3-methylglutaconyl-coenzyme A 3-MG-CoA hydratase EC 4.2.1.18 catalyses the fifth step in the leucine degradation pathway, the reversible hydration of 3-MG-CoA to 3-hydroxy-3-methyl-glutar
Trang 13-methylglutaconyl-CoA hydratase and its role in leucine metabolism
Matthias Mack1, Ute Schniegler-Mattox2, Verena Peters3, Georg F Hoffmann3, Michael Liesert4, Wolfgang Buckel4and Johannes Zschocke2
1 Institut fu¨r Technische Mikrobiologie der Hochschule Mannheim, Germany
2 Institut fu¨r Humangenetik, Ruprecht-Karls-Universita¨t Heidelberg, Germany
3 Abteilung fu¨r Allgemeine Pa¨diatrie, Ruprecht-Karls-Universita¨t Heidelberg, Germany
4 Labor fu¨r Mikrobiologie der Philipps-Universita¨t-Marburg, Germany
In humans, isolated deficiencies of each of the six
dif-ferent steps within the leucine degradation pathway
(Fig 1) cause their own characteristic disease [1] The
enzymes of this pathway are primarily located in the
mitochondria Together with the corresponding genes
and their associated metabolic disorders they are
sum-marized in Table 1 3-methylglutaconyl-coenzyme A
(3-MG-CoA) hydratase (EC 4.2.1.18) catalyses the fifth
step in the leucine degradation pathway, the reversible
hydration of 3-MG-CoA to
3-hydroxy-3-methyl-glutaryl-CoA (HMG-CoA) Reduced or absent 3-MG-CoA hydratase activity causes a metabolic block (Fig 1) and as a result, 3-MG-CoA accumulates within the mitochondrial matrix [2,3] 3-MG-CoA is hydrolyzed in the mitochondrion by a yet unknown acyl-CoA hydrolase to form 3-methylglutaconic acid and free CoA, followed by export of
3-methylglutacon-ic acid from the mitochondrion Reduced 3-MG-CoA hydratase activity also produces increased levels of 3-methylglutaric acid and 3-hydroxyisovaleric acid
Keywords
leucine metabolism; 3-methylglutaconic
aciduria type I;
3-methylglutaconyl-coenzyme A hydratase; AUH
Correspondence
M Mack, Institut fu¨r Technische
Mikrobiologie der Hochschule Mannheim,
Windeckstr 110, 68163 Mannheim,
Germany
Fax: +49 6212926420
Tel: +49 6212926496
E-mail: m.mack@hs-mannheim.de
(Received 14 September 2005, revised
3 March 2006, accepted 7 March 2006)
doi:10.1111/j.1742-4658.2006.05218.x
The metabolic disease 3-methylglutaconic aciduria type I (MGA1) is char-acterized by an abnormal organic acid profile in which there is excessive urinary excretion of 3-methylglutaconic acid, 3-methylglutaric acid and 3-hydroxyisovaleric acid Affected individuals display variable clinical manifestations ranging from mildly delayed speech development to severe psychomotor retardation with neurological handicap MGA1 is caused by reduced or absent 3-methylglutaconyl-coenzyme A (3-MG-CoA) hydratase activity within the leucine degradation pathway The human AUH gene has been reported to encode for a bifunctional enzyme with both RNA-binding and enoyl-CoA-hydratase activity In addition, it was shown that muta-tions in the AUH gene are linked to MGA1 Here we present kinetic data
of the purified gene product of AUH using different CoA-substrates The best substrates were (E)-3-MG-CoA (Vmax¼ 3.9 UÆmg)1, Km¼ 8.3 lm,
kcat¼ 5.1 s)1) and (E)-glutaconyl-CoA (Vmax¼ 1.1 UÆmg)1, Km¼ 2.4 lm,
kcat¼ 1.4 s)1) giving strong evidence that the AUH gene encodes for the major human 3-MG-CoA hydratase in leucine degradation Based on these results, a new assay for AUH activity in fibroblast homogenates was developed The only missense mutation found in MGA1 phenotypes, c.719C>T, leading to the amino acid exchange A240V, produces an enzyme with only 9% of the wild-type 3-MG-CoA hydratase activity
Abbreviations
ARE, A + U-rich elements; Gct, glutaconate CoA-transferase; HMG-CoA, 3-hydroxy-3-methylglutaryl-CoA; MBP, maltose binding protein; MGA (MGA1), 3-methylglutaconic aciduria (type I); 3-MG-CoA, 3-methylglutaconyl-CoA; MTP, mitochondrial trifunctional protein.
Trang 23-Methylglutaric acid is synthesized by the action of
an as yet unspecified dehydrogenase on accumulating
3-MG-CoA (Fig 1) whilst 3-hydroxyisovaleric acid is
produced via the enzymatic hydration of
3-methylcro-tonyl-CoA (crotonase, EC 4.2.1.17) (Fig 1) [2] Conse-quently, humans with reduced or absent 3-MG-CoA hydratase activity show excessive urinary excretion of 3-methylglutaconic acid, 3-hydroxyisovaleric acid and
Fig 1 The metabolic pathway of (S)-leucine
( L -leucine) and isovalerate Enzymes
involved are as follows: 1, EC 2.6.1.42,
branched chain amino transferase 1; 2, EC
1.2.4.4 ⁄ 2.3.1.168 ⁄ 1.8.1.4, branched chain
2-keto acid dehydrogenase complex; 3, EC
1.3.99.10, isovaleryl-CoA dehydrogenase;
4, EC 6.4.1.4, 3-methylcrotonyl-CoA
carboxylase 1; 5, EC 4.2.1.18,
3-methylgluta-conyl-CoA hydratase; 6, EC 4.1.3.4,
3-hydroxy-3-methylglutaryl-CoA lyase; 7, EC
2.8.3.–, isovalerate-CoA-transferase; 8,
crot-onase, EC 4.2.1.17 9, unknown.
Table 1 Enzymes, genes, and associated diseases of the human leucine degradation pathway.
Branched chain keto acid dehydrogenase E1,
alpha ⁄ beta subunits
Trang 33-methylglutaric acid [2] Human 3-MG-CoA
hydra-tase deficiency is known as type I 3-methylglutaconic
aciduria (MGA1, MIM 250 950) It has been found
in association with variable phenotypes ranging from
apparently normal development to severe psychomotor
retardation with progressive neurological symptoms
[4] At present, three additional forms of MGA in
humans have been recognized [5] These diseases are
not associated with reduced 3-methylglutaconyl-CoA
hydratase levels and the excretion of 3-MG and
3-methylglutaric acid is secondary Type II MGA
(MIM 302060), also referred to as Barth syndrome, is
a cardiomyopathy associated with neutropenia and
growth retardation and caused by mutations in the
gene encoding tafazzin (TAZ, previously denoted
G4.5) [6] Type III (MIM 258501) or Costeff
syndrome, is a disorder caused by mutations in the
OPA3 gene [7], leading to bilateral optic atrophy
Finally, type IV (MIM 250951) comprises a
heteroge-neous group of patients with progressive neurological
symptoms [5] The molecular basis for type IV MGA
is unknown, however, experiments with the fungus
Aspergillus nidulans carrying null alleles in the known
genes for 3-methylglutaconyl-CoA hydratase and
3-methylcrotonyl-CoA carboxylase strongly suggest
that a second route for 3-MG biosynthesis exists [8]
Interestingly, certain patients with Smith–Lemli–Opitz
syndrome also show abnormally increased plasma
levels of this compound, further challenging our
under-standing of 3-methylglutaconic acid metabolism [9]
Lastly, pregnancy was reported as a possible cause of
MGA [10] For a long time it had been unclear which
enzyme was responsible for the hydratase step within
leucine degradation A 3-MG-CoA hydratase was
partially purified from bovine⁄ ovine liver [11] It was
established that this enzyme catalyses the syn-addition
of water to (E)-3-MG-CoA leading to (S)-HMG-CoA
[12] Another enoyl-CoA hydratase, mitochondrial
crotonase, is not active using HMG-CoA and
measur-ing the reverse (dehydration) reaction [13]
Mitochond-rial trifunctional protein (MTP) is the main enoyl-CoA
hydratase in long chain fatty acid b-oxidation [14]
This enzyme, however, is unlikely to be involved in
leucine degradation since MTP deficiency (MIM
143450, MIM 600890) is not associated with increased
urinary excretion of 3-methylglutaconic acid A protein
was purified from human brain cells by affinity
chro-matography using the immobilized
RNA-oligonucleo-tide (AUUUA)5 or ‘AU’ followed by cloning of the
corresponding gene [15] Interestingly, the gene showed
sequence similarity to enoyl-CoA-hydratases-1
(2-trans-enoyl-CoA-hydratases; EC 4.2.1.17) and its gene
product had weak enoyl-CoA-hydratase activity using
crotonyl-CoA as a substrate [15] The gene encoding this bifunctional protein was named AUH (‘AU bind-ing homolog of enoyl-CoA hydratase’) The RNA-binding activity of the human protein and also of the murine homologue was investigated further, its biologi-cal function, however, remained unclear [16,17] The three-dimensional structure of AUH was determined at 2.2 A˚ resolution and regarding its hydratase activity a high affinity for short-chain substrates was predicted [18]
The first pure preparation of a 3-MG-CoA hydra-tase was obtained from the bacterium Acinetobacter
sp which aerobically grows on isovalerate as sole car-bon and energy source Isovalerate is activated by
a CoA-transferase (2.8.3.–) to give isovaleryl-CoA (Fig 1) Isovalerate is metabolized via isovaleryl-CoA,
an intermediate of the oxidative (S)-leucine degrada-tion pathway [19] The gene for 3-MG-CoA hydratase
in Acinetobacter sp was partially cloned The transla-ted nucleotide sequence had weak similarities to enoyl-CoA-hydratases (30% identity) and also human AUH
It was shown by two independent groups, that MGA1 patients with reduced or absent hydratase activity have mutations within the AUH gene [13,20]
In addition it was shown that AUH has 3-MG-CoA hydratase activity using HMG-CoA as a substrate and measuring the dehydration reaction AUH locates on chromosome 9q22.31
The present work was initiated to kinetically charac-terize AUH on its presumed natural substrate 3-MG-CoA using a new strategy for its synthesis and developing a new assay In addition, a mutant form of AUH (A240V) derived from an MGA1 patient was tested using 3-MG-CoA
Results
Overexpression of AUH in Escherichia coli and purification of the corresponding gene product The gene for AUH which was cloned from a cDNA library by Nakagawa et al [15] encodes 339 amino acids specifying a 40-kDa protein (AUHp40) Western blot analysis of brain extracts consistently revealed a
32 kDa AUH protein and it was thus assumed that the mature form of human AUH in brain has a molecular weight of 32 kDa (AUHp32) [15] For the kinetic characterization of AUH described in the work
at hand, AUH was overproduced in Escherichia coli as
a maltose binding protein fusion (MBP-AUH) The complete AUH gene (producing MBP-AUHp40 in
E coli) but also a truncated form of AUH (producing MBP-AUHp32 in E coli) were ligated into the
Trang 4bacterial expression vector pMAL-c2 Consequently,
two different forms of AUH, namely MBP-AUHp40
and MBP-AUHp32 could be isolated from the
corres-ponding E coli strains AUHp40 and
MBP-AUHp32 were purified to apparent homogeneity In a
subsequent step, the MBP portion of both fusion
teins was removed by proteolysis and the resulting
pro-teins AUHp40 and AUHp30 were again purified by
chromatography Thus, four different pure fractions of
AUH could be generated: MBP-AUHp40, AUHp40,
MBP-AUHp32 and AUHp32 Using the substrate
3-MG-CoA no difference in enzymatic activity was
detected between the four AUH forms MBP-AUHp40,
AUHp40, MBP-AUHp32 and AUHp32 Since the
purification procedure for MBP-AUHp40 was highly
reproducible, the kinetic data were collected using
purified MBP-AUHp40
Enzymatic synthesis of 3-MG-CoA and
glutaconyl-CoA
The substrates 3-MG-CoA and glutaconyl-CoA were
synthesized using recombinant glutaconate
CoA-trans-ferase from the glutamate fermenting bacterium
Acid-aminococcus fermentans [21,22] This enzyme catalyses
the transfer of coenzyme A from a CoA-donor to a
CoA-acceptor (Fig 2) In addition to its natural
sub-strate (R)-2-hydroxyglutarate, glutaconate
CoA-trans-ferase uses glutaconate, 3-methylglutaconate and other
short chain carboxylic acids as CoA-acceptors Because
the CoA-acceptor 3-methylglutaconate was not
com-mercially available, it was produced by alkaline
hydroly-sis of the corresponding dimethylester HPLC analyhydroly-sis
of the enzymatically produced 3-MG-CoA revealed five
signals (Fig 3) The compounds producing the signals
were analyzed by mass spectrometry The compounds
producing the first two signals (peak 1 and peak 2) had
molecular masses corresponding to unreacted
acetyl-CoA and free coenzyme A The compounds producing
the following signals (peak 3, peak 4 and peak 5) were
found to all have the same relative molecular mass of
893 matching the calculated molecular mass of 3-MG-CoA (893.647) Thus, three 3-MG-3-MG-CoA isomers were produced using the enzyme glutaconate CoA-transferase (Fig 4) The three different forms of 3-MG-CoA were separated by HPLC, collected and their concentration was determined using an enzymatic 5,5¢-dithiobis-2-ni-trobenzoate-based assay Subsequently, the 3-MG-CoA isomers were tested using AUH (Fig 3) It was found, that peak 5 (2 mm) was readily converted to (S)-HMG-CoA In addition, free CoA was detected Peak 4 (2 mm) produced significantly less HMG-CoA and also, in this reaction, a substantial amount of free CoA was found Peak 3 (0.5 mm) gave mainly free CoA and only small amounts of HMG-CoA Peak 5, being the best sub-strate, should correspond to (E)-3-MG-1-CoA, the inter-mediate of the leucine degradation pathway (Fig 4) Peak 4 is most likely to correspond to (E)-3-MG-5-CoA Peak 3 is probably (Z)-3-MG-5-CoA
Glutaconyl-CoA was prepared accordingly Also in this reaction two compounds were produced by glut-aconate CoA-transferase The molecules were separ-ated by HPLC, analyzed by mass spectrometry and were found to both have the same relative molecular mass of 881 corresponding to glutaconyl-CoA (881.247) Peak 1 was dominant and most likely was glutaconyl-1-CoA Peak 2 probably was glutaconyl-5-CoA The two isomers were separated from each other However, upon repeated analysis of the isolated compounds, the same two signals appeared The two isomers seem to interconvert into each other making a separation impossible Therefore, a mixture of the two isomers had to be used in the following studies
Kinetic constants for AUH on different CoA-substrates
Besides (E)-3-MG-1-CoA, the potential substrates glu-taconyl-CoA and HMG-CoA as well as crotonyl-CoA, 3-hydroxybutyryl-CoA and 3-methylcrotonyl-CoA were used for the kinetic characterization of AUH The data are summarized in Table 2
Overexpression of a mutant form of AUH and its activity on (E)-3-MG-1-CoA
Mutations in AUH are linked to the metabolic disease MGA1 Most published patients have been homozy-gous or compound heterozyhomozy-gous for null mutations expected to completely remove protein function [13,20] One patient was compound heterozygous for a null mutation and a missense mutation A240V (c.719C>T) This mutant form of AUH was overproduced as an
Fig 2 General mechanism for coenzyme A-transferases The CoAS–
moiety is transferred from the carboxyl group of the CoA-donor
(R1-COO – ) to the carboxyl group of the CoA-acceptor (R2-COO – ).
In the case of glutaconate CoA-transferase from A fermentans,
CoAS – transiently is bound to the c-carboxyl group of bE54 [24].
Trang 5MBP fusion (MBP-AUHp40mut) in E coli according
to the wild-type enzyme and tested using the substrate
(E)-3-MG-1-CoA In these experiments, the specific
activity of MBP-AUHp40mut was 9% (0.068 UÆmg)1
protein) in comparison to the wild-type enzyme
(0.76 UÆmg)1protein) Hence, the mutation A240V
cau-ses a significant loss of enzyme activity
Direct nonisotopic assay of 3-MG-CoA hydratase
in cultured human skin fibroblasts
The nonisotopic 3-MG-CoA hydratase assay that was
developed during this work was evaluated for its use
in testing homogenates derived from human skin fibro-blast cultures Different cell cultures derived from type
I MGA patients and from wild-type controls were grown, fibroblast homogenates were prepared and tes-ted using 3-MG-CoA (10 lm) It was not possible to detect HMG-CoA in this assay probably due to rapid further processing of this common intermediate by enzymes present in the fibroblast homogenate (e.g by 3-hydroxy-3-methylglutaryl-CoA lyase; EC 4.1.3.4) Therefore, 3-MG-CoA (10 lm) was replaced by gluta-conyl-CoA (10 lm), which had been shown in our work to be an excellent substrate for AUH The product of this reaction, 3-hydroxyglutaryl-CoA, was
mAU
0 200 400 600
Acetyl CoA
CoA
(Z)-3-methylglutaconyl-5-CoA
MW 893,36
(E)-3-methylglutaconyl-5-CoA
MW 893,36
(E)-3-methylglutaconyl-1-CoA
MW 893,59
0 200 400 600
1000 mAU
HMG-CoA 800
CoA
(E)-3-methyl-glutaconyl-1-CoA
AUH AUH AUH
0 20 40 60 80 100 120 mAU
CoA
(Z)-3-methyl- glutaconyl-5-CoA
200 400 600 800 1000 mAU
0
CoA HMG-CoA
(E)-3- methyl- glutaconyl-5-CoA
1
2
3
4 5
HMG-CoA
B A
Fig 3 Isomers of 3-MG-CoA as substrates for human 3-MG-CoA hydratase (AUH) (A) 3-MG-CoA was enzymatically synthesized by incuba-ting 100 m M 3-methylglutaconate in 100 m M potassium phosphate pH 7.0 with 1 m M acetyl-CoA (reaction volume 1 mL) Synthesis was started by addition of 0.25 mg glutaconate-CoA-transferase from A fermentans The reaction was analyzed by HPLC and the CoA deriva-tives were detected by their absorbance at 260 nm Five signals were found, analyzed by mass spectrometry and assigned to be free CoA (peak 1), acetyl-CoA (peak 2) (Z)-3-MG-5-CoA (peak 3) (E)-3-MG-5-CoA (peak 4) and (E)-3-MG-1-CoA (peak 5) The determined relative molec-ular masses (MW) of the 3-MG-CoA compounds producing the signals are shown Acetyl-CoA was purchased from Sigma Aldrich (A 2056) and does contain traces of free CoA (peak 1) Nothing is known about the fronting and the peak shoulder of peak 1, however, the compound
is described by the supplier as only approximately 95% pure The tailing of acetyl-CoA (peak 2) most likely also is due to impurities of the commercially available compound If acetyl-CoA (Sigma Aldrich A 2056) only (without the addition of fibroblast homogenate) is applied to the HPLC-system the same picture appears Thus, it seems, that the fronting, the shoulder and the tailing is due to acetyl-CoA and not due to any other compound (B) The peaks 3, 4, and 5 were isolated by HPLC and used as substrates for AUH The AUH assay contained 2 m M of the respective isomers of 3-MG-CoA in a total volume of 25 lL The reaction was started by addition of AUH (1 lg), incubated for 1 h and the CoA products were HPLC-detected by their absorbance at 260 nm Peak 3 produced small amounts of HMG-CoA and large amounts of free CoA Peak 4 produced HMG-CoA and also large amounts of free CoA Peak 5 produced large amounts of HMG-CoA, but also free CoA.
Trang 6readily detectable (Fig 5) We investigated fibroblast
homogenates from two controls and fibroblast
homo-genates from three patients with established MGA
type 1 Patient 1 [20] was homozygous for a mutation
leading to a stop codon at residue 197 (R197X) and
patient 2 [20] was homozygous for a mutation at the
splice acceptor site of intron 8 (IVS8–1G>A) Patient
3 [20] was compound heterozygous for a missense mutation A240V (c.719C>T) in exon 7 and an inser-tion mutainser-tion c.613–614insA This inserinser-tion causes a frameshift that starts at Met205 and leads to the intro-duction of a stop codon after four amino acids The intra-assay variation, estimated by measuring four fibroblast homogenates in a single experiment, was 3.9%, the interassay variation was 5.4% (n¼ 3 days) The fibroblast material from all MGA1 patients pro-duced significantly less (4–16 mUÆmg)1 protein, mean¼ 8 mUÆmg)1protein) of 3-hydroxyglutaryl-CoA
as compared to the two controls (72 mUÆmg)1protein and 80 mUÆmg)1protein) These results show, that the test measuring the hydratase reaction of AUH indeed
is useful for the direct analysis of fibroblast cultures derived from patients The residual activity within the patient material may be due to other enzymes in the fibroblast protein mixture No other specific soluble human enzyme, however, is known to accept glutaco-nyl-CoA as a substrate and to produce 3-hydroxyglut-aryl-CoA
Fig 4 Possible isomeric products of
3-MG-CoA produced by recombinant
glutaconate CoA-transferase (Gct) from
A fermentans Gct was used to produce
3-MG-CoA from (E,Z)-3-methylglutaconate
and acetyl-CoA (A) Gct transfers CoAS – to
either the 1-carboxyl- (left) or the 5-carboxyl
group (right) of (E)-3-methylglutaconate.
(E)-3-MG-1-CoA (left) was the best substrate
for human 3-MG-CoA hydratase (AUH).
According to this scheme (E)-3-MG-5-CoA
(right) bound to Gct isomerizes to give
(E)-3-MG-1-CoA (B) The production of
(Z)-3-MG-5-CoA by Gct is probably due to
the possible trans-conformation of the
5-carboxyl group of (Z)-3-methylglutaconate.
Table 2 Kinetic constants of AUH (human 3-methylglutaconyl-CoA
hydratase).
Substrate
K m
(l M )
V maxa
(UÆmg)1)
k cat
(s)1)
k cat ⁄ K m
(l M )1Æs)1)
(E)-3-Methylglutaconyl-1-CoA 8.3 3.9 5.1 0.6
(R,S)-3-Hydroxy-3-methylglutaryl-CoA
a Specific activities (UÆmg)1) are in lmolÆmin)1x mg protein.
Trang 7In the present study, we characterized the 3-MG-CoA
hydratase reaction of leucine catabolism at the protein
and DNA levels and developed a novel assay for
enzyme analysis in a diagnostic setting The human
AUH protein was first recognized by its ability to bind
A + U-rich elements (ARE) of mRNA Surprisingly,
AUH showed sequence similarity to enoyl-CoA
hydra-tases, suggesting that this enzyme had another function
in cellular metabolism Indeed, AUH had enoyl-CoA
hydratase activity, which was described as an additional
intrinsic function of the protein Its role in intermediary
metabolism, however, was not clear [15] It was
subse-quently shown, that the metabolic disorder MGA1
caused by reduced 3-MG-CoA hydratase activity is
associated with mutations in the AUH gene This
indi-cated that AUH, in addition to its RNA binding
func-tion, must play an important role in leucine catabolism
AUH was overproduced in E coli and characterized
by measuring the reverse reaction, the dehydration of
HMG-CoA to 3-MG-CoA The reaction was followed
photometrically [4]
In order to measure AUH in the forward reaction
(hydratase activity), it was necessary to synthesize
3-MG-CoA, which is not commercially available
Glut-aconate CoA-transferase (Gct) from A fermentans
proved to be useful for the enzymatic production of this compound Earlier, Gct was reported to be specific for (E)-glutaconate and to be completely inactive with (Z)-glutaconate [21] The activity of Gct using the CoA-donor acetyl-CoA and the CoA-acceptor 3-meth-ylglutaconate was relatively low Our data suggest that Gct produces three isomers The production of (Z)-3-MG-5-CoA is probably due to the possible trans-conformation of the C5-carboxyl group of (Z)-3-methylglutaconate (Fig 4) Most of (Z)-3-MG-5-CoA was hydrolyzed by AUH to give free CoA and (Z)-3-methylglutaconate, trace amounts of HMG-CoA, however, were detected (E)-3-MG-5-CoA was a better substrate for hydration with AUH An explanation for this may be, that upon binding to AUH, (E)-3-MG-5-CoA is isomerized to give (E)-3-MG-1-(E)-3-MG-5-CoA (Fig 4A)
An intrinsic isomerase activity has also been reported for 4-hydroxybutyryl-CoA-dehydratase of Clostridium aminobutyricum [23] The isomerization reaction, how-ever, obviously takes time and the acyl-CoA-hydrolase reaction is favored over the hydratase reaction produ-cing free CoA and (E)-3-methylglutaconate The best substrate for AUH was (E)-3-MG-1-CoA (Km¼ 8.3 lm, Vmax¼ 3.9 UÆmg)1, kcat¼ 5.1), which is the intermediate of the leucine degradation pathway Surprisingly, also with this substrate, large amounts
of free CoA were produced Enzyme assays for the
mAU
0
20
40
60
80
100
120
140
1 4
mAU
0 20 40 60 80 100 120 140 mAU
0 20 40 60 80 100 120 140
1
1 2 3 4
glutaconyl-CoA
glutaconyl-CoA
3-hydroxy-glutaryl-CoA
3-hydroxy- glutaryl-CoA
glutaconyl-CoA
Fig 5 Direct nonisotopic assay of 3-MG-CoA hydratase (AUH) in cultured human skin fibroblasts 3-MG-CoA hydratase was tested in fibro-blast homogenates using glutaconyl-CoA (10 l M ) as a substrate The reaction was started by the addition of fibroblast homogenate (55 mg fibroblast proteinÆL)1), incubated for 1 h and the products of the reaction were HPLC-detected by their absorbance at 260 nm (A) As a con-trol, the assay mixture was incubated without the addition of fibroblast homogenates Two different cell cultures derived from a wild-type control (B) and an MGA1 patient (C, homozygous for mutation IVS8-1G>A in the AUH gene) were grown and fibroblast homogenates were prepared in phosphate-buffered saline (55 mg proteinÆmL)1) The compounds producing the signals (peak 1, peak 2, peak 3 and peak 4) were analyzed by mass spectrometry Peak 1 is free CoA, peak 2 probably is glutaryl-CoA, peak 3 is 3-hydroxyglutaryl-CoA and peak 4 is the sub-strate glutaconyl-CoA The fibroblast homogenate derived from the MGA1 patient produces significantly less (9%) of 3-hydroxyglutaryl-CoA confirming AUH deficiency.
Trang 8characterization of AUH have been carried out with
crotonyl-CoA as a substrate or by measuring the
dehy-dratase reaction using HMG-CoA In these
experi-ments, an acyl-CoA-hydrolase activity of AUH was
not detected This is the first report showing kinetic
data for purified AUH, although a 3-MG-CoA
hydra-tase activity was found earlier in fibroblast and
lym-phocyte lysates measuring the hydration reaction [3]
At that time, the substrate [5-14C]3-MG-CoA was
pre-pared by incubation of 3-methylcrotonyl-CoA with
3-methylcrotonyl-CoA-carboxylase in the presence of
NaH14CO3 In this work, Kmvalues for the hydration
of [5-14C]3-MG-CoA of 6.9 lm (fibroblast) and 9.4 lm
(lymphocyte), respectively, were reported [3] The
formation of [5-14C]3-methylglutaconate from [5-14
C]3-methylglutaconyl-CoA was interpreted as nonspecific
hydrolysis Our results suggest that CoA-hydrolysis is
an intrinsic function of AUH
Experiments with the substrate HMG-CoA and
AUH showed that the hydration reaction (kcat¼ 5.1)
is favored by a factor of 20 over the dehydration
reac-tion (kcat¼ 0.26) which is consistent with the main
role of AUH in leucine catabolism, the hydration of
3-MG-CoA Comparing the turnover numbers of
glu-taconyl-CoA (kcat¼ 1.4) and 3-MG-CoA (kcat¼ 5.1)
it is obvious, that 3-MG-CoA is a better substrate
Crotonyl-CoA (kcat¼ 6.8) and 3-hydroxybutyryl-CoA
(kcat¼ 1.7) have higher Km values (12 and 55 mm),
indicating that the missing carboxylate reduces affinity
to the active site Also in this case, the hydration
reac-tion was favored over the dehydrareac-tion reacreac-tion (factor
of 4)
The mutant enzyme MBP-AUHp40mut (A240V),
identified in one MGA1 patient, had a clearly reduced
3-MG-CoA hydratase activity (9% of the wild-type
enzyme) This finding provides further evidence
con-firming that AUH is indeed the main hydratase in the
human leucine degradation pathway and that
muta-tions leading to reduced hydratase activity are
respon-sible for the MGA1 phenotype
The need to differentiate patients with AUH
defi-ciency from patients with other forms of MGA
requires the availability of a sensitive and specific
enzyme assay Our data show that the hydratase
reaction of AUH is favored over the dehydratase
reaction (factor of 20) Hence, measuring the
for-ward reaction in fibroblast homogenates of
patient-derived cells should increase the sensitivity of an
AUH test The product of this reaction, however, is
the common intermediate HMG-CoA, which is
quickly degraded by, e.g
3-hydroxy-3-methylglutaryl-CoA lyase (EC 4.1.3.4), to give acetyl-3-hydroxy-3-methylglutaryl-CoA and
acetoacetate As glutaconyl-CoA is a very good
substrate for AUH and since the product of the hy-dratase reaction, 3-hydroxyglutaryl-CoA, is not an intermediate within human metabolism, we hypothes-ized that glutaconyl-CoA may be used as a substrate for testing AUH activity in a routine setting Indeed,
we were able to show that AUH activity in fibro-blasts can be determined by monitoring the forma-tion of 3-hydroxyglutaryl-CoA The producforma-tion of a small amount of 3-hydroxyglutaryl-CoA in a patient homozygous for a null mutation in the AUH gene may be due to the action of another mitochondrial hydratase, e.g crotonase This will need to be taken into consideration when the assay is used in a diag-nostic setting Nevertheless, we believe that the novel assay may be a superior method for confirmation of AUH deficiency in fibroblast homogenates
In summary, our data show that the main biological function of AUH in human metabolism is the hydra-tion of (E)-3-MG-CoA to (S)-HMG-CoA in the leu-cine degradation pathway
Experimental procedures
Production and purification of human AUH
in E coli The production of AUHp40 (precursor form), AUHp32 (mature form), AUHp40A240 V and AUHp32A240 V in
E coliwas performed using the pMAL-c2 bacterial expres-sion vector [17] The different forms of AUH were pro-duced as fusions to the MBP of E coli The purification of the gene products was carried out as previously described [17] Protein was estimated using the method of Bradford [24]
Production and purification of glutaconate CoA-transferase from A fermentans in E coli
The production of glutaconate CoA-transferase from
A fermentansin E coli and its subsequent purification was carried out as described earlier [22]
Site-directed mutagenesis Plasmids corresponding to constructs a and b [17] were modified using the Stratagene QuikChange Site-Directed Mutagenesis Kit and the mismatch oligonucleotides AUH
GC-3¢ and AUH RP 5¢-GCCATCGAGGACTCGCACA GAGAATATGAGCT-3¢ (the c.719C>T mutation leading
to the amino acid exchange A240V is underlined) The AUH genes were proof-sequenced and no secondary muta-tions were detected
Trang 9Mass spectrometry
The CoA-esters were separated by HPLC, ionized by ESI
and detected by TOF The HPLC system consisted of a
HP1100 series binary-gradient pump, a vacuum degasser (all
from Hewlett-Packard), and a CTC HTS PAL autosampler
(CTC) The dry sample was dissolved in water and 20 lL
injected onto a 4· 40-mm Grom-Sil120 ODS-4 HE column
(3-lm particle diameter; Grom) The samples were separated
from interfering compounds by a gradient between solution
B (acetonitrile +1 vol% formic acid) and solution A (water
+ 1 volume % formic acid) The gradient (1 mLÆmin)1) was
as follows: 0–5 min, 0% B to 83% B; 6–8 min, 100% B All
gradient steps were linear, and the total analysis time,
inclu-ding equilibration, was 10 min A splitter between the
HPLC column and the mass spectrometer was used, and
100 lLÆmin)1 of eluent was introduced into the mass
spec-trometer A LCT TOF (time-of-flight) mass spectrometer
(Micromass) was used in the negative and positive
electro-spray ionization (ESI) mode Nitrogen was used as the
neb-ulizing gas The capillary voltage was 3 kV, the source
temperature was set at 120C, and the optimal cone-voltage
energy was 45 V
Enzymatic synthesis of 3-MG-CoA and
glutaconyl-CoA
Alkaline hydrolysis of dimethyl (E,Z)-3-methylglutaconate
(Sigma-Aldrich, Deisenhofen, Germany) in 1 m NaOH for
30 min under reflux followed by exchange of Na+ against
H+ with the ion exchanger Dowex 50 W· 8 (H+
-form, Serva, Heidelberg, Germany) yielded
(E,Z)-3-methylgluta-conic acid Its CoA-derivative 3-MG-CoA was enzymatically
synthesized by incubating 100 mm 3-methylglutaconate in
100 mm potassium phosphate pH 7.0 with 1 mm
acetyl-CoA The reaction (1 mL) was started by addition of
0.25 mg recombinant glutaconate CoA-transferase (EC
2.8.3.12) from A fermentans which was purified from an
overproducing E coli strain [22] After 1 h at room
tempera-ture the reaction was stopped by addition of an equal
vol-ume of 8 m guanidinium chloride (1 mL) and the pH was
adjusted to 3–4 using 1 m HCl The desired product
3-MG-CoA was separated from unreacted compounds using a
Waters Sep – Pak tC18 column (1 mL cartridge, 100 mg
sorbent) (Waters, Eschborn, Germany) The column was
first treated with 10 volumes 0.1% (v⁄ v) trifluoroacetic acid
in H2O, 50% acetonitrile (v⁄ v) The acetonitrile component
also contained 0.1% trifluoroacetic acid Subsequently, the
column was washed with 10 volumes 0.1% (v⁄ v)
trifluoro-acetic acid in H2O The samples were loaded and the column
was washed with 10 volumes 0.1% (v⁄ v) trifluoroacetic acid
in H2O The CoA-ester was eluted with 10 volumes 0.1%
(v⁄ v) trifluoroacetic acid in H2O, 50% acetonitrile (v⁄ v)
con-taining 0.1% trifluoroacetic acid After evaporation in vacuo,
the CoA-ester was dissolved in water and stored at)20 C
For analytic and preparative purposes a Phenomenex Syn-ergi 4 l Polar-RP 80 A column (5 lm) was used at a flow rate of 1 mLÆmin)1 (Phenomenex, Aschaffenburg, Ger-many) The eluents were 0.1% trifluoroacetic acid (v⁄ v) in
H2O (solution A) and 0.085% trifluoroacetic acid (v⁄ v) in acetonitrile (solution B) The columns were equilibrated for
5 min with solution A After injection of the CoA derivative
a linear gradient was applied from 0 to 8% solution B within
3 min in order to remove impurities The separation was achieved with a gradient from 8 to 13% solution B within
20 min Afterwards the column was regenerated with 100% solution B for 15 min followed by 100% sol A for 10 min The CoA esters were detected by their absorbance at
260 nm, evaporated in vacuo and dissolved in H2O The CoA esters were identified by their masses using mass spectro-metry The concentration of 3-MG-CoA and glutaconyl-CoA was determined enzymatically in a cuvette (1 mL) containing 100 mm potassium phosphate pH 7.0, 200 mm sodium acetate, 1 mm 5,5¢-dithiobis-2-nitrobenzoate and
1 mm oxaloacetate, k¼ 412 nm, e ¼ 14.2 mm)1Æcm)1[22,25]
An increase in absorbance after addition of 3-MG-CoA was due to free CoASH If acetyl-CoA was present, a further increase followed the addition of 10 lg citrate synthase (Roche, Mannheim, Germany) The final increase after addi-tion of 10 lg glutaconate CoA-transferase was proporaddi-tional
to the concentration of 3-MG-CoA The CoA-substrate glutaconyl-CoA was prepared accordingly from glutaconic acid (Fluka 49360) and acetyl-CoA
Other CoA-substrates (R,S)-HMG-CoA, crotonyl-CoA, 3-methylcrotonyl-CoA and 3-hydroxybutyryl-CoA were obtained from Sigma-Aldrich
Assay of 3-MG-CoA hydratase The assay contained in a total volume of 25 lL 50 mm Tris HCl pH 7.4, 10 mm EDTA, 1 mgÆmL)1bovine serum albu-min and 0.05–0.2 mm 3-MG-CoA The reaction was started
by addition of the enzyme (1 lg) The products of the reac-tion were analyzed by HPLC as described above and mass spectrometry The kinetic constants Km (lM) and Vmax
(UÆmg protein)1) were evaluated with the Michaelis-Menten equation and Lineweaver-Burk plots using the Microsoft Excel program The turnover numbers, kcat (s)1), were calculated with the subunit molecular mass (78.4 Da) of MBP-AUHp40
Direct nonisotopic assay of 3-MG-CoA hydratase
in cultured human skin fibroblasts Fibroblasts were grown and harvested as described elsewhere [26] Cells were suspended in 200 lL phosphate-buffered
Trang 10saline by repeated pipetting and sonicated three times on ice
for 15 s at 8 W at 45-s intervals An aliquot of the fibroblast
homogenate (55 mgÆL)1fibroblast protein) was added to the
3-MG-CoA hydratase assay Protein was estimated using the
method of Bradford [24] The 3-MG-CoA hydratase assay
mixture contained, in a final volume of 25 lL, 100 mm
Tris-HCl (pH 8.0), 10 mm EDTA, 1 gÆL)1bovine serum albumin
and 10 lm 3-MG-CoA or glutaconyl-CoA After incubation
at 37C for 60 min, the reaction was terminated by the
addi-tion of 2.5 lL of 2 m HCl The samples were homogenized,
and the assay tubes were placed on ice After 5 min, the
homogenates were brought to pH 6 with 2 mm KOH, 1 mm
Mes (pH 6) and centrifuged at 21 000 g for 10 min at 4C
The supernatant was transferred to an HPLC vial The
products of the reaction were detected at 260 nm using the
HPLC system described above for the synthesis of the
CoA-esters The assay was linear with an incubation time up to at
least 60 min with up to 70 mgÆL)1total protein
Acknowledgements
We are grateful to C Moroni and J Nakagawa for
sharing the cDNA encoding human AUH This work
received financial support from the Deutsche
Fors-chungsgemeinschaft, Grant number Zs 17⁄ 4–2
References
1 Sweetman L & Williams JC (2001) Branched chain
organic acidurias The Metabolic and Molecular Bases of
Inherited Disease(Scriver, C R, Beaudet, A L, Sly, W S
& Valle, D, eds), pp 2125–2163 McGraw-Hill, New
York
2 Duran M, Beemer FA, Tibosch AS, Bruinvis L, Ketting
D & Wadman SK (1982) Inherited 3-methylglutaconic
aciduria in two brothers – another defect of leucine
metabolism J Pediatr 101, 551–554
3 Narisawa K, Gibson KM, Sweetman L, Nyhan WL,
Duran M & Wadman SK (1986) Deficiency of
3-methyl-glutaconyl-coenzyme A hydratase in two siblings with
3-methylglutaconic aciduria J Clin Invest 77, 1148–1152
4 Shoji Y, Takahashi T, Sawaishi Y, Ishida A, Matsumori
M, Enoki M, Watanabe H & Takada G (1999)
3-Methyl-glutaconic aciduria type I: clinical heterogeneity as a
neurometabolic disease J Inherit Metab Dis 22, 1–8
5 Gibson KM, Elpeleg ON, Jakobs C, Costeff H & Kelley
RI (1993) Multiple syndromes of 3-methylglutaconic
aciduria Pediatr Neurol 9, 120–123
6 Bione S, D’Adamo P, Maestrini E, Gedeon AK,
Bolhu-is PA & Toniolo D (1996) A novel X-linked gene, G4.5
is responsible for Barth syndrome Nat Genet 12, 385–
389
7 Anikster Y, Kleta R, Shaag A, Gahl WA & Elpeleg O
(2001) Type III 3-methylglutaconic aciduria (optic
atrophy plus syndrome, or Costeff optic atrophy syndrome): identification of the OPA3 gene and its founder mutation in Iraqi Jews Am J Hum Genet 69, 1218–1224
8 Rodriguez JM, Ruiz-Sala P, Ugarte M & Penalva MA (2004) Fungal metabolic model for type I 3-methylgluta-conic aciduria J Biol Chem 279, 32385–32392
9 Kelley RI & Kratz L (1995) 3-Methylglutaconic acide-mia in Smith–Lemli–Opitz syndrome Pediatr Res 37, 671–674
10 Walsh R, Conway H, Roche G, Naughten E & Mayne
PD (1997) 3-Methylglutaconic aciduria in pregnancy Lancet 349, 776
11 Hilz H, Knappe J, Ringelmann E & Lynen F (1958) Methylglutaconase, eine neue Hydratase, die am Stoff-wechsel verzweigter Carbonsa¨uren beteiligt ist Biochem
Z 329, 476–489
12 Messner B, Eggerer H, Cornforth JW & Mallaby R (1975) Substrate stereochemistry of the hydroxymethyl-glutaryl-CoA lyase and the methylglutaconyl-CoA hydratase reactions Eur J Biochem 53, 255–264
13 Ijlst L, Loupatty FJ, Ruiter JP, Duran M, Lehnert W & Wanders RJ (2002) 3-Methylglutaconic aciduria type I
is caused by mutations in AUH Am J Hum Genet 71, 1463–1466
14 Wanders RJLIJ, Poggi F, Bonnefont JP, Munnich A, Brivet M, Rabier D & Saudubray JM (1992) Human trifunctional protein deficiency: a new disorder of mito-chondrial fatty acid beta-oxidation Biochem Biophys Res Commun 188, 1139–1145
15 Nakagawa J, Waldner H, Meyer-Monard S, Hofsteenge
J, Jeno P & Moroni C (1995) AUH, a gene encoding an AU-specific RNA binding protein with intrinsic enoyl-CoA hydratase activity Proc Natl Acad Sci USA 92, 2051–2055
16 Brennan LE, Nakagawa J, Egger D, Bienz K & Moroni
C (1999) Characterisation and mitochondrial localisa-tion of AUH, an AU-specific RNA-binding enoyl-CoA hydratase Gene 228, 85–91
17 Nakagawa J & Moroni C (1997) A 20-amino-acid autonomous RNA-binding domain contained in an enoyl-CoA hydratase Eur J Biochem 244, 890–899
18 Kurimoto K, Fukai S, Nureki O, Muto Y & Yokoyama
S (2001) Crystal structure of human AUH protein, a single-stranded RNA binding homolog of enoyl-CoA hydratase Structure (Cambridge) 9, 1253–1263
19 Liesert M (2000) Untersuchungen Zur Biochemie der Vererbbaren Stoffwechselerkrankung Glutarazidurie Typ I, PhD Thesis, Philipps-Universita¨t, Marburg
20 Ly TBN, Peters V, Gibson KM, Liesert M, Buckel W, Wilcken B, Carpenter K, Ensenauer R, Hoffmann GF, Mack M et al (2003) Mutations in the AUH gene cause 3-methylglutaconic aciduria type I Hum Mutat 21, 401– 407