ARCHIVES OF BIOCHEMISTRY AND BWHYSICS Vol. 269, No. 1, February 15, pp. ZOl-20’7,1989
of Pyruvate Carboxylase by Sequestration Coenzyme A with Sodium Benzoate’
ANN D. GRIFFITH, AND Depadment
DOUGLAS M. CYR, SHARON GEORGE C. TREMBLAY2 and Biophysics, 117 Merrill Hall, University Kinqstcm, Rhode Island 0288 1
G. EGAN, of Rhode Island,
Received August 29,19X8, and in revised form October 21,198X
Pyruvate-dependent CO2 fixation by isolated mitochondria was strongly inhibited by sodium benzoate. Pyruvate carboxylase was identified as a site of inhibition by limiting flux measurements to assays of pyruvate carboxylase coupled with malate dehydrogenase. Benzoate reduced pyruvate-dependent incorporation of [14C]KHC03 into malate and pyruvate-dependent malate accumulation by 74 and 72%) respectively. Aspartatedependent malate accumulation was insensitive to benzoate, ruling out malate dehydrogenase as a site of action. Inhibition by benzoate was antagonized by glycine, which sharply accelerated conversion of benzoate to hippurate. Assays of coenzyme A and its acyl derivatives revealed inhibition to correlate with depletion of acetyl CoA and accumulation of benzoyl CoA. Depletion of acetyl CoA was sufficient to account for >50% reduction in pyruvate carboxylase activity. Competition between acetyl CoA and benzoyl CoA for the activator site on pyruvate carboxylase was insignificant. Results support the interpretation that the observed inhibition of pyruvate carboxylase occurred primarily by depletion of the activator, acetyl CoA, through sequestration of coenzyme A (0 1089 Aeademie Press, Inc. during benzoate metabolism.
The capacity of adult human liver for the conversion of sodium benzoate to its glytine conjugate, hippurate, is prodigious. Lewis observed that 10 g of sodium benzoate, administered orally, could be recovered quantitatively as urinary hippurate within 6 h (1). Since glycine can be synthesized by the body, it was suggested that administration of benzoate might combat ammonia toxicity by promoting the disposal of substantial quantities of waste nitrogen as hippurate (2). Indeed, the administration of benzoate has moderated episodes of hyperammonemia in infants and children with genetic defects in the urea cycle (3). Our interest in benzoate-induced alterations in metabolism began with the ob-
servations that, contrary to expectation based on clinical reports, benzoate failed to counteract ammonia accumulation in suspensions of hepatocytes (4), and actually exacerbated ammonia toxicity in laboratory animals (5,6). Since glycine is limiting in the conversion of benzoate to hippurate (4, 7, 8), accumulation of an intermediate might account for potentiation of ammonia toxicity by benzoate (5). Accumulation of henzoyl CoA was subsequently observed in experiments with isolated hepatocytes (9) and intact animals (10). Sequestration of coenzyme A could result in potentiation of ammonia toxicity by interfering with ureagenesis at two intramitochondrial sites dependent on acetyl CoA: (i) synthesis of N-acetylglutamate, an essential activator of the ammonia-dependent carbamoyl-phosphate synthetase (EC 126.96.36.199) and (ii) activation of pyruvate carhoxylase (EC 188.8.131.52), a major source of the carbon
’ This work was supported by a grant (DK33536) from the National Institutes of Health, U.S. Public Health Service. ’ To whom correspondence should be addressed. 201
0003.9861/89 $3.00 Cngyriyht All rights
c\ 1%3!l hg Academic Press. Inc of reproduction in any form reserved
chain for regeneration of aspartate. Both sites are localized with the enzymes for hippurate synthesis in the mitochondrial matrix. Results of studies on the influence of benzoate metabolism on pyruvate carboxylase activity are reported herein. MATERIALS
Muteri&. Radiolabeled chemicals and Aquasol liquid scintillation cocktail were purchased from New England Nuclear (Boston, MA) solvents from Fisher Scientific (Medford, MA) and all other chemicals from Sigma Chemical (St. Louis, MO). Male SpragueDawley rats (COBS-CD) were obtained from Charles River Laboratories (Wilmington, MA) and maintained on Purina Rat Chow No. 5012 ad lib&am. The rats weighed 250-300 g and were fasted 18-22 h before use. Assay for pyruvate carboxylase. Rats were decapitated and the liver quickly homogenized in 10 vol of ice-cold 0.28 M mannitol made 1 mM in EDTA. The pellet from centrifugation at 800s for 5 min was discarded; mitochondria were harvested from the supernatant fluid and two resuspensions by centrifugation at 97709 for 10 min. A suspension of washed mitochondria at 40-60 mg protein/ml homogenizing medium was used to assay pyruvate-dependent incorporation of [‘4CJKHC03 into acid-stable radiolabeled product over 10 min at 30°C in the reaction mixture employed by Stirk et al (11). Reactions were terminated with 0.5 ml 1.5 N HCl. The acidified mixture was baked to dryness over a steam bath, the residue extracted with 2 ml water, and the extract diluted with 6.5 ml Aquasol for determination of acid-stable radiolabeled product. In assays with mitochondrial extracts, the washed mitochondrial pellet was suspended in water to yield 20-30 mg protein/ml and sonieated in an ice bath with three pulses of 20 s each. Insoluble matter was removed by centrifugation at 140,OOOgfor 30 min and a 1:5 dilution of the supernatant fluid in water was assayed for pyruvate-dependent incorporation of [14C]KHC03 into acid-stable product in 2 ml containing: ATP, 4 mM; Hepes3 buffer, 0.1 M, pH 7.8; sodium pyruvate, 10 mM; MgSO,, 10 mM; acetyl CoA and benzoyl CoA, as indicated; NADH, 0.25 mM; malate dehydrogenase (EC 184.108.40.206), 20 units; [14C]KHC03, 10 mM, 8 &i; enzyme preparation, 100 pl. Malate accumulation. Accumulation of malate was determined by measuring appearance of NADPH upon incubation of neutralized aliquots of the acidsoluble fraction of the assay mixture for pyruvate carboxylase with Tris buffer (0.1 M), MgCl, (4 mM),
’ Abbreviation used: Hepes, 4-(2-hydroxyethyl)-lpiperazineethanesulfonic acid.
ET AL. NADP (1.67 mM), and malic enzyme (EC 220.127.116.11, 1 unit) in 3 ml at pH 7.6. Enzyme-dependent generation of NADPH was recorded speetrophotometrically at 340 nm until equilibrium was reached. Generation of NADPH was stoichiometrically equivalent to known amounts of added malate in the range of concentrations reported. Benzoate (1 mM) did not interfere with the assay. Liberation of$xed CO,from malate. Incubation conditions were the same as those described for assay of pyruvate carboxylase in isolated mitochondria, except the reaction volume was doubled. Unreacted [‘4C]C0, was completely removed from the acidified reaction mixture by distillation into 20% KOH in a desiccator under vacuum at room temperature for 48 h. The acid-soluble fraction was then adjusted to pH 7.6 with KOH and diluted to 5.5 ml with water. Equal aliquots were assayed for malate content (described above) and radioactivity released as [‘4C]C02 by malic enzyme under the same conditions. For recovery of [14C]C02, the reaction mixture was contained in a 20ml scintillation vial sealed with a rubber stopper through which was suspended a plastic center well containing fluted filter paper moistened with 0.3 ml of 20% KOH. When the companion spectrophotometric measurement indicated that the reaction was at equilibrium, 0.3 ml 55 mM KHC03 was injected through the stopper into the reaction mixture to provide carrier for [‘4C]C0, distillation, immediately followed by 0.3 ml 6 N HCl. The [‘4C]C0, released by acidification was allowed to distill into the center well for 30 min at 37”C, with shaking. The center wells and contents were transferred to scintillation vials and rinsed with 2 ml water, and the wells and rinsings suspended in the gel formed upon addition of 6.5 ml Aquasol for assay of radioactivity. Assay for hippurate synthesis. Mitochondria were incubated under the same conditions employed for assay of CO2 fixation except that [l-‘4C]benzoate (1 mM, l-2 pCi) was substituted for radiolabeled bicarbonate (carrier bicarbonate was included at 20 mM). Radiolabeled hippurate was isolated from the neutralized acid-soluble fraction of the reaction mixture by cocrystallization with carrier (4). Assay of coenzyme A and derivatives. Assay of coenzyme A and acyl CoA derivatives followed the procedure of Corkey et al. (12). Reaction mixtures were cooled in ice for 1 min and the mitochondria were isolated by centrifugation at 13,SOOgfor 30 s. The mitochondrial pellet was acidified with ice-cold HC104 to 0.3 M, and finally brought to pH 4-5 with KOH. The supernatant fluid obtained by centrifugation was analyzed by HPLC using a Beckman Model 332 system equipped with a &-silica (ODS) guard column (37-53 pM particle size) and a 0.46 X 25-cm ODS analytical column (5 yM particle size). The uv detector was set at 254 nm. Aliquots of 50 ~1 were assayed for benzoyl CoA by using a phosphate buffer (0.2 M, pH 4.0):methanol (6.5:3.5) mobile phase at a flow rate of 0.5 ml/
INHIBITION OF PYRUVATE CARBOXYLASE BY SODIUM BENZOATE
and strongly inhibited by 1 mM benzoate (Fig. 1). Maximum inhibition of CO2 fixaASSAY OF PYRUVATE CARBOXYLASE IN tion was observed at concentrations of ISOLATED MITOCHONDRIA benzoate as low as 0.20 mM and the inhibiIncorporation of [14C]KHC03 tion was largely prevented by the addition of glycine (Fig. 2). Consistent with a mechinto acid-stable product (nmol/mg protein. min) anism that entails sequestration of coenAddition or deletion zyme A, glycine was shown in separate as26.9 AZ0.8 None (control) says to stimulate hippurate synthesis in - Pyruvate + malate isolated mitochondria some 30-fold, pre0.7 -+ 0.2 (5 mM) sumably liberating free coenzyme A in the + Phenylpyruvate process (Table II). Restoration of CO2 fix11.7 + 1.3 (2 mM) ation was specific in that glycine prevented 7.6 +0.5 + Glyoxylate (3 mM) inhibition by benzoate, but not by val+ Cyanohydroxyproate (Table II). Valproate is known to se0.5 -+ 0.1 cinnamate (5 mM) quester coenzyme A (16), but it is not meNote. Mitochondria (4-6 mg protein) were incutabolized to a glycine conjugate (1’7). bated 10 min at 30°C in 2-ml reaction mixtures, pH Direct measurements of coenzyme A 7.5, of the following composition: KCl, 120 mM; potasand its acyl derivatives confirmed the sium phosphate, 2.5 mM; Hepes buffer, 10mM;sodium above interpretations. Addition of benzopyruvate, 5 mM; [‘“C]KHCOa (0.03-0.3 PCi), 20 mM, and ate (1 mM) to the reaction mixture for asadditions or deletions as specified. Values are aversay of pyruvate carboxylase caused an avages Y!ZSE (n = 4). erage decline of 53 and 61% in the levels of free coenzyme A and acetyl CoA, respectively, and 57% of this loss could be accounted for in accumulated benzoyl CoA min; benzoyl CoA was eluted in 20 min. In a separate assay of the same reaction mixtures, the buffer:meth(Table III). Glycine was without signifianol ratio was changed to 9:l and the flow rate to 1.5 cant effect when added alone, but it largely TABLE I
ml/min to elute free coenzyme A and acetyl CoA at retention times of 12 and 38 min, respectively. Peaks were identified by cochromatography with authentic standards.
RESULTS Evidence that incorporation of [‘“ClKHC03 into acid-stable radiolabeled product was a measure of pyruvate carboxylase activity in intact mitochondria is shown in Table I. Incorporation was fully dependent upon the addition of pyruvate. Substitution of an alternative energy source (malate) resulted in a 97% reduction in formation of acid-stable radiolabeled product. Incorporation was reduced 55% by phenylpyruvate, a known inhibitor of pyruvate carboxylase (13) and 71% by glyoxylate, which has been shown to promote CO, transfer from the carboxybiotin moiety of the enzyme to water (14). Cyanohydroxycinnamate, an inhibitor of the transport of pyruvate into the mitochondrial matrix (15), reduced product formation by 98%. Fixation of CO2 by isolated mitochondria was linear for 15 min and promptly
( min )
FIG. 1. Kinetics of pyruvate-dependent COzfixation. Reaction conditions were the same as control conditions described in Table 1 except that benzoate was added as shown and the reaction was terminated at the times indicated.
1.0 ( mM )
FIG. 2. Inhibition of pyruvate-dependent CO, fixation by sodium benzoate. Reaction conditions were the same as control conditions described in Table I except that benzoate and glycine were included as indicated.
prevented the changes brought about by benzoate. While CO, fixation is clearly dependent on pyruvate carboxylase activity (Table I), the acid-stable product measured is a mixture of citrate, malate, and fumarate (18, 19). Accordingly, sequestration of coenzyme A could reduce formation of acid-stable product at one or both of two reactions
dependent upon acetyl CoA: pyruvate carboxylase and citrate synthase. To establish inhibition at pyruvate carboxylase with certainty, an alternate assay was employed in which only the CO2 fixed by the coupled action of pyruvate carboxylase and malate dehydrogenase was measured. This was accomplished by quantifying the radioactive CO2 released when the neutralized acid-soluble fraction from assay of pyruvate carboxylase was subjected to a second incubation period with NADP and malit enzyme to liberate COZ fixed into malate. Malate generated from citrate via the reactions of the tricarboxylic acid cycle would not be radioactive because the radiolabeled carbon introduced at citrate synthase would be lost as [‘4C]C0, in the cycle. The assay underestimates pyruvate carboxylase activity, but allows unambiguous interpretation of the effect of benzoate on the enzyme. As was observed for CO, fixation, malate accumulation was dependent upon added pyruvate (Table IV). Malate accumulation corresponded to 39% of the total pool of acid-stable radiolabeled product resulting from pyruvate-dependent CO2 fixation. When benzoate was included in the reaction mixture, malate accumulation was reduced by 72%) CO, fixation into acid-stable product was inhibited by 64%, and that fraction liberated as [‘4C]C02 by subsequent incubation with malic enzyme was depressed 74%. Since the specific activity of the malate pool was not known, values for [14C]C02 liberation
None Benzoate Valproate
26.9 +- 0.8 8.6 + 0.4 9.6 t 0.6
26.3 310.9 21.4 -I- 1.0 10.2 t 0.6
0.17 10.02 -
5.6 f 0.7 -
Note. Reaction conditions are described in the legend to Table I. Benzoate and valproate were added reaction concentration of 1 mM, where indicated. Bicarbonate was 20 mM throughout, but [‘%]KHCOa replaced with [1-‘%]benzoate (l-2 &i) when hippurate synthesis was measured. Radiolabeled hippurate isolated from the acid-soluble fraction by cocrystallization with carrier. Values are averages f SE (n = nanomoles precursor incorporated into product per milligram mitochondrial protein per minute.
at a was was 4) in
THE EFFECT OF BENZOATE ON COENZYME A LEVELS IN ISOLATED MITOCHONDRIA
Addition None Benzoate (1 mM) Benzoate and glycine (1 mM) Glycine
1.88 k 0.20
0.36 f 0.04
0.89 f 0.06
0.14 + 0.03
0.69 f 0.09
2.02 f 0.06 1.93 (n = 2)
0.29 + 0.04 0.33 (n = 2)
0.16 f 0.02 N.D.
Note. Reaction conditions were the same as control conditions described in Table I except that benzoate and glycine were added as indicated. Procedures for separation and quantification of coenzyme A and acyl derivates by HPLC are given under Materials and Methods. Values are averages k SE (n = 3) in nanomoles/per milligram mitochondrial protein. N.D. = not detectable.
are given in dpm/mg mitochondrial protein. It is also germane that, to the extent that malate equilibrates with fumarate, randomization of the radiolabeled carbon leads to an underestimation of the COZ fixed, as determined by assay with malic enzyme. Previous work indicates the reaction is at equilibrium and the label equally distributed between the C1 and C4 carboxyl groups (18). These results demonstrate that benzoate inhibits product formation
by the coupled action of pyruvate carboxylase and malate dehydrogenase. As with other measures, glycine provided protection against benzoate toxicity. The unlikely possibility that benzoate acts at malate dehydrogenase was ruled out with measurements of aspartate-dependent malate accumulation (glutamic oxalacetic transaminase coupled to malate dehydrogenase). Accumulation of malate (average t SE in nmol/mg protein; n = 3) when aspartate (5 mM) and a-ketoglutarate (5 mM) were substituted for pyruvate was 101 + 10 in the absence of benzoate and 109 + 15 in its presence (1 mM); 90% of this value was dependent upon addition of aspartate. Whether the mechanism of inhibition of pyruvate carboxylase by benzoate might include competition between benzoyl CoA and acetyl CoA for the acetyl CoA binding site was examined in assays with mitochondrial extracts. Benzoyl CoA at reaction concentrations of 150 and 512 yM did not significantly influence activation of pyruvate carboxylase by acetyl CoA at concentrations ranging from 5 to 400 pM (data not shown); these ranges include ratios of benzoyl CoA to acetyl CoA reported herein (Table III). DISCUSSION
In studies with isolated hepatocytes, McCune et al. (20) observed that benzoate in-
LIBERATION OF COa FIXED INTO MALATE BY INCUBATION WITH MALIC ENZYME
None (control) - Pyruvate + Benzoate + Benzoate and glycine
Malate accumulation (nmol/mg protein) 60 f 3 N.D. 17f3 39 i- 3
[‘4C]C02 fixed (nmol/mg protein) 154 i21 1.2t 0.2 56 + 12 111 f16
[%]CO, liberated (dpm/mg protein) 2029 f 210 38+ 4 518k 33 1212 + 140
Note. Malate accumulation and [‘%]CO, fixation occurred under the conditions for assay of pyruvate carboxylase described in the legend to Table I. Benzoate and glycine were added at 1 mM, where indicated. Liberation of [‘%]CO, fixed during the first incubation period was determined by incubating the neutralized acid-soluble fraction with NADP and malic enzyme during a second incubation period. Similarly, malate content was assayed by measuring NADPH generation during the second incubation period. Values shown are averages k SE (n = 3); N.D. = not detectable.
hibited gluconeogenesis from lactate, and suggested interference with pyruvate carboxylase through sequestration of coenzyme A as a possible mechanism. Subsequently, the same mechanism was proposed to explain inhibition of the urea cycle and the orotate pathway for de ncwo pyrimidine biosynthesis (5, 9), with evidence for accumulation of benzoyl CoA (9). In the present study, direct evidence is reported for inhibition of pyruvate carboxylase by sodium benzoate, with an associated accumulation of benzoyl CoA and depletion of acetyl CoA and free coenzyme A. The values reported for mitochondrial free CoA and acetyl CoA in the absence of benzoate (Table III) are similar to those of Corkey et al. (1.95 and 0.57 nmol/mg mitochondrial protein, respectively), whose assay we employed (12). The K, value for acetyl CoA binding to pyruvate carboxylase in intact mitochondria has been estimated at 210-290 pM (21). With the same assumption of 0.8 ~1 water/mg mitochondrial protein used by Von Glutz and Walter (21), our results (Table III) yield an intramitochondrial concentration of acetyl CoA of 450 yM in the absence of benzoate and 175 pM in its presence. Since the higher concentration is at or very near saturation for acetyl CoA (21), it may be concluded from these observations that the benzoate-induced depletion of acetyl CoA reported herein can account for a loss in pyruvate carboxylase activity in excess of 50%. The additional possibility that benzoyl CoA might compete with acetyl CoA for the activator site on pyruvate carboxylase was ruled out in assays with mitochondrial extracts. Whether inhibition of pyruvate carboxylase alone is responsible for potentiation of ammonia toxicity by benzoate remains to be established. Interference with ureagenesis at N-acetylglutamate synthetase is an equally plausible site of action (5, 9) and a recent study revealed that benzoate depressed hepatic levels of N-acetylglutamate in mice challenged with ammonia (22). The relative importance of these two sites of action requires further study. Clinical use of benzoate in the treatment of hyperammonemia preceded cognate studies with laboratory animals. In tests with laboratory animals, it was observed
that benzoate can potentiate ammonia toxicity, impede ammonia removal by the urea cycle, and inhibit gluconeogenesis and pyrimidine biosynthesis (4-6, 9, 20, 22). Inconsistencies between reports on the clinical effectiveness of benzoate and laboratory findings are unresolved. The dose of the drug may be quite critical within a fairly narrow range (5, 23) and nitrogen load may also be a deciding factor. To date, benzoate therapy has been restricted to treatment of hyperammonemia associated with urea cycle defects, in which case the nitrogen load is carefully restricted, and a major target of benzoate toxicity is genetically impaired. The laboratory findings support caution in extending benzoate therapy to the treatment of hyperammonemia associated with other diseases. Sequestration of coenzyme A as benzoyl CoA is an indication of an inadequate supply of glycine for hippurate synthesis. The ability of the body to replenish the glycine spent in hippurate synthesis is poor (10,24, 25), and it has been concluded from one study that supplementation with nitrogenfree analogs of glycine is required to effectively utilize hippurate synthesis as an alternative pathway for the disposal of waste nitrogen (26). Piridoxilate, a cardiac drug available in Europe, was proposed for such use in combination with benzoate (26). The drug is an adduct of pyridoxine and glyoxylate, and it is cleaved by the body to yield glyoxylate, an excellent nonnitrogenous precursor to glycine. We have seen that benzoate can potentiate ammonia toxicity (5, 6), perhaps by inhibiting pyruvate carboxylase (this report), and that glyoxylate is also an effective inhibitor of pyruvate carboxylase (Table I). A common site of action raises the possibility that the combination of benzoate and Piridoxilate might be much more toxic than either drug alone. Accordingly, it would seem prudent to avoid such combination therapy until its effectiveness is supported by tests with a suitable animal model. REFERENCES 1. LEWIS, H. B. (1914) J Biol. Chem. l&225-231. 2. BRUSILOW, S. W., VALLE, D. L., AND BATSHAW, M. L. (1979) Lancet 2,452-454.
3. BATSHAW,M.L.,THOMAS,G.H.,ANDBRUSILOW, S.W.(1981)Pediatrics68,290-297. 4. WENDLER,P.A.,ANDTREMBLAY,G.C.(~~~~) Bicthem. Biophys. Res. Cmnmun 105,1341-1346. 5. MASWOSWE,S. M., CYR, D. M., GRIFFITH, A. D., ANDTREMBLAY,G.C.(~~~~) Biochem. Biophys. Res. Commun. 138,369-373. 6. O'CONNOR,J. E., RIBELLES,M., AND GRISOLIA,S. (1982) Eur. J. Pediatr. 138,186-187. 7. GRIFFITH,W.H.(~~Z~)J. Biol. Chem. 82,415-427. 8. QuIcK,A.J.(~~S~)J. BioL Chem. 92,65-85. 9. CYR,D. M., MASWOSWE,S. M., AND TREMBLAY, G. C. (1987) J. Inherited Metab. Dis. 10, Suppl. 2,308-310. 10. BELIVEAU, G. P., AND BRUSILOW,S. W. (1987)J. Nutr. 117,36-41. 11. STIRK,J. H., BARTLETT, K., SHERRATT,H., AND STANLEY, A. (1983) Biochem Sot. Tr~ns. 11, 286-287. 12. CORKEY,B. E., BRANDT, M., WILLIAMS, R., AND WILLIAMSON,J. R.(1981) Anal. B&hem. 118, 30-41. 13. KREBS, H. A., AND DEGASQUET, P. (1964) Bie them. .I 90,149-154.
14. GOODALL,G. J., BALDWIN, G. S., WALLACE,J. C., AND KEECH,D.B.(~~~~) B&hem. J. 199,603609. 15. HALESTRAP,A. P. (1975) B&hem. J 148,85-96.
BY SODIUM BENZOATE
16. BECKER, C. M., AND HARRIS, R. A. (1983) Arch. Biochem. Biophys. 223,381-391.
17. KUKINO,K.,ANDMATSUMOTO, I.,ascitedin MATSUMOTO,I., KUHARA, T., AND YOSHINO, M. (1976) Biomed. Mass Spectrum. 3,235-240. 18. HAYNES, R. C., JR. (1965) J. Biol. Chem. 240,41034106. 19. WALTER, P., PAETKAU, V., AND LARDY, H. A. (1966) J. Biol. Chem. 241,2523-2532. 20. MCCUNE, S. A., DURANT, P. J., FLANDERS,L. E., AND HARRIS, R. A. (1982) Arch. Biochem. Bicphys. 214,124-133. 21. VON GLUTZ, G., AND WALTER, P. (1976) FEBS Lett. 72,299-303. 22. O'CONNOR,J. E., COSTELL,M., AND GRISOLIA, S. (1987) Biochem. Biophys. Res. Commun. 145, 817-824. 23. COLOMBO,J.P., BACHMANN,C., PFISTER,U., AND GRADWOHL,M. (1988) B&hem. Biophys. Res. Cmmun. 151,872-877. 24. QURESHI,I.,ROULEAU,T.,LETARTE,J.,ANDOUELLET, R. (1986) Biochem. Int. 12.839-846. 25. JACKSON,A. A.,BADALOO, A.V.,FORRESTER,T., HIBBERT,J.M.,AND PERSAUD,~.(~~~~) Brit. J. Nutr. 58,207-214.
26. COUDE,F.X.,COUDE,M.,GRIMBER,G.,PELET, A., AND CHARPENTIER,C. (1984) Clin Chim. Acta 136,211-217.