Contents - Previous - Next

This is the old United Nations University website. Visit the new site at

Estimation of "Available" Amino Acids in Foods

It cannot be assumed that all the amino acids released from food proteins by digestion in acid (or alkali) will be absorbed and utilized when the food is eaten. This is illustrated by the results in the first and fourth columns of table 4. Excessively high temperatures during the drying of milk caused some destruction of "total" Iysine

TABLE 4. Lysine Levels (mq/q N) in Four Samples of Milk Powder

Sample Total value after acid hydrolysis (104) Lysine reacting with FDNB Lysine released by in vitro enzymatic digestion Value obtained by growth assay with the rat
1. Good quality 500 513 519 506
2. Slightly damaged 475 400 388 381
3. Scorched 425 238 281 250
4. Severelyscorched 380 119 144 125

Source: Ref. 53. content (i.e., the Iysine liberated by acid hydrolysis), but the change in the value of the milk as a source of Iysine for rats (as measured in a growth assay) was very much greater (53) than could be accounted for by this change.

Foods such as milk with a high reducing-sugar content are particularly susceptible to damage when heated through a moisture content range of 50-250 g/kg (54, 55). In such material, even storage at tropical temperatures can be damaging. The dibasic amino acid Iysine is particularly affected by reactions involving the e-NH2 group (56-57). Processing can also damage the sulphur amino acids of foods, and the full extent of the damage through oxidation of the amino acids cystine and methionine will not be detected by specific procedures for measuring the sulphur amino acids in proteins that themselves determine the oxidized forms, cysteic acid and methionine sulphone. Severe heat treatment in some forms of processing can damage the protein of foods, and the availability of al) the essential amino acids may be affected in large part as a result of a decrease in the digestibility of the proteins (57).

In Vitro Enzymatic Methods

The major cause of the reduced nutritional) value of heat-damaged protein foods appears to be impaired enzymatic hydrolysis of the component proteins; thus, much work has been done to estimate the release of amino acids by in vitro enzymic digestion. The methods have been comprehensively reviewed (58, 59). Early experiments showed useful correlation of in vitro results with those from animal feeding experiments, even though the absolute in vitro values were very much lower. Increasingly sophisticated procedures to imitate the action of the mammalian digestive system have been developed. They have been used to demonstrate the physiological basis for some of the differences in the nutritive value of different materials.

However, the tests are, in general, too complex for the routine evaluation of samples.

A procedure suitable for routine screening has been described by Ford (60), who performed microbiological assays for amino acids with Streptococcus zymogenes, an organism that has considerable proteolytic power. When various dried animal protein materials were given a mild pretreatment with papain and then assayed with this organism for methionine activity, the results agreed wel) with results of animal growth assays for methionine (61). S. zymogenes can be used for assay of the other essential amino acids except Iysine, for which the organism does not have a requirement (62). Ford (63, 64) has reviewed the various microbiological assays that are used to screen for proteins for availability of selected essential amino acids, and has compared availability data on the same protein when measured by chemical and microbial assays. Additional comparisons were made, when possible, with chick bioassay data. He concluded that the use of microbiological assays for protein quality grading, and for assessing the biological availability of individual amino acids, can be of great value to the plant breeder and provide a necessary check on the results of total amino acid analysis.

The availability of all of the essential amino acids within a food protein can be measured with the enzymaticultrafiltrate digest (EUD) (65). This assay involves digesting a protein sample with pepsin-trypsin-pancreatin and then determining the available amino acids by analysing the ultrafiltrate of the multi-enzyme digest. This procedure is also briefly considered in chapter 3 (p. 32).

Microbiological techniques such as those using the protozoan Tetrahymena pyriformis W for determination of overall protein quality are discussed in chapter 3 (p. 33). However, T. pyriformis W has also been used for assessment of available Iysine and methionine (63, 64, 66). Animal techniques for available amino acids also exist; these, however, are discussed in chapter 4 (pp. 52-531.

Available Lysine

Processing damage in dried milk seems to be explained by the reaction of lactose with the free e-NH2 groups of Iysine units in the protein, with the result that the affected Iysine units no longer undergo the Van Slyke reaction or form a dinitrophenyl (DNP) derivative with fluorodinitrobenzene (FDNB), and are no longer nutritionally "available" as judged by animal growth assays 153, 54). This loss in availability can be seen by comparing the results for different dried milk samples in the second and fourth columns of table 4.

Fluorodinitrobenzene-reactive Iysine has been measured with many different procedures (57). A suitable procedure is described in chapter 8 (p. 95). The main technical problem is to minimize the loss of DNP-lysine during acid hydrolysis of the treated protein, as the DNP-lysine is susceptible to reduction of nitro groups. The other problem has been to separate and measure the DNP-lysine with an adequate degree of precision with a procedure simple enough for use in a non-specialized quality control laboratory; chromatographic procedures are probably the most specific for this separation (67,68).

A further problem lies in the interpretation of the results obtained. FDNB-reactive Iysine appears to be a good indicator of nutritionally available )ysine in oilseed products such as peanut flours (56, 69) and in milk powders (53, 70). Fructose-lysine derivatives give very little DNP-lysine when reacted with FDNB and then hydrolysed with acid (71). With protein-rich materials such as meat and fish products that do not contain appreciable levels of carbohydrates, very severe conditions during the drying process can lead to significant falls in the level of FDNB-reactive Iysine and in the nutritional value of the product. However, such severe treatment results in a large reduction in the availability of all other amino acids in addition to Iysine. Growth assay values for individual amino acids may decrease more than would be expected by the drop in FDNBreactive Iysine (57, 72, 73).

This type of processing damage is thought to be due to the formation of many cross linkages in the protein, involving Iysine and other amino acids, that greatly hinders normal enzymic attack. When digestibility of proteins is reduced, it is not justified to assume that "FDNB-reactive" is synonymous with "nutritionally available" Iysine. On the other hand, measurement of FDNB-reactive Iysine is more sensitive than measurement of total amino acid composition, which may remain less changed despite severe nutritional damage, as can be seen from the dried milk data in table 4.

Although the FDNB procedure has proved to be a useful indicator of protein quality for fish flours (74, 75) and for meat products (56, 76), some results with herring stored and cooked in various ways have not shown a high correlation between FDNB values and nutritional evaluation (77). This is possibly due to partial proteolysis of the raw materials. Ordinary FDNB methods do not measure either free lysine or Iysine that is nitrogen-terminal in a peptide chain. An alternative method, designed to overcome this difficulty, measures "total" Iysine and "FDNB-unreactive" Iysine - i.e., Iysine as such, released from materials after treatment with FDNB and then digestion with acid - (78, 79). Where the Iysine reactions are of a protein-protein nature, the approach generally seems to work well, but even here, measurement of FDNB-reactive Iysine is not always adequate, for example e(gama-glutamyl)-lysine is biologically fully available while e(beta-aspartyl)-lysine is not (80). Unfortunately, the procedure fails to measure adequately the type of damage resulting under mild conditions from reaction of reducing sugars with Iysine units (71). A rapid dye-binding procedure has recently been advocated for the determination of reactive Iysine in foodstuffs (81, 82). Dyebinding procedures are considered in more detail later in this chapter.

Trinitrobenzenesulphonic acid (TNBS) is an alternative reagent to FDNB. TNBS has the technical advantage that it is less dangerous to the user, is water-so)uble, and measures free Iysine. On the other hand, it gives Iysine derivatives that are sensitive to destruction during digestion in acid, and, with some types of material, there is the decisive disadvantage that it still measures a large proportion of the Maillard compounds formed between Iysine and sugar units as "TNBS-reactive" (71).

Finally, the furosine procedure (83) should be briefly described. Lactosyl Iysine in damaged milk upon hydrolysis with HCI is converted to both furosine and Iysine in specific proportions; analysis for furosine can thus be used as a direct indicator of damaged or blocked Iysine.

Chemical methods for determination of available amino acids in foodstuffs have recently been reviewed by Carpenter (82).

Enzymatic Methods for Determining Available Lysine

Because of the inability of chemical assays that use dye-binding procedures or 1-fluoro-2,4-dinitrobenzene (FDNB) to measure accurately available Iysine in many foods, especially high-carbohydrate cereals, alternate methods for estimating available Iysine have been investigated.

Methods for determining the Iysine content of acid hydrolysates using the enzyme Iysine decarboxylase have been described (84, 85). The enzyme is only specific for L-lysine, which has a free e-amino group, and thus can be used to predict biological availability. Lysine decarboxylase from Bacterium cadaveris decarboxylates only L lysine, producing CO2 and cadaverine, both of which can easily be measured and used in determining the Iysine content of a protein hydrolysate. The technique has been modified by immobilizing the Iysine decarboxylase enzyme from B. cadaveris or Escherichia cold B on the surface of a CO2-specific electrode and then determining the quantity of CO: released from a hydrolysate using the electrode response as a measure of the available Iysine (86, 87). Since measurements are made using protein hydrolysates, it is obvious that if the reactive group can be freed from the e-amino group of Iysine by the hydrolysis, then total Iysine rather than available Iysine would be measured by the procedure. insufficient data have as yet been accumulated using this technique to allow its general validity to be ascertained, particularly with respect to how well it compares with bioassay data. The procedure, however, appears to show considerable promise.

Methods for Measuring Available Sulphur-Containing Amino Acids

It has been indicated earlier (p. 11 ) that some of the oxidized forms of cystine and methionine are unavailable, or at best only partially available, for the rat. Methionine and methionine sulphoxide, for example, were shown to be equally available to the rat, but the more highly oxidized methionine sulphone was completely unavailable (88). Many reports have shown that experimental or commercial processes that promote oxidation, for example, drying in the presence of air or bleaching to decolourize, can convert methionine, cystine, and cysteine residues in the protein to their oxidized forms and thereby reduce the nutritive quality of that protein (22, 89-91 ).

In order to measure the available cysteine in processed proteins, Clelands reagent (dithiothreitol, DTT) was used (22) to convert all cystine to cysteine, and the cysteine produced by DTT treatment as well as that inherent in the protein was measured using 5,5'-dithiobis-2-nitrobenzoic acid (DTNB). Cysteic acid does not react with DTNB, and therefore only unoxidized or "available" cysteine is measured.

A method for differentiating between methionine and its oxidized forms (methionine sulphone and sulphoxide) was first described by McCarthy and Sullivan (92). These authors described a colorimetric assay that could detect the loss of methionine in proteins due to heat treatment by specifically determining the actual methionine content of the protein before and after heat treatment. This procedure was later modified (93) to eliminate the possible interference from histidine in the measurement of methionine. It was also demonstrated (22, 90) that the early colorimetric method (92) for available methionine could explain the poor nutritional quality of heat-treated casein measured by rat assay, as the oxidized methionine was non-available.

Because the colorimetric assay (92) for available methionine required an enzymatic hydrolysis to release the methionine prior to colorimetric analysis, methods to measure the amount of methionine directly using the intact protein have been sought. Inasmuch as the methionine residue on a protein was capable of reducing dimethyl sulphoxide (Me2SO) to dimethyl sulphide (Me2S), the methionine content of various peptides could be measured (94) by exposing them to Me2SO and measuring the amount of Me2S formed. The Me2SO was evolved into the headspace above the peptide solution and was then measured using GLC techniques. This procedure was then applied to food proteins that had part of their methionine in the oxidized form, either from H202 treatment or from typical industrial processing procedures. The results for available methionine correlated well (94) with rat PER data.

An alternative procedure for measuring the available methionine content of intact proteins was developed by Ellinger and Duncan (95), who treated the protein with cyanogen bromide (CNBr) and then measured the amount of methylthiocyanate (MeSCN) produced as a consequence of the CNBr reacting with methionine residues. Again, only unoxidized or "available" methionine was measured, as the oxidized forms were unreactive. Mackenzie (96) utilized an improved CNBr assay to determine the available methionine in pea protein.

Of the procedures so far described for determining the methionine content of food proteins, those that determine methionine directly on the intact protein, and therefore that eliminate the variable and timeconsuming enzymic hydrolysis step, have a definite advantage. For any of these procedures to be fully acceptable, however, as techniques for assaying available methionine, cysteine, and cystine in foods, compara tive data between available su)phur amino acids and biological assay values on the same samples are essential. Despite the paucity of data, it is now c)ear that the sulphur amino acids are liable to oxidation during the processing of food proteins, and this reactivity can be a major factor leading to loss of nutritional quality in some foods.

While not a measure of available sulphur amino acids, the use of total sulphur determination as an indicator of the combined level of the two sulphur-containing amino acids, cystine and methionine, has been studied by several workers (97-99). In genera), significant correlations have been found between the contents of "cystine + methionine" and total sulphur. However, a large part of the correlation derives from each measure's being correlated with the total nitrogen (99, 100). The usefulness of the S:N ratio in the prediction of "cystine + methionine" (g/kg crude protein) is less certain; within the limited range of such values found in breeding studies of grain legumes (101), high and significant correlation coefficients have not been established.

Dye-Binding Procedures

Dye-binding procedures are rapid and inexpensive methods of analysis that can be successfully semiautomated (102) and promise to be very useful when applied for protein quality process control where damage involving the binding of Iysine amino groups may occur (81).

Azo dyes combine with the free basic amino groups of Iysine, histidine, and arginine, and with terminal amino groups of the protein chain (103). Good correlations were obtained between dye-binding capacity (DBC) with Acid Orange 10 and chick growth (104) measured with heated soy bean meal, and between Acid Orange 12 binding and mouse growth with a series of heated barleys ( 105). Although positive correlations have been obtained between DBC and animal growth tests or FDNB-reactive Iysine with fish and meat meals, the scatter is still considerable (76, 104, 106). The use of DBC values as indicators of heat damage requires constant amino acid composition of the raw material, and this composition can vary in fish and meat meals (107, 108). Oilseeds tend to have a more constant amino acid composition.

A measure specific for lysine is obtained from the difference in the DBC of a sample before and after treatment with propionic anhydride, which inactivates the lysine groups but has no effect on arginine and histidine (109).

A disadvantage of azo dyes is that they fail to detect damage from protein-sugar reactions under mild conditions (108), probably because the early Maillard products, such as deoxyfructosyl Iysine units, are still basic in nature. Azo dyes cannot therefore be relied on for process control of milk powders.

The reactive dye Remazol Brilliant Blue R has been used as an indicator of the reactive Iysine content of heated milk (110) and whey protein (111). Under the conditions of the test, this dye is thought to react only with the free amino acid group of Iysine and the thiol group of cystine and may detect even early Maillard damage ( 108). However, the method involves gel filtration and is less convenient than azodye procedures.

The phthalein dye Cresol Red shows increased binding with soy bean meals subjected to increasing heat treatment (112), i.e., the opposite result to that obtained with azo dyes and reactive dyes. Good correlations have been found between the dye absorption of overheated soy bean meals and their protein quality for chicks ( 113, 114). However, the test is an empirical one and cannot be used as a general indicator of processing damage to protein foods (108).

Other Methods

With some materials such as cottonseed meals, simple tests of nitrogen solubility have proved useful indirect indicators of the amount of nutritional damage caused by processing (76, 115). With other materials, such as soy milks, this approach has not proved useful (116).

Contents - Previous - Next