent the equivalent of several milligrams of N per kilogram. Problem 8 is observed in several reported studies. Indeed, in the studies of Marchini et al. (1993), the leucine balances ranged from -100 µmol/(kg · d) when the FAO diet was consumed to +180 µmol/(kg · d) for the egg diet, values equivalent to changes in body weight over the 3 wk of the study of -1.5 kg to +2.5 kg. Whether such changes in body weight occurred is not mentioned by the authors.

As for problem 9, the fact that the tracer in stable isotope studies is not massless but may be infused at significant rates compared with either dietary intakes or measured oxidation rates is a serious problem. In some circumstances, the tracer infusion rate can be of the same magnitude as the oxidation rate that it is purporting to trace. For example, in a valine infusion (Pelletier et al. 1991a), the infusion rate was 80% of the oxidation rate being studied; in the lysine studies reported by Meredith et al. (1986) on the lowest intakes of lysine, the infusion rate was 72% of the reported oxidation rate, equal to three times the rate of lysine intake from the diet at the lowest dietary intake studied. In other leucine infusion studies (Pelletier et al. 1991b), 8.2 µmol/(kg · h) of leucine was infused, equivalent over 12 h to an intake of 13 mg/kg leucine in individuals fed 13 mg leucine/d. Clearly, the way in which the infusion intake is assumed to influence the system under study and the consequent way that it is treated in calculating oxidation rates and balance can markedly influence the results; this has been also commented on by Fuller and Garlick (1994).

A second problem arises from the fact that the tracer represents an additional level of intake above the dietary level to which the subjects have adapted; thus, what is studied is not what has been previously fed. In the 24-h studies (El-Khoury et al. 1995), the diet was adjusted on the infusion day by removing an amount of leucine equal to that infused (~10 mg). Although this means an overall (24-h) leucine intake that is the same as that during the adaptation period, the diet supplied less leucine than the nominal intake during the feeding period because of the tracer infused in the postabsorptive period. In fact, at the lowest dietary level of 14 mg/(kg · d), this reduced the actual intake during the feeding period (10 h) to ~8 mg (food + tracer), the other 6 mg (tracer) supplied during the postabsorptive 14-h period. The likely effect of this would be to reduce the capacity for postprandial net protein synthesis through leucine limitation and to increase oxidation and negative balance in the postabsorptive state, with an overall leucine balance that is more negative than otherwise. El-Khoury et al. (1995) did show a gradual increase in leucine oxidation during the fasting period, as the excess was oxidized, and a fall during feeding because of the high efficiency of utilization on the very low leucine diet. Clearly, attempting to study the limits of adaptation of leucine oxidation and balance with a nutritionally significant amount of the tracer amino acid is highly unsatisfactory. Fuller and Garlick (1994) concluded that this problem is likely to give rise to an appreciable overestimate of the amino acid requirement.

Problem 10 represents a "model" problem, resulting from an inadequate understanding of the kinetic model under analysis and especially the uncertainty about the true precursor isotopic abundance compared with what is measured. Currently, only leucine and the other BCA allow any solution in terms of their keto acids in plasma that derive from intracellular actual precursor pools, and even in these cases, the extent of any error remains unknown. Price et al. (1994) attempted to test the assumptions empirically by comparing rates of leucine oxidation with measured N excretion rates in the fed and fasted states in subjects fed protein intakes ranging from 0.33 to 2.2 g/(kg · d). Leucine oxidation rates varied in proportion to N excretion over the entire range of N excretion rates, although the amounts were lower than expected; the ratio of predicted to measured N excretion for the entire group was 0.79 (sd = 0.23, n = 38). In other words, leucine oxidation resulted in a 25% overestimate of balance compared with nitrogen. Point 11 relates to those studies in which balance is not measured at all but some other parameter such as a change in postprandial oxidation rate, which is assumed to directly indicate the requirement intake. This is considered further below.

In summary then, as far as ¹³C leucine oxidation studies are concerned, although not entirely free from potential problems, they do appear to allow reasonable estimates of whole-body protein balance when applied with care, particularly when intakes of leucine are not especially low. The 24-h ¹³C leucine balances currently employed by Young and colleagues involve far fewer assumptions and uncertainties compared with the original tracer studies used to validate the MIT scoring pattern, which, in my view, can be generally discounted. Nevertheless, at low leucine intakes, the tracer problem remains a serious one. Uncertainty remains concerning whether the apparent inability of 14 mg/(kg · d) to maintain balance is due to excessive postabsorptive oxidation due to the tracer.

	INDICATOR AMINO ACID OXIDATION "BREAK POINT" STUDIES

These studies are important because they represent the only stable isotope data on lysine (Duncan et al. 1996, Zello et al. 1993) apart from the MIT data. Furthermore, the indicator amino acid oxidation method is described by its authors as free from many of the problems associated with other tracer balance studies (see Zello et al. 1995). Because no dietary adaptation was used, it has the advantage that more multiple-level measurements can be made on each individual than in studies requiring periods of adaptation to each intake. The method derives from studies in growing animals fed amino acid mixtures; the level of one amino acid is reduced below the requirement for adequate postprandial net protein synthesis. At this point, oxidation of all other amino acids can be assumed to increase so that if one of them is labeled (the indicator amino acid), oxidation will be observed to increase. Zello et al. (1993) and Duncan et al. (1996) reported rates of ¹³C phenylalanine oxidation in subjects fed diets with fixed phenylalanine intakes but varying lysine intakes; the "breakpoint" at which oxidation begins to increase corresponded to lysine requirements of 37 and 45 mg/(kg · d) with estimated safe intakes of 60 and 67mg/(kg · d). Although there are design flaws in the Toronto studies, there is no simple explanation for the very high requirement they appear to indicate.

One problem is that, unlike in the growing animal in which growth rate, i.e., net protein accretion, is the actual determinant of adequacy and is directly measured by this method, in adults, adequacy can be demonstrated ultimately only by balance, i.e., maintenance of the lean body mass with time. Postprandial net protein accretion relates to overall balance in a complex way as we have shown. Thus contrary to what this group argues, at maintenance, it is only an indirect measure.

A second problem lies with the use of ¹³C phenylalanine as a tracer. The ¹³CO₂ is generated from tyrosine, and yet Zello et al. (1993) calculated phenylalanine oxidation from the isotopic enrichment of plasma phenylalanine, not tyrosine [although Duncan et al. (1996) use ¹³CO₂ production rates to identify the breakpoints]. However, although this markedly underestimates the actual oxidation rate of phenylalanine (the reported values are one tenth the expected values), the authors argue subsequently (Zello et al. 1995) that their experimental design is unlikely to result in variation in true precursor labeling over the various lysine intake levels. Thus the breakpoint should not be markedly influenced by their approach.

Another puzzle about these studies is the fact that a plateau enrichment in ¹³CO₂ was reported. The authors primed the phenylalanine pool but not the tyrosine pool, which in any case is expanded by additional dietary tyrosine. Deuterated tyrosine took >8 h to reach plateau when deuterated phenylalanine was infused (Clarke and Bier 1982) to the extent that Thompson et al. (1989) showed that only with priming with deuterated tyrosine would tyrosine reach plateau within 4 h. ¹³C tyrosine should be the same, and it is extremely surprising that a plateau in ¹³CO₂ was reported within 120 min.

However, the major design difference between the Toronto studies and all of the other studies of amino acids requirements and the likely explanation for the results is their assumption that adaptation does not influence the requirements. In fact, they fed generous levels of lysine at 60 mg/(kg · d) throughout their studies apart from the test day. This was also done when they tested the influence of different protein levels [0.8 and 1.0g(/kg · d) on the requirement (Duncan et al. 1996)], which will negate any adaptive change in lysine oxidation with variation in protein intake. Thus the simplest explanation for their results is that their dietary design is setting the requirement they are trying to measure.

It should be stressed that the estimated safe intakes of 60 and 67 mg/(kg · d) reported from these studies are by far the highest estimates of lysine requirements ever reported and, if true, have profound practical implications in both developed and developing countries. For example, the lysine intake of healthy UK vegetarians studied in the UK food intake survey (see Jackson and Margetts 1993) can be estimated at ~40 mg/(kg · d), similar to the mean value for the whole of India (see Pellett 1996). On the basis of the Toronto studies, these vegetarians and the whole of India would have very high prevalence rates of inadequacy. However, as far as the Western vegetarians are concerned, Jackson and Margetts (1993) point out that they have morbidity rates lower than the average omnivore. This makes it very difficult to accept that the Toronto studies are true estimates of minimum requirements. What is needed in my view is a study with ¹³C leucine as the indicator amino acid (much simpler) and with subjects habituated to low lysine diets such as those who were able to maintain body weights and fitness for 60 d when fed wheat-based diets of 18 mg lysine/(kg · d) (Bolourchi et al. 1968). Without such studies, in my view, it has to be concluded that no reliable kinetic data exist for a minimum lysine requirement.

Human synthesis of indispensable amino acids.  A potentially confounding factor in the discussion of obligatory metabolic demands is the extent to which de novo synthesis of indispensable amino acids is possible as a result of bacterial processes in the colonic microflora. It is quite clear that urea salvage is an extensive process in humans at all ages; in infants, it can account for half of the urea produced (Langran et al. 1991). The biological importance of urea recycling depends on the extent to which the salvaged nitrogen returns to the systemic amino nitrogen pool in the form of indispensable amino acids synthesized de novo by the colonic microflora or as ammonia. Established wisdom is that, with the absence of amino acid transporters in the colon, uptake of amino acids liberated after proteolysis of bacterial protein in the colonic lumen is minimal. In fact, studies some years ago (Rikimaru et al. 1984) suggested that ¹⁵N from urea could return to the systemic circulation as ¹⁵N-labeled indispensable amino acids including lysine. We have recently confirmed these findings, showing transfer of ¹⁵N from orally administered ¹⁵N urea to circulating lysine (Yeboah et al. 1996) and other amino acids (Gibson et al. 1997) in infants recovering from severe protein energy malnutrition. This is unlikely to reflect transamination from ammonia to lysine because no evidence of transamination has ever been found for lysine in mammalian tissues. For example Torrallardona et al. (1993a and 1993b) reported a lack of transfer of ¹⁵N from ¹⁵N ammonium chloride to systemic ¹⁵N lysine in germ-free rats, or in rats given tracer when coprophagy was prevented compared with transfer in normal rats. Thus transfer of ¹⁵N from urea to systemic lysine implies uptake from the colon of bacterially synthesized lysine. Amino acid transporters are not generally thought to be abundant in the colon, but a transporter of dipeptide that is expressed in the colon has been described (Dantzig et al. 1994). The extent of this process in the adult is unknown; our studies of the relationship between leucine oxidation and urea excretion suggest it may be less important than in the infant (Price et al. 1994). Nevertheless, the possibility of de novo synthesis of indispensable amino acids raises difficult additional problems in trying to understand the magnitude of the obligatory metabolic demands.

Functional indicators of adequacy: the anabolic drive.  A feature of the Millward and Rivers review (Millward and Rivers 1988 and 1989) was the concept that dietary protein exerted direct identifiable regulatory influences on metabolism, which could serve as functional indicators of adequacy. Work in growing animals had indicated that dietary protein intakes in excess of those associated with maximum efficiency of protein utilization were associated with hormonal responses [insulin, thyroid hormones and insulin-like growth factor-1 (IGF-1)) and consequent metabolic responses (increased rates of linear bone growth). This was defined as an "anabolic drive" of dietary protein, exerted by protein intakes in excess of minimal needs (see Millward and Rivers 1989). This concept has been further developed as part of a more general protein-stat model for growth control in which the anabolic drive of dietary protein especially on bone growth is linked to the coordinated regulation of whole-body growth, including appetite regulation, through an amino static mechanism (Millward 1995). The essential details of this mechanism are shown in Figure 8.

However, in the absence of rapid growth, as with most of the human life cycle, an anabolic drive is more difficult to identify than in growing animals and has not proved as useful as we initially thought. Indeed, our studies in adults fed a wide range of protein intakes was the opposite of what we predicted in the 1988 review, and in marked contrast to animal data. We failed to identify any marked hormonal variation in either insulin, IGF-1 or triiodothyronine (T₃) (see Pacy et al. 1994). This apparent resistance of hormonal status to the level of protein intake was consistent, however, with the resistance of protein turnover to protein intake (see Fig. 3).

Protein turnover: adaptation and accommodation.  Tracer balance studies have an advantage over nitrogen balance studies in that they allow measurement of protein turnover, which may help validate the significance of balance. Young et al. (1987) discussed the possibility that if subjects receiving low intakes, who have achieved balance and apparent adaptation, have reduced whole-body protein turnover, then they have accommodated, not adapted, with an adverse consequence for their metabolic flexibility in times of stress. In support of this, Young and Marchini (1990) reported marked reductions in whole-body protein synthesis in subjects fed low levels of specific IAA.

In fact, although the animal data for reductions in protein turnover with low protein diets are quite clear (Jepson et al. 1988, Waterlow et al. 1978), evidence in humans would suggest that whole-body protein turnover is not nutritionally sensitive. Certainly our studies (Fig. 3) show no influence of protein intake within the range of intakes likely to be consumed. No change in protein turnover was observed even in chronically undernourished Indian laborers (Soares et al. 1991). Marchini et al. (1993) reported no significant changes in protein synthesis in the postabsorptive or postprandial state when diets with the FAO, MIT or egg amino acid patterns were fed. Thus any influence of protein intake on protein turnover is probably below the detection limits of current methods.

The one exceptional body of data is the large reduction in protein synthesis observed in subjects fed very low levels of individual amino acids in the MIT studies (Young and Marchini 1990). Because such responses are observed only when the measurements are made with a tracer, which is the same as the limiting amino acid, and are not observed in response to an identical dietary protocol when turnover is measured by a nonlimiting labeled amino acid (e.g., Zello et al. 1993), a methodological problem associated with either compartmentation or the non-steady state is suggested.

Thus, although protein intake and consequent amino acid supply influence the feed-fast responses of protein synthesis and proteolysis that mediate the diurnal cycle of gains and losses, there is little evidence of average daily rates changing markedly so that protein turnover is not a useful indicator of nutritional adequacy. This is in fact disappointing because there are no other obviously quantifiable indicators of adequacy. Although the case is often made that maintenance of the immune system is a function of protein intake, no quantifiable responses have been identified.

	CONCLUSIONS

From the above it should be obvious to all that the problem of defining amino acid requirements is inherently difficult. Figure 9 attempts to summarize my understanding of what can and cannot be defined in relation to amino acid requirements, and especially the problem of assessing nutritional adequacy of proteins by scoring. The scheme is drawn for lysine, but the principle applies generally. A requirement for rapid growth at 1 mo of age can be defined on the basis of the factorial method proposed by Dewey et al. (1996) for growth and maintenance, using the revised values for maintenance shown in Table 4. A minimum obligatory metabolic demand can be defined, and this is low, possibly represented by the adjusted values in Table 4. The problem arises in assessing the nutritional value of intakes in the adaptive range between these two extremes. We have mechanisms allowing considerable adaptation of metabolic demand to match intakes in this adaptive range but because we cannot easily predict whether complete adaptation will occur, we cannot be confident of the nutritional value of the protein. This means that we cannot score the protein. Protein scoring as a means of predicting net protein utilization (NPU) in animals was adopted after it was demonstrated that it worked (more or less), i.e., score was closely correlated with experimentally determined NPU within a relatively simple animal model in which tissue growth dominated needs. Adult humans have a much more complex metabolic demand, which cannot be defined as a fixed quantity. In the absence of suitable indicators of adequacy, which allow judgements to be made about optimal requirements that are independent of balance, a scoring pattern for assessment of nutritional value in the adaptive range cannot be defined and is unlikely to be useful. Although a scoring pattern based on the adjusted FAO values in Table 4 may well represent a minimum nutritional value, without long-term studies of the extent to which diets with intakes as low as this allowed balance, health and fitness, actual adequacy cannot be judged.

Thus the key test of adequacy of either protein or amino acid intake must be the long-term response in terms of the specific function of interest. In relation to the general issue of protein needs, what would be undoubtedly of interest is a controlled trial repeating Chittenden's experiments in which he reported that reduced protein intakes improved the strength of athletes (see Milllward et al. 1994). Although N balance and stable isotope studies conducted with both protein level and quality as variables would be of great interest, the main outcome test of adequacy would be the simple one of performance. This is undoubtedly what John Rivers would have concluded.

	FOOTNOTES

	LITERATURE CITED

Abstract References

• Allison J.O.B., Anderson J. A., Seeley R. D. Some effects of methionine on the utilization of nitrogen in the adult dog. J. Nutr. 1947; 33:361-370
• Anggard E. Nitric oxide: mediator, murderer and medicine. Lancet 1994; 343:1199-1206[Medline]
• Ballevre O., Cadenhead A., Calder A. G., Rees W. D., Lobley G. E., Fuller M. F., Garlick P. J. Quantitative partition of threonine oxidation in pigs: effect of dietary threonine. Am. J. Physiol. 1990; 259:E483-E491[Abstract/Free Full Text]
• Bender, A. E. (1961) . Determination of the nutritive value of proteins by chemical analysis. In: Progress in Meeting Protein Needs of Infants and Preschool Children. Publication No. 843, pp. 407-415. National Academy of Sciences and National Reasearch Council, Washington, DC.
• Benevenga N. J., Gahl M. J., Crenshaw T. D., Fink M. D. Protein and amino acid requirements for maintenance and growth of laboratory rats. J. Nutr. 1994; 124:451-453
• Bergstrom J., Furst P., Vinnars E. Effect of a test meal with and without protein on muscle and plasma free amino acids. Clin. Sci. (Lond.) 1990; 79:331-337[Medline]
• Bolourchi S., Friedeman C. M., Mickelson O. Wheat flour as a source of protein for adult human subjects. Am. J. Clin. Nutr. 1968; 21:827-835[Abstract]
• Calloway D. H., Odell A.C.F., Margen S. Sweat and miscellaneous nitrogen losses in human balance studies. J. Nutr. 1971; 101:775-786
• Clarke J.T.R., Bier D. M. The conversion of phenylalanine to tyrosine in man. Direct measurement by continuous intravenous tracer infusions of L-[ring-²H₅]phenylalanine and L-[1-¹³C] tyrosine in the postabsorptive state. Metabolism 1982; 31:999-1005[Medline]
• Clugston, G., Dewey, K. G., Fjeld, C., Millward, D. J., Reeds, P., Scrimshaw, N. S., Tontisirin, K., Waterlow, J. C., & Young, V. R. (1996) Report of the working party on protein and amino acid requirements. Eur. J. Clin. Nutr. (suppl. 1): S193-S195.
• Collin-Vidal C., Cayol M., Obled C., Ziegler F., Bommelaer G., Beaufrere B. Leucine kinetics are different during feeding with whole protein or oligopeptides. Am. J. Physiol. 1994; 267:E907-E914[Abstract/Free Full Text]
• Costa G., Ullrich L., Kantor F., Holland J. F. Production of elemental nitrogen by certain mammals including man. Nature (Lond.) 1968; 281:546-551
• Dantzig, A. H., Hoskins, J., Tabas, L. B., Bright, S.. Shepard, R. L., Jenkins, I. L., Duckworth, D. C., Sportsman. J. R., Maekensen, D., Rosteck, P. R., Jr. & Skatrud, P. L. (1994) Association of intestinal peptide transport with a protein related to the cadherin superfamily. Science (Washington, DC) 264: 430-433.
• Dewey, K. G., Beaton, G., Fjeld, C., Lonnerdal, B. & Reeds, P. (1996) Protein requirements of infants and children. Eur. J. Clin. Nutr. 50 (suppl. 1): S119-S147.
• Duncan A. M., Ball R. O., Pencharz P. B. Lysine requirements of adult males is not affected by decreasing dietary protein. Am. J. Clin. Nutr. 1996; 64:718-725[Abstract/Free Full Text]
• Edwards C. H., Booker L. K., Rumph C. H., Wright W. G., Ganapathy M. S., Ganapathy S. Utilization of wheat by adult man: nitrogen metabolism, plasma amino acids and lipids. Am. J. Clin. Nutr 1971; 24:181-193[Abstract]
• El Khoury, A. E., Fukagawa N. K., Sanchez M., Tsay R. H., Gleason R. E., Chapman T. E., Young V. R. Validation of the tracer-balance concept with reference to leucine: 24-h intravenous tracer studies with L-[1-¹³C]leucine and [¹⁵N-¹⁵N]urea. Am. J. Clin. Nutr. 1994a; 59:1000-1011[Abstract/Free Full Text]
• El Khoury, A. E., Fukagawa N. K., Sanchez M., Tsay R. H., Gleason R. E., Chapman T. E., Young V. R. The 24-h pattern and rate of leucine oxidation, with particular reference to tracer estimates of leucine requirements in healthy adults. Am. J. Clin. Nutr. 1994b; 59:1012-1020[Abstract/Free Full Text]
• El Khoury, A. E., Sanchez M., Fukagawa N. K., Gleason R. E., Tsay R. H., Young V. R. The 24-h kinetics of leucine oxidation in healthy adults receiving a generous leucine intake via three discrete meals. Am. J. Clin. Nutr. 1995; 62:579-590[Abstract/Free Full Text]
• FAO/WHO (1973) Energy and Protein Requirements. Report of a Joint FAO/WHO Ad Hoc Expert Committee. Who Technical Report Series No. 522, WHO, Geneva, Switzerland.
• FAO/WHO (1991) Protein Quality Evaluation. Report of a Joint FAO/WHO Expert Consultation. FAO Food and Nutrition Paper 51, FAO, Geneva, Switzerland.
• FAO/WHO/UNU (1985) Energy and Protein Requirements. Report of a Joint Expert Consultation. WHO Technical Report Series No. 724, WHO, Geneva, Switzerland.
• Fereday, A., Gibson, N., Cox, M., Halliday, D., Pacy, P. J., & Millward, D. J. (1994) Postprandial utilisation of wheat protein in normal adults. Proc. Nutr. Soc. 53: 201 (abs.).
• Fereday, A., Gibson, N., Cox, M., Halliday, D., Pacy, P. J. & Millward, D. J. (1997) Postprandial protein utilisation of wheat protein from a single meal in normal adults. Proc. Nutr. Soc. 56: 80 (abs.).
• Forbes G. B. Another source of error in the metabolic balance method. Nutr. Rev. 1973; 31:297-300[Medline]
• Fuller M. F., Garlick P. J. Human amino acid requirements: can the controversy be resolved? Annu. Rev. Nutr. 1994; 14:217-241[Medline]
• Fuller M. F., McWilliam R., Wang T. C., Giles L. R. The optimum dietary amino acid pattern for growing pigs; requirements for maintenance and for tissue protein accretion. Br. J. Nutr. 1989; 62:255-267[Medline]
• Fuller M. F., Milne A., Harris C. I., Reid T. M., Keenan R. Amino acid losses in ileostomy fluid on a protein-free diet. Am. J. Clin. Nutr. 1994; 59:70-73[Abstract/Free Full Text]
• Gahl M. J., Finke M. D., Crenshaw T. D., Benevenga N. J. Use of a four-parameter logistic equation to evaluate the response of growing rats to ten levels of each indispensable amino acid. J. Nutr. 1991; 121:1720-1729
• Gibson, N. R., Ah-Sing, E., Badalloo, A., Forrester,A., Jackson A. A. & Millward, D. J. (1997) Transfer of ¹⁵N from urea to the circulating dispensable and indespensable amino acid pool in the human infant. Proc. Nutr. Soc. 56: 79 (abs.).
• Hegsted D. M. Variation in requirements of nutrients: amino acids. Fed. Proc. 1963; 22:1424-1430[Medline]
• Hegsted, D. M. (1973) The amino acid requirements of rats and human beings. In: Proteins in Human Nutrition (Porter, J.W.G. & Rolls, B. A., eds.), pp. 275-292. Academic Press, London, UK.
• Hegsted D. M. Balance studies. J. Nutr 1976; 106:307-311
• Jackson A. A., Margetts B. M. Protein intakes in the adult population of the UK. Int. J. Food Sci. Nutr. 1993; 44:95-104
• Jepson M. M., Bates P. C., Millward D. J. The role of insulin and thyroid hormones in the regulation of muscle growth and protein turnover in response to dietary protein. Br. J. Nutr 1988; 59:397-415[Medline]
• Jones E. M., Bauman C. A., Reynolds M. S. Nitrogen balances in women maintained on various levels of lysine. J. Nutr. 1956; 60:549-559
• Kurzer M. S., Calloway D. H. Nitrate and nitrogen metabolism in men. Am. J. Clin. Nutr 1981; 34:1305-1313[Abstract/Free Full Text]
• Langran M., Moran B. J., Murphy J. L., Jackson A. A. Adaptation to a diet low in protein: effect of complex carbohydrate upon urea kinetics in normal man. Clin. Sci. (Lond.) 1991; 82:191-198[Medline]
• Marchini J. S., Cortiella J., Hiramatsu T., Chapman T. E., Young V. R. Requirements for indispensable amino acids in adult humans: longer-term amino acid kinetic study with support for the adequacy of the Massachusetts Institute of Technology amino acid requirement pattern. Am. J. Clin. Nutr. 1993; 58:670-683[Abstract/Free Full Text]
• Meguid M. M., Matthews D. E., Bier D. M., Meredith C. N., Soeldner J. S., Young V. R. Leucine kinetics at graded leucine intakes in young men. Am. J. Clin. Nutr. 1986a; 43:770-780[Abstract/Free Full Text]
• Meguid M. M., Matthews D. E., Bier D. M., Meredith C. N., Young V. R. Valine kinetics at graded valine intakes in young men. Am. J. Clin. Nutr. 1986b; 43:781-786[Abstract/Free Full Text]
• Meredith C. N., Wen Z.-M., Bier D. M., Matthews D. E., Young V. R. Lysine kinetics at graded lysine levels in young men. Am. J. Clin. Nutr. 1986; 43:787-794[Abstract/Free Full Text]
• Millward D. J. Amino acid requirements in adult man. Am. J. Clin. Nutr. 1990; 51:492-493[Free Full Text]
• Millward, D. J. (1992) The metabolic basis of the amino acid requirement. In: Protein-Energy Interactions (Scrimshaw N. S. & Schurch, B., eds.), pp.31-57. I/D/E/C/G Nestle Foundation, Lausanne Switzerland.
• Millward D. J. Stable isotope tracer studies of amino acid balance and human indispensable amino acid requirements. Am. J. Clin Nutr. 1993; 57:81-86[Free Full Text]
• Millward D. J. Can we define indispensable amino acid requirements and assess protein quality in adults? J. Nutr. 1994; 124:1509-1516
• Millward D. J. A protein-stat mechanism for the regulation of growth and maintenance of the lean-body mass. Nutr. Res. Rev. 1995; 8:93-120
• Millward, D. J. (1999) The nutritional value of plant based diets in relation to human amino acid and protein requirements. Proc. Nutr. Soc. (In press).
• Millward D. J., Bowtell J. L., Pacy P., Rennie M. J. Physical activity, protein metabolism and protein requirements. Proc. Nutr. Soc. 1994; 53:223-240[Medline]
• Millward D. J., Fereday A., Gibson N., Pacy P. J. Postprandial protein metabolism. Ballieres Clin. Endocrinol. Metab. 1996; 10:533-549[Medline]
• Millward D. J, Fereday A., Gibson N. R. & Pacy P. J. (1997) Efficiency of post-prandial protein utilisation in man. In: Amino Acid and Protein Metabolism in Health and Disease: Nutritional Implications (Tessari, P., Soeters, P. B., Pittoni, G. & Tiengo, A., eds.), pp. 79-91. Smith-Gordon, London, UK.
• Millward D. J., Jackson A. A., Price G., Rivers J. P. W. Human amino acid and protein requirements: current dilemmas and uncertainties. Nutr. Res. Rev. 1989; 2:109-132
• Millward D. J., Pacy P. J. Postprandial protein utilisation and protein quality assessment in man. Clin. Sci. (Lond.) 1995; 88:597-606[Medline]
• Millward D. J., Price G. M., Pacy P.J.H., Halliday D. Maintenance protein requirements: the need for conceptual revaluation. Proc. Nutr. Soc. 1990; 49:473-487[Medline]
• Millward D. J., Rivers J.P.W. The nutritional role of indispensible amino acids and the metabolic basis for their requirements. Eur. J. Clin. Nutr. 1988; 42:367-393[Medline]
• Millward D. J., Rivers J. P. The need for indispensable amino acids: the concept of the anabolic drive. Diabetes Metab. Rev. 1989; 5:191-211[Medline]
• Millward D. J., Roberts S. R. Protein requirement of older individuals. Nutr. Res. Rev. 1996; 9:67-88
• Millward D. J., Waterlow J. C. Letter to the editor. Eur. J. Clin. Nutr. 1996; 50:832-833[Medline]
• Munro, H. N. (1964) General aspects of the regulation of protein metabolism by diet and by hormones. In: Mammalian Protein Metabolism (Munro, H. N. & Allison, J. B., eds.), vol. 1 pp. 381-481. Academic Press, New York, NY.
• Osborne T. B., Mendel L. B. The amino-acid minimum for maintenance and growth, as exemplified by further experiments with lysine and tryptophan. J. Biol. Chem. 1916; 25:1-8[Free Full Text]
• Pacy P. J., Price G. M., Halliday D., Quevedo M. R., Millward D. J. Nitrogen homeostasis in man: the diurnal responses of protein synthesis and degradation and amino acid oxidation to diets with increasing protein intakes. Clin. Sci. (Lond.) 1994; 86:103-116[Medline]
• Pelletier V., Marks L., Wagner D. A., Hoerr R. A., Young V. R. Branched-chain amino acid interactions with reference to amino acid requirements in adult men: valine metabolism at different leucine intakes. Am. J. Clin. Nutr. 1991a; 54:395-401[Abstract/Free Full Text]
• Pelletier V., Marks L., Wagner D. A., Hoerr R. A., Young V. R. Branched-chain amino acid interactions with reference to amino acid requirements in adult men: leucine metabolism at different valine and isoleucine intakes 1-13. Am. J. Clin. Nutr. 1991b; 54:402-407[Abstract/Free Full Text]
• Pellett, P. L (1996) World essential amino acid supply with special attention to South-East Asia. Food Nutr. Bull. 17: 204-234.
• Price G. M., Halliday D., Pacy P. J., Quevedo M. R., Millward D. J. Nitrogen homeostasis in man: 1. Influence of protein intake on the amplitude of diurnal cycling of body nitrogen. Clin. Sci. (Lond.) 1994; 86:91-102[Medline]
• Quevedo M. R., Price G. M., Halliday D., Pacy P. J., Millward D. J. Nitrogen homeostasis in man: 3. Diurnal changes in nitrogen excretion, leucine oxidation and whole body leucine kinetics during a reduction from a high to a moderate protein intake. Clin. Sci. (Lond.) 1994; 86:185-193[Medline]
• Rikimaru T., Fujita Y., Okuda T., Kajiwara N., Date C., Heywood P. F., Alpers M. P., Koishi H. Utilization of urea nitrogen in Papua New Guinea highlanders. J. Nutr. Sci. Vitaninol. 1984; 31:393-402
• Rose W. C. The amino acid requirements of adult man. Nutr. Abstr. Rev. 1957; 27:631-647
• Said A. K., Hegsted D. M. Response of adult rats to low dietary levels of essential amino acids. J. Nutr. 1970; 100:1363-1376
• Scrimshaw N. S. Human protein requirements: a brief update. Food Nutr. Bull. 1996; 17:185-190
• Scrimshaw N. S., Taylor Y., Young V. R. Lysine supplementation of wheat gluten at adequate and restricted energy intakes in young men. Am. J. Clin. Nutr. 1973; 26:965-972[Abstract]
• Sinclair A. J. John Rivers (1945-1989): his contribution to research on polyunsaturated fatty acids in cats. J Nutr. 1994; 124:2513S-2519S
• Soares M. J., Piers L. S., Shetty P. S., Robinson S., Jackson A. A., Waterlow J. C. Basal metabolic rate, body composition and whole-body protein turnover in Indian men with differing nutritional status. Clin. Sci. (Lond.) 1991; 81:419-425[Medline]
• Thompson G. N., Pacy P.J.H., Merritt H., Ford G. C., Read M. A., Cheng K. N., Halliday D. Rapid measurement of whole body and forearm protein turnover using a [²H₅]phenylalanine model. Am. J. Physiol. 1989; 256:E631-E639[Abstract/Free Full Text]
• Tolbert B., Watts J. H. Phenylalanine requirements of women consuming a minimal tyrosine diet and the sparing effect of tyrosine on the phenylalanine requirement. J. Nutr. 1963; 80:111-117
• Torrallardona, D., Harris, C. I., Milne, E. & Fuller, M. F. (1993a) Contribution of intestinal microflora to lysine requirements in non-ruminants. Proc. Nutr. Soc. 52: 153 (abs.).
• Torrallardona. D., Harris, C. I., Milne, F., Ronaasen, V., Coatcs, M. F. & Fuller. M. F. (1993b) Contribution of intestinal microflora to lysine requirements in rats. Proc. Nutr. Soc. 52: 368 (abs.).
• Wallace W. M. Nitrogen content of the body and its relation to retention and loss of nitrogen. Fed. Proc. 1959; 18:1125-1130[Medline]
• Wang T. C., Fuller M. F. The optimum dietary amino acid pattern for growing pigs. 1. Experiments by amino acid deletion. Br. J. Nutr. 1989; 62:77-89[Medline]
• Waterlow, J. C., Garlick, P. J. & Millward, D. J. (1978) Protein Turnover in Mammalian Tissues and the Whole Body. Elsevier/North Holland, Amsterdam, The Netherlands.
• Yeboah, N., Ah-Sing, E., Badalloo, A., Forrester, T., Jackson, A. & Millward, D. J. (1996) Transfer of ¹⁵N from urea to the circulating lysine pool in the human infant. Proc. Nutr. Soc. 55: 37 (abs.).
• Yoshida, A. (1983) Specificity of amino acids for the nutritional evaluation of proteins. In: Proceedings of the International Association of Cereal Chemists Symposium on Amino Acid Composition and Biological Value of Cereal Proteins (Lasztity, R. & Hidvegi, M., eds.), pp. 163-182. Akademiai Kiado, Budapest, Hungary.
• Yoshida, A. & Moritoki, K. (1974) Nitrogen sparing action of methionine and threonine in rats receiving a protein free diet. Nutr. Rept. Int. 9:159-168.
• Young V. R. Nutritional balance studies: indicators of human requirements or adaptive mechanisms. J. Nutr. 1986; 116:700-703
• Young V. R., Bier D. M., Pellet P. L. A theoretical basis for increasing current estimates of the amino acid requirements in adult man with experimental support. Am. J. Clin. Nutr. 1989; 50:80-92[Abstract/Free Full Text]
• Young V. R., El Khoury, A. E. Can amino acid requirements for nutritional maintenance in adult humans be approximated from the amino acid composition of body mixed proteins? Proc. Natl. Acad. Sci. U.S.A. 1995; 921:300-304
• Young, V. R., Gucalp, C., Rand, W. M., Matthews, D. E. & Bier, D. M. (1987) Leucine kinetics during three weeks at submaintenance-to-maintenance intakes of leucine in men: adaptation and accommodation. Hum. Nutr. Clin. Nutr. 41C:1-18.
• Young V. R., Marchini J. S. Mechanisms and nutritional significance of metabolic responses to altered intakes of protein and amino acids with reference to nutritional adaptation in humans Am. J. Clin. Nutr. 1990; 51:270-289
• Young V. R., Pellet P. L. Current concepts concerning indispensable amino acid needs in adults and their implications for international nutrition planning. Food Nutr. Bull. 1990; 12:289-300
• Zello G. A., Pencharz , P. B., Ball R O. Lysine requirement in young adult males. Am. J. Physiol. 1993; 264:E677-E685[Abstract/Free Full Text]
• Zello G. A., Wykes L. J., Ball R. O., Pencharz P. B. Recent advances in methods of assessing dietary amino acid requirements for adult humans. J. Nutr. 1995; 125:2907-2915
• Zhao X.-H., Wen Z.-M., Meredith C. N., Matthews D. E., Bier D. M., Young V. R. Threonine kinetics at graded threonine intakes in young men. Am. J. Clin. Nutr. 1986; 43:795-802[Abstract/Free Full Text]

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