Amino Acids

Amino acids are the building blocks of proteins. As the name says, they are compounds that carry an amino group and an acid group. In the twenty amino acids of proteins the acid group is a carboxyl group (–COOH), and the amino group and the carboxyl group are attached to the same carbon atom, called the α-carbon, which is why they are known as α-amino acids [1]. The α-carbon also carries a hydrogen atom and a side chain, and it is the side chain that makes one amino acid different from another.

A cell builds its proteins from twenty amino acids, linked into chains in the order written in its genes. The genetic code reads that order in groups of three letters, and most of its 64 triplets stand for one of the twenty [2]. Two more amino acids are inserted by the genetic code, but only in some organisms: selenocysteine, at a triplet that normally means “stop” [3], and pyrrolysine, found in some archaea and bacteria [4].

Chemistry allows many more. In a computer study, researchers assembled a library of 1,913 possible α-amino acids and drew 100 million random sets of twenty from it. When the sets were compared by how widely and evenly they covered three calculated properties of the side chains (size, charge and hydrophobicity), only six were as good as the natural set or better [5]. In 1968 Francis Crick suggested that the number twenty and the choice of amino acids were “at least in part” a historical accident [2]; more recent work argues that the set is close to the best one possible [6] [5]. This page describes the twenty amino acids – their structures, names, properties and roles in the body – and comes back to the question of why it is these twenty.

The 20 amino acids

The chart groups the twenty by the character of their side chains. Each cell gives the three-letter and one-letter symbols, the name and the molar mass, and leads to the page about that amino acid.

Molar masses in g/mol, from the CRC Handbook of Chemistry and Physics [7]. Grouping by side chain follows a common textbook convention; other books place glycine, cysteine, tyrosine and tryptophan in different groups.

Structure and the zwitterionic form

Tables, and the chart above, draw amino acids in an uncharged form, H2N–CH(R)–COOH, where R is the side chain. This is a convention. In water near pH 7 the carboxylic acid group has given up a proton and the amino group has taken one up, so the molecule carries a positive and a negative charge at the same time: R–CH(NH3+)–COO−. Such a molecule is called a zwitterion, a dipolar ion or an amphion; the uncharged drawing is kept because it is simpler to read [1].

L-alanine drawn in the uncharged form H2N-CH(CH3)-COOH and as the zwitterion H3N+-CH(CH3)-COO-
L-Alanine, (S)-2-aminopropanoic acid. Left: the conventional uncharged form used in tables. Right: the zwitterion, the form that predominates in water near pH 7. A solid wedge marks a bond pointing towards the viewer.

The acidity constants show why. In glycine, the simplest amino acid, the pKa of the carboxyl group is 2.34 and that of the amino group 9.58 [7]. Between these two values most molecules are zwitterions. Below pH 2.34 most of them carry an extra proton and are positive ions (glycinium); above pH 9.58 most have lost a proton and are negative ions (glycinate) [1]. In between lies the isoelectric point (pI), the pH at which the molecules carry no net charge on average: 5.97 for glycine [7].

Seven of the twenty – aspartic acid, glutamic acid, cysteine, tyrosine, histidine, lysine and arginine – have a third group in the side chain that can gain or lose a proton. Their isoelectric points therefore spread from 2.77 for aspartic acid to 10.76 for arginine [7]. These values belong to free amino acids in water. Inside a folded protein the same groups are affected by neighbouring charges, hydrogen bonds and burial away from water, and their pKa values can shift a long way [8].

L- and D-amino acids

Nineteen of the twenty amino acids are chiral. Their α-carbon carries four different groups, so each of them exists in two forms that are mirror images of each other and cannot be superimposed, like a left and a right hand. Glycine is the exception: its side chain is a second hydrogen atom [1]. Isoleucine and threonine have a second stereocentre in the side chain [6].

The two forms are named L and D. The letters describe the arrangement of groups around the α-carbon by formal comparison with L- or D-serine and, through serine, with glyceraldehyde [1]. Proteins are built from L-amino acids, so for an amino acid from a protein the prefix is usually left out: “alanine” means L-alanine [1]. Ribosomes, the machines that build proteins, use L-amino acids together with achiral glycine [1]. Doig argues that the helices and sheets of a folded protein form properly only if its amino acids are all of the same handedness, and that the choice between L and D itself “appears to be arbitrary” [6].

L-alanine and D-alanine as mirror images, and L-cysteine drawn with the same arrangement as L-alanine
L-Alanine and D-alanine are mirror images. L-Cysteine has the same spatial arrangement as L-alanine, yet by the CIP rules it is (R), because the CH2SH group outranks the carboxyl group. A solid wedge marks a bond pointing towards the viewer.

Chemists also use a second system, (R) and (S), which ranks the groups around a stereocentre by the atomic numbers of their atoms (the Cahn–Ingold–Prelog, or CIP, rules). For the protein amino acids L nearly always corresponds to (S). The important exception is cysteine: the sulfur atom of its side chain outranks the oxygen atoms of the carboxyl group, so L-cysteine is (R) [1]. Threonine, with two stereocentres, is (2S,3R) [1]. Structures on this site are labelled in both systems.

Abbreviations

Each amino acid has a three-letter symbol and a one-letter symbol [1]. The three-letter symbol is usually the first three letters of the name and is written with one capital letter: Gln, not GLN or gln. A symbol without a prefix means the L-form [1].

The one-letter code was introduced for long sequences: to compare related proteins, to print them in tables and to handle them by computer. Gamow and Yčas mentioned the idea in 1958, Šorm and colleagues systematised it in 1961, and Dayhoff and Eck used one-letter symbols in their compilations of protein sequences. The international recommendations were approved in 1968 [1]. Because the code is harder to follow for readers who do not know it, the recommendations advise against it in ordinary text [1].

How the letter was chosenLetters
First letter of the name, where no other amino acid starts with itC cysteine · H histidine · I isoleucine · M methionine · S serine · V valine
First letter shared by several names: it went to the most common and structurally simplest oneA alanine · G glycine · L leucine · P proline · T threonine
Sound of the nameF phenylalanine · R arginine
Shape: the double ring of the molecule and the “bulky” letterW tryptophan
Assigned; the recommendations give no reasonD aspartic acid · E glutamic acid · N asparagine · Q glutamine
Of the letters left, the closest in the alphabet to the initial letterK lysine · Y tyrosine
For partly determined sequencesB aspartic acid or asparagine (Asx) · Z glutamic acid or glutamine (Glx) · X unknown or other (Xaa)

After the IUPAC–IUB recommendations [1].

Some letters were left out on purpose. U is easily mistaken for V in handwriting, O for G, Q, C, D or zero in poor printouts, and J is missing from several alphabets [1]. The free letter U was later proposed for selenocysteine [3].

Classification of amino acids

There is no single classification of amino acids; each answers a different question. Chemistry sorts them by their side chains, nutrition by whether the body can make them, and metabolism by what their carbon skeletons become. This section deals with the side chains; the other two classifications come later on this page.

By side chain

The chart at the top of the page uses a grouping found in many textbooks. Nine side chains are made mostly of carbon and hydrogen and are counted as nonpolar. Alanine, valine, leucine and isoleucine carry plain hydrocarbon chains; phenylalanine and tryptophan carry aromatic rings; methionine has a sulfur atom inside its chain; and in proline the side chain bends back and joins the amino group, closing a five-membered ring. Glycine, whose side chain is a single hydrogen atom, is usually placed in the same group. Six side chains are polar but uncharged at pH 7: the hydroxyl groups of serine, threonine and tyrosine, the thiol group of cysteine and the amide groups of asparagine and glutamine. Two side chains end in a carboxyl group and are negatively charged at pH 7 (aspartic and glutamic acids), and three carry a nitrogen group that can take up a proton (lysine, arginine and histidine) [1].

The borders are not sharp. Histidine is counted among the basic amino acids, yet the pKa of its side chain, 6.04, lies so close to neutral that at pH 7 only about one histidine side chain in ten carries a proton (calculated from the pKa [7]). The side chains of cysteine and tyrosine, listed as uncharged, lose a proton at higher pH, with pKa values of 8.14 and 10.10 [7].

By hydropathy

Side chains can also be placed on a single scale, from those that prefer water to those that avoid it. The most widely used scale was published by Kyte and Doolittle in 1982. They combined two kinds of data: how readily model compounds leave water for the vapour phase, and how often each amino acid is found buried inside the structures of proteins. For the property they coined the word hydropathy, “strong feeling about water”; their index runs from +4.5 for isoleucine to −4.5 for arginine [9].

Bar chart of the Kyte-Doolittle hydropathy index of the 20 amino acids, from isoleucine +4.5 to arginine -4.5, coloured by side-chain group
Hydropathy index of the 20 amino acids. Positive values mean water-repelling, negative values water-loving. Data from Kyte and Doolittle, Table 2 [9]; the value for cysteine is given there for cysteine and cystine together. Colours show the groups of the chart above: cysteine, a polar amino acid in the chart, is among the most water-repelling, while proline and tryptophan, nonpolar in the chart, come out slightly water-loving.

The authors were open about how much judgement went into the numbers. Several values were adjusted as “the result of personal bias and heated discussion between the authors”; arginine was “arbitrarily assigned to the lowest point”; and glycine, which “does not have strong feelings about water”, was given a value near the middle [9]. They also showed that this hardly mattered. When the twenty were reduced to three clusters – water-loving, intermediate and water-repelling, each with one average value – the hydropathy profiles of proteins came out almost the same [9].

Properties: pKa, pKb and isoelectric points

The table gives the acid dissociation constants and isoelectric points of the twenty at 25 °C, as compiled in the CRC Handbook of Chemistry and Physics [7]. The handbook labels the constants pKa, pKb and pKc; other tables call them pK1, pK2 and pKR. In both cases all three are acid constants: the “b” here means the second constant, not a base constant [7]. The last column shows how much of each amino acid dissolves in a kilogram of water. The range is wide, from more than 1.6 kg for proline to half a gram for tyrosine.

Amino acidpKa
α-COOH
pKb
α-NH3+
pKc
side chain
pISolubility,
g/kg water
Glycine Gly2.349.58–5.97239
Alanine Ala2.339.71–6.00166.9
Valine Val2.279.52–5.9688
Leucine Leu2.329.58–5.9823.8
Isoleucine Ile2.269.60–6.0234.2
Proline Pro1.9510.47–6.301625
Phenylalanine Phe2.189.09–5.4827.9
Tryptophan Trp2.389.34–5.8913.2
Methionine Met2.169.08–5.7456
Serine Ser2.139.05–5.68250
Threonine Thr2.208.96–5.6090.6
Cysteine Cys1.9110.288.145.07very soluble
Asparagine Asn2.168.73–5.4125.1
Glutamine Gln2.189.00–5.6542
Tyrosine Tyr2.249.0410.105.660.51
Aspartic acid Asp1.959.663.712.775.04
Glutamic acid Glu2.169.584.153.228.6
Lysine Lys2.159.1610.679.745.8
Arginine Arg2.039.0012.1010.76182.6
Histidine His1.709.096.047.5943.5

Values at 25 °C, including solubility, from the CRC Handbook [7], which compiles them from several reference works. For cysteine the handbook gives 10.28 for the amino group and 8.14 for the thiol group.

Isoelectric points

The isoelectric point (pI) is the pH at which the molecules of an amino acid carry no net charge on average. For an amino acid without an ionisable side chain it lies halfway between the two pKa values: for glycine (2.34 + 9.58) / 2 = 5.96, against 5.97 in the table. For acidic amino acids the pI is the average of the two lowest constants, for basic ones the average of the two highest. Calculated in this way, the values come close to those in the table but do not always match them: 2.83 instead of 2.77 for aspartic acid, 10.55 instead of 10.76 for arginine. The handbook draws its pKa and pI values from more than one source [7], so the columns need not agree to the last digit.

Free amino acids and amino acids in a chain

In a protein the α-carboxyl and α-amino groups are used up in peptide bonds, except at the two ends of the chain, and the groups that remain behave differently. In alanine pentapeptides used as models of protein chains, the terminal carboxyl group has a pKa of 3.67 and the terminal amino group 8.00 [8], compared with 2.34 and 9.58 in free glycine [7]. Side chains change less: 6.54 for histidine and 10.40 for lysine in the peptides [8], against 6.04 and 10.67 in the free amino acids [7]. In a folded protein, buried groups and neighbouring charges can shift the values much further [8]. The same paper is a reminder of how uncertain such numbers are: for the side chain of arginine, published values range from 11.6 to 12.6 [8].

How amino acids were discovered

The first amino acid was found by accident. In 1806 the chemists Vauquelin and Robiquet left the juice of asparagus shoots standing for a few days and found crystals in it; they had isolated asparagine [10]. Modern protein chemistry, however, is dated from 1820, when Braconnot boiled gelatin with sulfuric acid to see whether protein, like starch, would yield sugar. A month later sweet crystals had formed. Braconnot called them “sucre de gélatine” and did not discover that they contained nitrogen, but they were glycine, the first amino acid obtained from a protein by acid hydrolysis [11].

Over the next hundred years the protein amino acids were found one after another, mostly by chemists who broke proteins down with acid and tried to separate the mixture. Leucine, glycine, alanine, tyrosine, phenylalanine, glutamic acid and serine were, in the words of the historians of the subject, Vickery and Schmidt, “more or less chance products of fractional crystallization”. The basic amino acids came to light once reagents were found that precipitate them. Isoleucine was discovered because Ehrlich wanted to explain why samples of “leucine” differed by a few degrees in optical rotation; tryptophan was tracked down by following a colour reaction; and methionine turned up in the search for a growth factor for bacteria, when the active fraction was found to contain sulfur [11].

Not every claim survived. Vickery and Schmidt accepted a substance as a protein amino acid only if a second investigator had isolated it too and its structure had been proved by synthesis. Even Fischer described a “diaminotrioxydodecanoic acid” in 1904 and withdrew it in 1917 [11].

Amino acidFirst describedFirst shown in a protein
Glycine Gly1820, Braconnot1820, Braconnot
Leucine Leu1819, Proust1820, Braconnot
Tyrosine Tyr1846, Liebig1849, Bopp
Serine Ser1865, Cramer1865, Cramer
Glutamic acid Glu1866, Ritthausen1866, Ritthausen
Aspartic acid Asp1827, Plisson1868, Ritthausen
Phenylalanine Phe1879, Schulze1881, Schulze and Barbieri
Alanine Ala1850, Strecker (synthesis)1888, Weyl
Lysine Lys1889, Drechsel1889, Drechsel
Arginine Arg1886, Schulze1895, Hedin
Histidine His1896, Kossel; Hedin1896, Kossel; Hedin
Cysteine Cys1810, Wollaston (as cystine)1899, Mörner (as cystine); 1935, Mirsky and Anson (as cysteine)
Valine Val1856, von Gorup-Besanez1901, Fischer
Proline Pro1900, Willstätter (synthesis)1901, Fischer
Tryptophan Trp1901, Hopkins and Cole1901, Hopkins and Cole
Isoleucine Ile1903–1904, Ehrlich1903–1904, Ehrlich
Methionine Met1922, Mueller1922, Mueller
Asparagine Asn1806, Vauquelin and Robiquet1932, Damodaran
Glutamine Gln1883, Schulze and Bosshard1932, Damodaran and colleagues
Threonine Thr1935, McCoy, Meyer and Rose1936, Meyer and Rose

In the order in which each was shown to be a component of proteins, after Vickery [10]; first descriptions after Vickery and Schmidt [11] and Vickery [10]. For isoleucine the two reviews give 1903 and 1904, the years of Ehrlich’s first and second papers.

The list of 1931 looks strange today. It contained 21 substances, among them hydroxyproline, thyroxine and an iodinated tyrosine, but neither asparagine, glutamine nor threonine [11]. By 1972 Vickery counted “the classical twenty” and moved hydroxyproline, thyroxine and more than a dozen others into a separate group: amino acids that are made from the twenty by enzymes after the protein chain has been built [10].

Asparagine and glutamine were missing for a simple reason. Boiling with acid, the standard way of breaking down a protein, turns both amides into aspartic and glutamic acids and ammonia. As early as 1873 chemists suspected that the ammonia in protein hydrolysates came from such amides, but the proof came only in 1932, when Damodaran, working in Chibnall’s laboratory at Imperial College in London, broke proteins down with enzymes instead of acid and isolated asparagine from the protein edestin and glutamine from gliadin of wheat [10].

Threonine came last, and it was found in a different way: not in a protein hydrolysate but in the diet of rats. From 1930 W. C. Rose at the University of Illinois fed young rats on mixtures of pure amino acids made up to match casein, the protein of milk. They did not grow unless a little whole protein was added, so something was missing. In 1935 it turned out that two things were: isoleucine and an unknown substance [10]. McCoy, Meyer and Rose isolated the unknown from a hydrolysate of 12 kilograms of fibrin and identified it as one of the four optical isomers of α-amino-β-hydroxy-n-butyric acid [10] [12]. About 0.6 per cent of it in the diet was enough. For the first time, animals grew on a diet in which purified amino acids were the only source of nitrogen [12].

The new amino acid, threonine, also caused a problem of names. Meyer and Rose named it “d(−)-threonine”, a name meant to describe both of its stereocentres; the “d” referred to the second one, which matches the sugar d(−)-threose, while the α-carbon has the same configuration as in the other protein amino acids. Other authors soon wrote “l(−)-threonine” or simply “l-threonine”, mixing up symbols for configuration and for the direction of optical rotation. The confusion that followed was one of the reasons for the general revision of amino acid nomenclature that Vickery proposed in 1947 [10]. More important for nutrition, once the full set of twenty was available in pure form, it became possible to leave out one amino acid at a time and see which ones an animal, or a person, cannot do without.

Essential and nonessential amino acids

The experiments that followed the discovery of threonine were made on people. In the 1940s and 1950s Rose and his colleagues studied healthy young men on carefully composed diets, leaving out one amino acid at a time, and followed their nitrogen balance – the difference between the nitrogen they took in and the nitrogen they lost. When any one of eight amino acids (isoleucine, leucine, lysine, methionine, phenylalanine, threonine, tryptophan or valine) was left out, the balance quickly became negative and recovered as soon as the amino acid was put back. The others could be left out without a noticeable effect [13].

Histidine was the puzzle. Men on Rose’s diets without histidine stayed in balance for weeks, probably because the body holds large reserves of it in haemoglobin and in the dipeptide carnosine. In a later study that lasted about a month, plasma histidine fell to half its normal level within a day, and after 5 to 20 days the nitrogen balance turned negative [13]. Today nine amino acids are classed as indispensable, or essential: histidine, isoleucine, leucine, lysine, methionine, phenylalanine, threonine, tryptophan and valine. Even so, the experts who set the international requirements wrote in 2007 that “the indispensability of histidine in healthy adults remains unresolved”. They kept it in the list, citing the fall in haemoglobin seen when a histidine-free diet was eaten for a long time [14].

Essential by definition

The word “essential” sounds absolute, but it rests on a definition. The one used since 1946 calls an amino acid indispensable if it “cannot be synthesized by the animal organism out of materials ordinarily available to the cells at a speed commensurate with the demands for normal growth” [15]. Every qualifier counts: what is ordinarily available, how fast, and for what kind of growth. Rose himself wrote in 1947 that classing an amino acid such as arginine or glutamic acid as dispensable or indispensable “is purely a matter of definition” [15].

Behind the definition is chemistry. Reeds pointed out that each of the nine has a structural feature whose synthesis mammalian enzymes cannot carry out: a branched carbon chain in leucine, isoleucine and valine, an aromatic ring in phenylalanine, an indole ring in tryptophan, an imidazole ring in histidine, the secondary alcohol group of threonine and the primary amino group at the end of the lysine chain [15]. Where only the carbon skeleton is missing, the body can attach an amino group to it; lysine and threonine cannot be made even that way. Reeds argued that by this strict metabolic test only these two (and perhaps tryptophan) are truly essential, and only glutamic acid and serine, which the body makes from ammonia and a carbon source, are truly nonessential. This is a way of thinking about the chemistry, not a replacement for the nutritional list of nine [15] [13].

Conditionally essential amino acids

Between the two groups lies a third. Conditionally indispensable amino acids are normally made in the body, but not always in sufficient amounts: the World Health Organization lists cysteine, tyrosine, taurine, glycine, arginine, glutamine and proline as examples [14]. Taurine, unlike the others, is not one of the twenty protein amino acids. Their synthesis depends on a precursor or on a few organs. Tyrosine is made only from phenylalanine, and cysteine takes its sulfur from methionine [13]; proline and arginine are made largely in the intestine [15]. Babies born prematurely have too little of the enzyme that makes cysteine, and some inherited disorders turn tyrosine or arginine into dietary necessities [13]. A study that used the oxidation of a labelled amino acid found that glycine becomes indispensable late in pregnancy [16]. Lists of conditionally indispensable amino acids differ between authors – a recent review, for instance, leaves taurine out [16] – which is another reminder that the categories are drawn by people, not by nature.

From the point of view of the body, all twenty are needed. They are used not only for protein but also for other molecules: glutathione is made from cysteine, glutamic acid and glycine, creatine from glycine, arginine and methionine, and nitric oxide from arginine [15]. The difference lies in whether the body can make enough of them itself.

How many amino acids does your body need?

The current international estimates were published in 2007 by a joint expert group of the World Health Organization, the Food and Agriculture Organization and the United Nations University [14]. For a healthy adult the average requirement for protein is 0.66 g per kilogram of body weight per day, and the safe level, which covers almost everyone, is 0.83 g/kg [14]. For a person weighing 70 kg, the safe level is about 58 g of protein a day (calculation). The essential amino acid requirements calculator gives both values, and the amounts of each essential amino acid, for any body weight and age. The requirements for the essential amino acids are given in the table, both per kilogram of body weight and per gram of protein in the diet.

Amino acidmg per kg of body weight per daymg per g of protein
Histidine His1015
Isoleucine Ile2030
Leucine Leu3959
Lysine Lys3045
Methionine + cysteine Met + Cys1522
  of which methionine Met1016
  of which cysteine Cys46
Phenylalanine + tyrosine Phe + Tyr2538
Threonine Thr1523
Tryptophan Trp46
Valine Val2639
Total184277

Average requirements of adults, from the WHO/FAO/UNU report, Table 23 [14]. The safe level of intake is about 24 per cent higher. The separate values for methionine (10.4 mg/kg) and cysteine (4.1 mg/kg) add up to 14.5 mg/kg, rounded to 15 for the pair [14].

These figures are roughly twice as high as those of 1985 for most of the essential amino acids. The earlier values were based mainly on Rose’s nitrogen balance studies, which are now thought to have underestimated the needs because some losses of nitrogen were not measured [14]. Leucine has the highest requirement; it is also the most abundant amino acid in body tissues and in food proteins [14].

The numbers describe healthy adults, and most of the data behind them come from young men [16]. The expert group did not set separate values for older people because the evidence was “fragmentary and conflicting” [14]. Newer studies with a different method, which follows the oxidation of a labelled amino acid, suggest that pregnant women need more protein, especially late in pregnancy, and that older adults may need nearly twice as much leucine as young men [16]. The authors of that review stress that a requirement is a minimum, not a recommended maximum [16].

What foods contain amino acids?

Every food that contains protein contains amino acids. What differs from food to food is the proportion of the essential ones and how much of them the body actually absorbs. To compare foods, nutritionists set the amino acid content of a food protein against the requirement pattern – the amounts per gram of protein in the table of the previous section.

The table below shows such a comparison from the WHO/FAO/UNU report. Per gram of protein, animal proteins, soya and potato supply more than 100 per cent of the requirement for every essential amino acid. The weak point of cereals is lysine: wheat and maize protein provide less than 60 per cent of the required amount, rice 86 per cent. Cassava and yam fall just short in lysine as well, and cassava also in the branched-chain amino acids. The effect shows in whole diets: averaged over 17 Indian states, a typical diet supplied 87 per cent of the lysine requirement, while a typical British diet supplied 140 per cent [14]. These figures compare amino acid content only; how well each protein is digested is not taken into account (see protein quality below).

Food proteinLysTrpThrMet + CysLeu + Ile + ValPhe + Tyr
Egg139293223225168301
Beef203213202182144275
Milk158417191164151271
Soya144217191114136281
Potato121240167131120243
Rice86224153176146305
Maize58117157132177314
Wheat57217127203122306
Cassava9219211512479135
Yam91213157125116265
UK diet140211177174143311
Indian diet87293143182132317

Amino acid content of food proteins and diets as a percentage of the adult requirement pattern; values below 100 are in bold. From the WHO/FAO/UNU report, Table 24 [14]. The comparison is per gram of protein and does not take digestibility into account.

A study of commercial protein isolates – the concentrated proteins sold as ingredients and supplements – measured the same differences directly [17]. Essential amino acids made up on average 26 per cent of plant proteins and 37 per cent of animal proteins, close to the 38 per cent of human muscle protein. Oat, lupin and wheat proteins fell below the WHO requirement pattern; among the plant isolates, only potato protein also reached the pattern for isoleucine and valine. On average, plant proteins contained about half as much lysine as animal proteins and less than half as much methionine, and the leucine content ranged from 5.1 per cent in hemp protein to 13.5 per cent in maize protein. The authors calculated that 50:50 blends of maize, hemp or brown rice protein (low in lysine) with soya, microalgae or pea protein (low in methionine) have a more complete composition, although a little more protein would still be needed to cover every essential amino acid. These are figures for isolated proteins, not for ordinary foods, and tryptophan was not measured [17].

Protein quality: PDCAAS and DIAAS

The amino acids in a food are useful only if they are absorbed. Since 1989 protein quality has been rated by the Protein Digestibility Corrected Amino Acid Score (PDCAAS): the content of the most limiting essential amino acid, compared with the requirement pattern and multiplied by the digestibility of the protein as a whole [18]. In 2013 an expert consultation of the Food and Agriculture Organization proposed a successor, the Digestible Indispensable Amino Acid Score (DIAAS). It uses the digestibility of each essential amino acid separately, measured at the end of the small intestine rather than from what is left in the faeces, and it allows scores above 100 for single foods, so that high-quality proteins are not all cut to the same value [18].

The report gives examples. Against the requirement pattern for young children, wheat scores 40, peas 64 and whole milk powder 122 [18]. A mixture of 400 g wheat, 100 g peas and 35 g milk powder scores 82 for adults, limited by lysine (here the digestibility values were predicted from data on pigs) [18]: the milk and the peas make up much of what the wheat lacks. As an example of how the score might be used on food labels, the consultation suggested that a protein should score at least 75 to be called a good source and at least 100 to be called an excellent one, adding that the cut-offs needed further consideration. It also noted that data on amino acid digestibility measured in humans are still limited [18].

What amino acids do in the body

Most amino acids in the body are not free but built into proteins, and the largest store of both kinds is skeletal muscle, which makes up 40–45 per cent of body mass [19]. The proteins of the body are broken down and rebuilt all the time, and the amino acids that are not reused are burned or converted to other compounds. On an ordinary Western diet an adult breaks down some 80–100 g of amino acids a day [20]. As fuel they are less efficient than glucose or fat: when amino acids are oxidised, a smaller share of their energy ends up in ATP [19]. Their main value lies elsewhere.

Raw material for other molecules

Many substances that are not proteins are made from amino acids. Tyrosine, made from phenylalanine, is the starting point for dopamine, noradrenaline and adrenaline, and for the thyroid hormones. Tryptophan gives rise to serotonin, melatonin and the vitamin niacin [19]. Glutathione, the main antioxidant inside cells, is a small peptide of cysteine, glutamic acid and glycine [19] [15]. Creatine is made from glycine, arginine and methionine [15], and arginine is the source of the signal molecule nitric oxide. Its discovery has been called “one of the most important and distinctive chapters in the history of amino acid metabolism” [20].

Signals

Some amino acids also act as signals. Leucine, glutamine and arginine activate signalling through mTOR, a protein kinase that controls the rate of protein synthesis in cells. Leucine in particular stimulates the synthesis of muscle protein, while the other two branched-chain amino acids, isoleucine and valine, do not [19]. A signal is not the same as a building block, though: in rats and piglets, adding leucine to the diet for a long time did not make them grow faster [19].

Amino acids and neurotransmitters

In the brain, amino acids serve as neurotransmitters in their own right. Glutamate is an excitatory transmitter, aspartate is also listed as one, and glycine is an inhibitory transmitter in the central nervous system [19]. Amino acids are also the precursors of other transmitters: GABA, dopamine, noradrenaline, adrenaline, histamine and serotonin [20].

The brain even uses a D-amino acid. D-Serine is made from L-serine by the enzyme serine racemase and helps to activate the NMDA receptor for glutamate [20]. This was a surprise. In the 1930s Hans Krebs had discovered an enzyme that breaks down D-amino acids, and wondered at an enzyme “for which practically no substrate is found in nature”; three decades later he still counted it among the “enzymes in search of a function” [20]. D-amino acids are not used in ribosomal protein synthesis [1], but the idea that one of them has a physiological role has been described as “revolutionary” [20].

More is not better

An excess of one amino acid can interfere with the use of another. Known pairs and groups that compete in this way are lysine and arginine, the three branched-chain amino acids, and threonine and tryptophan; large excesses of single amino acids are toxic [19]. A balanced supply matters more than a large amount of any one of them – which is worth keeping in mind when reading about amino acid supplements.

Amino acid catabolism

Amino acids that the body does not use for protein or for other molecules are broken down. Their catabolism is usually described in two stages: first the α-amino group is removed and its nitrogen is turned into a form that can be excreted, then the carbon skeleton that remains is converted into compounds that the cell can burn or use to make glucose or fat [21]. According to one standard textbook, the oxidation of amino acids supplies between 10 and 15 per cent of the energy produced by metabolism in animals [21].

What happens to the nitrogen

Excess nitrogen is converted, mainly into urea, by the urea cycle [21]. This pathway was described by Krebs and Henseleit in 1932 – before the better-known Krebs cycle [20]. An adult makes about 30 g of urea a day [20]. Amino acids are broken down in several organs, so their nitrogen has to travel between them. In the fasting state, muscle releases amino acids into the blood, and more than half of the α-amino nitrogen it releases is in the form of just two of them, alanine and glutamine [22]. Alanine makes up only 7–10 per cent of muscle protein but at least 30 per cent of this output, because muscle makes it from pyruvate. The liver takes up the alanine and turns its carbon back into glucose, which can return to the muscle: the glucose–alanine cycle [22]. In this way nitrogen reaches the liver without raising the level of ammonia in the blood [20]. Glutamine is taken up mostly by the intestine and the kidneys, and the branched-chain amino acids are burned mainly outside the liver, in muscle and brain [22].

Glucogenic and ketogenic amino acids

The carbon skeletons of the twenty amino acids are funnelled into only seven compounds: pyruvate, α-ketoglutarate, succinyl-CoA, fumarate, oxaloacetate, acetyl-CoA and acetoacetate. Amino acids that give pyruvate or one of the four intermediates of the citric acid cycle can be used to make glucose and are called glucogenic (or glycogenic). Those whose skeletons end up as acetyl-CoA or acetoacetate give rise to ketone bodies and can be turned into fat, but cannot be used for net synthesis of glucose; they are called ketogenic. Some amino acids are both [21].

This distinction matters most during fasting. In the studies reviewed by Felig, after a night without food about three quarters of the glucose released by the liver still came from its glycogen store. When fasting lasts longer than a day, the share made by gluconeogenesis – the synthesis of new glucose, partly from amino acids released by muscle – increases. In prolonged starvation the brain switches to ketone bodies as its main fuel, which reduces the need for glucose and so spares body protein [22].

Which amino acid belongs where is less settled than textbook tables suggest. A comparison of eleven biochemistry textbooks found that most of them list leucine and lysine as the only purely ketogenic amino acids, but two count lysine as both glucogenic and ketogenic, which leaves leucine as the only purely ketogenic one. Threonine is glucogenic in many of the books and both glucogenic and ketogenic in others. Tryptophan is listed as both in almost all of them, but appears as purely ketogenic in one figure, probably because of a colouring mistake. One textbook classified only nineteen amino acids: asparagine and glutamine were missing from its table, and hydroxyproline was included [21].

The reasons are historical. The convertibility of amino acids into carbohydrate or fat was worked out in nutritional studies between the 1920s and 1950s, and the classification rests on reactions studied under conditions far from normal life, such as feeding single amino acids to dogs in which diabetes had been induced; and some of the pathways described may be incomplete or only alternative routes [21]. Felig, reviewing amino acid metabolism in humans in 1975, put it simply: “virtually all amino acids other than leucine may be categorized as glycogenic” [22]. One textbook quoted in the comparison says the same in other words: whether an amino acid is regarded as glucogenic, ketogenic or both “depends partly on the eye of the beholder” [21].

Beyond the twenty

The twenty are the amino acids of the standard genetic code, but they are not the only amino acids in living things. By the 1960s reviews counted 90 and then 183 amino acids found in plants and animals, most of them never built into proteins [10].

The 21st and 22nd amino acids

In 1986 two groups found, independently, that the genes for an enzyme of the mouse, glutathione peroxidase, and for an enzyme of the bacterium Escherichia coli, formate dehydrogenase, contain a UGA codon in the middle of the coding sequence. UGA normally means “stop”, but here it codes for selenocysteine, a cysteine with selenium in place of sulfur. Selenocysteine has its own transfer RNA, which is first loaded with serine; the serine is then converted into selenocysteine on the RNA itself. Selenoproteins were found in all three domains of life, and those known in 1991 all catalysed oxidation–reduction reactions. The authors called selenocysteine “the 21st amino acid” and proposed the letter U for it [3].

In 2002 a second stop codon, UAG, was found to code for pyrrolysine, a derivative of lysine, in enzymes that methane-producing archaea of the genus Methanosarcina use to break down methylamines. It also has a dedicated transfer RNA and its own enzyme that loads the RNA, and genes for both were found in a bacterium as well. The authors described pyrrolysine as “the 22nd genetically encoded natural amino acid” [4].

Amino acids changed after the protein is made

Many proteins contain amino acids that the genetic code does not specify. They are made from the standard twenty by enzymes after the chain has been built. Vickery listed at least sixteen of them in 1972, among them hydroxyproline, hydroxylysine, the thyroid hormone thyroxine and several methylated lysines [10]. Cystine is a related case. Two cysteine residues in a protein can be joined through their sulfur atoms by a disulfide bond, forming a cystine unit; cysteine is the amino acid built into the chain, and the bond is probably formed afterwards [10].

Amino acids that are not in proteins

Other amino acids work in the body without ever entering a protein. β-Alanine is a β-amino acid: its amino group sits on the second carbon from the carboxyl group rather than the first [19]. Taurine, 2-aminoethanesulfonic acid, has a sulfonic acid group in place of the carboxyl group [7]. GABA, 4-aminobutanoic acid, has its amino group further still from the acid group [7]. Ornithine and citrulline are intermediates of the urea cycle [13], and D-serine, as described above, acts in the brain [20]. GABA has its own page on this site.

Why these twenty?

The question raised at the start of this page has no final answer, but it has two well-argued sides. In 1968 Francis Crick proposed what he called the frozen accident theory. Once many proteins depended on the genetic code, any change in the meaning of a codon would have altered so many of them at once that it would have been lethal, so the code became fixed. In his view, early life used only a few amino acids, perhaps glycine, alanine, serine and aspartic acid, and others were added later, with tryptophan and methionine among the last. Both the number twenty and the actual amino acids, he wrote, were “at least in part due to historical accident”, and the code could have frozen at a local optimum rather than the best possible one [2]. Crick was candid about the weakness of his idea: it is “too accommodating”, and “in a loose sort of way it can explain anything” [2].

The opposite view is that the twenty were chosen for good reasons. The computer study mentioned at the start compared the natural set with 100 million random sets of twenty and found only six that covered size, charge and hydrophobicity as well or better. None of those six had a lower calculated enthalpy of formation, summed over the twenty, than the natural set, which the authors read as a sign that energy also mattered. They concluded that the natural set reflects “considerable natural selection” [5]. The properties used were calculated, not measured, and the authors themselves noted that three properties are a simplification [5].

Doig, in a review that argues this side, starts from what the first proteins had to do [6]. Their catalytic work could be done by small helper molecules and metal ions, so the amino acids were selected mainly to make chains that fold into soluble structures with tightly packed cores. Branched side chains such as those of valine, leucine and isoleucine lose less freedom when a protein folds than straight chains do, which may explain why norvaline and norleucine, their straight-chain relatives, are absent. Cysteine, he suggests, was chosen for binding metals. It was selected in a world without oxygen, where disulfide bridges would hardly form; their use in proteins came much later, after the Great Oxidation Event – in Doig’s words, a “lucky accident”. The cost of making each amino acid, counted in molecules of glucose and ATP in the table Doig uses, is often surprising: by that count leucine costs one ATP, its isomer isoleucine eleven, and methionine is the most expensive. In Doig’s view, the only true frozen accident may be the use of isoleucine rather than its stereoisomer, alloisoleucine [6]. His title sums up the position: frozen, but no accident.

The two views are less far apart than they seem. Crick explained why the code stopped changing; the later work suggests that by the time it did, the set it had settled on was already very good. Doig cites several studies that agree on a likely early set of about ten of the twenty – alanine, aspartic acid, glutamic acid, glycine, isoleucine, leucine, proline, serine, threonine and valine [6] – and the fact that selenocysteine and pyrrolysine were later fitted into stop codons shows that the code was never completely closed [3] [4].

In short

  • Proteins are built from twenty α-amino acids; selenocysteine and pyrrolysine are added by the genetic code in some organisms.
  • Nineteen of the twenty are chiral, and ribosomes use only their L-forms; glycine is not chiral.
  • In water near pH 7 amino acids exist as zwitterions; their isoelectric points range from 2.77 (aspartic acid) to 10.76 (arginine).
  • Nine amino acids are essential for adults, but the line between essential and nonessential is partly a matter of definition.
  • Animal proteins, soya and potato meet the requirements for all essential amino acids; cereals are low in lysine.
  • Amino acids are also the raw material for hormones, neurotransmitters, glutathione, creatine and nitric oxide.
  • Why life uses these twenty is still debated: partly historical accident, partly a set that is close to the best possible.

References

  1. IUPAC-IUB Joint Commission on Biochemical Nomenclature (JCBN). Nomenclature and symbolism for amino acids and peptides (Recommendations 1983). Pure and Applied Chemistry 56(5):595–624 (1984). https://doi.org/10.1351/pac198456050595
  2. Crick FHC. The origin of the genetic code. Journal of Molecular Biology 38(3):367–379 (1968). https://doi.org/10.1016/0022-2836(68)90392-6
  3. Böck A, Forchhammer K, Heider J, Leinfelder W, Sawers G, Veprek B, et al. Selenocysteine: the 21st amino acid. Molecular Microbiology 5(3):515–520 (1991). https://doi.org/10.1111/j.1365-2958.1991.tb00722.x
  4. Srinivasan G, James CM, Krzycki JA. Pyrrolysine encoded by UAG in Archaea: charging of a UAG-decoding specialized tRNA. Science 296(5572):1459–1462 (2002). https://doi.org/10.1126/science.1069588
  5. Ilardo M, Meringer M, Freeland S, Rasulev B, Cleaves HJ II. Extraordinarily adaptive properties of the genetically encoded amino acids. Scientific Reports 5:9414 (2015). https://doi.org/10.1038/srep09414
  6. Doig AJ. Frozen, but no accident – why the 20 standard amino acids were selected. The FEBS Journal 284(9):1296–1305 (2017). https://doi.org/10.1111/febs.13982
  7. Haynes WM, Lide DR, Bruno TJ, editors. Properties of amino acids. In: CRC Handbook of Chemistry and Physics, 97th edition, pp. 7-1–7-2. Boca Raton: CRC Press (2016).
  8. Thurlkill RL, Grimsley GR, Scholtz JM, Pace CN. pK values of the ionizable groups of proteins. Protein Science 15(5):1214–1218 (2006). https://doi.org/10.1110/ps.051840806
  9. Kyte J, Doolittle RF. A simple method for displaying the hydropathic character of a protein. Journal of Molecular Biology 157(1):105–132 (1982). https://doi.org/10.1016/0022-2836(82)90515-0
  10. Vickery HB. The history of the discovery of the amino acids. II. A review of amino acids described since 1931 as components of native proteins. Advances in Protein Chemistry 26:81–171 (1972). https://doi.org/10.1016/S0065-3233(08)60140-0
  11. Vickery HB, Schmidt CLA. The history of the discovery of the amino acids. Chemical Reviews 9(2):169–318 (1931). https://doi.org/10.1021/cr60033a001
  12. McCoy RH, Meyer CE, Rose WC. Feeding experiments with mixtures of highly purified amino acids. VIII. Isolation and identification of a new essential amino acid. Journal of Biological Chemistry 112(1):283–302 (1935).
  13. Laidlaw SA, Kopple JD. Newer concepts of the indispensable amino acids. The American Journal of Clinical Nutrition 46(4):593–605 (1987). https://doi.org/10.1093/ajcn/46.4.593
  14. Joint WHO/FAO/UNU Expert Consultation. Protein and amino acid requirements in human nutrition. WHO Technical Report Series 935. Geneva: World Health Organization (2007). https://iris.who.int/handle/10665/43411
  15. Reeds PJ. Dispensable and indispensable amino acids for humans. The Journal of Nutrition 130(7):1835S–1840S (2000). https://doi.org/10.1093/jn/130.7.1835S
  16. Paoletti A, Courtney-Martin G, Elango R. Determining amino acid requirements in humans. Frontiers in Nutrition 11:1400719 (2024). https://doi.org/10.3389/fnut.2024.1400719
  17. Gorissen SHM, Crombag JJR, Senden JMG, Waterval WAH, Bierau J, Verdijk LB, van Loon LJC. Protein content and amino acid composition of commercially available plant-based protein isolates. Amino Acids 50(12):1685–1695 (2018). https://doi.org/10.1007/s00726-018-2640-5
  18. FAO. Dietary protein quality evaluation in human nutrition. Report of an FAO Expert Consultation. FAO Food and Nutrition Paper 92. Rome: Food and Agriculture Organization of the United Nations (2013). ISBN 978-92-5-107417-6.
  19. Wu G. Amino acids: metabolism, functions, and nutrition. Amino Acids 37(1):1–17 (2009). https://doi.org/10.1007/s00726-009-0269-0
  20. Brosnan JT. Amino acids, then and now – a reflection on Sir Hans Krebs’ contribution to nitrogen metabolism. IUBMB Life 52:265–270 (2001).
  21. D’Andrea G. Classifying amino acids as gluco(glyco)genic, ketogenic, or both. Biochemical Education 28(1):27–28 (2000). https://doi.org/10.1111/j.1539-3429.2000.tb00007.x
  22. Felig P. Amino acid metabolism in man. Annual Review of Biochemistry 44:933–955 (1975). https://doi.org/10.1146/annurev.bi.44.070175.004441