- Email: [email protected]
Altered acyl-CoA metabolism in riboflavin deficiency
Biochimica et Biophysica Acta,
1006(1989)335-343
335
Elsevier BBALIP53254
Altered acyl-CoA metabolism in riboflavin deficiency K. V e i t c h ~*, J.-P. D r a y e ~,2, j. V a m e c q ~, A . G . C a u s e y 3, K. B a r t l e t t 3, H.S.A. S h e r r a t t a n d F. V a n H o o f ~,2
4
t Laboratoire de Chimie Physiologique, L C.P.& UCL, and 2 Department of Pediatric Neurology, Universitd Catholique de Louvain, Brussels (Belgium), ~ Departments of Child Health & Clinical Biochemistry, and ~ Department of Pharmacological Sciences, The Medical School, The University, Newcastle upon Tyne (U.K.)
(Received 9 May 1989)
Key words' Riboflavindeficiency;Mitochondrion;Peroxisome;AcyI-CoA;Flavoprotein;Organicacid; (Rat) We have recently described the effects of r[bof|avin deficiency on the metabolism of dicarboxylic acids (Draye et al. (1988) Eur. J. Biochem. 178, IK3-|89). As both mitochondria and peroxisomes are thought to be involved, we have exam|ned the activities of var|ous enzymes in these organelles in the livers of riboflavin-deficient rats° Mitochondrial ~-ox|dat|on of fatty adds was severely depressed due to loss of activity of the tiwee fatty acyI-CoA dehydrogenases, whereas there was an enhancement of peroxisomal fl-ox|dation due to an increased activiff of the FAD-dependent fatty acyl-CoA oxidase, although the activities of other peroxisoma| flavoproteins, D-amino acid oxidase and glycolate oxidase, were lowered. Hepatoc)te ~r~home~ry revealed an increase in the numbers of peroxisomes, indicating a proliferation induced by the deficiency. The mitoebondrial acyl-CoA dehydrogenases involved in branched-chain amino acid melabolism were also severely decreased leading to characteristic organic acidurias. There was some loss of activiff of the flavin-dependent sections of the electron transport chain (complexes ~ and II), but these were probably not sufficient to affect normal function in vivo. The specificity of these effects allows the use of the riboflavin-deficient rat as a model for the study of dicarboxylate metabolism.
|nU'eduction The measurement of organic acids excreted in urine is often used to detect errors in metabolism, which may present clinically with a variety of symptoms. Defects in /]-oxidation produce an unusual or excessive excretion of dicarboxylic acids and/or glycine conjugates in the urine. Although several animal models have been proposed for the study of dicarboxylic aciduria, the riboflavin-deficient rat [1,2] would be particularly useful since several conditions which produce organic acidurias respond to therapeutic administration of riboflavin (see Ref. 3). Defective fl-oxidation in riboflavin deficiency is due to low activities of the three FAD-dependent acylCoA dehydrogenases [4-7] which catalyse the first intramitochondrial step in the/t-oxidation spiral. However, the peroxisomal/~..oxidation system [81 has also been suggested to play a role in the processing of dicarboxylic acids [9], so it is important to know the
Correspondence (and * present address): K. Veitch, Hormone and Metabolic Research Unit, ICP&UCL 75.29, avenue Hippocrate 75, B-1200 Brussels,Belgium.
condition of this activity in any model system. The 'rate-limiting' step in this process is the FAD-dependent acyl-CoA oxidase [10]. Catalase [11,12] and flavindependent peroxisomal oxidases [12-14] are depressed in riboflavin-deficient rats. However, in those studies describing the effects of the deficiency on peroxisomal fl-oxidation, these other peroxisomal activities were either unaffected or not measured [5,15]. There are no studies in which all these peroxison:al oxidases have been assayed in riboflavin deficiency. We have reported str~dies of the in vivo metabolism of exogenous dicarboxylic acids in riboflavin-deficient rats [16,!71. !1[ere we report the relevant mitochondrial and r~eroxisomal enzyme activities in this model, which were undertaken to elucidate the relative roles of the two organelles in dicarboxylate metabolism. These studies included examination of the sections of the mitochondrial electron transport chain which are flavin-dependent (complexes I and II), and which have been described i, the literature to be affected inconsistently by the vitamin deficiency [4,13]. Our results may, in part, explain the discrepancies between these previous reports. Part of this work has been presented to the Belgian Biochemical Society [l 8].
0005-2760/89/$03.50 © ElsevierSciencePublishersB.V.(BiomedicalDivision)
336 Materials and Methods The sources of chemicals have been given previously [7,17,19]. A semi-synthetic riboflavin-deficient diet was purchased from ICN Pharmaceuticals (Cleveland, OH, U.S.A.) as a powder which was mixed with a 5~ gelatine solution (100 ml/kg) and dried to form 'cakes'. 'Diet-control' animals received the same diet with 20 mg/kg riboflavin added during preparation. Control animals were fed a standard rat chow (A-03, U.A.R., Epinay, France). Except where specified, all animals were maintained on a 12 h light/dark cycle with food and water ad libitum. r
",ration of animals Young male Wistar rats were used in two series of experiments, in the first, six animals (mean body weight 99 :t: 2 g) were housed individually in metabolism cages. Three control animals received chow diet, and three tests the riboflavin-deficient di~t. At weekly intervals all the animals were fasted overnight (18 h), urine being collected over this period. In the second series 18 rats (84 ± 1.3 g) were housed in three groups of six in sawdust-lined cages, with no attempt made to prevent coprophagia other than by daily cleaning. These groups received the control, diet-control and riboflavin-deft. cient diets, respectively, ad libitum. After 6 weeks, all animals were fasted overnight, during which period urine was collected, before killing by decapitation. Livers were rapidly removed, weighed and used to prepare 10~ homogenates and mitochondrial fractions; small liver samples were fixed for electron microscopy as described below. Blood was collected from the tail vein for the measurement of the erythrocyte glutathione reductase activity coefficient as an index of riboflavin deficiency
[20].
Mitochondria Mitochondrial fractions were prepared as described in Ref. 19, and finally re,suspended in 0.3 M mannitol/10 mM Hepe.s/l mM EGTA (pH 7.4) at 25-50 mg protein per ml. The rates of mitochondrial oxidation were measured polarographically in a 3.0 ml volume [21], with I0 mM succinat¢, I0 mM glutamate (+ 1 mM malate) and 33 ~M palmitoylcarnitine as substrates, stimulated by the addition of I ~mol ADP. When assaying fi-oxidation, 5 mM malonate and 1.5 mg/ml defatted bovine serum albumin were included to ensure complete oxidation to acctoacetate, which directly measures the flux through /)-oxidation [22]. State 3, state 4, respiratory corJtrol ratio and A D P / O ratio were determined [23l. Mitocbondrial oxidation rates were also measured spectrophotometrically by the rate of reduction of ferricyanide at 420-475 run [24] in an Aminco DW-2 spectrophotometer using 20 mM succinate, 20 mM glutamate
( + 2 mM malate), 25 ~M palmitoylcarnitine, 25 /tM palmitoyl-CoA (+1 mM L-carnitine) or 50 ltM octanoylcarnitine as substrates. The oxidation of 1 mM 3-[U-14C]methyl-2-oxopentanoate, prepared according to Ref. 19, was assayed [25] with radio-HPLC analysis of mitochondrial extracts. Mitochondrial enzymes, citrate synthase [26], succinate dehydrogenas¢, with phenazine ethosulphate instead of phenazine methosulphate [27], glutamate dehydrogenase [28], NADH dehydrogenase [29] and carnitine palmitoyltransferase [30] were assayed by published methods. Acyl-CoA dehydrogenases were prepared from mitochondrial extracts and assayed in 25 ~tM FAD to protect the enzymes during extraction as described [7]. Citrate synthase was measured in the 10~ liver homogenates as a marker enzyme for mitochondria. Peroxisomes Peroxisomal activities were measured in 10% liver homogenates and stored at - 8 0 ° C until use. Peroxisomal iS-oxidation was assayed at 37°C as the cyanide-insensRive reduction of NAD + with 50 ~tM palmitoyl-CoA as substrate [31]. Acyl-CoA oxidase, Damino-acid oxidase and glycolate oxidase were assayed at pH 7.5 by the H202-dependent iminoquinone formation catalysed by peroxidase [32] with 50/tM lauroylCoA, 25 mM D-proline and 10 mM glycolate as substrates, respectively, in 25 #M of the appropriate ravin cofactor. Urate oxidase and catalase were assayed [33]. All assays, except peroxisomal r-oxidation (37°C) and catalase (0 o C) were performed at 30 o C. Activities are expressed as units per gram of liver or per mg mitochondrial protein, 1 unit being the amount of enzyme required to catalyse the reaction of 1 p mol of substrate or the formation of 1/tmol of product per min under the described conditions. Catalase is expressed in Baudhuin units, 1 unit being the amount of enzyme causing the destruction of 90~ of the substrate in 1 min in a volume of 50 ml under the assay conditions [33]. Protein was measured in sonicated mitochondria and liver homogenates [34], with bovine serum albumin as standard. Organic acids Urine samples were prepared and analysed by gas - ],, chromatography [t I, with pentadecanoic acid as internal standard. The two peaks corresponding to 2-methylbutyrylglycine and isovalerylglycine comigrated in this system, the latter being found in much greater quantities. As both 2-methylbutyryl-CoA and isobutyryl-CoA are substrates for the same acyl-CoA dehydrogenase [35], no further attempt was made to separate the two glycine conjugates, the appearance of isobuty~lglycine indicating a defect in this activity. The identity of the organic acids was confirmed by using an ion-trap detector (Finnigan-Mat model 700).
337
Electron microscopy Samples of liver were immediately fixed in 0.1 M sodium cacodylate/270 (w/v) glutaraldehyde (pH 7.4). Post-fixation was made in 270 (w/v) osmium tetroxide, dehydration in ethanol, and embedding in the mixture as described by Spurr [36]. Ultrathin sections, stained by uranyl acetate and lead citrate were examined with a Philips EM-301 electron microscopy. Micrographs were taken from a minimum of three different sections, each cut at least 5/~m from the preceding one. Magnification (about 15000 x ) was accurately determined with a calibration grid (E.F. Fullam, Schenectady, NY) and stereological analysis performed [37]. A multipurpose test grid similar to that described in Ref. 38 was used for the determination of volume fraction or membrane area, the number of particles being calculated [39]. Statistical analysis of the differences between group means was performed by Student's t-test. Where shown, analyses of linear regression and correlation coefficients were performed with the individual data from 18 preparations, using six animals from each experimental group. Results
Animals and organic acids After only 1 week, both groups receiving the test diet weighed significantly less than chow-fed controls (Fig. 1). The difference in weight between the two control groups, approx. 12%, was mair~tained throughout the experimental period, diet controls weighing 87% of chow-fed controls after 6 weeks ( P < 0.001). The vitamin-deficient animals were more severely affected, weighing 42 and 49% (both P < 0.001) of chow-fed controls and diet-controls, respectively, after 6 weeks. Overnight fasting causing a 6-8% decrease in body weight in each group. After 6 weeks, the erythrocyte 300
O 200
g 0
100
Z < uJ 1
I
I
I
I
I
I
0
1
2
3
4
5
6
WEEKS
ON DIET
Fig. 1. The mean body weights of the three groups of animals during the feeding of the control and test diets. Each point is the mean ol six animals re: con|:~<~J~ ~o), diet controls (A) and riboflavin-deficient group (Ilia. $.EM. ba~'s are not shown, L:zi~g smaller than the symbol:, used. Alter 1 w:~ek all groups were significantly different from each other at each ~me point ( P < 0.01). o , control group; A, diet control group; ~, ~,:~:,oflavin-~r.ficientgroup.
glutathione reductase activity coefficient was significantly increased in the riboflavin-deficient group (1.24 _+0.05, 1.22 + 0.15 and 2.18 _+0.18 for controls, dietcontrols and riboflavin-deficients, respectively, means _+ S.E. for four animals in each group), values greater than 1.35 indicating riboflavin deficiency [20]. Liver weights were greater in riboflavin-deficient animals when expressed as a percentage of total body weight (3.06 _+0.07 and 4.94 _+0.28% for controls and tests, respectively, P < 0.01), but protein content was not significantly affected (250 _+ 6 and 231 _+ 3 mg/g liver). There were no differences between the two control groups. The organic acid excretion over this period was monitored in the first series (Fig. 2). An organic aciduria progressively developed during the 6 weeks of study in the riboflavin-deficient animals. Compounds which were excreted in abnormally high amounts included saturated and unsaturated medium-chain dicarboxylates, glutaric acid and the glycine conjugates of butyrate, isobutyrate and isovalerate (and 2-methylbutyrate). The amounts excreted after 6 weeks in the second series are shown in Table I. Most of the dicarbc, xylates excreted were unsaturated, as octenedioic and decenedioic acids. Such large amounts of dicarboxylates indicate a major defect of fl-oxidation in the liver [1].
Mitochondria The impairment in the rate of fl-oxidation caused by riboflavin-deficiency was confirmed (Tables lI&Ill), particularly that of palmitoylcarnitine which was decreased to 36% of controls (Table II). Succinate oxidation was 30% of control rates, and glutamate oxidation was further decrea3ed (56% of controls). The mitochondria from livers of riboflavin-deficient rats were not damaged as shown by the small effects on respiratory control, A D P / O ratios and nearly complete inhibition (96-98%) of ferricyanide reduction by 1 # g / m l antimycin (not shown). These effects were confirmed by the spectrophotometric assay of ferricyanide reduction which collects electrons from the respiratory chain at the level of eytochrome c, and which excludes the activity of complex IV from the rate (Table III). With this method, succinate oxidation was decreased by 50%, while glutamate/malate oxidation was 62% of controls, whi:h is not markedly different from the polarographic data. The effect on succinate can be explained by the activity of succ~aa~e dehydrogenase, the FAD-dependent enzyme. This was decreased to 47% of co=trois in riboflavin-deficier,t preparations (Table.~ IV), and there was a highly significant co~re~atioJ~ betLween this activi'~y a~,d th:. rate of succinate oxidation assayed spectrophotomz~,';cal!y (r--~0.903, P<0.001). The differences observed between the polarographic and spectrophotomettic data confirm the observation that succinate dehydro-
338 1200 1200 10( 1000600800200400Adipic c ~ ai ~ ~ .,,,,
c8 @ ¢0 O)
0 ft
1
2
3
4
5
6
1200 1000 800 600 400 200 0
1 2 3 4 5 6
:t oot,.oo,o/
Butyrylglycine
1 2 3 4 5 6
_
E O)
~ >.
100 80 60 40 20 0
0
120 100 80 60 40 20
Isobutyrylglyoine
........
~
2 3 4
56
1
sciatic ac,a
400 3OO 2OO tO0 "--~' o L----~:'~'~'~'~r'~-1
2 3 4
S
23
4
2000 . Oecenedlolc p,~ e,. ~ / \'o tOO0 0
6
12
34
1 2 3 4 5 6
56
56
600 50O 400 300 200 100 0
1 2 3 4
56
WEEKS ON DIET Fig. 2, Excretion of the major organic acids found in urin~~during the experimental period. Urine samples were collected at weekly intervals during an 18 h fast. Each point is the mean of three animals fed control chow diet (o), or riboflavin-deficient diet (O).
genase is not the rate-limiting step in succinate oxidation when it proceeds as far as molecular oxygen [4], and suggest that such control occurs in complex IV subsequent to cytochrome c. Glutamate dehydrogenase activity was increased by the deficiency (Table IV). However, glutamate oxidation is NADH-dependent, The FMN-dependent NADH dehydrogenase was decreased to 53% by the deficiency (Table IV), and there were significant correlations of TABLB I The m ~ w orgem~c¢~Ms excreted in the w~ne d~rmg an oeernigkt fast 08 k), oftee 6 weeks on the expeelmental diets
Valuesare the means:l:S,E for six animalsin eachgroup expressedas pg/mg creatinine. Compound
Control group
Isolmtyrylglycine Butyrylglycine lsovaleuIglycinea Su~inate (31uta~ ¢cid
<3 <3 43.6:1:14,8 61.$± 14,4
Ac~'.'pi¢aci~ Octe.nedioic acid (ul~saturated suberatc) Su~ctic ~..id Oe~;~'aedio~cacid
(unsatu~t~xlsebacat©) Sebacic acid
"~ 3
S,9± 3.1 .~.5+ 4,4
4,2 ± ~,0
Riboflavin. deficient group 38,6+ 231 ~ 502 ± 16,3± 163 ±
9,8 58 72 6,8
52 7~0 :l: 166 1943 :l:601 174 ~- 34
2"/7 -I-80
2084 +621
27.2+11.4
649 +_184
a Total of isovaleryiglycine aad 2-methyibutyrylglycine.
this activity with the rates of glutamate/malate oxidation assayed polarographically (r ffi 0.83, P < 0.001) and spectrophotometrically (r - 0.63, P < 0.005). 0-Oxidation of fatty acids also depends on NADH dehydrogenase through the action of the 3-hydroxyacylCoA dehydrogenase, the third step in the r-oxidation spiral, which reduces NAD +. The rate of r-oxidation in vivo is partially controlled by the N A D H / N A D + ratio [41]. However, if the rate of oxidation of glutamate/ malate indicates NADH dehydrogenase activity, then the deficient mitochondria would have sufficient capacity to cope with NADH generated by fi-oxidation (Table !I). This suggests that the rate of r-oxidation is impaired by the low activities of the three FAD-dependent fatty acyl-CoA dehydrogenases (Fig. 3), the decreases in both short-chain and long-chain activities were about the same as in the two assays of oxidation rate. As the rates of oxidation of palmitoylcarnitine and paimitoyl-CoA were equally affected (Table Ill), carnifine palmitoyltransferase (CPT) was not apparently limiting in these assays, indeed CPT activity increased by 158~ in the deficient animals (Table IV). 3-Methyl-2-oxopentanoate formed by transamination of isoleucine, is oxidat~vely decarboxylated in the n~'~ochondrial matrix to 2-methylbutyryI-CoA, which is subsequently oxidised via intermediates such as propionylCoA and methylmalonyl-CoA to acetyl-CoA and succinyl-CoA (see Ref. 19). 2-MethylbutyryI-CoA dehydrogenase activity was markedly decreased in riboflavin deficiency (Fig. 3), associated with excessive excretion
339 TABLE I|
The effects of riboflavin-deficiency on hefatic mitochondriol oxygen uptake Values are the means +_S.E. of s;x preparations in each expressed as nanogram aioms O per rnin per mg protein. Respiratory control ratio (RCR) is the ratio of the rates of oxidation of sta ie 3 on state 4. * P < 0.05; * * P < 0.01; * * * P < 0.001 vs. controls. Substrate
Group
State 3
State 4
RCR
ADP/O
control diet control ribofl.-deficient
163 ± 9 155 + 7 131 -+ 8 *
48 + 2 47 + 1 39 + 3 *
3.45 + 0.21 3.30 + 0.19 3.44 + 0.23
1.62 ± 0.05 1.65 ± 0.06 1.80 + 0.08
control diet control ribofl.-deficiem
109 + 7 106 + 6 61 + 1 * * *
26 + 1 26 ± 1 21 5:1 * * *
4.25 + 0.37 4.19_+ 0.24 2.96 ± 0.11 * *
2.54 ± 0.08 2.43 ± 0.07 2.47 ± 0.14
control diet control ribofl.-deficient
69 5:5 76 + 7 25 5:1 * * *
22 :t: 3 26 5: ! 20 5:1
3.43 + 0.38 2.95 ± 0.17 1.25 + 0.07 * * *
2.40 ± 0.09 2.38 ± 0.07 2.69 ± 0.4
Succinate
Glutamate ( + malate)
Palmitoylcarnitine
.c_
TABLE I!1
@
The effects of riboflavin deficiet~cv ot~ spe~tropholometriccdly mea.sured hepatic mitochondrial oxidations Values are means + S.E. for the six preparations in Table lI, expressed as nmoles Fe(CN)63- reduced per minute per mg protein (Figures in parentheses are the percentages of the control values.) * * P < 0.01; • * * P < 0.001 vs. Controls. Substrate
Control group
Diet control group
Riboflavin-deficient group
Succinate Glutamate (+malate) Palmitoylcarnitine PalmitoylCoA Octanoyleamitine
437+26
425_+16
(97)
219_+14 * * * (50)
2055:17
194_+19
(95)
127_+12 **
87+ 8
84+ 5
(97)
65 5 : 7
67 + 3 (103)
16 5 : 2 * * * (25)
88 ± 6
84+ 5
25 5 : 3 * * * (28)
(62)
25
f: 20 15
c
•~ 2o
~-C= 10 O 1:3 o
E
0
jjj_ I
ql
e-
21+ 2"**
40
Butyryl
Uctanoyl Palmitoyl Isovaleryl
Isobulyryl
1.0 0.5
0.0
(24) AcyI-CoA ester substrate
(95)
of both 2-methylbutyrylglycine (measured with the iontrap detector, not shown) and isobutyrylglycine (Fig. 2 and Table I). This decreased activity causes a 10-fold greater accumulation of 2-methylbutyryl-CoA in
Fig. 3. The activities of the five mitochondrial acyi-CoA dehydrogenases in the three experimental groups. Each value is the mean of six preparations for control group (open bars), diet control group (hatched bars) and riboflavin-deficient group (closed bars), with S.E.M. bars as shown.
mitochondria from riboflavin-deficient rats incubated with 3-oxo-2-[UJ4C]methylpentanoate compared with raitochondria from control and diet-control animals
TABLE IV
Mitochondriai enzyme activ~zies Values are the means+S.F, for the six prepalations in each group as given in Table If, expressed as U / m g protein. (Figures in parentheses are percentages of control vak~es.) * * * P < 0.001 vs. controls. Enzyme
~.i;o~rol groul~
Diet cch~r,,I group
RJboflavi~-deficient grou~
Succinate dehydrogenase Glutamate dehydrogenas(: N A D H dehydrogenase CPT Citrate synthase
0.048 :J: 0.003 0.179 + 0.0~ ~ 2.49 :t 0.11 1.5i :1:0.11 0.210_+_ 0.013
0.046 5:0.003 (96) 0.211 + 0.018 (18~) 2.45 5:0.14 (99) 1.38 +0.04 (91) 0.208 + 0.011 (99)
0.023 :t 0.002 * ~' * (4'1) 0.395 5:0.025 * * * (221) 1.31 5:0.08 * * * (53) 3.89 5:0.44 * * * (258) 0.239+0.018 (114)
340 TABLE V
Acyl.CoA intermediates accumulated in liver mitochondria during 3 rain incubation with I mM 3-[U-14C]methyl-2-oxope ntanoate Values shown are the means ± S.E. for four preparations from controls and riboflavin deficient rats and the mean of two diet controls, expressed as d.p.m, per mg mitochondrial protein. ** P < 0.005, * * * P < 0.001 vs. controls. AcyI-CoA Ester
2.MethylbutyrylCoA 2°Methylbut-2-enoyl-
CoA PropionyI-CoA AcelyI.CoA MelhylnmlonyI.CoA SuccinyI.CoA
Control group
Diet control
Riboflavin deficient
group
group
464 ± 23
540
4 543 ± 345 * * *
93 ± 10 140 ± 13 245 ± 22
79 119 349 735 517
n,d. n,d, 166 ± 31 190± 60 **
555~:44 546 ± 54
n.d.
n,d,. not detected,
(Table V). Assuming that all the metabolism of 3.oxo-2[U.t4C]methylpentanoate occurs in the matrix, and a mitochondrial content of CoA of 2 nmol/mg protein and a matrix volume of 0.5 p l / m g protei~t, it was estimated from the UV HPLC trace [19] mat the matrix concentration of 2.methylbutyryI-CoA was 2.3 mM. 2.methylbut-2-enoyl-CoA, propionyl-CoA and succinylCoA were not detected in incubations with riboflavindeficient mitochondria, and the acetyl-CoA and methylmalonyI-CoA concentrations were 68 and 34S$ of controis, respectively (Table V). Therefore, the mitochondfia from the riboflavin-deficient rats appeared to function normally, apart from selective defects in flavoprotein-dependent activities, most notably of the five acyI-CoA deh)-drogenases. The activities were expressed relative to mitochondrial prorein, which appeared to be valid as the specific activity of citrate synthase was the same for all three groups (Table IV). The specific activity of citrate synthase in liver homogenates was significantly increased (P < 0.001) in riboflavin-deficient preparations (16.4 + 0.8, 16.2 + 1.0 and 22.9 + 0.8 U / g wet weight, for controls, diet-controls and deficients, respectively), but total activity per liver was unchanged (143 + 9, 130 + 8 and 134 + 12 U per liver, respectively).
Peroxisomes Peroxisomes are another site for the/l-oxidation of fatty acyl-CoA esters [8]. The 'rate-limiting' step is also catalysed by an FAD-dependent enzyme, acyl-CoA oxidase [1O]. However, in contrast to the mitochondria the rate of peroxisomal /]-oxidation was greater in homogenates of rivers from riboflavin-deficient rats (Table VI}. It was confirmed that this effect was not due to cfifferent amounts of apoenzyme being present by preincubatb,~ in 100/AM FAD for 10 min. This treatment
was equally effective in increasing activity in controls and tests. There was a similar ivcrease in actual acylCoA oxidase (Table VI), but this was not due to a general increase in peroxisomal flavoproteins, as shown by the marked decreases in activities of D-amino-acid oxidase (FAD) and glycolate oxidase (FMN). The qualitative differences between these activities in controls and tests were unaffected by addition of the relevant co-factor, the data shown being obtained with 25 /~M F A D / F M N . Catalase activity per gram tissue was also decreased in the livers of riboflavin-deficient animals, but the total activity per liver was unaffected. There was no effect on urate oxidase. The very similar dat~ obtained from the two control groups confirmed that these effects were not due to differences in dietary composition, other than the absence of riboflavin for the tests.
Morphometry The surprising observation of an increase in peroxisomal/]-oxidation was supported to some extent by the morphometric analysis presented in Table VII. The fractional volume of ~:he hepatocyte occupied by peroxisomes was not si~i~nificantly affected ( < 10% increase), but there was t~n increase in peroxisomal membrane area (28%), and there was an 82% increase in the number of peroxisomes. This increase in the number of peroxisomes, which were smaller than those found in control tissue, is consistent with peroxisomal proliferation accoiding to current concepts of peroxisomal biogenesis that fission of 'parent' organelles forms smaller 'progeny' [42]. The total fractional volume of the hepatocytes due to mit~hondria was markedly increased (29%) consistent with the biochemical data. This increase in mitochondrial volume content together with the 14% increase in mitochondrial citrate synthase TABLE Vl
Peroxisomal en=yme activities assayed in liver homogenates Means+S.E. for six animals in each group, figures in parentheses being percentages of control values. * P < 0.05, * * P < 0.01, * * * P < 0.001 vs. controls. Activity
Control group
fl-Oxidation a 1.15+0.06 AcyI-CoA oxidase b 1.64+0.04 ~Amino-acid oxidase b 0.79+0.09 Glycolate oxidase b 1,174-0.07 Urate oxidase b 2.30+0.11 Catalase c 59.6 -t-3.2
Diet control Riboflavin deficient group group 1.08+0.08
2.23+0.21 * * * (194)
1.50±0.13
2.72+0.31 **
0.73+0.08
0.22+0.07 * * *
1,28±0.08 2.27+0.09 59.3 +2.2
(166) (27)
0.20+0.03 * * * (17) 2.41 +0.14 (105) 44.2 +4.2 * (74)
pmol N A D + reduced per rain per g liver at 37 o C. b Units per g liver. c 8audhuin units per g liver (see g e l 33).
341 TABLE Vll
Morphornetric data fr,~m control and riboflavin.deficient rat hepatocyws The livers were fixed and prepared as described in Materials and Methods. The r.nalysed section areas in hepatocytes were 1399, 1417 and 1593 pm 2 for control, diet control and riboflavin-deficient rats, respectively. Values are given 5: S.E. of the mean for the analysed sample. The estimation of particle number was obtained by the application of the following equation as described by Weibel and Gomez [391: N~=(Na)3/2/[~(V~)I/2, where N~ is the number of particles per unit volume; Na is the number of profiles per unit area in section; l~ the volume fraction of particles, and p a shape factor. Peroxisomes are considered as spheres (fl ffi 1.382). Control group
Diet control group
Riboflavin-deficient group
Fractional volume (% of the cytoplasmic volume) mitochondria 23.5 5:1.5 22.0 5:1.2 30.2 peroxisomes 1.6 5:0.2 1.5 5:0.2 1.7 lipid droplets 1.1 5:0.7 0.2 5:0.2 2.8
5:0.9 5:0.3 5:0.6
Membrane area per unit volume (/A m2/p m~) peroxisomes 0.152 :t: 0.028 0.132 ± 0.018
0.194:t:0.018
Number of particles per cubic micron peroxisomes 0.107 0.115
0.195
activity (Table Vl) explains the 40% increase in specific activity in liver homogenates. Electron microscopic examination did not reveal any other consistent difference between any of the three groups, although hepatocytes from riboflavin-deficient animals were richer in small elements of smooth endoplasmic reticulum. Discussion
Riboflavin-deficiency in rats has been widely studied since the observation that such animals utilise food inefficiently [43], :.ndicated by an abnormally high respiratory quotient [44]. Subsequent studies were concerned with the effects of the vitamin-deficiency on mitochondrial oxidative phosphorylation, particularly the oxidation of NADH [13,45,46]. These early studies demonstrated a rapid loss of flavin coenzymes from the liver, FAD and FMN falling to 29 and 10%, respectively of control values after 5-6 weeks of deficiency [14] and 25 and 5%, respectively, after 8-10 weeks [13]. In these studies there was no effect on the FMN-dependent NADH dehydrogenase activity (complex I), a result inconsistent with the dramatic effect on the coenzyme concentration Succinate dehydrogenase (complex II) decreased by 75%, with a 50% decrease in succinate oxidation rate [13], an effect confirmed by Ref. 4, although other workers [46] found no effect on the rates of succinate oxidation. These inconsistencies may be explained by our d~ta on succinate and glutamate oxidations, and the respective enzyme activities presented in Tables II-IV. A 50% loss of succinate dehydrogenase activity had a relatively
small effect on the rate of suceinate oxidation when measured with oxygen as final electron acceptor (20%), but 50% when re1. icyanide was the final acceptor. This suggests that the activity of the electron transfer chain subsequent to cytochrome c is rate-limiting. Early studies using polarographic or manometric techniques may therefore have failed to detect small changes in rate of succinate oxidation associated with relatively large losses of enzyme activity. We observed a marked decrease in the rate of oxidation of glutamate (in the presence of 1 mM malate), which we attribute to a loss of N A D H dehydrogenase activity, with an excellent correlation between these two activities. Some groups reported decreased glutamate oxidation [6,15], while others found no change [4,46]. In studies using inhibitors of NADH-ubiquinone reductase, the inhibition of glutamate oxidation by intact mkochondria, and of NADH oxidation after freezethawing have identical dose' responses (Veitch, K. and Hue, L., unpublished data). These results, ar.d those reported in this study, suggest that the flux lhrough complex I is the rate-limiting step in glutamate oxidation and that the observed decrease in this rate ira riboflavin deficiency is due to the loss of N A D H dehydrogenase, consistent with the fall in hepatic FMN content. However, these effects on the respiratory chain complexes are much less than those observed on hepatic fatty acid oxidation, which is affected within 1 week of deficiency [4]. The urinary excretion of organic acids (Fig. 2) is further evidence of the sensitivity of fl-oxidation to the vitamin deficiency, with an increased dicarboxylic aciduria after 1-2 weeks. There is (Fig. 3) loss of activities of the three fatty acid-CoA dehydrogenases [4-7], hence the observed decrease in /t-oxidation. It should be noted that these activities were assayed in mitochondrial extracts containing 25 #M FAD (see Ref. 7) which did not reactivate the medium- and long-chain activities as has been reported [5], in total agreement with [6]. Short- and medium-chain acyl-CoA dehydrogenases are inhibited in hypoglycin-poistmcd rats, which excrete large quantities of unsaturated medium-chain dicarboxylates [47]. This excretion is due to the omega-oxidat;on of linoleic acid metabolites which cannot be fl-oxidised due to the impaired mitochondrial enzyme activities [48]. The octenedioic and decenedioic acids excreted by riboflavin-deficient rats are presumably derived from the same source. The organic aciduria also revealed the sensitivity of amino-acid metabolism to the deficiency, the appearance of glycine conjugates of isobutyrate, isovalerate and 2-methylb~,tsrate being at least as rapid as the dicarbozylic, aciduria. The activities of the two branched-chain acyl-CoA dehydrogenases were more severely affected than the straight-chain activities, and the oxidation of 3-[U-14C]methyl-2-oxopentanoate was
342 drastically impaired. The elevated ratio of 2-methylbutyryI-CoA to 2-methyl-2-enoyi-CoA (controls - 4.9 + 0.8, riboflavin-deficients > 100) indicates a block in the flux at the point of isobutyryl-CoA (2-methylbutyrylCoA) dehydrogenase. The branched-chain 2-oxo-acid dehydrogenase complex, which contains subunits of the FAD-dependent lipoyl dehydrogenase (EC 1.6.4.3), was su:*ficiently active to generate high concentrations of 2-methylbutyryl-CoA by the oxidative decarboxylation of 3-methyl-2-oxopentanoate. All the major organic acids excreted are derived from metabolic p ~ in which the accumulated intermediates are CoA-esters. Glutaric acid is excreted due to the loss of activity of glutaryI-CoA dehydrogenase [1]; alutaryI-CoA being formed in the catabolism of lysine, hydroxylysine and tryptophane. Intracellular accumulation of CoA esters has been suggested as one of the signals for peroxisomal proliferation [49], consistent with an ine,ease in numbers of peroxisomes (Table VII). In early studies [13,14], D-amino acid oxidase and glycolate oxid~tse were used as markers for FAD. and FMN.depo~dent activities, respectively, before these enzymes had been located in the peroxisomes, and shown to be dramatically decreased. Furthermore, decreased catalase activity has been reported in riboflavin-deficient mice [50] and rats [11,12]. In studies on the effects on peroxisomal B-oxidation there were either no significant effects on catalase [5,15] or the enzyme was not assayed [6], other pero~somal activities were not measured. We have confirmed that while there are losses of other flavoprotein peroxisomal oxidases and catalase, there is an enhanced peroxisomal/~-oxidation capacity due to an increase in acyI-CoA oxidase. This specific increase, probably as a result of the decreased mitochondrial ,0-oxidation, susgests a role for peroxisomes in clearing accumulated CoA-esters. In con~usion, riboflavin deficiency in rats decreases the activities of the mitcchondrial electron transpor~ chain (complexes 1 and I!), but these effects are probably not sufficient to affect normal function in vivo. However, the cellular metabolism of CoA esters is severely perturbed due to the loss of several, but not ell, flavin-dependent dehydrogenases. These effects are rapidly apparent with the excretion of the nonmetabolised substrates as dicarboxyfic acids or glycine conjugates in the urine. Furthermore, despite decreased activities of ~ m e peroxisomal flavo-oxidases, there is an increased capacity for peroxisomal/3-oxidation. The loss of mitochondrial ~0-oxidation capacity leads to a dicarboxyfic aciduria, apparently due to insufficient oxidation of medium.chain fatty acids, confirming the relevant importance of the mitochondrial system compared with the peroxisomal one. The relative specificity of these effects mean~ that the riboflavin-deficient rat is suitable from the study of the meta'~(~fism of dicarboxyfi¢ acids [51]. This aspect of peroxisomal metabofism
has been studied in more detail using riboflavin-deficient rats [16,17]. Acknowledgements We thank Mrs. T.-T. Muller for her excellent technical assistance. This work was supported by grants from the Belgian Fonds National de la Recherche Scientifique (FNRS), by grant 84/90-'74 of the Action de Recherche Concert~e du Premier Ministre, US Public Health Services (grant 2RO1 DK9235). Equipment for HPLC was provided by Action for the Crippled Child, and the Muscular Dystrophy Group of Great Britain. J.V. is Charg~ de Recherches of the Belgian FNRS. K.V. was recipient of a Royal Society (London) European Exchange Fellowship.
References 1 Goodman, S.I. (1981) Am. J. Ciin. Nutr. 34, 2434-2437. 2 Gregersen, N. and Koivraa, S. (1982) J. lnher. Metab. Dis. 5, Suppl. 1, 17-18. 3 Gregersen, N. (1985) J, lnher. Metab. Dis. 8 (Suppl. 1), 65-69. 4 Hoppel, C., DiMarco, J.P. and Tandler, B. (1979) J. Biol. Chem. 254, 4164-4170. $ Sakurai, T., Miyazawa, S., Furuta, S. and Hashimoto, T. (1982) Lipids 17, 598-604. 6 Ross, N.B. and Hoppel, C.L. (1987) Biochem. J. 244, 387-391. 7 Veitch, K., Draye, J.-P., Van Hoof, F. and Sherratt, H.S.A. (1988) Biochem, J. 254, 477-481. 8 Lazarow, P.B. (1978) J. Biol. Chem. 253, 1522-1528. 9 Morlensen, P.B. and Gregersen, N. (1982) Biochim. Biophys. Acta 710, 477-484. 10 Osumi, T. and Hashimoto, T. (1979) J. Biochem. 85, 131-139. 11 Okayasu, T., Kameda, K., Ono, T. and Hashimoto, T. (1977) Biochim. Biophys. Acta 489, 397-402. 12 lee, S.-S., Ye, J., Jones, D.P. and McCormick, D.B. (1983) Biochem. Biophys. Res. Commun. 117, 788-793. 13 Butch, H.B., Hunter, F.E,, Combs, A.M. and Schutz, B.A. (1960) J. Biol. Chem. 235, 1540-1544. 14 Pmsky, L., Burch, ll.B., Bejrablaya, D, Lowry, O.H. and Combs, A.M. (1964) J. Biol. Chem. 239, 2691-2695. 15 Brady, P,S. and Hoppel, C,L. (1985) Biochem. J. 229, 717-721. 16 Draye, J..P., Veitch, K., Vamecq, J. and Van Hoof, F. (1987) Arch. Int. Physiol. Biochim. 95, B195, 17 Draye, J.-P., Veitch, K., Vamecq, J. and Van Hoof, F. (1988) Eur. J. Biochem. 178, 183-189. 18 Veitch, K., Draye, J..P., Vamecq, J. and Van Hoof, F. (1987) Arch. Int. Physiol. Biochim. 95, B99. 19 Causey, A.G., Middleton, B. and Bartlett, K. (1986) Biochem. J. 235, 343-350. 20 Tillotson, J.A, and Sauberlich, H.E. (1971) J. N,ttr. 101,1454-1466. 21 Turnbull, D.M., Bartlett, K., Younan, S.I.M. and Sherratt, H.S.A. (1984) Biochem. Pharmacol. 33, 475-481. 22 Sherratt, H.S.A. and Osmundsen, H. (1976) Biochem. Pharmacol. 25, 743-750. 23 Estabrook, R.W. (1967) Methods Enzymol. 10, 41-47. 24 Osmundsen, H. and Bremer, J. (1977) Biochem. J. 164, 621-633. 25 Causey, A.G. and Bartlett, K. (1989) Methods Enzymoi., in press. 26 Shepherd, D. and Garland, P.D. (1969) Methods Enzymol. !3, 11-16. 27 Hoppel, C. and Cooper, C. (1968) Biochem. J. 107, 367-375. 28 Fisher, H.F. (1985) Methods Enzymol. 113, 16-27.
343 29 Wilson, M.A. and Cascarano, 3. ~1972) Biochem. 3. 129, 209-218. 30 Miyazawa, S., Ozasa, H., Osumi, T. and Hashimoto, T. (1983) J. Bioehem. 94, 529-542. 31 Lazarow, P.B. (1981) Methods Enzymol. 72, 315-319. 32 Osumi, T. and Hashimoto, T. (1978) Biochem. Biophys. Res. Commun. 83, 479-485. 33 Baudhuin, P., Beaufay, H., Rahman-Li, Y., Sellinger, O.Z., Wattiaux, R., Jacques, P. and de Duve, C. (1964) Biochem. J. 92, 179-184. 34 Lowry, O.H., Rosebrough, N.J., Farr, A.L. and Randall, R.3. (1951) J. Biol. Chem. 193, 265-275. 35 Ikeda, Y. and Tanaka, K. (1983) J. Biol. Chem. 258, 1066-1076. 36 Spurr, A.R. (1969) J. Ultrastruct. Res. 26, 31-43. 37 Baudhuin, P. (1974) Methods Enzymol. 328, 3-20. 38 Weibel, E.R., Staubli, W., Gnagi, H.R. and Hess, F.A. (1969) J. Cell Biol. 42, 68-91. 39 Weibel, E.R. and Gomez, D.M. (1962) J. Appl. Physiol. 17, 343-348. 40 Von Jagow, G. and Klinsenberg, M. (1970) Eur. J. Biochem. 12, 583-592.
41 Latipaa, P.M., Karl
1006(1989)335-343
335
Elsevier BBALIP53254
Altered acyl-CoA metabolism in riboflavin deficiency K. V e i t c h ~*, J.-P. D r a y e ~,2, j. V a m e c q ~, A . G . C a u s e y 3, K. B a r t l e t t 3, H.S.A. S h e r r a t t a n d F. V a n H o o f ~,2
4
t Laboratoire de Chimie Physiologique, L C.P.& UCL, and 2 Department of Pediatric Neurology, Universitd Catholique de Louvain, Brussels (Belgium), ~ Departments of Child Health & Clinical Biochemistry, and ~ Department of Pharmacological Sciences, The Medical School, The University, Newcastle upon Tyne (U.K.)
(Received 9 May 1989)
Key words' Riboflavindeficiency;Mitochondrion;Peroxisome;AcyI-CoA;Flavoprotein;Organicacid; (Rat) We have recently described the effects of r[bof|avin deficiency on the metabolism of dicarboxylic acids (Draye et al. (1988) Eur. J. Biochem. 178, IK3-|89). As both mitochondria and peroxisomes are thought to be involved, we have exam|ned the activities of var|ous enzymes in these organelles in the livers of riboflavin-deficient rats° Mitochondrial ~-ox|dat|on of fatty adds was severely depressed due to loss of activity of the tiwee fatty acyI-CoA dehydrogenases, whereas there was an enhancement of peroxisomal fl-ox|dation due to an increased activiff of the FAD-dependent fatty acyl-CoA oxidase, although the activities of other peroxisoma| flavoproteins, D-amino acid oxidase and glycolate oxidase, were lowered. Hepatoc)te ~r~home~ry revealed an increase in the numbers of peroxisomes, indicating a proliferation induced by the deficiency. The mitoebondrial acyl-CoA dehydrogenases involved in branched-chain amino acid melabolism were also severely decreased leading to characteristic organic acidurias. There was some loss of activiff of the flavin-dependent sections of the electron transport chain (complexes ~ and II), but these were probably not sufficient to affect normal function in vivo. The specificity of these effects allows the use of the riboflavin-deficient rat as a model for the study of dicarboxylate metabolism.
|nU'eduction The measurement of organic acids excreted in urine is often used to detect errors in metabolism, which may present clinically with a variety of symptoms. Defects in /]-oxidation produce an unusual or excessive excretion of dicarboxylic acids and/or glycine conjugates in the urine. Although several animal models have been proposed for the study of dicarboxylic aciduria, the riboflavin-deficient rat [1,2] would be particularly useful since several conditions which produce organic acidurias respond to therapeutic administration of riboflavin (see Ref. 3). Defective fl-oxidation in riboflavin deficiency is due to low activities of the three FAD-dependent acylCoA dehydrogenases [4-7] which catalyse the first intramitochondrial step in the/t-oxidation spiral. However, the peroxisomal/~..oxidation system [81 has also been suggested to play a role in the processing of dicarboxylic acids [9], so it is important to know the
Correspondence (and * present address): K. Veitch, Hormone and Metabolic Research Unit, ICP&UCL 75.29, avenue Hippocrate 75, B-1200 Brussels,Belgium.
condition of this activity in any model system. The 'rate-limiting' step in this process is the FAD-dependent acyl-CoA oxidase [10]. Catalase [11,12] and flavindependent peroxisomal oxidases [12-14] are depressed in riboflavin-deficient rats. However, in those studies describing the effects of the deficiency on peroxisomal fl-oxidation, these other peroxisomal activities were either unaffected or not measured [5,15]. There are no studies in which all these peroxison:al oxidases have been assayed in riboflavin deficiency. We have reported str~dies of the in vivo metabolism of exogenous dicarboxylic acids in riboflavin-deficient rats [16,!71. !1[ere we report the relevant mitochondrial and r~eroxisomal enzyme activities in this model, which were undertaken to elucidate the relative roles of the two organelles in dicarboxylate metabolism. These studies included examination of the sections of the mitochondrial electron transport chain which are flavin-dependent (complexes I and II), and which have been described i, the literature to be affected inconsistently by the vitamin deficiency [4,13]. Our results may, in part, explain the discrepancies between these previous reports. Part of this work has been presented to the Belgian Biochemical Society [l 8].
0005-2760/89/$03.50 © ElsevierSciencePublishersB.V.(BiomedicalDivision)
336 Materials and Methods The sources of chemicals have been given previously [7,17,19]. A semi-synthetic riboflavin-deficient diet was purchased from ICN Pharmaceuticals (Cleveland, OH, U.S.A.) as a powder which was mixed with a 5~ gelatine solution (100 ml/kg) and dried to form 'cakes'. 'Diet-control' animals received the same diet with 20 mg/kg riboflavin added during preparation. Control animals were fed a standard rat chow (A-03, U.A.R., Epinay, France). Except where specified, all animals were maintained on a 12 h light/dark cycle with food and water ad libitum. r
",ration of animals Young male Wistar rats were used in two series of experiments, in the first, six animals (mean body weight 99 :t: 2 g) were housed individually in metabolism cages. Three control animals received chow diet, and three tests the riboflavin-deficient di~t. At weekly intervals all the animals were fasted overnight (18 h), urine being collected over this period. In the second series 18 rats (84 ± 1.3 g) were housed in three groups of six in sawdust-lined cages, with no attempt made to prevent coprophagia other than by daily cleaning. These groups received the control, diet-control and riboflavin-deft. cient diets, respectively, ad libitum. After 6 weeks, all animals were fasted overnight, during which period urine was collected, before killing by decapitation. Livers were rapidly removed, weighed and used to prepare 10~ homogenates and mitochondrial fractions; small liver samples were fixed for electron microscopy as described below. Blood was collected from the tail vein for the measurement of the erythrocyte glutathione reductase activity coefficient as an index of riboflavin deficiency
[20].
Mitochondria Mitochondrial fractions were prepared as described in Ref. 19, and finally re,suspended in 0.3 M mannitol/10 mM Hepe.s/l mM EGTA (pH 7.4) at 25-50 mg protein per ml. The rates of mitochondrial oxidation were measured polarographically in a 3.0 ml volume [21], with I0 mM succinat¢, I0 mM glutamate (+ 1 mM malate) and 33 ~M palmitoylcarnitine as substrates, stimulated by the addition of I ~mol ADP. When assaying fi-oxidation, 5 mM malonate and 1.5 mg/ml defatted bovine serum albumin were included to ensure complete oxidation to acctoacetate, which directly measures the flux through /)-oxidation [22]. State 3, state 4, respiratory corJtrol ratio and A D P / O ratio were determined [23l. Mitocbondrial oxidation rates were also measured spectrophotometrically by the rate of reduction of ferricyanide at 420-475 run [24] in an Aminco DW-2 spectrophotometer using 20 mM succinate, 20 mM glutamate
( + 2 mM malate), 25 ~M palmitoylcarnitine, 25 /tM palmitoyl-CoA (+1 mM L-carnitine) or 50 ltM octanoylcarnitine as substrates. The oxidation of 1 mM 3-[U-14C]methyl-2-oxopentanoate, prepared according to Ref. 19, was assayed [25] with radio-HPLC analysis of mitochondrial extracts. Mitochondrial enzymes, citrate synthase [26], succinate dehydrogenas¢, with phenazine ethosulphate instead of phenazine methosulphate [27], glutamate dehydrogenase [28], NADH dehydrogenase [29] and carnitine palmitoyltransferase [30] were assayed by published methods. Acyl-CoA dehydrogenases were prepared from mitochondrial extracts and assayed in 25 ~tM FAD to protect the enzymes during extraction as described [7]. Citrate synthase was measured in the 10~ liver homogenates as a marker enzyme for mitochondria. Peroxisomes Peroxisomal activities were measured in 10% liver homogenates and stored at - 8 0 ° C until use. Peroxisomal iS-oxidation was assayed at 37°C as the cyanide-insensRive reduction of NAD + with 50 ~tM palmitoyl-CoA as substrate [31]. Acyl-CoA oxidase, Damino-acid oxidase and glycolate oxidase were assayed at pH 7.5 by the H202-dependent iminoquinone formation catalysed by peroxidase [32] with 50/tM lauroylCoA, 25 mM D-proline and 10 mM glycolate as substrates, respectively, in 25 #M of the appropriate ravin cofactor. Urate oxidase and catalase were assayed [33]. All assays, except peroxisomal r-oxidation (37°C) and catalase (0 o C) were performed at 30 o C. Activities are expressed as units per gram of liver or per mg mitochondrial protein, 1 unit being the amount of enzyme required to catalyse the reaction of 1 p mol of substrate or the formation of 1/tmol of product per min under the described conditions. Catalase is expressed in Baudhuin units, 1 unit being the amount of enzyme causing the destruction of 90~ of the substrate in 1 min in a volume of 50 ml under the assay conditions [33]. Protein was measured in sonicated mitochondria and liver homogenates [34], with bovine serum albumin as standard. Organic acids Urine samples were prepared and analysed by gas - ],, chromatography [t I, with pentadecanoic acid as internal standard. The two peaks corresponding to 2-methylbutyrylglycine and isovalerylglycine comigrated in this system, the latter being found in much greater quantities. As both 2-methylbutyryl-CoA and isobutyryl-CoA are substrates for the same acyl-CoA dehydrogenase [35], no further attempt was made to separate the two glycine conjugates, the appearance of isobuty~lglycine indicating a defect in this activity. The identity of the organic acids was confirmed by using an ion-trap detector (Finnigan-Mat model 700).
337
Electron microscopy Samples of liver were immediately fixed in 0.1 M sodium cacodylate/270 (w/v) glutaraldehyde (pH 7.4). Post-fixation was made in 270 (w/v) osmium tetroxide, dehydration in ethanol, and embedding in the mixture as described by Spurr [36]. Ultrathin sections, stained by uranyl acetate and lead citrate were examined with a Philips EM-301 electron microscopy. Micrographs were taken from a minimum of three different sections, each cut at least 5/~m from the preceding one. Magnification (about 15000 x ) was accurately determined with a calibration grid (E.F. Fullam, Schenectady, NY) and stereological analysis performed [37]. A multipurpose test grid similar to that described in Ref. 38 was used for the determination of volume fraction or membrane area, the number of particles being calculated [39]. Statistical analysis of the differences between group means was performed by Student's t-test. Where shown, analyses of linear regression and correlation coefficients were performed with the individual data from 18 preparations, using six animals from each experimental group. Results
Animals and organic acids After only 1 week, both groups receiving the test diet weighed significantly less than chow-fed controls (Fig. 1). The difference in weight between the two control groups, approx. 12%, was mair~tained throughout the experimental period, diet controls weighing 87% of chow-fed controls after 6 weeks ( P < 0.001). The vitamin-deficient animals were more severely affected, weighing 42 and 49% (both P < 0.001) of chow-fed controls and diet-controls, respectively, after 6 weeks. Overnight fasting causing a 6-8% decrease in body weight in each group. After 6 weeks, the erythrocyte 300
O 200
g 0
100
Z < uJ 1
I
I
I
I
I
I
0
1
2
3
4
5
6
WEEKS
ON DIET
Fig. 1. The mean body weights of the three groups of animals during the feeding of the control and test diets. Each point is the mean ol six animals re: con|:~<~J~ ~o), diet controls (A) and riboflavin-deficient group (Ilia. $.EM. ba~'s are not shown, L:zi~g smaller than the symbol:, used. Alter 1 w:~ek all groups were significantly different from each other at each ~me point ( P < 0.01). o , control group; A, diet control group; ~, ~,:~:,oflavin-~r.ficientgroup.
glutathione reductase activity coefficient was significantly increased in the riboflavin-deficient group (1.24 _+0.05, 1.22 + 0.15 and 2.18 _+0.18 for controls, dietcontrols and riboflavin-deficients, respectively, means _+ S.E. for four animals in each group), values greater than 1.35 indicating riboflavin deficiency [20]. Liver weights were greater in riboflavin-deficient animals when expressed as a percentage of total body weight (3.06 _+0.07 and 4.94 _+0.28% for controls and tests, respectively, P < 0.01), but protein content was not significantly affected (250 _+ 6 and 231 _+ 3 mg/g liver). There were no differences between the two control groups. The organic acid excretion over this period was monitored in the first series (Fig. 2). An organic aciduria progressively developed during the 6 weeks of study in the riboflavin-deficient animals. Compounds which were excreted in abnormally high amounts included saturated and unsaturated medium-chain dicarboxylates, glutaric acid and the glycine conjugates of butyrate, isobutyrate and isovalerate (and 2-methylbutyrate). The amounts excreted after 6 weeks in the second series are shown in Table I. Most of the dicarbc, xylates excreted were unsaturated, as octenedioic and decenedioic acids. Such large amounts of dicarboxylates indicate a major defect of fl-oxidation in the liver [1].
Mitochondria The impairment in the rate of fl-oxidation caused by riboflavin-deficiency was confirmed (Tables lI&Ill), particularly that of palmitoylcarnitine which was decreased to 36% of controls (Table II). Succinate oxidation was 30% of control rates, and glutamate oxidation was further decrea3ed (56% of controls). The mitochondria from livers of riboflavin-deficient rats were not damaged as shown by the small effects on respiratory control, A D P / O ratios and nearly complete inhibition (96-98%) of ferricyanide reduction by 1 # g / m l antimycin (not shown). These effects were confirmed by the spectrophotometric assay of ferricyanide reduction which collects electrons from the respiratory chain at the level of eytochrome c, and which excludes the activity of complex IV from the rate (Table III). With this method, succinate oxidation was decreased by 50%, while glutamate/malate oxidation was 62% of controls, whi:h is not markedly different from the polarographic data. The effect on succinate can be explained by the activity of succ~aa~e dehydrogenase, the FAD-dependent enzyme. This was decreased to 47% of co=trois in riboflavin-deficier,t preparations (Table.~ IV), and there was a highly significant co~re~atioJ~ betLween this activi'~y a~,d th:. rate of succinate oxidation assayed spectrophotomz~,';cal!y (r--~0.903, P<0.001). The differences observed between the polarographic and spectrophotomettic data confirm the observation that succinate dehydro-
338 1200 1200 10( 1000600800200400Adipic c ~ ai ~ ~ .,,,,
c8 @ ¢0 O)
0 ft
1
2
3
4
5
6
1200 1000 800 600 400 200 0
1 2 3 4 5 6
:t oot,.oo,o/
Butyrylglycine
1 2 3 4 5 6
_
E O)
~ >.
100 80 60 40 20 0
0
120 100 80 60 40 20
Isobutyrylglyoine
........
~
2 3 4
56
1
sciatic ac,a
400 3OO 2OO tO0 "--~' o L----~:'~'~'~'~r'~-1
2 3 4
S
23
4
2000 . Oecenedlolc p,~ e,. ~ / \'o tOO0 0
6
12
34
1 2 3 4 5 6
56
56
600 50O 400 300 200 100 0
1 2 3 4
56
WEEKS ON DIET Fig. 2, Excretion of the major organic acids found in urin~~during the experimental period. Urine samples were collected at weekly intervals during an 18 h fast. Each point is the mean of three animals fed control chow diet (o), or riboflavin-deficient diet (O).
genase is not the rate-limiting step in succinate oxidation when it proceeds as far as molecular oxygen [4], and suggest that such control occurs in complex IV subsequent to cytochrome c. Glutamate dehydrogenase activity was increased by the deficiency (Table IV). However, glutamate oxidation is NADH-dependent, The FMN-dependent NADH dehydrogenase was decreased to 53% by the deficiency (Table IV), and there were significant correlations of TABLB I The m ~ w orgem~c¢~Ms excreted in the w~ne d~rmg an oeernigkt fast 08 k), oftee 6 weeks on the expeelmental diets
Valuesare the means:l:S,E for six animalsin eachgroup expressedas pg/mg creatinine. Compound
Control group
Isolmtyrylglycine Butyrylglycine lsovaleuIglycinea Su~inate (31uta~ ¢cid
<3 <3 43.6:1:14,8 61.$± 14,4
Ac~'.'pi¢aci~ Octe.nedioic acid (ul~saturated suberatc) Su~ctic ~..id Oe~;~'aedio~cacid
(unsatu~t~xlsebacat©) Sebacic acid
"~ 3
S,9± 3.1 .~.5+ 4,4
4,2 ± ~,0
Riboflavin. deficient group 38,6+ 231 ~ 502 ± 16,3± 163 ±
9,8 58 72 6,8
52 7~0 :l: 166 1943 :l:601 174 ~- 34
2"/7 -I-80
2084 +621
27.2+11.4
649 +_184
a Total of isovaleryiglycine aad 2-methyibutyrylglycine.
this activity with the rates of glutamate/malate oxidation assayed polarographically (r ffi 0.83, P < 0.001) and spectrophotometrically (r - 0.63, P < 0.005). 0-Oxidation of fatty acids also depends on NADH dehydrogenase through the action of the 3-hydroxyacylCoA dehydrogenase, the third step in the r-oxidation spiral, which reduces NAD +. The rate of r-oxidation in vivo is partially controlled by the N A D H / N A D + ratio [41]. However, if the rate of oxidation of glutamate/ malate indicates NADH dehydrogenase activity, then the deficient mitochondria would have sufficient capacity to cope with NADH generated by fi-oxidation (Table !I). This suggests that the rate of r-oxidation is impaired by the low activities of the three FAD-dependent fatty acyl-CoA dehydrogenases (Fig. 3), the decreases in both short-chain and long-chain activities were about the same as in the two assays of oxidation rate. As the rates of oxidation of palmitoylcarnitine and paimitoyl-CoA were equally affected (Table Ill), carnifine palmitoyltransferase (CPT) was not apparently limiting in these assays, indeed CPT activity increased by 158~ in the deficient animals (Table IV). 3-Methyl-2-oxopentanoate formed by transamination of isoleucine, is oxidat~vely decarboxylated in the n~'~ochondrial matrix to 2-methylbutyryI-CoA, which is subsequently oxidised via intermediates such as propionylCoA and methylmalonyl-CoA to acetyl-CoA and succinyl-CoA (see Ref. 19). 2-MethylbutyryI-CoA dehydrogenase activity was markedly decreased in riboflavin deficiency (Fig. 3), associated with excessive excretion
339 TABLE I|
The effects of riboflavin-deficiency on hefatic mitochondriol oxygen uptake Values are the means +_S.E. of s;x preparations in each expressed as nanogram aioms O per rnin per mg protein. Respiratory control ratio (RCR) is the ratio of the rates of oxidation of sta ie 3 on state 4. * P < 0.05; * * P < 0.01; * * * P < 0.001 vs. controls. Substrate
Group
State 3
State 4
RCR
ADP/O
control diet control ribofl.-deficient
163 ± 9 155 + 7 131 -+ 8 *
48 + 2 47 + 1 39 + 3 *
3.45 + 0.21 3.30 + 0.19 3.44 + 0.23
1.62 ± 0.05 1.65 ± 0.06 1.80 + 0.08
control diet control ribofl.-deficiem
109 + 7 106 + 6 61 + 1 * * *
26 + 1 26 ± 1 21 5:1 * * *
4.25 + 0.37 4.19_+ 0.24 2.96 ± 0.11 * *
2.54 ± 0.08 2.43 ± 0.07 2.47 ± 0.14
control diet control ribofl.-deficient
69 5:5 76 + 7 25 5:1 * * *
22 :t: 3 26 5: ! 20 5:1
3.43 + 0.38 2.95 ± 0.17 1.25 + 0.07 * * *
2.40 ± 0.09 2.38 ± 0.07 2.69 ± 0.4
Succinate
Glutamate ( + malate)
Palmitoylcarnitine
.c_
TABLE I!1
@
The effects of riboflavin deficiet~cv ot~ spe~tropholometriccdly mea.sured hepatic mitochondrial oxidations Values are means + S.E. for the six preparations in Table lI, expressed as nmoles Fe(CN)63- reduced per minute per mg protein (Figures in parentheses are the percentages of the control values.) * * P < 0.01; • * * P < 0.001 vs. Controls. Substrate
Control group
Diet control group
Riboflavin-deficient group
Succinate Glutamate (+malate) Palmitoylcarnitine PalmitoylCoA Octanoyleamitine
437+26
425_+16
(97)
219_+14 * * * (50)
2055:17
194_+19
(95)
127_+12 **
87+ 8
84+ 5
(97)
65 5 : 7
67 + 3 (103)
16 5 : 2 * * * (25)
88 ± 6
84+ 5
25 5 : 3 * * * (28)
(62)
25
f: 20 15
c
•~ 2o
~-C= 10 O 1:3 o
E
0
jjj_ I
ql
e-
21+ 2"**
40
Butyryl
Uctanoyl Palmitoyl Isovaleryl
Isobulyryl
1.0 0.5
0.0
(24) AcyI-CoA ester substrate
(95)
of both 2-methylbutyrylglycine (measured with the iontrap detector, not shown) and isobutyrylglycine (Fig. 2 and Table I). This decreased activity causes a 10-fold greater accumulation of 2-methylbutyryl-CoA in
Fig. 3. The activities of the five mitochondrial acyi-CoA dehydrogenases in the three experimental groups. Each value is the mean of six preparations for control group (open bars), diet control group (hatched bars) and riboflavin-deficient group (closed bars), with S.E.M. bars as shown.
mitochondria from riboflavin-deficient rats incubated with 3-oxo-2-[UJ4C]methylpentanoate compared with raitochondria from control and diet-control animals
TABLE IV
Mitochondriai enzyme activ~zies Values are the means+S.F, for the six prepalations in each group as given in Table If, expressed as U / m g protein. (Figures in parentheses are percentages of control vak~es.) * * * P < 0.001 vs. controls. Enzyme
~.i;o~rol groul~
Diet cch~r,,I group
RJboflavi~-deficient grou~
Succinate dehydrogenase Glutamate dehydrogenas(: N A D H dehydrogenase CPT Citrate synthase
0.048 :J: 0.003 0.179 + 0.0~ ~ 2.49 :t 0.11 1.5i :1:0.11 0.210_+_ 0.013
0.046 5:0.003 (96) 0.211 + 0.018 (18~) 2.45 5:0.14 (99) 1.38 +0.04 (91) 0.208 + 0.011 (99)
0.023 :t 0.002 * ~' * (4'1) 0.395 5:0.025 * * * (221) 1.31 5:0.08 * * * (53) 3.89 5:0.44 * * * (258) 0.239+0.018 (114)
340 TABLE V
Acyl.CoA intermediates accumulated in liver mitochondria during 3 rain incubation with I mM 3-[U-14C]methyl-2-oxope ntanoate Values shown are the means ± S.E. for four preparations from controls and riboflavin deficient rats and the mean of two diet controls, expressed as d.p.m, per mg mitochondrial protein. ** P < 0.005, * * * P < 0.001 vs. controls. AcyI-CoA Ester
2.MethylbutyrylCoA 2°Methylbut-2-enoyl-
CoA PropionyI-CoA AcelyI.CoA MelhylnmlonyI.CoA SuccinyI.CoA
Control group
Diet control
Riboflavin deficient
group
group
464 ± 23
540
4 543 ± 345 * * *
93 ± 10 140 ± 13 245 ± 22
79 119 349 735 517
n,d. n,d, 166 ± 31 190± 60 **
555~:44 546 ± 54
n.d.
n,d,. not detected,
(Table V). Assuming that all the metabolism of 3.oxo-2[U.t4C]methylpentanoate occurs in the matrix, and a mitochondrial content of CoA of 2 nmol/mg protein and a matrix volume of 0.5 p l / m g protei~t, it was estimated from the UV HPLC trace [19] mat the matrix concentration of 2.methylbutyryI-CoA was 2.3 mM. 2.methylbut-2-enoyl-CoA, propionyl-CoA and succinylCoA were not detected in incubations with riboflavindeficient mitochondria, and the acetyl-CoA and methylmalonyI-CoA concentrations were 68 and 34S$ of controis, respectively (Table V). Therefore, the mitochondfia from the riboflavin-deficient rats appeared to function normally, apart from selective defects in flavoprotein-dependent activities, most notably of the five acyI-CoA deh)-drogenases. The activities were expressed relative to mitochondrial prorein, which appeared to be valid as the specific activity of citrate synthase was the same for all three groups (Table IV). The specific activity of citrate synthase in liver homogenates was significantly increased (P < 0.001) in riboflavin-deficient preparations (16.4 + 0.8, 16.2 + 1.0 and 22.9 + 0.8 U / g wet weight, for controls, diet-controls and deficients, respectively), but total activity per liver was unchanged (143 + 9, 130 + 8 and 134 + 12 U per liver, respectively).
Peroxisomes Peroxisomes are another site for the/l-oxidation of fatty acyl-CoA esters [8]. The 'rate-limiting' step is also catalysed by an FAD-dependent enzyme, acyl-CoA oxidase [1O]. However, in contrast to the mitochondria the rate of peroxisomal /]-oxidation was greater in homogenates of rivers from riboflavin-deficient rats (Table VI}. It was confirmed that this effect was not due to cfifferent amounts of apoenzyme being present by preincubatb,~ in 100/AM FAD for 10 min. This treatment
was equally effective in increasing activity in controls and tests. There was a similar ivcrease in actual acylCoA oxidase (Table VI), but this was not due to a general increase in peroxisomal flavoproteins, as shown by the marked decreases in activities of D-amino-acid oxidase (FAD) and glycolate oxidase (FMN). The qualitative differences between these activities in controls and tests were unaffected by addition of the relevant co-factor, the data shown being obtained with 25 /~M F A D / F M N . Catalase activity per gram tissue was also decreased in the livers of riboflavin-deficient animals, but the total activity per liver was unaffected. There was no effect on urate oxidase. The very similar dat~ obtained from the two control groups confirmed that these effects were not due to differences in dietary composition, other than the absence of riboflavin for the tests.
Morphometry The surprising observation of an increase in peroxisomal/]-oxidation was supported to some extent by the morphometric analysis presented in Table VII. The fractional volume of ~:he hepatocyte occupied by peroxisomes was not si~i~nificantly affected ( < 10% increase), but there was t~n increase in peroxisomal membrane area (28%), and there was an 82% increase in the number of peroxisomes. This increase in the number of peroxisomes, which were smaller than those found in control tissue, is consistent with peroxisomal proliferation accoiding to current concepts of peroxisomal biogenesis that fission of 'parent' organelles forms smaller 'progeny' [42]. The total fractional volume of the hepatocytes due to mit~hondria was markedly increased (29%) consistent with the biochemical data. This increase in mitochondrial volume content together with the 14% increase in mitochondrial citrate synthase TABLE Vl
Peroxisomal en=yme activities assayed in liver homogenates Means+S.E. for six animals in each group, figures in parentheses being percentages of control values. * P < 0.05, * * P < 0.01, * * * P < 0.001 vs. controls. Activity
Control group
fl-Oxidation a 1.15+0.06 AcyI-CoA oxidase b 1.64+0.04 ~Amino-acid oxidase b 0.79+0.09 Glycolate oxidase b 1,174-0.07 Urate oxidase b 2.30+0.11 Catalase c 59.6 -t-3.2
Diet control Riboflavin deficient group group 1.08+0.08
2.23+0.21 * * * (194)
1.50±0.13
2.72+0.31 **
0.73+0.08
0.22+0.07 * * *
1,28±0.08 2.27+0.09 59.3 +2.2
(166) (27)
0.20+0.03 * * * (17) 2.41 +0.14 (105) 44.2 +4.2 * (74)
pmol N A D + reduced per rain per g liver at 37 o C. b Units per g liver. c 8audhuin units per g liver (see g e l 33).
341 TABLE Vll
Morphornetric data fr,~m control and riboflavin.deficient rat hepatocyws The livers were fixed and prepared as described in Materials and Methods. The r.nalysed section areas in hepatocytes were 1399, 1417 and 1593 pm 2 for control, diet control and riboflavin-deficient rats, respectively. Values are given 5: S.E. of the mean for the analysed sample. The estimation of particle number was obtained by the application of the following equation as described by Weibel and Gomez [391: N~=(Na)3/2/[~(V~)I/2, where N~ is the number of particles per unit volume; Na is the number of profiles per unit area in section; l~ the volume fraction of particles, and p a shape factor. Peroxisomes are considered as spheres (fl ffi 1.382). Control group
Diet control group
Riboflavin-deficient group
Fractional volume (% of the cytoplasmic volume) mitochondria 23.5 5:1.5 22.0 5:1.2 30.2 peroxisomes 1.6 5:0.2 1.5 5:0.2 1.7 lipid droplets 1.1 5:0.7 0.2 5:0.2 2.8
5:0.9 5:0.3 5:0.6
Membrane area per unit volume (/A m2/p m~) peroxisomes 0.152 :t: 0.028 0.132 ± 0.018
0.194:t:0.018
Number of particles per cubic micron peroxisomes 0.107 0.115
0.195
activity (Table Vl) explains the 40% increase in specific activity in liver homogenates. Electron microscopic examination did not reveal any other consistent difference between any of the three groups, although hepatocytes from riboflavin-deficient animals were richer in small elements of smooth endoplasmic reticulum. Discussion
Riboflavin-deficiency in rats has been widely studied since the observation that such animals utilise food inefficiently [43], :.ndicated by an abnormally high respiratory quotient [44]. Subsequent studies were concerned with the effects of the vitamin-deficiency on mitochondrial oxidative phosphorylation, particularly the oxidation of NADH [13,45,46]. These early studies demonstrated a rapid loss of flavin coenzymes from the liver, FAD and FMN falling to 29 and 10%, respectively of control values after 5-6 weeks of deficiency [14] and 25 and 5%, respectively, after 8-10 weeks [13]. In these studies there was no effect on the FMN-dependent NADH dehydrogenase activity (complex I), a result inconsistent with the dramatic effect on the coenzyme concentration Succinate dehydrogenase (complex II) decreased by 75%, with a 50% decrease in succinate oxidation rate [13], an effect confirmed by Ref. 4, although other workers [46] found no effect on the rates of succinate oxidation. These inconsistencies may be explained by our d~ta on succinate and glutamate oxidations, and the respective enzyme activities presented in Tables II-IV. A 50% loss of succinate dehydrogenase activity had a relatively
small effect on the rate of suceinate oxidation when measured with oxygen as final electron acceptor (20%), but 50% when re1. icyanide was the final acceptor. This suggests that the activity of the electron transfer chain subsequent to cytochrome c is rate-limiting. Early studies using polarographic or manometric techniques may therefore have failed to detect small changes in rate of succinate oxidation associated with relatively large losses of enzyme activity. We observed a marked decrease in the rate of oxidation of glutamate (in the presence of 1 mM malate), which we attribute to a loss of N A D H dehydrogenase activity, with an excellent correlation between these two activities. Some groups reported decreased glutamate oxidation [6,15], while others found no change [4,46]. In studies using inhibitors of NADH-ubiquinone reductase, the inhibition of glutamate oxidation by intact mkochondria, and of NADH oxidation after freezethawing have identical dose' responses (Veitch, K. and Hue, L., unpublished data). These results, ar.d those reported in this study, suggest that the flux lhrough complex I is the rate-limiting step in glutamate oxidation and that the observed decrease in this rate ira riboflavin deficiency is due to the loss of N A D H dehydrogenase, consistent with the fall in hepatic FMN content. However, these effects on the respiratory chain complexes are much less than those observed on hepatic fatty acid oxidation, which is affected within 1 week of deficiency [4]. The urinary excretion of organic acids (Fig. 2) is further evidence of the sensitivity of fl-oxidation to the vitamin deficiency, with an increased dicarboxylic aciduria after 1-2 weeks. There is (Fig. 3) loss of activities of the three fatty acid-CoA dehydrogenases [4-7], hence the observed decrease in /t-oxidation. It should be noted that these activities were assayed in mitochondrial extracts containing 25 #M FAD (see Ref. 7) which did not reactivate the medium- and long-chain activities as has been reported [5], in total agreement with [6]. Short- and medium-chain acyl-CoA dehydrogenases are inhibited in hypoglycin-poistmcd rats, which excrete large quantities of unsaturated medium-chain dicarboxylates [47]. This excretion is due to the omega-oxidat;on of linoleic acid metabolites which cannot be fl-oxidised due to the impaired mitochondrial enzyme activities [48]. The octenedioic and decenedioic acids excreted by riboflavin-deficient rats are presumably derived from the same source. The organic aciduria also revealed the sensitivity of amino-acid metabolism to the deficiency, the appearance of glycine conjugates of isobutyrate, isovalerate and 2-methylb~,tsrate being at least as rapid as the dicarbozylic, aciduria. The activities of the two branched-chain acyl-CoA dehydrogenases were more severely affected than the straight-chain activities, and the oxidation of 3-[U-14C]methyl-2-oxopentanoate was
342 drastically impaired. The elevated ratio of 2-methylbutyryI-CoA to 2-methyl-2-enoyi-CoA (controls - 4.9 + 0.8, riboflavin-deficients > 100) indicates a block in the flux at the point of isobutyryl-CoA (2-methylbutyrylCoA) dehydrogenase. The branched-chain 2-oxo-acid dehydrogenase complex, which contains subunits of the FAD-dependent lipoyl dehydrogenase (EC 1.6.4.3), was su:*ficiently active to generate high concentrations of 2-methylbutyryl-CoA by the oxidative decarboxylation of 3-methyl-2-oxopentanoate. All the major organic acids excreted are derived from metabolic p ~ in which the accumulated intermediates are CoA-esters. Glutaric acid is excreted due to the loss of activity of glutaryI-CoA dehydrogenase [1]; alutaryI-CoA being formed in the catabolism of lysine, hydroxylysine and tryptophane. Intracellular accumulation of CoA esters has been suggested as one of the signals for peroxisomal proliferation [49], consistent with an ine,ease in numbers of peroxisomes (Table VII). In early studies [13,14], D-amino acid oxidase and glycolate oxid~tse were used as markers for FAD. and FMN.depo~dent activities, respectively, before these enzymes had been located in the peroxisomes, and shown to be dramatically decreased. Furthermore, decreased catalase activity has been reported in riboflavin-deficient mice [50] and rats [11,12]. In studies on the effects on peroxisomal B-oxidation there were either no significant effects on catalase [5,15] or the enzyme was not assayed [6], other pero~somal activities were not measured. We have confirmed that while there are losses of other flavoprotein peroxisomal oxidases and catalase, there is an enhanced peroxisomal/~-oxidation capacity due to an increase in acyI-CoA oxidase. This specific increase, probably as a result of the decreased mitochondrial ,0-oxidation, susgests a role for peroxisomes in clearing accumulated CoA-esters. In con~usion, riboflavin deficiency in rats decreases the activities of the mitcchondrial electron transpor~ chain (complexes 1 and I!), but these effects are probably not sufficient to affect normal function in vivo. However, the cellular metabolism of CoA esters is severely perturbed due to the loss of several, but not ell, flavin-dependent dehydrogenases. These effects are rapidly apparent with the excretion of the nonmetabolised substrates as dicarboxyfic acids or glycine conjugates in the urine. Furthermore, despite decreased activities of ~ m e peroxisomal flavo-oxidases, there is an increased capacity for peroxisomal/3-oxidation. The loss of mitochondrial ~0-oxidation capacity leads to a dicarboxyfic aciduria, apparently due to insufficient oxidation of medium.chain fatty acids, confirming the relevant importance of the mitochondrial system compared with the peroxisomal one. The relative specificity of these effects mean~ that the riboflavin-deficient rat is suitable from the study of the meta'~(~fism of dicarboxyfi¢ acids [51]. This aspect of peroxisomal metabofism
has been studied in more detail using riboflavin-deficient rats [16,17]. Acknowledgements We thank Mrs. T.-T. Muller for her excellent technical assistance. This work was supported by grants from the Belgian Fonds National de la Recherche Scientifique (FNRS), by grant 84/90-'74 of the Action de Recherche Concert~e du Premier Ministre, US Public Health Services (grant 2RO1 DK9235). Equipment for HPLC was provided by Action for the Crippled Child, and the Muscular Dystrophy Group of Great Britain. J.V. is Charg~ de Recherches of the Belgian FNRS. K.V. was recipient of a Royal Society (London) European Exchange Fellowship.
References 1 Goodman, S.I. (1981) Am. J. Ciin. Nutr. 34, 2434-2437. 2 Gregersen, N. and Koivraa, S. (1982) J. lnher. Metab. Dis. 5, Suppl. 1, 17-18. 3 Gregersen, N. (1985) J, lnher. Metab. Dis. 8 (Suppl. 1), 65-69. 4 Hoppel, C., DiMarco, J.P. and Tandler, B. (1979) J. Biol. Chem. 254, 4164-4170. $ Sakurai, T., Miyazawa, S., Furuta, S. and Hashimoto, T. (1982) Lipids 17, 598-604. 6 Ross, N.B. and Hoppel, C.L. (1987) Biochem. J. 244, 387-391. 7 Veitch, K., Draye, J.-P., Van Hoof, F. and Sherratt, H.S.A. (1988) Biochem, J. 254, 477-481. 8 Lazarow, P.B. (1978) J. Biol. Chem. 253, 1522-1528. 9 Morlensen, P.B. and Gregersen, N. (1982) Biochim. Biophys. Acta 710, 477-484. 10 Osumi, T. and Hashimoto, T. (1979) J. Biochem. 85, 131-139. 11 Okayasu, T., Kameda, K., Ono, T. and Hashimoto, T. (1977) Biochim. Biophys. Acta 489, 397-402. 12 lee, S.-S., Ye, J., Jones, D.P. and McCormick, D.B. (1983) Biochem. Biophys. Res. Commun. 117, 788-793. 13 Butch, H.B., Hunter, F.E,, Combs, A.M. and Schutz, B.A. (1960) J. Biol. Chem. 235, 1540-1544. 14 Pmsky, L., Burch, ll.B., Bejrablaya, D, Lowry, O.H. and Combs, A.M. (1964) J. Biol. Chem. 239, 2691-2695. 15 Brady, P,S. and Hoppel, C,L. (1985) Biochem. J. 229, 717-721. 16 Draye, J..P., Veitch, K., Vamecq, J. and Van Hoof, F. (1987) Arch. Int. Physiol. Biochim. 95, B195, 17 Draye, J.-P., Veitch, K., Vamecq, J. and Van Hoof, F. (1988) Eur. J. Biochem. 178, 183-189. 18 Veitch, K., Draye, J..P., Vamecq, J. and Van Hoof, F. (1987) Arch. Int. Physiol. Biochim. 95, B99. 19 Causey, A.G., Middleton, B. and Bartlett, K. (1986) Biochem. J. 235, 343-350. 20 Tillotson, J.A, and Sauberlich, H.E. (1971) J. N,ttr. 101,1454-1466. 21 Turnbull, D.M., Bartlett, K., Younan, S.I.M. and Sherratt, H.S.A. (1984) Biochem. Pharmacol. 33, 475-481. 22 Sherratt, H.S.A. and Osmundsen, H. (1976) Biochem. Pharmacol. 25, 743-750. 23 Estabrook, R.W. (1967) Methods Enzymol. 10, 41-47. 24 Osmundsen, H. and Bremer, J. (1977) Biochem. J. 164, 621-633. 25 Causey, A.G. and Bartlett, K. (1989) Methods Enzymoi., in press. 26 Shepherd, D. and Garland, P.D. (1969) Methods Enzymol. !3, 11-16. 27 Hoppel, C. and Cooper, C. (1968) Biochem. J. 107, 367-375. 28 Fisher, H.F. (1985) Methods Enzymol. 113, 16-27.
343 29 Wilson, M.A. and Cascarano, 3. ~1972) Biochem. 3. 129, 209-218. 30 Miyazawa, S., Ozasa, H., Osumi, T. and Hashimoto, T. (1983) J. Bioehem. 94, 529-542. 31 Lazarow, P.B. (1981) Methods Enzymol. 72, 315-319. 32 Osumi, T. and Hashimoto, T. (1978) Biochem. Biophys. Res. Commun. 83, 479-485. 33 Baudhuin, P., Beaufay, H., Rahman-Li, Y., Sellinger, O.Z., Wattiaux, R., Jacques, P. and de Duve, C. (1964) Biochem. J. 92, 179-184. 34 Lowry, O.H., Rosebrough, N.J., Farr, A.L. and Randall, R.3. (1951) J. Biol. Chem. 193, 265-275. 35 Ikeda, Y. and Tanaka, K. (1983) J. Biol. Chem. 258, 1066-1076. 36 Spurr, A.R. (1969) J. Ultrastruct. Res. 26, 31-43. 37 Baudhuin, P. (1974) Methods Enzymol. 328, 3-20. 38 Weibel, E.R., Staubli, W., Gnagi, H.R. and Hess, F.A. (1969) J. Cell Biol. 42, 68-91. 39 Weibel, E.R. and Gomez, D.M. (1962) J. Appl. Physiol. 17, 343-348. 40 Von Jagow, G. and Klinsenberg, M. (1970) Eur. J. Biochem. 12, 583-592.
41 Latipaa, P.M., Karl