31 FORMATION, NUTRITIONAL VALUE, AND SAFETY OF D-AMINO ACIDS
Mendel Friedman Western Regional Resarch Center Agricultural Research Service, U.S. Department of Agriculture, 800 Buchanan, Albany, California 94710
A8STRACT The extent of racemization of L-amino acid residues to D-isomers in food proteins increases with pH, time, and temperature. The nutritional utilization of different D-amino acids vary widely, both in animals and humans. In addition, some D-amino acids may be deleterious. For example, although D-phenylalanine is nutritionally available as a source of L-phenylalanine, high concentrations of D-tyrosine inhibit the growth of mice. The antimetabolic effect of D-tyrosine can be minimized by increasing the L-phenylalanine content of the diet. Similarly, L-cysteine has a sparing effect on L-methionine when fed to mice; however, D-cysteine does not. The wide variation in the utilization of D-amino acids is exemplified by the fact that D-lysine is not utilized as a source of L-lysine, whereas the utilization of D-methionine as a source of the L-isomer for growth is dose-dependent, reaching 7&% of the value obtained with L-methionine. Both D-serine and the mixture of L-L and L-D isomers of lysinoalanine induce histological changes in the rat kidneys. D-tyrosine, D-serine, and lysinoalanine are produced in significant amounts under the influence of even short periods of alkaline treatment. Unresolved is whether the biological effects of D-amino acids vary, depending on whether they are consumed in the free state or as part of a food protein. Possible, metabolic interaction, antagonism, or synergism among D-amino acids jn vivo also merits further study. The described results with mice complement related studies with other species and contribute to the understanding of nutritional and toxi col ogi ca 1 consequences of i ngesti ng D-ami no acids. Such an understanding will make it possible to devise food processing conditions to minimize or prevent the formation of undesirable D-amino acids in food proteins and to prepare better and safer foods. INTRODUCTION Processed proteins are increasingly used to meet human dietary
M. Friedman (ed.), Nutritional and Toxicological Consequences of Food Processing © Springer Science+Business Media New York 1991
447
needs. Alkali treatment of casein, corn, and soy proteins brings about desirable changes in flavor, texture, and solubility. Such treatments which are used to prepare protein isolates also destroy toxins and trypsin inhibitors. Treating food protein with alkali and heat may, however, produce undesirable changes in the constituent amino acids. Such changes may include crosslinking, browning reactions, and racemization (1-92). Racemization of L-amino acids to D-isomers in food proteins often impair biological value and safety of foods by (a) forming nonutilizable forms of amino acids; (b) creating L-O, D-L, and 0-0 peptide bonds that may resist hydrolysis by proteolytic enzymes; and (c) forming unnatural amino acids that may be nutritionally antagonistic or act as antimetabolites (14-17). In this paper, I present a limited overview of our studies on the factors which influence racemization of L-amino acid residues in food proteins and on the biological utilization and safety of selected D-amino acids. CHEMISTRY OF RACEMIZATION Since the early part of this century (17), alkali and heat treatments have been known to racemize amino acids. As a result of food processing using these treatments, D-amino acids are continuously consumed by animals and man. Levels of D-aspartic acid detected in some commercial foods are shown in Table 1. Because all of the amino acid residues in a protein undergo--ratemization simultaneously, but at differing rates, assessment of the extent of racemization in a food protein requires quantitative measurement of at least 3& optical isomers, 18 Land 18 D. Analytically, this is a difficult problem not yet solved. Racemization of an amino acid proceeds by removal of a proton from the tt-carbon atom to form a carbanion intermediate. The trigonal carbon atom of the carbanion, having lost the original asymmetry of th~ tt-carbon, recombines with a proton from the envi ronment to regenerate a tetrahedral structure. The reaction is written as: L-amino acid
It
...--
~
O-amino acid
It'
where kra~ and krac are the first-order rate constants forward and reverse racemization of the stereoisomers.
(1 )
for the
The product is racemic if recombination can take place equally well on either side of the carbanion, giving an equimolar mixture of L- and D-isomers. Recombination may be biased, if the molecule has more than one asymmetric center, resulting in an equilibrium mixture slightly different from a 1:1 enantiomeric ratio. lhe following equation was derived to describe the kinetic course of the racemization process (65-66):
448
k L~D
k'
-dI/dt • kL - k'D if Dt =o«4=o' then Lt=o .. L + D and 4=0 - L .. D
-dI/dt • kL - k' (4'"'0 - L) .. kL - k'4=o + k'L • (k + k')L - k'4=o dI/dt = - (k + k')L + k'Lt'"'O dI/dt + (k + k')L = k'Lt=o dI/dt elk + k')t + (k + k')L e (k+k')t • k'4-o elk +k')t dVdt (Le(k + k'lt) = k'4=o elk + k'lt
fd
(Le) (Le(k + k')t)
=
f(k'4=o elk + k')t'·dt
Le(k +k')t. (kI(k + k')ILt=o elk + k')t + constant L" (k'/(k + k'114=o + constant 'e-(k + k')t At t = 0, L" Lt=o and e-(k + k'lt .. 1 4'"'0
=(
k'/(k + k' II 4-0 + ronstant
4=0 -( k'/(k + k' 114=0 .. constant 4=0(1 - k'/(k + k'll
= ronstant
L" (k'/(k + k')]Lteo + Lt-o kI(k + k'l e-(k +k'lt if4=o"'L+D L -I k'/(k + k'II(L + DI + (L + DI kI(k + k'l e-(k + k'lt lI(L + DI • k'/(k + k'l + IkI(k + k'lle-(k - k'lt L(k + k'l - k'(L + DI .. [kllk + k')J--(k + k'lt (L+DHk+k')
D(k'/k [ L - L+D
~
Lk - at'
'1
(L + OJ(k + k'l
e -(k + k'lt - (!¥LlJk'ftll ] [ Lll1L[(l + ( Ll
In [ I -
· [I -l%L!W~{kl J (~'fm'] . -(k + k'It
1
(!l'1-1 In [ I1 + _ (QlLI(k'/k) .. (k + k'It if k'/k • I' In [ I +
1-
(~] )
I
(eq. 2)
449
The foll owing operational equat Ion i'.i uSeful to relate the apparent racemization rate constant (k) to the extent of racemization (OIL) during a specific time period, e.g. 3 hr: ln k
[1 f- OIL]
In[lf-OIL]
1 - OIL 3 hr
1 - OIL
0 hr
= ---------,---------
( 2a)
2·k(3hr)
An alternative mathematical analysis of the first-order kinetics of racemization gives equation 3. However, equation 2 is operationally more useful to measure racemization rates in food proteins (36, 37, 59, 60, 65, 66) than equation 3 because we can measure D/L ratios of amino acids with a 1-3% error compared to a 15% error when concenlralions of D only are me.1sured, as required by equation 3. (eq. 3) where De equals equilibriuRi value of D and Dt equals D at time t. Because the structural and electronic factors which facilitate the formation and stabili7ation of the carbanion intermediate are unique for each amino acid, it follows that the reaction rate for the isomerization of each amino acid is also unique. Thus, the inductive strengths of the R-substituents have been invoked to explain dHfering racemi1ation of amino acid residues in food proteins as influenced by pH, time (t), and temperilture, as illustrated in Figures 1-5. Plotling racemization for individual amino acids for soybean proteins against the 1nductlve parameters clearly demonstrates strong correlations (Figures 6,7). The reader should consult the cited papers for detailed interpretations. 06
0.4
....--.
0.2
"0 0
0
f-+
I
c:
....=.....
8'
..J
-0.2 -0,4
II
-06
...~
-OB
00
-10
Q
..J
o A
o
-12 -14
7.0
II
9.0
A
..
110
130
Effective pH at 65"
Figure 1. 450
Effect of pH of aspartic acid (0), phenylalanine glutamic acid (0) racemilation in casein. Ref. 37.
(11),
and
1.8 16 1.4
rlg
U
!: !.
1.0
"
0..
~
..J
LEU
0.4
•
Figure 2.
TIME (HRI at 65"
II
nme
course of amino acid racemization of caseinn in NaOH at 65 C. Ref. 37.
:::i
•
o-se~/
."
I:
2 C
~
ii
I
Q
•
CIt !D
"-
~
0
Z
3
•
-..
g 30
. ..
.- • .-• / .......... .. L:/ l • // "
45
Q "Q
0.1 N
0
15
III I
Q
.- .--
I
25
./O-TYf
//
e ........e
35
-;;I:
~
oJ
./
.........
45
55
65
75
85
95
Temperature I'C)
Figure 3.
Effect of temperature on O-serine, lysinoalanine content of alkali-treated
Ref. 3&.
D-tyrosine, and soybean protein.
451
0.2 0.0
...
"j
..c:
3.3
3.2
3.1
3.0
2.9
lIT 1 103 (oK-I) rigure 4.
452
Temperature dependence (Arrhenius plots) for the racemization of amino acid residues in casein. Ref. 31.
0.9
9
1o
:1
0.' 0.7
;::
= Q
~ 0Z
i
C
...> ...... Q
0.6
6
0.5
5 V
0.4
Q 4 Z
C
'"
. 0
0.3 0.2 0.1
,
10
12
11
13
14
pH
Figure 5.
Effecl of pH 011 lhe relationship between the extent of hydrolysis of peptide bonds by trypsin in casein and the content of lysinoalan1ne and D-amino acids. Ref. 41.
0.65
asp-
0CJ¥
0.55 0.45
•
035
lO2!!J ..".8
j
glu-
0.15 Q05
0 -005 -0.15
-025
o Figure
0.2
Q4
0.6
CT*
6. Relationship between the inductive constant l«' "') of leucine, valine, phenylalanine, glutamic acid, and aspartic acid side chains of casein and the racemization rate constants relative to alanine tn 0.1 N NaOH at 65 0 C for 3 hr. Ref. 37.
463
1.0
PHE
o
..
GLU• ( . - ' • MET GLN ALA /TYR
LV'
/
• LYS
.ILE -1.0
/ . VAL
o
.5
1.0
(j"'*
Figure 1.
454
Relationship between the inductive constants (a*) of the amino acid side chains of 50ybean protein and the logarithms of the racemization rate constants (k) relative to that of alanine (kala)' Ref. 45.
BIOLOGICAL UTILIZATION OF D-AMINO ACIDS Two pathways are available for the biological utilization of D-amino acids (55-56): (a) racemases or epimerases may convert D-amino acids directly to L-isomers or to (DL) mixtures; or (b) D-amino acid oxidases may catalyze oxidative deam1nation of the a.-amino group to form a.-keto acids, which can then be specifically ream1nated to the L-form. Although both pathways may operate in microorganisms, only the latter activity has been demonstrated in mammals. The amounts and specificities of D-amino acid oxidase are known to vary in different animal species. In some, the oxidase system may be rate-limiting in the utilization of a D-amino acid as a source of the L-isomer. In this case, the kinetics of transamination of D-enantiomers would be too slow to support optimal gf'owth. In addition, growth depression could result from nutritionally antagonistic or toxic manifestations of D-enantiomers exerting a metabolic burden on the organism. The biological utilization of sulfur D-amino acids is conveniently discussed in terms of the transamination and transsulfuration pathways (F.Ig.8). D-methionine can be transformed to the L-isomer ~ia oxidative deamlnation followed by reamination. L-Methionine may be incorporated into proteins. It acts as the methyl group donor via S-adenosylmethionine; is a precursor for cysteine, cystine, and taurine; and participates in transamination reactions in which 3-methylthiopropionate is an intermediate. Not all of the details in the cited pathways, and their possible relationship to inherited metabo 1ic di seases such as cystathionuria and homocystinuria, have been completely elucidated (B). D-Methionine lhe reported wide variation in the biological utilization of sulfur D-amino acids for growth of various animal species may be related to the relative activities of D-amino acid oxidases (12,32, 55-57, 91).However, the reported relative oxidation rates of D-methionine by D-amino acid oxidase in kidney homogenates from man, monkeys, chickens, frogs, rats, and mice do not support this hypothesis. For example, the reported rate of oxidation of D-methionine by the oxidase in mouse kidneys is about one
third
to
one
half
of
corresponding
rates
for
rats
or man.
Nevertheless, we have observed that the nutritional utilization of D-methionine in mice is at least as good or better than reported bioavailabil1ties of D-methionine for either rats or man. In fact, D-methionine appears to be poorly utilized by human~ (12,54,75,91}.when consumed either orally or during total parenteral nutrition. One factor giving rise to inconsistencies regarding the utilization of D-methionine is the dose dependency (Figure 9) of the apparent potency of Dmethionine relative to its L-isomer, 1.~., the dietary concentration of the D-form for any given growth response relative to that of the L-form which would produce the same growth response. This dose-dependency is a result of the non-linear nature of the dose-response curves. This complicates attempts to compare results from some earlier studies with other animal species, which often report data based on a single substitution of the D- for the L-isomer. The extent of oxidation of D-methionine (and other D-amino acids) by D-amino acid oxidase may not be the limiting step governing
455
o-Methionine
I t
o-amino-acid oxidase OI-keto-'I'-Methyl-thiolbutyrate
+
transamination 3-Methylthloproplonate"- L-Methionlne -Proteins
~ + ATP S-Adenosylmethionina methyl transfetase ~ S-AdenosVJhomocystelna / "+ methyl,ted acceptor Homocysteine ~Homocystine cystathionine + Serine p-synthass
I
+
thi 'I'_cystBth:: \onina Inorganic Sulfur
Proteins
I
~
N_AceIVIcys.I.ln~tei\~ Cyst.ne
I
-
~
lanthionine .......
aminotransfetase
'I'-Glutamyl cycle cysteine dioxygfHIase
1
Mercaptopyruvate
Cysteine sulfinlc acid OXidBtiOn/
cysteine Bu/finic acid dacarboxy/ass
Cysteic acid
~ypotaurlne
--Sulfinylaldehyde
cysreic acid
decarboxylase
Taurine _Cholyltaurine
Figure 8.
Transamination and lranssulfuration palhways of D- and Lmethionine. Ref. 32. 80a. II
'-It
Figure 9.
456
0.1t
t.I7
Relationship of weight gain to percent L- or D-methionine in amino acid diets fed to mice. Ref. 32.
utilization. Other factors that could influence utilization include: rates of transport, action of intestinal enzymes and bacteria, rates of absorption, renal clearance, and possible toxic effects.
In discussing D-cystine, it is informative to review first the utilization of the L-isomer. Although L-cystine is not an essential amino acid for rodents, less L-methionine is needed for growth if the diet contains L-cystine (19, 32, 66, 82). The mechanism of this so-called sparing effect needs further clarification. One possibility is that L-cystine serves as a source of L-cysteine and other sulfur amino acids, thus minimizing the need for L-methionine to serve as a reservoir for these amino acids. Our results show that L-cystine has a low level of apparent methionine- like activity in mice (32). Growth response was 6.5% when it was substituted for an optimum level of L-methionine (1.17% in an amino acid diet devoid of any other sulfur amino acids). This may reflect a limited sparing effect for depletion of endogenous L-methionine, which may then be reutilized in higher priority pathways. L-cystine is somewhat ,"ore efficient in sparing D-methionine than L-methionine in amino acid diets containing a low level (0.29%) of L- or D-methionine (32). Supplementation of the D-isomer with an equal sulfur equivalent of L-cystine nearly doubled growth, bringing the overall response essentially equal to that produced by L-methionine in the presence of L-cystine. Since D-methionine must first be converted to the L-form before it can be utilized, these results illustrate a marked lessening in demand for l~thionine when L-cystine is available. Thus, even though D-methionine alone is somewhat less efficiently utilized in mice than is the L-isomer, the combination of L-cystine with either Lor D-methionine provides essentially equivalent nutritive value. In contrast, supplementation of an amino acid diet containing 25% of the optima I level (0.29%) of L-methionine with the same concentration of D-cystine increased growth from 58.7 to only 70.7% of the weight gain with optimal (1.17%) L-methionine. Increasing D-cystine to 1.11% reduced relative growth from 56.7 to 43.5%, which may imply that excess D-cystine in the diet is toxic. D-Cysteine Table 2 shows that L-cysteine had a sparing effect on L-methionine, but that D-cysteine did not. In fact, D-cysteine imposed a metabolic burden indicated by depressed growth when fed with less than optimal levels of L-methionine. The 24% decrease in weight gain of the D-cysteine-plus- L-methionine compared to L-methionine alone could mean that D-cysteine was nutritionally antagonistic or toxic. The mechanism(s) of the growth-depressing effects of D-cysteine and D-cystine remains to be elucidated, as do the reported toxic manifestations resulting from excess consumption of L-methionine. Cavallanl et al. (17) and Krijgsheld et al. (57) report that D-cystelne aprears lo be a more efficient precursor of inorganic sulfate in the blood of rats than L-cysteine. These findings prompted Krijgsheld et a1. to suggest that therapeutic administration of D-cysteine could be beneficial in cases of sulfate depletion. However, some caution may be in order, in view of our finding of a growth-depressing effect of D-cysteine in mice.
457
Lanthionine Isomers Lanthionine (S(CH2CH(NH2)COOH)2) isomers are formed during exposure of food and other proteins to alkali and heat (23-26, 41, 64), 72, 79). Such treatments generate dehydroalanine side chains from half-cystine and serine. Reaction of the SH group of cysteine and the double bond of dehydroalanine gives rise to one pair of optically active isomers (enantiomers) and one diastereoisomeric (meso) form. These fonns are correctly named as follows: (R)-L-lanthionine or S-[(2R)-2-amino ·2-carboxyethyl] L-cysteine (S)-D-Lanthionine or S-[(2S)-2-amino-2-carboxyet~yl]-D-cysteine (S)-L-lanthionine or S-[(2S)-2-amino-2-carboxyethyl]-L-cysteine (meso-lanthionine). Although food processing produces a mixture of all three isomeric lanthionines, it would be of interest to know the biological utilization of the individual isomers as a source of L-methionine. However, only the mixture of all three was available to us. Table 2 shows that DL + ~eso-Lanthionine has a moderate sparing effect on L-methionine, as evidenced by 27% greater weight gain ~Ihen the two amino acids were fed together, than from suboptimal L-methionine alone. The metabolic pathway for the utilization of lanthionine probably involves cystathionase-catalyzed transformation to fonn cysteine and pyruvic acid, analogous to the observed transfonnation of cystathionine to cysteine and tt-ketobutyric acid (32, 49). Stereochemical analysis of the cleavage shows that L-lanthionine will produce one molecule of L-cysteine; D-lanthionine, one molecule of D-cysteine: and ~-lanthionine, an equal mixture of both isomers. Complete cleavage of an equimolar mixture of all three isomers will, therefore, fonn an equimolar mixture of D-and L-cysteine, and only the latter, as our results have shown, will be utIlized nutritionally in the presence of L-methionine. D-Phenylalanine Table 3 and Figure 10 show that the relative growth of animals fed L-and D-Phe ranged from 28.3 to 81.3%, depending on the concentrations selected for comparison. Inspection of Figure 10 reveals that near maximum growth may be expected with sufficient D-Phe in the diet. The data suggest the absence of any marked antinutritional effects or toxicity from feeding either Phe isomer at twice the optimum dietary level. The dose dependence of the relative response of the two isomers in a range beyond that producing maximum growth with L-Phe complicates attempts to compare results from earlier studies with other speCies, which often report data based on a single substitution of the D- for the L-isomer. Nevertheless, the &2.1% relative response shown in Table 3 with a D-Phe concentration equal to the level of L-Phe In the complete amino add diet fonnulation (1.51%) is similar to the reported relative efficacy of &8% in rats and 75% in poultry (9,33). Although the semi-logarithmic plot shown in Figure 10 offers a useful graphic presentation of the relative potencies of the phenylalanine isomers, it does not limit correctly, ~., the point corresponding to 0% cannot be an extrapolation of the other data. An alternative way to represent the data would be to use the following simple exponential function which fits the data and limits correctly at both ends. 458
Table 1.
D-Aspartic acid content in commercial food products OIL ASP 0.095
D-ASP D-ASPtl-ASP 0.09
Baby formula (sov protein)
0.108
0.10
Simulated hacon (soy protein)
0.143
0.13
Corn chips
0.164
0.14
Dairy creamer (sodium caseinate)
0.208
0.17
Commercial Product Texturized SOY Drotein
Ref. 66.
Table 2.
Bioavailability of amino acid derivatives of l-cysteine wilh and without l-methionine in mice Relative Percent in Diet a
1 est
Substance
Mean Weight Gain b
Gain Relative to Methionine alone
(g)
(%)
L-Mr.thion1ne
None L-Cyste ine D-Cysteine Dl + Meso-lanthionine
100 100 100
_2L
f.~
4.0 -3.6 -4,2 -3.2
7.4 13.2 5.6 9.4
100
1"1B
76 127
a Relative 1001 is the molar equivalent of 1.17g l-methionine per 100g diet. b Weight gain after 14 days. Ref. 32, 34.
Table 3.
Percent in diet
Utilization of land D-phenylalanine for body weight gain in mice Relative percent in diet
Body Weight Gain a (g) l-Phenlllalanine
0 0.38 0.76 1. 51 3.02
0 25 50 100 200
-4.6 7.4 13.4 12.8 13.6
Relative Response for for D-Phenylalanineb (I) D-Phenlllalan1ne -4.6
-1.2 1.8 6.2 10.2
28.3 35.6 62.1 81.3
a Weight gain after 14 days. b Net weight gain for D-form divided by net weight gain for l-form, times 100. Ref. 33. 469
Table 4.
Weighl gain of mice fed an amino acid diet supplemented with D-tyrosine Diet
Percent D-Tyr in diet
D~lyr:L-Phe
0.76% L-Phe in diet
Wt.gain a g
o
o
0.42
1: 2
0.83
I :1
1.55
2: 1
13.5 13.5 8.8 1.5
1.51% L-Phe 1n diet
o
o
0.42 0.83 1. 55
1 :4 1: 2 1: 1
a Weight gain after 14 days.
13.8 14.3
15.0 10.0
Ref. 33.
Table 5 Weight gain of mice fed a casein diet supplemented with D-tyrosine Diet D-Tyr in diet (%) 10%
protein
o
1 :4 1: 2 1: 1
0.37 0.74 1.48
2:1
o
o
{).37
1 :4
1:2
0.'11
1.48
1: 1
2:1
2.97
a Weight
gain after 14 days.
Mean wt gaina (g)
o
0.19
460
D-lyr:L-Phe
Ref. 33.
8.8 9.7 9.0
10.2 2.2
14.7 13.H
13.8 10.8 7.0
Assume that, 13.6 - gain 13.6 - (-4.6)
=
e -KP
(3);
1n
13.6 - gain 13.6 - (-4.6)
-KP
(4)
where: 13.6 is the maximum weight gain with L-Phe (lable 3); gain is the weight gain with 0 and 25% L-Phe and with all concentrations of D-Phe; -4.6 is the weight gain (loss) with 0% Phe: K is a constant: and P is the relative % of either Phe isomer (35). Equation 4 predicts that a plot of the left-hand ptlrt against P should give a straight line with 01 s lope of -K. Such plots (Figure 11) based on the data in Table 3 yield a value of K "0.0428 for l-Phe and 0.00844 for D-Phe. Thus, a measure of the relative potency of D-Phe compared to that of L-Phe, within the range of the growth response observed for the D-isomer, is 0.00844/0.0428 or approximately 20%. Other workers have reported the following related observations. Healthly young adults can invert about one-third of D-Phe to its l-isomer (86). When a young human female ingested deuterium-labeled 0and l-Phe (8 mglkg body weight), the alllino acid appeared more rapidly .lnd abundantly and disappeared more slowly from plasma with the 0- than with the L-isomer (58). Although D- and L-isomers may be absorbed at the same rate from the intestine of rats, tissue accumulation of the isomers varies (88). The D- and l-Phe isomers differ in their ability to stimulate pancreatic secretions in dogs (67). lhe presence of other D-ami no acids decreases the ut il i zalian of D··Phe, presumably because of the increased competition for available 0 amino-acid oxidase needed to transform D- to l-isomers (56). Finally, D-Phe produces analgesia in mice and humans by potent iating the action of enkephalins in the brain through inhibition of the enzyme that degrades brain opiolds (6). Metabolic studies of Phe, including the data shown in Figure 10, indicate that the D-isomer, because it must first be inverted, is utilized more slowly. This leads to the speculation that D-Phe might not have the same effect on phenylketonuria In children as the L-isomer (1). Should this be the case, the slow and uniform release of L-Phe from the O-isomer could possibly provide a basis for the utilization of D-Phe in the treatment of phenylketonuria. Animal and human studies are needed to demonstrate this possibility. D-Tyrosine The amino acid L-Tyr is an in vivo precursor for brain catecholamine, dopamine and norepinephrine; for the biogenic amine tyramine, and for the ubiquitous pigment melanin (7). In !,itro, the phenolic hydroxyl group activates the benzene ring of tyrosine, making it more sl1!;ceptible to chemical and radiation-induced modification than the corresponding ring in phenylalanine (40). The inductive nature of the phenolic group enhances such food processing factors as high pH and temperature to bring about the rapid racemization of L-Tyr residues in proteins (36, 59). D-Tyr has been shown to prevent and reverse stress-induced chronic hypertension, which may indicate possible effects on the central nervous system (80). Nutritionally, L-Tyr is classified as a semi-essential amino acid since it Is synthesized in vivo from L-'Phe (72). Combinations of L-Tyr and L-Phe are complementary in supporting the growth of mice (33) Thus, under conditions where L-Phe may be limiting, L-Tyr can supply 481
I. 1.
'2 10
..!!!
z
:c •
D•
PhlNlylalonln_
0
~
::I:
8~ ·2
U
1.4
1.5
1.6
1.7
1.1
l.t
2.0
2.1
2.2
2.3
2.A
LOG PHENYLALANINE IN DIET
figure 10. Growth response of mice fed 0- and L-phenylalanine.
Ref. 33.
1.0 0.9 0.1 0.7
0.6 C
0.5
1i
CII
l:CII .......
•~
1 ......
• •
c.; c.;
0.4
\ L.pHENYLALANINE
D.PHENYLALANINE
0 0.3
0.2
\
\ \
o 100 200 RELATIVE PERCENT IN DIET (P in equation 4)
Figure 11
462
Growth reponse of mice fed 0- and L··pheny1alanine.
Ref. 35 ..
about half the requirement of L-Phe alone, a value similar to those previously reported for humans, chicks, and rats. Table 4 shows that with D-Tyr in an amino acid diet, growth inhibition was severe at a D-Tyr:L-Phe ratio of 2:1, but was much more moderate when the ratio was 1: 1. Similar results were obtained wi th a casein based diet supplemented with D-Tyr (Table 5). One or more of the following six mechanisms may explain this inhibition of weight gain by dietary D-Tyr: Tyrosyl ribonucleic acid synthethase was found to catalyze the incorporation of D-Tyr into tyrosyl-tRNA, an aminoacyl adenyl ate derivative similar to that formed by L-Tyr (16, 90). Since the structural features of an amino acid after combining with tRNA do not seem to be significant in the specificity of its incorporation into protein, D-Tyr ingestion should lead to the formation of faulty peptides and proteins. These, in turn, could interfere with normal metabolic processes and be responsible for the observed growth inhibition. A related possibility is the suppression of normal protein synthesis by D-Tyr through competitive inhibition of L-Tyr or L-Phe incorporation into aminoacyl-tRNA, and thus, into proteins. A third possibility is an interference by D-Tyr in the biosynthesis or biological action of vital neurotransmitters such as dopamine. The observed hypotensive effect of D-Tyr is postulated to involve this effect (80). A fourth possibility is hydroxylation of L-Phe to L-Tyr.
an
interference
by
D-lyr
in
the
A fifth possibility is that D-Tyr at concentrations provided in the present study could overload metabolic pathways needed to eliminate or detoxify excess D-amino acids. However, when the dietary level of L-Phe was increased, these same concentrations of D-Tyr failed to bring about marked growth inhibition. A sixth possibility is competition of D-tyrosine with L-tyrosine or L-phenylalanine for membrane transport (4). A discussion of membrane transport of amino acids, which appeared in an article in Nutrition Reviews (4) is relevant to the theme of this paper, and accordingly it is quoted here: "The catalytic transport of amino acids across the plasma membrane tends to have lower structural specificity than is usually characteristic of enzymatic reactions. Furthermore, the stereospecificity varies among amino acid transport systems and among membranes. In some cases, stereospecificity is high, as appears to be characterist1c of the inner mitochondrial membrane. A study with the Ehrlich ascites tumor cell demonstrated differences in selectivity. In that study, there was only about threefold preference for L-methionine over its D-antipode. Furthermore, another study showed that a transport system that was able to discriminate between glutamic acid antipodes had little ability to discriminate between aspartic acid antipodes. "Associated with this low stereoselectivity of transport is the familiar circumstance in higher animals that the transport of amino acids is mainly carried out by 'public' systems, able to handle a wide range of ami no acids, rather than by the occas iona 1 'pri vate' systems restricted to a single amino acid. The probability is very high that not only phenylalanine and tyrosine, but also tryptophan, the 463
branched-chain amino acids and several others have a large proportion of their inter-organ flows mediated by a common transport system designated L. This possibility occasions the currently re-emphasized suspicion that in untreated phenylketonuria, the high ci rculating L-phenyla lanine levels interfere to critical degrees with the passage of several important amino acids across the blood-brain barrier. Tews et a1. (85) have given special attention to analogue inhibition of amino acid transport across this barrier, arising from dietary amino acid imbalances. "In connection with competition for membrane transport, Friedman and Gumbmann (33) are correct in focusing attention on the interference of D-tyrosine with L-phenylalanine as well as with L-tyrosine metabolism. They cite a study pertinent to possible competition by D-tyrosine with transport of L-amino acids. When a young woman ingested deuterium-labeled D-and L-phenylalanine (B mg/kg body weight), the D-isomer appeared more rapidly and disappeared more slowly from the plasma than the L-isomer (58). The higher tolerance curve for the 0than for the L-tyrosine brings attention to the important differences in the rates at which 0- and L-tyrosine are withdrawn from the circulation. In the case of D-phenylalanine, but not for D-tyrosine, inversion to the L-form limits and compensates for the competitive effects arising from elevated levels of the D- form". ·Although the specific mechanism by which D-tyrosine becomes an antimetabolite for the mouse is still not clear, these findings warn investigators to avoid the use of DL-tyrosine in experimental and clinical nutrition. They also dramatize the competition between enantiomorphs of some amino acids. In the case of D-tyrosine, the precise mechanismas can be elucidated only by further study." In conclusion, the cited data demonstrate that D-Tyr, unlike L-Tyr, has no sparing effect for L-Phe. In fact, a metabolic stress, in the form of growth inhibition in mice, may become evident when D-Tyr is present in the diet at equal or greater molar concentrations to L-Phe.
This acute effect remains to be defined toxicologically, and the potential for sub-chronic and chronic toxicity following exposure to lower levels of D-Tyr remains unknown. D-Tryptophan The relative potency of D-tryptophan (as defined in Tables Ii and 7) compared to the L-isomer was strongly dose-dependent, being inversely related to dietary concentration and ranging from 29% to 1i4%. A plot of this data showed that growth was linearly related to D-tryptophan. The maximum growth obtainable for L-tryptophan occurred at or slightly less than 0.174% in the diet. By increasing the dietary concentration of D-tryptophan up to 0.52% (three times the highest level shown in Table Ii), it was possible to demonstrate for D-tryptophan that growth also passed through a maximum, one which equalled 82% of that achieved with the L-isomer. This occurred at approximately 0.44% D-tryptophan for a relative potency of 25% (data not shown). This level in the short term assay appeared to be well tolerated. However, additional studies are needed, both In mice and other species, to define the effects of more prolonged dietary exposure to D-trytophan.
464
Table £I.
Growth of mice fed 0- and 1-tryptophan
Amino acid tested
Percent Mean weight In diet gain (g)a
L-tryptophan
0 0.01'2 0.044 0.087 0.174
-3.5 + 0.29 -1.8:;: 0.37 2.2 0.49 12.4 + 1.25 15.& 0.&8
0 0.022 0.044 0.087 0.174
-3.5 -2.8 -1.8 -0.4 3.8
D-tryptophan
Percent equivalent In dlet b
Relative potency (%)c
± ±
± 0.29
+ 0.49 ;. 0.37 + 0.81 0.58
±
a Weight gain after 14 days. Range tes l.
N = 5.
0.014 0.024 0.033 0.051
&4
55 38 29
Means ± S.E. Duncan's Multiple
b Percent equivalent in the diet i~ thilt concentration of L-amino acid which would produce the same growth ohserved for the D-amino acid (dp.tp.rmined graphically). c Re1alive potency Is the percent equivalent in the diet divided by the percp.nt actually fed X 100.
Table 7. Relative percent In diet a
o
50 100 150 200 250 300
Growth of Mice fed D- and L-Tryptophan for 14 days 80dy weight gain (g)b D-tryptophan
L-tryptophan
-2.8 ± 0.4 0.2 + 0.& 4.4 +1.5 10.& ;: 1.3 11.4 +1.0 12.2+1.2 11.4 1. 1
-2.8 ± 0.4 14.5 ± 0.4
±
(N
= 5).
Relative weight gain (D/L)
17 .3
41.& 77 .5
82.1 8&.7 82.1
a 100% relative equals 0.174% in thp. diet of either D or L-tryptophan. b Means ± S.E. Duncan's Multiple Range test.
465
Considerable species, variation is known to exist for the nutritive value of D-tryptophan. In chicks fed amino acid diets, the relative biologital activity of the 0- to l-lsomer has been, reported to be 20%, which is similar to the relative potency we find for mice (61,70,71) This 1S in marked contrast to rats als.o fed amino acid diets where the stimulation of growth was equivalent to that produced by L-tryptophan, for a biological actIvity of 100%. It should be noted that O-tryptophan supplementation of a protein based diet containing 0.085% protein-bound L-tryptophan resulted in growth stimulation in mice nearly equal to that produced by L-trytophan supplementation. Plasma analysis indicated that the rate of conversfon of 0 to L closely reflected the difference in response between chicks and rats. Essent ia lly no convers ion of D-tryptophan to the L-isomer could be detected in chick plasma. and most of the D-isomer administered was excreted unchanged. In the rat. plasma levels of the L-isomer rapidly increased upon administration of D-tryptophan, and only lJ of the dose was excreted in the urine. Lysinoalanine Isomers Lysinoalanine (HOOCCH(NHl)CH2CH2CH2CH2NHCH2CH(NH2)COOH,(LAL». is an unnatural amino acid (formed as shown in figures 12-14) has been identified in hyrolyzates of processed food proteins. in particular those subjected to alkali (19, '21,25,27,31,38,39,41,42,44,46, 47, 53, 60, 68, 74, 81. 83, 89). Lysinoalanine has two asymmetric carbon atollls. Four stereoisomers are therefore possible: LL. LO. OL. and ~O. The following systematic name is now for the isomer formed in the greatest quantity during food processing (25): N e: -(mc2-~2-carboxyethyl)-trlYSine.
A more correct name with R and lysine would be: Nfi
~
designation for the substituents of
-l(2I!S)-2-ami~2-carboxyethylJ -!:
lysine.
In analogy with the two isomers derived from L-lysine. O-lysine would give rise to the following correctly named diastereoisomers:
~- [(2I!S)-2-amino-2-carboxyethyll-Q-lYSine. It is also worth noting tbat the reaction between the t-NH2 group of lysine and the double of dehydrothreonine can, in principle, give rise to five symmetric' centers and 32 stereoisomers (25). In rats, histological changes in the kidneys have been identified which are related to dietary exposure to this substance, either isolated or as part of intact proteins. The lesions are located in the epithelial cells of the straight portion of the proximal renal tubules and are characterized by enlargement of the nucleus and cytoplasm, increased nucleoprotein tontent, and disturbances in DNA synthesis and mitosis. 8ecause of these observations. contern has ari sen about the safety of foods which may contain LAL and related dehydroalanine-derived amino acids which are known to produce similar lesions. However, since the mechanism by which these compounds damage the rat kidney is unknown, it 1s difficult to assess the risk to human health caused by their presence in the diet. 488
LAL has two asyrrmetric carbon atoms making possible four separate diastereoisomeric forms, LL, LO, OL, and ~O. Its structure suggests that it should have excellent chelating potenial for metal ions, a property which may have relevance to its toxic action. Accordingly, we have examined LAL (a mixture of the LL and 1.0 isomers) for its affinity towards a series of metal ions, of which copper (II) was chelated the most strongly (39, 42,74), On this basis, we have suggested a possible mechanism for kidney damage in the rat involving LAL's interaction with copper within the epithelial cells of the proximal tubules.
"
OH~ ,I
I
CH,-C-P I 1 Y N=C-P
Y
1l
1~ ~® H I
CH,..:C-P I I Y N-C-P
H!t.oSHIFT
>
,.....1
-..y
~o) CH,-C!..P NH-C-P
II
!H@ 0
( s..
CIJIUNION INTERMEDIATE
I
NH-C-P
2. P·ELlMINATION 18)
!(-Y$~
P·HlMINATION IA)
I:;> I
Y
I RACEMIZATION
I
NH-C-P
CH, -C-P
OH
Mendel Friedman Western Regional Resarch Center Agricultural Research Service, U.S. Department of Agriculture, 800 Buchanan, Albany, California 94710
A8STRACT The extent of racemization of L-amino acid residues to D-isomers in food proteins increases with pH, time, and temperature. The nutritional utilization of different D-amino acids vary widely, both in animals and humans. In addition, some D-amino acids may be deleterious. For example, although D-phenylalanine is nutritionally available as a source of L-phenylalanine, high concentrations of D-tyrosine inhibit the growth of mice. The antimetabolic effect of D-tyrosine can be minimized by increasing the L-phenylalanine content of the diet. Similarly, L-cysteine has a sparing effect on L-methionine when fed to mice; however, D-cysteine does not. The wide variation in the utilization of D-amino acids is exemplified by the fact that D-lysine is not utilized as a source of L-lysine, whereas the utilization of D-methionine as a source of the L-isomer for growth is dose-dependent, reaching 7&% of the value obtained with L-methionine. Both D-serine and the mixture of L-L and L-D isomers of lysinoalanine induce histological changes in the rat kidneys. D-tyrosine, D-serine, and lysinoalanine are produced in significant amounts under the influence of even short periods of alkaline treatment. Unresolved is whether the biological effects of D-amino acids vary, depending on whether they are consumed in the free state or as part of a food protein. Possible, metabolic interaction, antagonism, or synergism among D-amino acids jn vivo also merits further study. The described results with mice complement related studies with other species and contribute to the understanding of nutritional and toxi col ogi ca 1 consequences of i ngesti ng D-ami no acids. Such an understanding will make it possible to devise food processing conditions to minimize or prevent the formation of undesirable D-amino acids in food proteins and to prepare better and safer foods. INTRODUCTION Processed proteins are increasingly used to meet human dietary
M. Friedman (ed.), Nutritional and Toxicological Consequences of Food Processing © Springer Science+Business Media New York 1991
447
needs. Alkali treatment of casein, corn, and soy proteins brings about desirable changes in flavor, texture, and solubility. Such treatments which are used to prepare protein isolates also destroy toxins and trypsin inhibitors. Treating food protein with alkali and heat may, however, produce undesirable changes in the constituent amino acids. Such changes may include crosslinking, browning reactions, and racemization (1-92). Racemization of L-amino acids to D-isomers in food proteins often impair biological value and safety of foods by (a) forming nonutilizable forms of amino acids; (b) creating L-O, D-L, and 0-0 peptide bonds that may resist hydrolysis by proteolytic enzymes; and (c) forming unnatural amino acids that may be nutritionally antagonistic or act as antimetabolites (14-17). In this paper, I present a limited overview of our studies on the factors which influence racemization of L-amino acid residues in food proteins and on the biological utilization and safety of selected D-amino acids. CHEMISTRY OF RACEMIZATION Since the early part of this century (17), alkali and heat treatments have been known to racemize amino acids. As a result of food processing using these treatments, D-amino acids are continuously consumed by animals and man. Levels of D-aspartic acid detected in some commercial foods are shown in Table 1. Because all of the amino acid residues in a protein undergo--ratemization simultaneously, but at differing rates, assessment of the extent of racemization in a food protein requires quantitative measurement of at least 3& optical isomers, 18 Land 18 D. Analytically, this is a difficult problem not yet solved. Racemization of an amino acid proceeds by removal of a proton from the tt-carbon atom to form a carbanion intermediate. The trigonal carbon atom of the carbanion, having lost the original asymmetry of th~ tt-carbon, recombines with a proton from the envi ronment to regenerate a tetrahedral structure. The reaction is written as: L-amino acid
It
...--
~
O-amino acid
It'
where kra~ and krac are the first-order rate constants forward and reverse racemization of the stereoisomers.
(1 )
for the
The product is racemic if recombination can take place equally well on either side of the carbanion, giving an equimolar mixture of L- and D-isomers. Recombination may be biased, if the molecule has more than one asymmetric center, resulting in an equilibrium mixture slightly different from a 1:1 enantiomeric ratio. lhe following equation was derived to describe the kinetic course of the racemization process (65-66):
448
k L~D
k'
-dI/dt • kL - k'D if Dt =o«4=o' then Lt=o .. L + D and 4=0 - L .. D
-dI/dt • kL - k' (4'"'0 - L) .. kL - k'4=o + k'L • (k + k')L - k'4=o dI/dt = - (k + k')L + k'Lt'"'O dI/dt + (k + k')L = k'Lt=o dI/dt elk + k')t + (k + k')L e (k+k')t • k'4-o elk +k')t dVdt (Le(k + k'lt) = k'4=o elk + k'lt
fd
(Le) (Le(k + k')t)
=
f(k'4=o elk + k')t'·dt
Le(k +k')t. (kI(k + k')ILt=o elk + k')t + constant L" (k'/(k + k'114=o + constant 'e-(k + k')t At t = 0, L" Lt=o and e-(k + k'lt .. 1 4'"'0
=(
k'/(k + k' II 4-0 + ronstant
4=0 -( k'/(k + k' 114=0 .. constant 4=0(1 - k'/(k + k'll
= ronstant
L" (k'/(k + k')]Lteo + Lt-o kI(k + k'l e-(k +k'lt if4=o"'L+D L -I k'/(k + k'II(L + DI + (L + DI kI(k + k'l e-(k + k'lt lI(L + DI • k'/(k + k'l + IkI(k + k'lle-(k - k'lt L(k + k'l - k'(L + DI .. [kllk + k')J--(k + k'lt (L+DHk+k')
D(k'/k [ L - L+D
~
Lk - at'
'1
(L + OJ(k + k'l
e -(k + k'lt - (!¥LlJk'ftll ] [ Lll1L[(l + ( Ll
In [ I -
· [I -l%L!W~{kl J (~'fm'] . -(k + k'It
1
(!l'1-1 In [ I1 + _ (QlLI(k'/k) .. (k + k'It if k'/k • I' In [ I +
1-
(~] )
I
(eq. 2)
449
The foll owing operational equat Ion i'.i uSeful to relate the apparent racemization rate constant (k) to the extent of racemization (OIL) during a specific time period, e.g. 3 hr: ln k
[1 f- OIL]
In[lf-OIL]
1 - OIL 3 hr
1 - OIL
0 hr
= ---------,---------
( 2a)
2·k(3hr)
An alternative mathematical analysis of the first-order kinetics of racemization gives equation 3. However, equation 2 is operationally more useful to measure racemization rates in food proteins (36, 37, 59, 60, 65, 66) than equation 3 because we can measure D/L ratios of amino acids with a 1-3% error compared to a 15% error when concenlralions of D only are me.1sured, as required by equation 3. (eq. 3) where De equals equilibriuRi value of D and Dt equals D at time t. Because the structural and electronic factors which facilitate the formation and stabili7ation of the carbanion intermediate are unique for each amino acid, it follows that the reaction rate for the isomerization of each amino acid is also unique. Thus, the inductive strengths of the R-substituents have been invoked to explain dHfering racemi1ation of amino acid residues in food proteins as influenced by pH, time (t), and temperilture, as illustrated in Figures 1-5. Plotling racemization for individual amino acids for soybean proteins against the 1nductlve parameters clearly demonstrates strong correlations (Figures 6,7). The reader should consult the cited papers for detailed interpretations. 06
0.4
....--.
0.2
"0 0
0
f-+
I
c:
....=.....
8'
..J
-0.2 -0,4
II
-06
...~
-OB
00
-10
Q
..J
o A
o
-12 -14
7.0
II
9.0
A
..
110
130
Effective pH at 65"
Figure 1. 450
Effect of pH of aspartic acid (0), phenylalanine glutamic acid (0) racemilation in casein. Ref. 37.
(11),
and
1.8 16 1.4
rlg
U
!: !.
1.0
"
0..
~
..J
LEU
0.4
•
Figure 2.
TIME (HRI at 65"
II
nme
course of amino acid racemization of caseinn in NaOH at 65 C. Ref. 37.
:::i
•
o-se~/
."
I:
2 C
~
ii
I
Q
•
CIt !D
"-
~
0
Z
3
•
-..
g 30
. ..
.- • .-• / .......... .. L:/ l • // "
45
Q "Q
0.1 N
0
15
III I
Q
.- .--
I
25
./O-TYf
//
e ........e
35
-;;I:
~
oJ
./
.........
45
55
65
75
85
95
Temperature I'C)
Figure 3.
Effect of temperature on O-serine, lysinoalanine content of alkali-treated
Ref. 3&.
D-tyrosine, and soybean protein.
451
0.2 0.0
...
"j
..c:
3.3
3.2
3.1
3.0
2.9
lIT 1 103 (oK-I) rigure 4.
452
Temperature dependence (Arrhenius plots) for the racemization of amino acid residues in casein. Ref. 31.
0.9
9
1o
:1
0.' 0.7
;::
= Q
~ 0Z
i
C
...> ...... Q
0.6
6
0.5
5 V
0.4
Q 4 Z
C
'"
. 0
0.3 0.2 0.1
,
10
12
11
13
14
pH
Figure 5.
Effecl of pH 011 lhe relationship between the extent of hydrolysis of peptide bonds by trypsin in casein and the content of lysinoalan1ne and D-amino acids. Ref. 41.
0.65
asp-
0CJ¥
0.55 0.45
•
035
lO2!!J ..".8
j
glu-
0.15 Q05
0 -005 -0.15
-025
o Figure
0.2
Q4
0.6
CT*
6. Relationship between the inductive constant l«' "') of leucine, valine, phenylalanine, glutamic acid, and aspartic acid side chains of casein and the racemization rate constants relative to alanine tn 0.1 N NaOH at 65 0 C for 3 hr. Ref. 37.
463
1.0
PHE
o
..
GLU• ( . - ' • MET GLN ALA /TYR
LV'
/
• LYS
.ILE -1.0
/ . VAL
o
.5
1.0
(j"'*
Figure 1.
454
Relationship between the inductive constants (a*) of the amino acid side chains of 50ybean protein and the logarithms of the racemization rate constants (k) relative to that of alanine (kala)' Ref. 45.
BIOLOGICAL UTILIZATION OF D-AMINO ACIDS Two pathways are available for the biological utilization of D-amino acids (55-56): (a) racemases or epimerases may convert D-amino acids directly to L-isomers or to (DL) mixtures; or (b) D-amino acid oxidases may catalyze oxidative deam1nation of the a.-amino group to form a.-keto acids, which can then be specifically ream1nated to the L-form. Although both pathways may operate in microorganisms, only the latter activity has been demonstrated in mammals. The amounts and specificities of D-amino acid oxidase are known to vary in different animal species. In some, the oxidase system may be rate-limiting in the utilization of a D-amino acid as a source of the L-isomer. In this case, the kinetics of transamination of D-enantiomers would be too slow to support optimal gf'owth. In addition, growth depression could result from nutritionally antagonistic or toxic manifestations of D-enantiomers exerting a metabolic burden on the organism. The biological utilization of sulfur D-amino acids is conveniently discussed in terms of the transamination and transsulfuration pathways (F.Ig.8). D-methionine can be transformed to the L-isomer ~ia oxidative deamlnation followed by reamination. L-Methionine may be incorporated into proteins. It acts as the methyl group donor via S-adenosylmethionine; is a precursor for cysteine, cystine, and taurine; and participates in transamination reactions in which 3-methylthiopropionate is an intermediate. Not all of the details in the cited pathways, and their possible relationship to inherited metabo 1ic di seases such as cystathionuria and homocystinuria, have been completely elucidated (B). D-Methionine lhe reported wide variation in the biological utilization of sulfur D-amino acids for growth of various animal species may be related to the relative activities of D-amino acid oxidases (12,32, 55-57, 91).However, the reported relative oxidation rates of D-methionine by D-amino acid oxidase in kidney homogenates from man, monkeys, chickens, frogs, rats, and mice do not support this hypothesis. For example, the reported rate of oxidation of D-methionine by the oxidase in mouse kidneys is about one
third
to
one
half
of
corresponding
rates
for
rats
or man.
Nevertheless, we have observed that the nutritional utilization of D-methionine in mice is at least as good or better than reported bioavailabil1ties of D-methionine for either rats or man. In fact, D-methionine appears to be poorly utilized by human~ (12,54,75,91}.when consumed either orally or during total parenteral nutrition. One factor giving rise to inconsistencies regarding the utilization of D-methionine is the dose dependency (Figure 9) of the apparent potency of Dmethionine relative to its L-isomer, 1.~., the dietary concentration of the D-form for any given growth response relative to that of the L-form which would produce the same growth response. This dose-dependency is a result of the non-linear nature of the dose-response curves. This complicates attempts to compare results from some earlier studies with other animal species, which often report data based on a single substitution of the D- for the L-isomer. The extent of oxidation of D-methionine (and other D-amino acids) by D-amino acid oxidase may not be the limiting step governing
455
o-Methionine
I t
o-amino-acid oxidase OI-keto-'I'-Methyl-thiolbutyrate
+
transamination 3-Methylthloproplonate"- L-Methionlne -Proteins
~ + ATP S-Adenosylmethionina methyl transfetase ~ S-AdenosVJhomocystelna / "+ methyl,ted acceptor Homocysteine ~Homocystine cystathionine + Serine p-synthass
I
+
thi 'I'_cystBth:: \onina Inorganic Sulfur
Proteins
I
~
N_AceIVIcys.I.ln~tei\~ Cyst.ne
I
-
~
lanthionine .......
aminotransfetase
'I'-Glutamyl cycle cysteine dioxygfHIase
1
Mercaptopyruvate
Cysteine sulfinlc acid OXidBtiOn/
cysteine Bu/finic acid dacarboxy/ass
Cysteic acid
~ypotaurlne
--Sulfinylaldehyde
cysreic acid
decarboxylase
Taurine _Cholyltaurine
Figure 8.
Transamination and lranssulfuration palhways of D- and Lmethionine. Ref. 32. 80a. II
'-It
Figure 9.
456
0.1t
t.I7
Relationship of weight gain to percent L- or D-methionine in amino acid diets fed to mice. Ref. 32.
utilization. Other factors that could influence utilization include: rates of transport, action of intestinal enzymes and bacteria, rates of absorption, renal clearance, and possible toxic effects.
In discussing D-cystine, it is informative to review first the utilization of the L-isomer. Although L-cystine is not an essential amino acid for rodents, less L-methionine is needed for growth if the diet contains L-cystine (19, 32, 66, 82). The mechanism of this so-called sparing effect needs further clarification. One possibility is that L-cystine serves as a source of L-cysteine and other sulfur amino acids, thus minimizing the need for L-methionine to serve as a reservoir for these amino acids. Our results show that L-cystine has a low level of apparent methionine- like activity in mice (32). Growth response was 6.5% when it was substituted for an optimum level of L-methionine (1.17% in an amino acid diet devoid of any other sulfur amino acids). This may reflect a limited sparing effect for depletion of endogenous L-methionine, which may then be reutilized in higher priority pathways. L-cystine is somewhat ,"ore efficient in sparing D-methionine than L-methionine in amino acid diets containing a low level (0.29%) of L- or D-methionine (32). Supplementation of the D-isomer with an equal sulfur equivalent of L-cystine nearly doubled growth, bringing the overall response essentially equal to that produced by L-methionine in the presence of L-cystine. Since D-methionine must first be converted to the L-form before it can be utilized, these results illustrate a marked lessening in demand for l~thionine when L-cystine is available. Thus, even though D-methionine alone is somewhat less efficiently utilized in mice than is the L-isomer, the combination of L-cystine with either Lor D-methionine provides essentially equivalent nutritive value. In contrast, supplementation of an amino acid diet containing 25% of the optima I level (0.29%) of L-methionine with the same concentration of D-cystine increased growth from 58.7 to only 70.7% of the weight gain with optimal (1.17%) L-methionine. Increasing D-cystine to 1.11% reduced relative growth from 56.7 to 43.5%, which may imply that excess D-cystine in the diet is toxic. D-Cysteine Table 2 shows that L-cysteine had a sparing effect on L-methionine, but that D-cysteine did not. In fact, D-cysteine imposed a metabolic burden indicated by depressed growth when fed with less than optimal levels of L-methionine. The 24% decrease in weight gain of the D-cysteine-plus- L-methionine compared to L-methionine alone could mean that D-cysteine was nutritionally antagonistic or toxic. The mechanism(s) of the growth-depressing effects of D-cysteine and D-cystine remains to be elucidated, as do the reported toxic manifestations resulting from excess consumption of L-methionine. Cavallanl et al. (17) and Krijgsheld et al. (57) report that D-cystelne aprears lo be a more efficient precursor of inorganic sulfate in the blood of rats than L-cysteine. These findings prompted Krijgsheld et a1. to suggest that therapeutic administration of D-cysteine could be beneficial in cases of sulfate depletion. However, some caution may be in order, in view of our finding of a growth-depressing effect of D-cysteine in mice.
457
Lanthionine Isomers Lanthionine (S(CH2CH(NH2)COOH)2) isomers are formed during exposure of food and other proteins to alkali and heat (23-26, 41, 64), 72, 79). Such treatments generate dehydroalanine side chains from half-cystine and serine. Reaction of the SH group of cysteine and the double bond of dehydroalanine gives rise to one pair of optically active isomers (enantiomers) and one diastereoisomeric (meso) form. These fonns are correctly named as follows: (R)-L-lanthionine or S-[(2R)-2-amino ·2-carboxyethyl] L-cysteine (S)-D-Lanthionine or S-[(2S)-2-amino-2-carboxyet~yl]-D-cysteine (S)-L-lanthionine or S-[(2S)-2-amino-2-carboxyethyl]-L-cysteine (meso-lanthionine). Although food processing produces a mixture of all three isomeric lanthionines, it would be of interest to know the biological utilization of the individual isomers as a source of L-methionine. However, only the mixture of all three was available to us. Table 2 shows that DL + ~eso-Lanthionine has a moderate sparing effect on L-methionine, as evidenced by 27% greater weight gain ~Ihen the two amino acids were fed together, than from suboptimal L-methionine alone. The metabolic pathway for the utilization of lanthionine probably involves cystathionase-catalyzed transformation to fonn cysteine and pyruvic acid, analogous to the observed transfonnation of cystathionine to cysteine and tt-ketobutyric acid (32, 49). Stereochemical analysis of the cleavage shows that L-lanthionine will produce one molecule of L-cysteine; D-lanthionine, one molecule of D-cysteine: and ~-lanthionine, an equal mixture of both isomers. Complete cleavage of an equimolar mixture of all three isomers will, therefore, fonn an equimolar mixture of D-and L-cysteine, and only the latter, as our results have shown, will be utIlized nutritionally in the presence of L-methionine. D-Phenylalanine Table 3 and Figure 10 show that the relative growth of animals fed L-and D-Phe ranged from 28.3 to 81.3%, depending on the concentrations selected for comparison. Inspection of Figure 10 reveals that near maximum growth may be expected with sufficient D-Phe in the diet. The data suggest the absence of any marked antinutritional effects or toxicity from feeding either Phe isomer at twice the optimum dietary level. The dose dependence of the relative response of the two isomers in a range beyond that producing maximum growth with L-Phe complicates attempts to compare results from earlier studies with other speCies, which often report data based on a single substitution of the D- for the L-isomer. Nevertheless, the &2.1% relative response shown in Table 3 with a D-Phe concentration equal to the level of L-Phe In the complete amino add diet fonnulation (1.51%) is similar to the reported relative efficacy of &8% in rats and 75% in poultry (9,33). Although the semi-logarithmic plot shown in Figure 10 offers a useful graphic presentation of the relative potencies of the phenylalanine isomers, it does not limit correctly, ~., the point corresponding to 0% cannot be an extrapolation of the other data. An alternative way to represent the data would be to use the following simple exponential function which fits the data and limits correctly at both ends. 458
Table 1.
D-Aspartic acid content in commercial food products OIL ASP 0.095
D-ASP D-ASPtl-ASP 0.09
Baby formula (sov protein)
0.108
0.10
Simulated hacon (soy protein)
0.143
0.13
Corn chips
0.164
0.14
Dairy creamer (sodium caseinate)
0.208
0.17
Commercial Product Texturized SOY Drotein
Ref. 66.
Table 2.
Bioavailability of amino acid derivatives of l-cysteine wilh and without l-methionine in mice Relative Percent in Diet a
1 est
Substance
Mean Weight Gain b
Gain Relative to Methionine alone
(g)
(%)
L-Mr.thion1ne
None L-Cyste ine D-Cysteine Dl + Meso-lanthionine
100 100 100
_2L
f.~
4.0 -3.6 -4,2 -3.2
7.4 13.2 5.6 9.4
100
1"1B
76 127
a Relative 1001 is the molar equivalent of 1.17g l-methionine per 100g diet. b Weight gain after 14 days. Ref. 32, 34.
Table 3.
Percent in diet
Utilization of land D-phenylalanine for body weight gain in mice Relative percent in diet
Body Weight Gain a (g) l-Phenlllalanine
0 0.38 0.76 1. 51 3.02
0 25 50 100 200
-4.6 7.4 13.4 12.8 13.6
Relative Response for for D-Phenylalanineb (I) D-Phenlllalan1ne -4.6
-1.2 1.8 6.2 10.2
28.3 35.6 62.1 81.3
a Weight gain after 14 days. b Net weight gain for D-form divided by net weight gain for l-form, times 100. Ref. 33. 469
Table 4.
Weighl gain of mice fed an amino acid diet supplemented with D-tyrosine Diet
Percent D-Tyr in diet
D~lyr:L-Phe
0.76% L-Phe in diet
Wt.gain a g
o
o
0.42
1: 2
0.83
I :1
1.55
2: 1
13.5 13.5 8.8 1.5
1.51% L-Phe 1n diet
o
o
0.42 0.83 1. 55
1 :4 1: 2 1: 1
a Weight gain after 14 days.
13.8 14.3
15.0 10.0
Ref. 33.
Table 5 Weight gain of mice fed a casein diet supplemented with D-tyrosine Diet D-Tyr in diet (%) 10%
protein
o
1 :4 1: 2 1: 1
0.37 0.74 1.48
2:1
o
o
{).37
1 :4
1:2
0.'11
1.48
1: 1
2:1
2.97
a Weight
gain after 14 days.
Mean wt gaina (g)
o
0.19
460
D-lyr:L-Phe
Ref. 33.
8.8 9.7 9.0
10.2 2.2
14.7 13.H
13.8 10.8 7.0
Assume that, 13.6 - gain 13.6 - (-4.6)
=
e -KP
(3);
1n
13.6 - gain 13.6 - (-4.6)
-KP
(4)
where: 13.6 is the maximum weight gain with L-Phe (lable 3); gain is the weight gain with 0 and 25% L-Phe and with all concentrations of D-Phe; -4.6 is the weight gain (loss) with 0% Phe: K is a constant: and P is the relative % of either Phe isomer (35). Equation 4 predicts that a plot of the left-hand ptlrt against P should give a straight line with 01 s lope of -K. Such plots (Figure 11) based on the data in Table 3 yield a value of K "0.0428 for l-Phe and 0.00844 for D-Phe. Thus, a measure of the relative potency of D-Phe compared to that of L-Phe, within the range of the growth response observed for the D-isomer, is 0.00844/0.0428 or approximately 20%. Other workers have reported the following related observations. Healthly young adults can invert about one-third of D-Phe to its l-isomer (86). When a young human female ingested deuterium-labeled 0and l-Phe (8 mglkg body weight), the alllino acid appeared more rapidly .lnd abundantly and disappeared more slowly from plasma with the 0- than with the L-isomer (58). Although D- and L-isomers may be absorbed at the same rate from the intestine of rats, tissue accumulation of the isomers varies (88). The D- and l-Phe isomers differ in their ability to stimulate pancreatic secretions in dogs (67). lhe presence of other D-ami no acids decreases the ut il i zalian of D··Phe, presumably because of the increased competition for available 0 amino-acid oxidase needed to transform D- to l-isomers (56). Finally, D-Phe produces analgesia in mice and humans by potent iating the action of enkephalins in the brain through inhibition of the enzyme that degrades brain opiolds (6). Metabolic studies of Phe, including the data shown in Figure 10, indicate that the D-isomer, because it must first be inverted, is utilized more slowly. This leads to the speculation that D-Phe might not have the same effect on phenylketonuria In children as the L-isomer (1). Should this be the case, the slow and uniform release of L-Phe from the O-isomer could possibly provide a basis for the utilization of D-Phe in the treatment of phenylketonuria. Animal and human studies are needed to demonstrate this possibility. D-Tyrosine The amino acid L-Tyr is an in vivo precursor for brain catecholamine, dopamine and norepinephrine; for the biogenic amine tyramine, and for the ubiquitous pigment melanin (7). In !,itro, the phenolic hydroxyl group activates the benzene ring of tyrosine, making it more sl1!;ceptible to chemical and radiation-induced modification than the corresponding ring in phenylalanine (40). The inductive nature of the phenolic group enhances such food processing factors as high pH and temperature to bring about the rapid racemization of L-Tyr residues in proteins (36, 59). D-Tyr has been shown to prevent and reverse stress-induced chronic hypertension, which may indicate possible effects on the central nervous system (80). Nutritionally, L-Tyr is classified as a semi-essential amino acid since it Is synthesized in vivo from L-'Phe (72). Combinations of L-Tyr and L-Phe are complementary in supporting the growth of mice (33) Thus, under conditions where L-Phe may be limiting, L-Tyr can supply 481
I. 1.
'2 10
..!!!
z
:c •
D•
PhlNlylalonln_
0
~
::I:
8~ ·2
U
1.4
1.5
1.6
1.7
1.1
l.t
2.0
2.1
2.2
2.3
2.A
LOG PHENYLALANINE IN DIET
figure 10. Growth response of mice fed 0- and L-phenylalanine.
Ref. 33.
1.0 0.9 0.1 0.7
0.6 C
0.5
1i
CII
l:CII .......
•~
1 ......
• •
c.; c.;
0.4
\ L.pHENYLALANINE
D.PHENYLALANINE
0 0.3
0.2
\
\ \
o 100 200 RELATIVE PERCENT IN DIET (P in equation 4)
Figure 11
462
Growth reponse of mice fed 0- and L··pheny1alanine.
Ref. 35 ..
about half the requirement of L-Phe alone, a value similar to those previously reported for humans, chicks, and rats. Table 4 shows that with D-Tyr in an amino acid diet, growth inhibition was severe at a D-Tyr:L-Phe ratio of 2:1, but was much more moderate when the ratio was 1: 1. Similar results were obtained wi th a casein based diet supplemented with D-Tyr (Table 5). One or more of the following six mechanisms may explain this inhibition of weight gain by dietary D-Tyr: Tyrosyl ribonucleic acid synthethase was found to catalyze the incorporation of D-Tyr into tyrosyl-tRNA, an aminoacyl adenyl ate derivative similar to that formed by L-Tyr (16, 90). Since the structural features of an amino acid after combining with tRNA do not seem to be significant in the specificity of its incorporation into protein, D-Tyr ingestion should lead to the formation of faulty peptides and proteins. These, in turn, could interfere with normal metabolic processes and be responsible for the observed growth inhibition. A related possibility is the suppression of normal protein synthesis by D-Tyr through competitive inhibition of L-Tyr or L-Phe incorporation into aminoacyl-tRNA, and thus, into proteins. A third possibility is an interference by D-Tyr in the biosynthesis or biological action of vital neurotransmitters such as dopamine. The observed hypotensive effect of D-Tyr is postulated to involve this effect (80). A fourth possibility is hydroxylation of L-Phe to L-Tyr.
an
interference
by
D-lyr
in
the
A fifth possibility is that D-Tyr at concentrations provided in the present study could overload metabolic pathways needed to eliminate or detoxify excess D-amino acids. However, when the dietary level of L-Phe was increased, these same concentrations of D-Tyr failed to bring about marked growth inhibition. A sixth possibility is competition of D-tyrosine with L-tyrosine or L-phenylalanine for membrane transport (4). A discussion of membrane transport of amino acids, which appeared in an article in Nutrition Reviews (4) is relevant to the theme of this paper, and accordingly it is quoted here: "The catalytic transport of amino acids across the plasma membrane tends to have lower structural specificity than is usually characteristic of enzymatic reactions. Furthermore, the stereospecificity varies among amino acid transport systems and among membranes. In some cases, stereospecificity is high, as appears to be characterist1c of the inner mitochondrial membrane. A study with the Ehrlich ascites tumor cell demonstrated differences in selectivity. In that study, there was only about threefold preference for L-methionine over its D-antipode. Furthermore, another study showed that a transport system that was able to discriminate between glutamic acid antipodes had little ability to discriminate between aspartic acid antipodes. "Associated with this low stereoselectivity of transport is the familiar circumstance in higher animals that the transport of amino acids is mainly carried out by 'public' systems, able to handle a wide range of ami no acids, rather than by the occas iona 1 'pri vate' systems restricted to a single amino acid. The probability is very high that not only phenylalanine and tyrosine, but also tryptophan, the 463
branched-chain amino acids and several others have a large proportion of their inter-organ flows mediated by a common transport system designated L. This possibility occasions the currently re-emphasized suspicion that in untreated phenylketonuria, the high ci rculating L-phenyla lanine levels interfere to critical degrees with the passage of several important amino acids across the blood-brain barrier. Tews et a1. (85) have given special attention to analogue inhibition of amino acid transport across this barrier, arising from dietary amino acid imbalances. "In connection with competition for membrane transport, Friedman and Gumbmann (33) are correct in focusing attention on the interference of D-tyrosine with L-phenylalanine as well as with L-tyrosine metabolism. They cite a study pertinent to possible competition by D-tyrosine with transport of L-amino acids. When a young woman ingested deuterium-labeled D-and L-phenylalanine (B mg/kg body weight), the D-isomer appeared more rapidly and disappeared more slowly from the plasma than the L-isomer (58). The higher tolerance curve for the 0than for the L-tyrosine brings attention to the important differences in the rates at which 0- and L-tyrosine are withdrawn from the circulation. In the case of D-phenylalanine, but not for D-tyrosine, inversion to the L-form limits and compensates for the competitive effects arising from elevated levels of the D- form". ·Although the specific mechanism by which D-tyrosine becomes an antimetabolite for the mouse is still not clear, these findings warn investigators to avoid the use of DL-tyrosine in experimental and clinical nutrition. They also dramatize the competition between enantiomorphs of some amino acids. In the case of D-tyrosine, the precise mechanismas can be elucidated only by further study." In conclusion, the cited data demonstrate that D-Tyr, unlike L-Tyr, has no sparing effect for L-Phe. In fact, a metabolic stress, in the form of growth inhibition in mice, may become evident when D-Tyr is present in the diet at equal or greater molar concentrations to L-Phe.
This acute effect remains to be defined toxicologically, and the potential for sub-chronic and chronic toxicity following exposure to lower levels of D-Tyr remains unknown. D-Tryptophan The relative potency of D-tryptophan (as defined in Tables Ii and 7) compared to the L-isomer was strongly dose-dependent, being inversely related to dietary concentration and ranging from 29% to 1i4%. A plot of this data showed that growth was linearly related to D-tryptophan. The maximum growth obtainable for L-tryptophan occurred at or slightly less than 0.174% in the diet. By increasing the dietary concentration of D-tryptophan up to 0.52% (three times the highest level shown in Table Ii), it was possible to demonstrate for D-tryptophan that growth also passed through a maximum, one which equalled 82% of that achieved with the L-isomer. This occurred at approximately 0.44% D-tryptophan for a relative potency of 25% (data not shown). This level in the short term assay appeared to be well tolerated. However, additional studies are needed, both In mice and other species, to define the effects of more prolonged dietary exposure to D-trytophan.
464
Table £I.
Growth of mice fed 0- and 1-tryptophan
Amino acid tested
Percent Mean weight In diet gain (g)a
L-tryptophan
0 0.01'2 0.044 0.087 0.174
-3.5 + 0.29 -1.8:;: 0.37 2.2 0.49 12.4 + 1.25 15.& 0.&8
0 0.022 0.044 0.087 0.174
-3.5 -2.8 -1.8 -0.4 3.8
D-tryptophan
Percent equivalent In dlet b
Relative potency (%)c
± ±
± 0.29
+ 0.49 ;. 0.37 + 0.81 0.58
±
a Weight gain after 14 days. Range tes l.
N = 5.
0.014 0.024 0.033 0.051
&4
55 38 29
Means ± S.E. Duncan's Multiple
b Percent equivalent in the diet i~ thilt concentration of L-amino acid which would produce the same growth ohserved for the D-amino acid (dp.tp.rmined graphically). c Re1alive potency Is the percent equivalent in the diet divided by the percp.nt actually fed X 100.
Table 7. Relative percent In diet a
o
50 100 150 200 250 300
Growth of Mice fed D- and L-Tryptophan for 14 days 80dy weight gain (g)b D-tryptophan
L-tryptophan
-2.8 ± 0.4 0.2 + 0.& 4.4 +1.5 10.& ;: 1.3 11.4 +1.0 12.2+1.2 11.4 1. 1
-2.8 ± 0.4 14.5 ± 0.4
±
(N
= 5).
Relative weight gain (D/L)
17 .3
41.& 77 .5
82.1 8&.7 82.1
a 100% relative equals 0.174% in thp. diet of either D or L-tryptophan. b Means ± S.E. Duncan's Multiple Range test.
465
Considerable species, variation is known to exist for the nutritive value of D-tryptophan. In chicks fed amino acid diets, the relative biologital activity of the 0- to l-lsomer has been, reported to be 20%, which is similar to the relative potency we find for mice (61,70,71) This 1S in marked contrast to rats als.o fed amino acid diets where the stimulation of growth was equivalent to that produced by L-tryptophan, for a biological actIvity of 100%. It should be noted that O-tryptophan supplementation of a protein based diet containing 0.085% protein-bound L-tryptophan resulted in growth stimulation in mice nearly equal to that produced by L-trytophan supplementation. Plasma analysis indicated that the rate of conversfon of 0 to L closely reflected the difference in response between chicks and rats. Essent ia lly no convers ion of D-tryptophan to the L-isomer could be detected in chick plasma. and most of the D-isomer administered was excreted unchanged. In the rat. plasma levels of the L-isomer rapidly increased upon administration of D-tryptophan, and only lJ of the dose was excreted in the urine. Lysinoalanine Isomers Lysinoalanine (HOOCCH(NHl)CH2CH2CH2CH2NHCH2CH(NH2)COOH,(LAL». is an unnatural amino acid (formed as shown in figures 12-14) has been identified in hyrolyzates of processed food proteins. in particular those subjected to alkali (19, '21,25,27,31,38,39,41,42,44,46, 47, 53, 60, 68, 74, 81. 83, 89). Lysinoalanine has two asymmetric carbon atollls. Four stereoisomers are therefore possible: LL. LO. OL. and ~O. The following systematic name is now for the isomer formed in the greatest quantity during food processing (25): N e: -(mc2-~2-carboxyethyl)-trlYSine.
A more correct name with R and lysine would be: Nfi
~
designation for the substituents of
-l(2I!S)-2-ami~2-carboxyethylJ -!:
lysine.
In analogy with the two isomers derived from L-lysine. O-lysine would give rise to the following correctly named diastereoisomers:
~- [(2I!S)-2-amino-2-carboxyethyll-Q-lYSine. It is also worth noting tbat the reaction between the t-NH2 group of lysine and the double of dehydrothreonine can, in principle, give rise to five symmetric' centers and 32 stereoisomers (25). In rats, histological changes in the kidneys have been identified which are related to dietary exposure to this substance, either isolated or as part of intact proteins. The lesions are located in the epithelial cells of the straight portion of the proximal renal tubules and are characterized by enlargement of the nucleus and cytoplasm, increased nucleoprotein tontent, and disturbances in DNA synthesis and mitosis. 8ecause of these observations. contern has ari sen about the safety of foods which may contain LAL and related dehydroalanine-derived amino acids which are known to produce similar lesions. However, since the mechanism by which these compounds damage the rat kidney is unknown, it 1s difficult to assess the risk to human health caused by their presence in the diet. 488
LAL has two asyrrmetric carbon atoms making possible four separate diastereoisomeric forms, LL, LO, OL, and ~O. Its structure suggests that it should have excellent chelating potenial for metal ions, a property which may have relevance to its toxic action. Accordingly, we have examined LAL (a mixture of the LL and 1.0 isomers) for its affinity towards a series of metal ions, of which copper (II) was chelated the most strongly (39, 42,74), On this basis, we have suggested a possible mechanism for kidney damage in the rat involving LAL's interaction with copper within the epithelial cells of the proximal tubules.
"
OH~ ,I
I
CH,-C-P I 1 Y N=C-P
Y
1l
1~ ~® H I
CH,..:C-P I I Y N-C-P
H!t.oSHIFT
>
,.....1
-..y
~o) CH,-C!..P NH-C-P
II
!H@ 0
( s..
CIJIUNION INTERMEDIATE
I
NH-C-P
2. P·ELlMINATION 18)
!(-Y$~
P·HlMINATION IA)
I:;> I
Y
I RACEMIZATION
I
NH-C-P
CH, -C-P
OH
