Introduction
When compared with the metabolism of carbohydrates and lipids, amino acid metabolism in man has been only sparsely studied in relation to overall energy metabolism. The roles of protein in starvation(Reference Felig, Owen and Wahren1, Reference Goodman, Belur and Lowell2) and malnutrition(Reference Smith, Pozefsky and Chhetri3, Reference Frenk4), however, have received more attention; however, in any case, our actual knowledge of N metabolism in man is far more limited than the detailed information available on energy partition between carbohydrate (glucose) and lipids, including their regulation systems. Curiously, the study of proteins has been neglected despite being a key nutritional source of energy. Probably, the present situation of limited knowledge is a compounded consequence of the relatively large number of different molecular species, their easy interconversion, the multiple catabolic paths followed by their hydrocarbon skeletons, the methodological difficulties of tracing the fate of N, its close relationship with protein turnover, a multiplicity of functional amino acid pools coupled with an active inter-organ metabolism, and last, but not least, an excessively sketchy knowledge of their catabolic paths and regulation in man (mammals).
Irrespective of the lack of information, the manipulation of protein content of diets has been actively developed for at least half a century, mainly by using hyperproteic very-low-energy diets for the treatment of obesity(Reference Scalfi, Alfieri and Borrelli5, Reference Gougeon, Hoffer and Pencharz6), with serious problems often arising from their application(Reference Doherty, Wadden and Zuk7). There has been a considerable utilisation of low-carbohydrate diets, in which the protein component is conspicuous(Reference Johnstone, Horgan and Murison8, Reference Willi, Oexmann and Wright9), but most of the discussion of their effects has been centred on their ketogenic (i.e. lipid, absence of carbohydrate) nature(Reference Yancy, Olsen and Guyton10), the amino acids essentially being considered potential gluconeogenic substrates(Reference Sankar and Sotero de Menezes11), with little impact on protein synthesis(Reference Masanés, Fernández-López and Alemany12). More recently, the use of hyperproteic diets(Reference Uebanso, Taketani and Fukaya13) is again on the rise, but we still lack the necessary basic knowledge to interpret the results obtained, largely because of few systematic analyses and the continued stress on their ketogenic nature(Reference Johnstone, Horgan and Murison8, Reference Willi, Oexmann and Wright9).
However, under conditions of abundant food supply, the main question is not how amino acids fare under conditions of low energy availability, but the contrary: how the metabolic machinery can override the strong protective mechanisms preventing N wasting under conditions of excess energy (i.e. lipid, carbohydrate and protein) intake. There are few studies on how amino acids are used as substrates, especially on the fate of N under conditions of excess energy intake or obesity. In the present review, the main questions posed by the combination of excess energy and amino-N availability are analysed under the light of the scarce information available.
Amino acids as energy substrates: amino-nitrogen sparing
One of the most significant differences between what our ancestors ate (i.e. the diet for which our digestive system and metabolic energy partitioning are geared and optimised) and the present-day diet is, in addition to the overwhelming abundance of lipid, the constant presence of protein, with a relatively high proportion of high-quality protein. The tandem lipid–animal protein is substituting progressively complex carbohydrate–low-quality plant protein as the main dietary staple. The proportion of protein energy v. total energy intake is not too much different today from the ancestral diet(Reference Frassetto, Schloetter and Mietus-Synder14), but the total amount of energy (after correcting for exercise) is higher, as is the proportion of essential amino acids(Reference Eaton and Eaton15), whilst the relationship between dietary amino acids v. glucose derived from dietary sources tends to change as a direct consequence of the substitution of starches by fats(Reference Brooks and Lampi16).
Our starvation resistance-prone mechanisms of adaptation preserve the use of amino acids as energy substrate when there is sufficient energy in the gut(Reference Oke and Loerch17). In addition, both amino acids and glucose are spared when (if) the availability of lipid is high(Reference Hart, Wolf and Zhang18, Reference Kawaguchi, Osatomi and Yamashita19). In consequence, a diet rich in energy and lipids, with a sizeable proportion of easily digestible protein, rich in essential amino acids, will necessarily result in difficulties to process and oxidise the amino acid surplus, since we are metabolically conditioned to actively preserve them. However, under excess-energy diets (including protein), it is no longer necessary to retain so much amino-N and essential amino acids. Amino acids are used for energy when in relative excess(Reference Stoll, Henry and Reeds20, Reference Wolfram and Scharrer21), but the even higher availability of energy from other sources strongly limits our metabolic machinery to do so(Reference Esteve, Rafecas and Fernández-López22).
Amino acid catabolism tends to retain 2-amino-N, largely because most amino acid hydrocarbon skeletons are oxidised after transamination (typically to 2-oxoglutarate/glutamate). However, a few amino acids yield directly ammonia (Table 1). A quantitative analysis of several common food proteins shows that, as expected, the theoretical direct yield of 2-amino-N is much higher than that of direct ammonium production (Table 2). Thus, the parity needed to synthesise our main N excretion product, urea, requires (in a dietary equilibrium) the additional conversion of a varying proportion of dietary 2-amino-N to ammonium to reach the required (1:1) balance. The ratio amino-N:ammonium-N in most dietary proteins is close to 2 (Table 2), which leaves a wide margin for preservation of N under conditions of starvation, but in the end requires the mineralisation of about half of all amino-N to ammonia under normal feeding conditions.
THF, tetrahydrofolate; [trans to], transamination to; UC, urea cycle; AcCoA, acetyl-coenzyme A; [KC], Krebs or tricarboxylic acid cycle; pyr, pyruvate; [1-C], one-carbon pathways (essentially via THF); succ, succinate; OAA, oxaloacetate; [dehydrogen to], dehydrogenation to; [oxid to], oxidation to; PNC, purine nucleotide cycle; [BO], β-oxidation pathway; AcAc, acetoacetate; AcAcCoA, acetoacetyl-coenzyme A.
* All values are estimations based on the scant data available from human subjects and other mammals; flow of amino acid catabolism via a given path depends largely on the size of amino-N pool, energy availability, metabolic needs and the relative abundance of the amino acid (essential amino acids).
* The data have been calculated from standard protein amino acid composition tables(131) and the yield in free ammonium and transaminable 2-amino groups resulting from the complete catabolism of these amino acids (Table 1). Since urea excretion requires equal proportions of amino and ammonium, any -NH2:NH3 ratio above 1·00 represents a relative excess of amino groups, which may be further converted to ammonium via the purine nucleotide cycle or (in certain tissues and physiological conditions) by glutamate dehydrogenase. The ammonium yield of the proteins listed may be underestimated, since most of the data gathered give a combined Glu + Gln (Glx) value in the overall analyses. From the analysis of whole rat protein(Reference Rafecas, Esteve and Fernández-López132) in which amide-N was analysed separately, we assumed, conservatively, and for the sake of these calculations only, that half the Glx values corresponded to Gln and half to Glu; this correction has been included in the calculations and is reflected in the data shown in the Table.
When active (exercise), muscle uses most of the body energy available: its standard feed is glucose, but blood lipids (fatty acids) limit glucose uptake and favour fatty acid oxidation (insulin resistance)(Reference Cahová, Vavrínková and Kazdová23). Some amino acids are oxidised in muscle (especially branched-chain(Reference Shinnick and Harper24)) in the postprandial state to save glucose, but excess energy hampers the process and its timing, since in fact there is no real scarcity of glucose. Amino-N conversion into ammonia is largely done in the liver and muscle through the purine nucleotide cycle(Reference Goodman and Lowenstein25), and its operation is both linked to active glycolysis(Reference Tornheim and Lowenstein26) and increased AMP levels (i.e. low ATP availability, partly compensated by the action of adenylate kinase)(Reference Atkinson and Walton27); thus, under excess energy availability, the cycle is largely idle. Glutamate dehydrogenases play a key role in ammonium metabolism in micro-organisms(Reference Núñez de Castro, Arias-Saavedra and Machado28), and in the muscle of invertebrates(Reference Batrel and Regnault29). However, Their role in ammoniagenesis in mammalian muscle is limited, because of the predominance of the purine nucleotide cycle(Reference Lowenstein30) in this role, and the low presence of the enzyme compared with the liver(Reference Arola, Palou and Remesar31), unaltered under starvation(Reference Remesar, Arola and Palou32). In the liver (and kidney), the activity of glutamate dehydrogenase is considerable(Reference Arola, Palou and Remesar31), but its function is clearly that of resynthesising glutamate from 2-ketoglutaratre and excess ammonia, as determined by direct studies and the analysis of its kinetic constrictions(Reference McGivan and Chappell33–Reference Bailey, Bell and Bell35).
Equilibrium between amino-nitrogen and ammonia for urea synthesis
Thus, in muscle there is no other major way to produce ammonia than the purine nucleotide cycle(Reference Lowenstein30). In consequence the muscle cannot use amino acids as an energy source in significant amounts(Reference Flanagan, Holmes and Sabina36) and to use their N to produce and release glutamine. This is an important question, since glutamine is the main form of blood transport of ammonia, towards the splanchnic bed(Reference Marliss, Aoki and Pozefsky37), i.e. intestine and kidney; there, glutaminases free the ammonia again for its ultimate disposal as urea(Reference Wu38), or as urinary ammonium ion(Reference Good and Burg39). Lower muscle synthesis of glutamine results, then, in lower splanchnic synthesis of carbamoyl-P, insufficient to maintain an adequate flow of urea synthesis (Fig. 1). Alternative sources of ammonium, such as the microbiota(Reference Karasawa and Nakata40, Reference Wrong and Vince41), which is in part derived from glutamine(Reference Anderson, Bennett and Alleyne42, Reference Weber, Veach and Friedman43), and the direct ammoniagenic amino acids cited above (serine, threonine, glycine), help maintain a steady albeit diminished rate of urea synthesis in the intestine–liver system, a rate insufficient to cope with the excess 2-amino-N pool generated by the diet and limited amino acid disposal.
High amino acids in conjunction with high energy availability can generate a paradoxical scarcity of ammonia, retaining a large and unshrinkable pool of amino-N because the mechanisms that protect its conversion to ammonia remain unaffected, and are both efficient and effective (and potentially crippling). In the metabolic syndrome (and in general, in energy-rich feeding), urea synthesis is decreased(Reference Barber, Viña and Viña44), but there is not – either – a massive accumulation of body-N(Reference Esteve, Rafecas and Remesar45). Amino acids tend to be preserved in spite of excess energy availability(Reference Serra, Gianotti and Palou46), but in any case, the excess N is eventually lost, albeit not in the canonical way of urea formation(Reference Esteve, Rafecas and Remesar47). A small but significant proportion of N is excreted as N2 gas(Reference Costa, Ullrich and Kantor48, Reference Cissik, Johson and Rokosch49) by means of, so far, unknown pathways. In addition there is an increased (but relatively small) loss of dietary amino-N in the form of urinary nitrate and nitrite(Reference Green, Ruiz de Luzuriaga and Wagner50). Obligatory N losses also include urinary losses of uric acid (from purine catabolism(Reference Sutton, Toews and Ward51)), creatinine and small proportions of peptides and amino acids, as well as the ammonium ion, excreted by the kidney (especially in acidosis)(Reference Desir, Bratusch-Marrain and DeFronzo52). Small amounts of ammonium may be also excreted by the lungs(Reference Robin, Travis and Bromberg53).
Muscle also accumulates fat, near mitochondrial clusters (C Cabot, K Pouillot, S Roy, MM Romero, R Vilà, MM Grasa, M Esteve, JA Fernández-López, M Alemany and X Remesar, unpublished results) and adapts itself to the utilisation of this main substrate (as well as to glucose, but to a lower extent)(Reference Turner, Bruce and Beale54, Reference Lam, Hatznikolas and Weir55). Exercise facilitates the massive utilisation of energy and streamlines the oxidation of fats(Reference Bruce, Thrush and Mertz56), but also restores in part the production of ammonium via the purine nucleotide cycle(Reference Graham, Bangsbo and Saltin57), thus increasing the flow of glutamine to the splanchnic bed. However, a large proportion of glucose, lipids and amino acids can not be taken up by any of the above cited systems, leaving them unused and in high serum concentrations, waiting for their storage as fat in the last-recourse energy pool: white adipose tissue.
The nitric oxide pathway
NO∙ is synthesised from arginine by NO∙ synthases, yielding citrulline(Reference Moncada, Palmer and Higgs58). Excess N availability increases the synthesis of ornithine(Reference Mouillé, Robert and Blachier59), including the intermediate step of acetyl-glutamate synthesis(Reference Tujioka, Lyou and Hirano60), which is also a key regulator of carbamoyl-P synthase 2, and thus also participates in the regulation of ammonium disposal(Reference Saheki, Ohkubo and Katsunuma61). In consequence, higher amino-N levels may increase those of arginine, irrespective of low carbamoyl-P (i.e. low ammonium) availability, shunting the NO∙ cycle from arginine to citrulline and leaving out ornithine (and the synthesis of urea) (Fig. 2 (a) and (b)) (Reference Moncada and Higgs62). In cells that do not have a fully operative urea cycle, the eventual regulation is even easier since it is largely dependent on arginine availability(Reference McCall, Boughton-Smith and Palmer63).
It may be expected, then, that under high energy and amino-N availability, the low ammonia concentrations(Reference Herrero, Angles and Remesar64) can not sustain an effective excretion of N through the urea cycle(Reference Roig, Esteve and Remesar65), indirectly favouring an increased activity of the NO∙ synthesis shunt. The excess NO∙ in blood vessels (derived from the activity of erythrocyte and endothelial NO∙ synthases)(Reference Kleinbongard, Schultz and Rassaf66) may initially raise the blood flow, at least locally, increasing the availability of oxygen and substrates to the surrounding cells(Reference Sureda, Tauler and Aguiló67).
NO∙ is highly reactive and interacts with specific proteins, such as guanylate cyclase(Reference Arnold, Mittal and Katsuki68), increasing the production of cyclic GMP which activates protein kinase G (PKG)(Reference Francis, Busch and Corbin69) which, in turn, relaxes the smooth muscle of small vessels and thus increases blood flow and lowers arterial tension(Reference Lincoln, Komalavilas and Cornwell70). This is the main recognised function of NO∙(Reference Sessa71), but NO∙ is also able to bind cysteine residues of other proteins, such as protein kinase A (PKA)(Reference Ferro, Coash and Yamamoto72), which may induce a phantom adrenergic stimulation (i.e. without the intervention of catecholamines or cAMP)(Reference Burgoyne and Eaton73).
Cytochrome c also efficiently oxidises NO∙ to nitrite(Reference Torres, Sharpe and Rosquist74). Most of the NO∙, however, rapidly reacts with oxyhaemoglobin, eventually oxidising NO∙ to nitrate(Reference Gow, Luchsinger and Pawloski75). Other highly reactive NOx compounds, such as peroxynitrite(Reference Huie and Padmaja76), are formed by further oxidation with reactive oxygen species. Part of these nitrogen oxides react with proteins, fatty acids and other compounds yielding nitro-derivatives(Reference Trostchansky and Rubbo77, Reference Rubbo and Radi78), often short-lived, but which can cause permanent structural changes(Reference Jain, Siddam and Marathi79).
Nitrite and other forms of nitrogen excretion
In the obese, the overall production and levels of NO∙ are increased(Reference Asl, Ghasemi and Azizi80), as is its loss in the air breathed(Reference Maniscalco, de Laurentiis and Zedda81), but there is a marked decrease in the excretion of urea(Reference Barber, Viña and Viña44). A significant part of the difference in the N balance is made up of N2 gas(Reference Esteve, Rafecas and Remesar45, Reference Esteve, Rafecas and Remesar47, Reference Cissik, Johnson and Hertig82). A possible source is the reaction of nitrite and free amino acids, which in an acidic medium yield N2 gas and 2-hydroxyacids(Reference Schmidt83); this reaction has been described to occur in the stomach lumen(Reference Yoshida and Kasama84). However, this reaction can hardly explain the large discrepancies found in N balances. There must be another – larger – source of N2 gas integrated in the amino acid metabolism, which so far has not been discovered. We can hypothesise the existence of an ‘emergence’ pathway, shunting the action of NO∙ synthases towards the reaction of arginine with nitrite, yielding citrulline and N2 gas under acidotic conditions. This way, nitrite, the main active product of NO∙ synthesis, would be rapidly removed and, at the same time, the excess 2-amino-N would be decreased at the expense of aspartate-derived arginine guanido-N; unfortunately, no enzyme has been found (so far) able to carry out this reaction, which nevertheless is known to proceed spontaneously under low pH conditions(Reference Schmidt83).
Glucocorticoids may elicit counteractive actions(Reference Worrall, Misko and Sullivan85) to inhibit NO∙ synthesis, but catecholamines increase its production(Reference Lin, Tsai and Huang86). It is unclear whether NO∙ overproduction in the obese can be a consequence of leptin-related catecholamine vasoconstriction(Reference Hall87), a consequence of enhanced NO∙ synthesis through activation of endothelial or inducible NO∙ synthases(Reference Elizalde, Rydén and van Harmelen88), or a lower bioavailability of NO∙ favouring its increased synthesis(Reference Blouet, Mariotti and Mathe89, Reference Williams, Wheatcroft and Shah90).
Nitrite is considered a stabilised form of NO∙(Reference Tsuchiya, Kanematsu and Yoshizumi91), which can yield NO∙ under hypoxic conditions by reacting with Hb(Reference Piknova, Keszler and Hogg92–Reference Lundberg, Weitzberg and Gladwin94), thus helping increase blood flow to hypoxic areas(Reference Ingram, Pinder and Bailey95). Nitrite is also a source of NO∙ in the alimentary canal(Reference Eckmann, Jaurent and Langford96, Reference Lundberg, Weitzberg and Lundberg97); it is largely the product of reduction by the oral biota of nitrate secreted by salivary glands(Reference Chen, Ren and Lu98). Nitrite-derived NO∙ also kills bacteria in the stomach(Reference Iijima, Fyfe and McColl99); in this acidic medium, nitrite reacts with free amino acids yielding N-nitroso-proline from arginine(Reference Ishibashi and Kawabata100), as well as N2 as indicated above(Reference Yoshida and Kasama84).
The ‘obese’ microbiota(Reference Cani, Possemiers and van de Wiele101, Reference Brignardello, Morales and Diaz102) is probably a consequence of this magnified effect of NOx(Reference Dykhuizen, Frazer and Duncan103); changes in the gut microbial ecosystem and composition also influence the relationships with the host through modulation of the immune response(Reference Manco, Putignani and Bottazzo104, Reference Vaarala, Atkinson and Neu105). The relative abundance of protein debris in the intestine, a consequence of diets rich in lipid and protein, together with relatively scarce fibre and polysaccharides, also affects the composition of the microbiota, increasing the share of amino acid-related catabolism in the process of formation of stool(Reference Harmon, Becker and Jensen106, Reference Hughes, Magee and Bingham107). The resulting higher pH, and the production of amines through amino acid decarboxylation(Reference Hughes, Magee and Bingham107), higher proportions of amines, ammonium, as well as amine- and sulfide-related catabolites may also help induce the development of intestinal cancer(Reference Bingham, Pignatelli and Pollock108).
Health consequences of hampered nitrogen excretion in the obese
The main problem posed by this question is the almost absolute lack of information about the human patterns of N excretion in overnutrition, obesity and the metabolic syndrome. It has been found that a relative increase in dietary protein at the expense of carbohydrates facilitates the loss of weight(Reference Lee, Lee and Bae109, Reference Pasiakos, Mettel and West110), but we only know the short-term macroscopic changes; the dynamics of 2-amino-N under these conditions has not been studied.
We have mechanisms to adjust amino acid catabolism to their relative abundance with respect to glucose(Reference Oke and Loerch17), but the large presence of lipids in the diet alters everything. Ketogenic diets favour the excretion of ammonium in the urine to counter the acidosis produced by ketone bodies(Reference Schloeder and Stinebaugh111, Reference Simon, Martin and Buerkert112), and increase liver gluconeogenesis from amino acids(Reference Sherwin, Hendler and Felig113), but the problem of conversion of amino-N to ammonium remains. The possible negative effects of a few truly hyperproteic (i.e. not ketogenic) diets(Reference Metges and Barth114), and their limited effect on body fat point both to a generalised inefficiency of the so-called ‘high-protein’ diets(Reference Johnstone115) and support the relative danger of their uncontrolled application.
One of the most important aspects of amino acid metabolism is the synergistic complementarity of the roles of a number of peripheral organs, the liver and the rest of splanchnic bed organs(Reference Aikawa, Matsutaka and Takezawa116, Reference Felig117). Hyperproteic diets may induce changes in their roles in the absence of energy overload, i.e. under conditions of active amino acid catabolism(Reference Morens, Gaudichon and Fromentin118). However, it is highly improbable that the finely adjusted inter-organ relationships could be maintained under the pressure of high-energy diets, as the low urea output seems to indicate; as a consequence the whole body is affected by an excess of 2-amino-N.
There are few data on human subjects supporting an increase in the synthesis of NO∙ in high-energy availability conditions, except for an increased breath release of NO∙(Reference Maniscalco, de Laurentiis and Zedda81) and the consequent formation of nitrite and nitrate(Reference Kim-Shapiro, Schechter and Gladwin93, Reference Lundberg, Weitzberg and Gladwin94). Perhaps the high levels of nitrite and the easy interconversion of nitrite and NO∙(Reference Iijima, Fyfe and McColl99, Reference Gladwin, Crawford and Patel119), a powerful vasodilator(Reference Rees, Palmer and Moncada120), may be related to the ‘obesity paradox’, i.e. a decreased severity of the consequences of heart failure in the obese(Reference Gruberg, Weissman and Waksman121, Reference Badheka, Rathod and Kizilbash122). The large presence of NOx in the alimentary canal and its profound influence on the microbiota has to produce, necessarily, changes in their properties and functions: at least a different way to cope with unused substrates and different relationships with the immune system-controlled intestinal barrier. The latter may be related to the higher levels of circulating lipopolysaccharide observed in the metabolic syndrome(Reference Brun, Castagliuolo and di Leo123), also linked to the maintenance of low-key inflammation(Reference Cancello and Clément124, Reference Matsuo, Hashizume and Shioji125) caused by increased intestinal bacteria activity(Reference Sabate, Jouet and Harnois126). These findings hint to the postulated excess 2-amino-N, in agreement with the higher availability of amino acids and energy to increase protein turnover(Reference Robinson, Jaccard and Persaud127, Reference Yuile, Lucas and Olson128) and to maintain a fully functional immune system(Reference Dunca and Schmidt129) observed in the metabolic syndrome.
Conclusions
Humans are fairly well prepared for amino-N scarcity: dietary protein utilisation is maximised, and amino acid catabolism is restricted in order to preserve body protein, and, with that, to maintain the ability to function and survive. However, the same mechanisms that make possible sparing amino acid catabolism for energy seriously hamper the metabolism of excess dietary amino-N under conditions of overfeeding and excess available energy. Insulin resistance limits the use of glucose when fats are readily available, and ample glucose (energy) availability practically blocks the removal of amino-N to form ammonia, the only, and narrow, canonical way to dispose of excess N. The obese excrete less urea than the lean, high-energy diets inhibit the urea cycle function, but also alter the glucose–alanine cycle and the operation of the purine nucleotide cycle; the path of conversion of amino-N to ammonium is severely restricted. This creates a surplus amino acid availability which enhances growth and protein synthesis, but protein turnover simply stores, and transamination changes, the hydrocarbon skeletons, preserving the amino groups. Consequently, non-conventional mechanisms are necessarily activated (there is no body storage of this surplus N). We do not know how this is accomplished, and only can suggest the possible implication of NO∙-increased synthesis, followed by higher nitrite (and nitrate) secretion/excretion, and including the production of N2 gas, through a mechanism so far unsolved.
The metabolic consequences of the imbalance between amino- and ammonia-N are far-reaching and should be studied in detail, since probably a number of unexplained phenomena of the metabolic syndrome sink their roots in the profound alteration of N homeostasis. The consequences of excess dietary protein and our inability to dispose of it have not been studied, but the indications obtained from animal studies and the few data available suggest that excess protein is harmful in the long term for humans.
Acknowledgements
The present review was supported by grant no. SAF2009-11739 of the Plan Nacional de Investigación en Biomedicina of the Government of Spain. There are no conflicting interests to disclose.