Two novel variants of human medium chain acyl-CoA
dehydrogenase (MCAD)
K364R, a folding mutation, and R256T, a catalytic-site mutation
resulting in a well-folded but totally inactive protein
Linda P. O’Reilly
1,
*, Brage S. Andresen
2
and Paul C. Engel
1
1 Department of Biochemistry and Conway Institute of Biomolecular and Biomedical Research, University College Dublin, Belfield, Dublin,
Ireland
2 Research Unit for Molecular Medicine, University Hospital, Skejby Sygehus, Aarhus, and Institute of Human Genetics, Aarhus University,
Denmark
Keywords
active site; enzyme deficiency; medium
chain acyl-CoA dehydrogenase (MCAD);
point mutations; protein folding
Correspondence
P. C. Engel, Department of Biochemistry,
Conway Institute, University College Dublin,
Belfield, Dublin 4, Ireland
Fax: +353 12837211
Tel: +353 17166764
E-mail:
Website: />Enzymes
Medium chain acyl-CoA dehydrogenase
(MCAD; EC 1.3.99.3); long chain acyl-CoA
dehydrogenase (LCAD; EC 1.3.99.13); short
chain acyl-CoA dehydrogenase (SCAD;
EC 1.3.99.2); glutaryl-CoA dehydrogenase

R256T, by contrast, is a well-folded protein that is nevertheless devoid of
catalytic activity. How the mutations specifically affect the catalytic activity
and the folding is further discussed.
Abbreviations
ACAD, acyl-CoA dehydrogenase; BCIP, 5-bromo-4-chloro indol-3-yl phosphate; DCPIP, 2,6-dichlorophenolindophenol; ETF, electron-
transferring flavoprotein; GCD, glutaryl-CoA dehydrogenase; INT, 2-(4-iodophenyl) 3-(4-nitrophenyl) 5-phenyl-tetrazolium chloride; IVD,
isovaleryl-CoA dehydrogenase; NBT, 2,2¢-di-p-nitrophenyl 5,5¢-diphenyl 3,3¢-(3,3¢-dimethoxy-4,4¢-diphenylene) ditetrazolium chloride; PES,
phenazine ethosulphate; SCAD, short-chain acyl-CoA dehydrogenase; VLCAD, very long chain acyl-CoA dehydrogenase.
FEBS Journal 272 (2005) 4549–4557 ª 2005 FEBS 4549
Medium chain acyl-CoA dehydrogenase (MCAD) is a
homotetrameric flavoprotein that catalyses one of the
recurrent steps in the b-oxidation of fatty acids.
MCAD functions by removing two reducing equiva-
lents from the fatty acyl substrate, and donating them
to electron-transferring flavoprotein (ETF), which ulti-
mately feeds them to the electron transport chain to
generate ATP [1]. Within the active site, the substrate
fatty acyl chain is sandwiched between the isoalloxa-
zine ring of the FAD prosthetic group and the carb-
oxyl group of the catalytic glutamic acid residue
(E376). This residue removes one hydrogen as a proton
from the C-2 position of the fatty acid thioester. The
other is simultaneously removed as a hydride ion by
the N-5 position of the isoalloxazine ring. With this
oxidation step the FAD becomes reduced to FADH
2
[2]. The reducing equivalents are then transferred from
the reduced flavin of MCAD to ETF, and the enoyl-
CoA is released to be further degraded via the b-oxida-
tion cycle [1].

(i.e. elevated hexanoylglycine in urine), have so far
remained clinically and developmentally normal [11].
The second mutation, K364R (MCAD1165AfiG),
was detected in homozygous form in a UK child of
Asiatic origin exhibiting biochemical indications of
MCAD deficiency.
In order to assess whether these newly discovered
missense mutations affect the ability of the MCAD
protein to fold, expression levels in Escherichia coli
were monitored, with and without the co-overexpres-
sion of GroEL and GroES [6,9,12]. Stability of the
mutant proteins in this model system was further
investigated by determining the effect of temperature
on the enzyme activity and structure. The MCAD
mutants were also purified so that the kinetic parame-
ters and stability of the homogeneous proteins could
be investigated. Interaction with the natural electron
acceptor, electron-transferring flavoprotein (ETF) was
also tested, to give a more complete picture of the bio-
chemical outcome of these point mutations.
Results
The effect of chaperonin on the ability of the
mutant proteins to fold
For a number of other MCAD point mutations,
co-overexpression of the GroELS genes has been
shown to rescue enzyme activity, suggesting that these
mutations may affect folding in vivo [6,12]. Therefore
the GroELS genes were overexpressed with those for
each of the MCAD mutants, K364R and R256T, and
for the wild-type enzyme, in E. coli cells grown both at

)1
, an increase of
almost 60% when compared with growth at 31 °C.
This figure may in part reflect upregulation of both
MCAD and chaperonin. It is clear, however, that,
under these more stressful (though physiological) con-
ditions of temperature, even the wild-type enzyme
depends upon chaperonin for optimal folding in this
recombinant overexpression system.
MCAD catalysis (R256T) and folding (K364R) mutants L. P. O’Reilly et al.
4550 FEBS Journal 272 (2005) 4549–4557 ª 2005 FEBS
No MCAD activity was detected in lysates of cells
expressing R256T, but, at both temperatures, the fold-
ing of this mutant protein appears to be even more
successful than for wild-type enzyme, as judged both
by the levels of MCAD tetramer present on the west-
ern blot and by the minimal effect of chaperonin
(Fig. 1B). This suggests that the R256T mutation
harms the function of the folded protein rather than
the ability of the protein to fold.
K364R, by contrast, appears to be a severe tempera-
ture-sensitive folding mutation. In the absence of chap-
eronin, no activity was detected at either growth
temperature (Fig. 1A). With chaperonin, the activity
expressed at 31 °C (2950 nmol ferriceniumÆmg
)1
Æh
)1
)
could be rescued to 25% of the wild-type figure, but

activity.
Kinetic parameters of the purified variant proteins
The mutant and wild-type proteins were purified to
homogeneity by anion exchange and dye affinity chro-
matography, with a novel substrate elution procedure
proving very effective in securing a pure protein prod-
uct. The results of kinetic analysis are displayed as
Michaelis–Menten plots in Fig. 3A. The K
m
, V
max
(as
determined by the Direct Linear and Wilkinson meth-
ods), and k
cat
values are shown in Table 1. The K
m
value of wild-type MCAD for octanoyl-CoA, 3.69 lm,
compares well with the published value of 3.4 lm
[13,14]. Although the K
m
is increased somewhat for
the K364R mutant protein (5.53 lm), the k
cat
is not
significantly decreased. R256T was also purified, but,
even at a final concentration of 21.4 lgÆmL
)1
in the
assay, this mutant protein still showed no catalytic

proteins
To assess whether these mutations affected the affinity
of the enzyme for the bound prosthetic group, FAD,
absorption spectra were obtained. As FAD absorbs
strongly at 450 nm, the ratio A
450
⁄A
280
gives a rough
indication of the amount of FAD bound to the puri-
fied protein (assuming no major change in the extinc-
tion coefficient of the bound cofactor in the case of
the mutants) [15]. The peak absorbance ratios showed
first of all that the wild-type enzyme as purified here
(ratio ¼ 7.2) is somewhat depleted of FAD because
the value should be 5.2–5.5. This probably reflects the
rigorous procedure to remove excess FAD added dur-
ing the purification, and certainly does not indicate the
level of contamination by other protein, to judge from
SDS ⁄ PAGE. The ratios for R256T (8.8) and K364R
(9.6), both handled in exactly the same way as the
wild-type enzyme, suggest that both may be slightly
weakened in their affinity for FAD, with K364R the
most affected. As this is an unstable protein, the FAD
binding may be less secure than with a more tightly
folded protein. The catalytic competence of the enzyme
is in any case not greatly affected, suggesting that any
FAD binding impairment cannot be the primary dele-
terious effect of this mutation. As the reduction in
affinity is not as great for R256T, FAD depletion is

Æmin
)1
) (as deter-
mined by Direct Linear and Wilkinson methods), and k
cat
(s
)1
) for
octanoyl-CoA oxidation.
K
m
V
max
k
cat
Wild-type MCAD
Direct linear 3.69 24.4 · 10
3
19
Confidence limits (68%) 3.23–4.03 23.8–24.9 · 10
3
Wilkinson 3.68 24.6 · 10
3
19.1
Standard error 0.305 621
K364R
Direct linear 5.38 25.5 · 10
3
19.8
Confidence limits (68%) 5.09–5.87 25.4–25.6 · 10

enzyme at fewer than 10% of the positions [17,18], has
been solved to 2.4 A
˚
(Fig. 5) [17]. In this structure,
R256 is in very close proximity to the catalytic residue
E376. During catalysis the E376 sidechain swings
towards the Ca atom of the substrate, in order to
abstract the proton [17]. It seems likely that the fixed
charge of the guanidino group of R256 stabilizes the
catalytic carboxylate in the correct position for cata-
lysis. It is thus not surprising that removal of the con-
served positive charge in the mutant R256T prevents
catalysis. Indeed this residue has recently been studied
in rat MCAD, where the arginine was mutated to
alanine, lysine, glutamine and glutamic acid. The
authors found that the lysine mutant exhibited signifi-
cantly reduced activity, whereas the other variants
were completely inactive [19].
As expected from the extensive sequence homology,
the recently solved structure of a mammalian SCAD
shows a high degree of similarity to MCAD [20]. At
the amino acid position in SCAD corresponding to
K364 in MCAD there is an arginine residue, R356 in
the SCAD sequence. It is striking that this conserva-
tive substitution is identical to the clinical mutation in
the present case. Interestingly, though, this position in
SCAD has also previously been identified as the site of
a clinical mutation R356W [20]. This was found in
a female newborn, who presented with hypotonia,
seizures and developmental delay [21]. Another highly

genase (VLCAD), an analogous mutation (R410H) has
also been found to cause disease in three compound
heterozygote patients [24,25]. To the authors’ know-
ledge, this residue is the most frequently mutated posi-
tion across the entire ACAD family. Modeling of this
residue, using the rat SCAD crystal structure, sugges-
ted disruption of the local bonding and steric hin-
drance by the introduced amino acid.
Discussion
In this paper we have described the protein folding,
enzyme function, and thermostability of two novel rare
MCAD mutants, R256T and K364R, which are strik-
ingly different in their molecular behavior. The first
mutation, R256T is a nonconservative substitution of a
strongly basic internal residue by a smaller and less
polar one. R256T was unaffected in its folding, with lev-
els of tetramer formation in E. coli cells comparable to
wild type, even at 37 °C. Likewise there was only a mod-
erate decrease in the thermostability of this protein.
Nevertheless, R256T is clearly an acute mutation,
because, regardless of its stability, the catalytic activity
was completely abolished, as determined by the ferrice-
nium assay, ETF assay, and the INT ⁄ PES activity stain.
As R256T was successfully purified by the same method
as wild-type MCAD, i.e. using affinity elution with the
substrate, this mutation is unlikely to impede acyl-CoA
substrate binding, but rather must affect catalysis, either
in the acceptance of reducing equivalents from the
substrate, or the donation of these equivalents to ETF.
Although K364R is a relatively conservative substi-

activity. As R256T has so far been found as a com-
pound heterozygote with the K304E mutation, this
could either mask or enhance the clinical manifestation
of the disease [6,26]. Although the individuals with this
mutation have remained asymptomatic, the analogous
mutations in glutaryl-CoA dehydrogenase (R257W,
compound heterozygote with P278S, and R257Q), are
known to be disease-causing, indicating that R256T
could also be potentially disease-causing [27]. As
K364R was found as a homozygote, our biochemical
studies are more directly applicable.
Regardless of the enzyme activity as determined by
biochemical testing, the actual outcome can vary from
individual to individual depending on the functional
overlap of VLCAD, LCAD, MCAD and SCAD, the
efficiency of the chaperonin-aided folding, the effi-
ciency of the detoxification of accumulated intermedi-
ates, and avoidance of exposure to the environmental
triggers. Certainly in the case of MCAD deficiency,
environmental factors appear to outweigh the genetic
factors [26]. The possibility that different mutations
alone may cause varying severity of disease, resulting
in the wide clinical manifestation, has been considered
[28]. However, subsequent biochemical and molecular
folding studies of the various point mutations have
revealed no clear correlation between the genotype and
phenotype [6]. This becomes most apparent in the
study of the homozygous K304E, where the entire clin-
ical spectrum of MCAD deficiency has been observed
[5,6,26,29], suggesting that other background factors

MCAD1165AfiG (K364R) mutations were introduced by
site-directed PCR-based mutagenesis [12,30]. Mutagenic
antisense oligonucleotides for MCAD  842GfiC(5¢-GAT
AAAACCACACCTGTAGTAGCTG-3¢) and MCAD
1165AfiG(5¢-CCTGTAGAAAGACTAATGAGGGATG
CC-3¢) (mutagenic substitutions are shown in bold), and
the antisense primer (5¢-GTAACGCCAGGGTTTTCCCA
GTCAC-3¢) were used to generate a megaprimer. This was
used in a secondary PCR, with the sense primer (5¢-GATC
CAGATCCTAAAGCTCCTGCT-3¢), to generate the full-
length fragment, which was then subcloned into the pWT
vector, using the EcoRI and HindIII sites. The expression
vectors were sequenced across the region encoding the
MCAD gene, to exclude PCR-based errors. Each mutant
MCAD vector was cotransformed into JM109 with either
pGroESL (encoding the chaperones GroES and GroEL)
[31] or pCap (empty vector control), and cultured as des-
cribed elsewhere [12].
Polyacrylamide gel electrophoresis and western
blotting
SDS ⁄ PAGE, native PAGE, and western blotting were
performed essentially as described previously, using
BCIP ⁄ NBT tablets for colour development [9].
Protein purification
The mutant protein and wild type were purified by anion
exchange, and dye-affinity chromatography, as described
elsewhere [32].
Activity staining
This method was initially modified from an activity assay
for short chain acyl-CoA dehydrogenase [33] by substitu-

variation due to the mutation) to define the concentration
of active sites. Michaelis–Menten plots were used only to
display the results.
Electron transferring flavoprotein (ETF) assay
This assay utilizes 2,6-dichlorophenolindophenol as the
final electron acceptor, in 50 mm KP
i
, 0.3 mm EDTA, 5%
glycerol, pH 7.6 buffer, at 25 °C [37].
Thermal stability of enzyme activity in cleared
bacterial lysates
One hundred microlitre samples of bacterial lysates contain-
ing mutant or wild-type MCAD (10 lgÆmL
)1
in 100 mm
KP
i
,5mm EDTA, pH 7.6) were dispensed into separate
Eppendorf flasks. Each was incubated for 10 min in a water
bath at the chosen temperature before removing to ice, and
sampling for activity [34] and western blot analysis [12].
Acknowledgements
We warmly acknowledge the help and collaboration of
Dr Simon Olpin of the Sheffield Children’s Hospital
who detected the patient with the MCAD1166AfiG
mutation and supplied material to make identification
of this mutation possible. This work was supported by
a grant from the March of Dimes Foundation (grant
number 1-FY-2003–688 to BSA) and also by Grant
1C ⁄ 2002 ⁄ 073 under the International Collaboration

acyl-CoA dehydrogenase (MCAD) deficiency in com-
pound heterozygous patients: is there correlation
between genotype and phenotype? Hum Mol Genet 6,
695–707.
7 Iafolla AK, Thompson RJ Jr & Roe CR (1994) Med-
ium-chain acyl-coenzyme A dehydrogenase deficiency:
clinical course in 120 affected children. J Pediatr 124,
409–415.
8 Coates PM & Roe CR (1995) Acyl Co-A dehydrogen-
ase deficiencies. In The Metabolic and Molecular Basis
of Inherited Disease (Scriver CR, Beaudet A, Sly WS
& Valle D, eds), pp. 1501–1533. McGraw-Hill, New
York.
9 Bross P, Andresen BS, Winter V, Krautle F, Jensen
TG, Nandy A, Kolvraa S, Ghisla S, Bolund L & Gre-
gersen N (1993) Co-overexpression of bacterial GroESL
chaperonins partly overcomes non-productive folding
and tetramer assembly of E. coli-expressed human med-
ium-chain acyl-CoA dehydrogenase (MCAD) carrying
the prevalent disease-causing K304E mutation. Biochim
Biophys Acta 1182, 264–274.
10 Nasser I, Mohsen AW, Jelesarov I, Vockley J,
Macheroux P & Ghisla S (2004) Thermal unfolding
of medium-chain acyl-CoA dehydrogenase and
iso(3)valeryl-CoA dehydrogenase: study of the effect of
genetic defects on enzyme stability. Biochim Biophys
Acta 1690, 22–32.
11 Albers S, Levy HL, Irons M, Strauss AW & Marsden
D (2001) Compound heterozygosity in four asympto-
matic siblings with medium-chain acyl-CoA dehydro-

acidemia. Biochemistry 37, 10325–10335.
17 Kim JJ, Wang M & Paschke R (1993) Crystal structures
of medium-chain acyl-CoA dehydrogenase from pig
liver mitochondria with and without substrate. Proc
Natl Acad Sci USA 90, 7523–7527.
18 Kelly DP, Kim JJ, Billadello JJ, Hainline BE, Chu TW
& Strauss AW (1987) Nucleotide sequence of medium-
chain acyl-CoA dehydrogenase mRNA and its expres-
sion in enzyme-deficient human tissue. Proc Natl Acad
Sci USA 84, 4068–4072.
19 Zeng J & Li D (2004) Expression and purification of
His-tagged rat mitochondria medium-chain acyl-CoA
dehydrogenase wild-type and Arg256 mutant proteins.
Protein Expression Purification 37, 472–478.
20 Battaile KP, Molin-Case J, Paschke R, Wang M, Ben-
nett D, Vockley J & Kim JJ (2002) Crystal structure of
rat short chain acyl-CoA dehydrogenase complexed with
acetoacetyl-CoA: comparison with other acyl-CoA
dehydrogenases. J Biol Chem 277, 12200–12207.
21 Corydon MJ, Vockley J, Rinaldo P, Rhead WJ, Kjeld-
sen M, Winter V, Riggs C, Babovic-Vuksanovic D,
Smeitink J, De Jong J et al. (2001) Role of common
MCAD catalysis (R256T) and folding (K364R) mutants L. P. O’Reilly et al.
4556 FEBS Journal 272 (2005) 4549–4557 ª 2005 FEBS
gene variations in the molecular pathogenesis of short-
chain acyl-CoA dehydrogenase deficiency. Pediatr Res
49, 18–23.
22 Goodman SI, Stein DE, Schlesinger S, Christensen E,
Schwartz M, Greenberg CR & Elpeleg ON (1998) Glu-
taryl-CoA dehydrogenase mutations in glutaric acidemia

28 Brackett JC, Sims HF, Steiner RD, Nunge M, Zimmer-
man EM, deMartinville B, Rinaldo P, Slaugh R &
Strauss AW (1994) A novel mutation in medium chain
acyl-CoA dehydrogenase causes sudden neonatal death.
J Clin Invest 94, 1477–1483.
29 Andresen BS, Jensen TG, Bross P, Knudsen I, Winter V,
Kolvraa S, Bolund L, Ding JH, Chen YT, Van Hove
JL et al. (1994) Disease-causing mutations in exon 11 of
the medium-chain acyl-CoA dehydrogenase gene. Am J
Hum Genet 54, 975–988.
30 Sarkar G & Sommer SS (1990) The ‘megaprimer’
method of site-directed mutagenesis. Biotechniques 8 ,
404–407.
31 Goloubinoff P, Gatenby AA & Lorimer GH (1989)
GroE heat-shock proteins promote assembly of foreign
prokaryotic ribulose bisphosphate carboxylase oligomers
in Escherichia coli. Nature 337, 44–47.
32 O’Reilly L, Bross P, Corydon TJ, Olpin SE, Hansen J,
Kenney JM, McCandless SE, Frazier DM, Winter V,
Gregersen N et al. (2004) The Y42H mutation in med-
ium-chain acyl-CoA dehydrogenase, which is prevalent
in babies identified by MS ⁄ MS-based newborn screen-
ing, is temperature sensitive. Eur J Biochem 271, 4053–
4063.
33 Williamson G & Engel PC (1984) Butyryl-CoA dehy-
drogenase from Megasphaera elsdenii: specificity of the
catalytic reaction. Biochem J 218, 521–529.
34 Lehman TC, Hale DE, Bhala A & Thorpe C (1990) An
acyl-coenzyme A dehydrogenase assay utilizing the
ferricenium ion. Anal Biochem 186, 280–284.