Human Molecular Genetics, 2012, Vol. 21, No. 15
doi:10.1093/hmg/dds175
Advance Access published on May 18, 2012
3435–3448
Molecular mechanisms of riboflavin
responsiveness in patients with ETF-QO variations
and multiple acyl-CoA dehydrogenation deficiency
Nanna Cornelius1,∗ , Frank E. Frerman3, Thomas J. Corydon2, Johan Palmfeldt1, Peter Bross1,
Niels Gregersen1 and Rikke K. J. Olsen1
1
The Research Unit for Molecular Medicine, Aarhus University Hospital and Department of Clinical Medicine
and 2Department of Biomedicine, Aarhus University, Denmark and 3Department of Pediatrics, University of
Colorado – Denver and Colorado Intellectual and Developmental Disabilities Research Center, Aurora, CO, USA
Received March 5, 2012; Revised and Accepted May 2, 2012
Riboflavin-responsive forms of multiple acyl-CoA dehydrogenation deficiency (RR-MADD) have been known
for years, but with presumed defects in the formation of the flavin adenine dinucleotide (FAD) co-factor rather
than genetic defects of electron transfer flavoprotein (ETF) or electron transfer flavoprotein-ubiquinone oxidoreductase (ETF-QO). It was only recently established that a number of RR-MADD patients carry genetic
defects in ETF-QO and that the well-documented clinical efficacy of riboflavin treatment may be based on
a chaperone effect that can compensate for inherited folding defects of ETF-QO. In the present study, we investigate the molecular mechanisms and the genotype – phenotype relationships for the riboflavin responsiveness in MADD, using a human HEK-293 cell expression system. We studied the influence of riboflavin
and temperature on the steady-state level and the activity of variant ETF-QO proteins identified in patients
with RR-MADD, or non- and partially responsive MADD. Our results showed that variant ETF-QO proteins
associated with non- and partially responsive MADD caused severe misfolding of ETF-QO variant proteins
when cultured in media with supplemented concentrations of riboflavin. In contrast, variant ETF-QO proteins
associated with RR-MADD caused milder folding defects when cultured at the same conditions. Decreased
thermal stability of the variants showed that FAD does not completely correct the structural defects induced
by the variation. This may cause leakage of electrons and increased reactive oxygen species, as reflected by
increased amounts of cellular peroxide production in HEK-293 cells expressing the variant ETF-QO proteins.
Finally, we found indications of prolonged association of variant ETF-QO protein with the Hsp60 chaperonin
in the mitochondrial matrix, supporting indications of folding defects in the variant ETF-QO proteins.
INTRODUCTION
Multiple acyl-CoA dehydrogenation deficiency (MADD), also
known as glutaric aciduria type II (GAII) (OMIM #231680), is
an autosomal recessively inherited disorder. In most cases,
MADD is caused by variations in one of two genes encoding
electron transfer flavoprotein (ETF) or electron transfer
flavoprotein-ubiquinone oxidoreductase (ETF-QO). The
ETF/ETF-QO proteins make up the electron transport chain
in the mitochondria that links dehydrogenation reactions of
12 flavoprotein dehydrogenases to oxidative phosphorylation
and thereby production of cellular energy in the inner mitochondrial membrane (1,2). The clinical features of MADD
are highly heterogeneous and have been classified into lethal
neonatal onset forms, often with severe congenital anomalies
(severe MADD or S-MADD) and milder forms (milder
MADD or M-MADD), typically with later onset and a large
spectrum of symptoms ranging from hypoglycemic encephalopathy, sometimes combined with cardiomyopathy and childhood death, to patients presenting with an isolated and
potentially progressive muscle myopathy as adults (2,3). A
genotype – phenotype relationship has been demonstrated in
∗
To whom correspondence should be addressed at: The Research Unit for Molecular Medicine, Aarhus University Hospital and Department of Clinical
Medicine, Brendstrupgaardsvej 100, Aarhus DK-8200, Denmark. Tel: +45 78455410; Fax: +45 86278402; Email: [email protected]
# The Author 2012. Published by Oxford University Press. All rights reserved.
For Permissions, please email: [email protected]
3436
Human Molecular Genetics, 2012, Vol. 21, No. 15
MADD, with S-MADD patients being homozygous or compound heterozygous for functional null variations of ETF/
ETF-QO, and M-MADD patients have at least one missense
variation, which may potentially give rise to some enzyme activity. Thus, M-MADD shows moderately decreased fatty acid
oxidation flux (14 – 52% of controls) in cultured M-MADD
fibroblasts (3 – 6), whereas cultured S-MADD fibroblasts
have fatty acid oxidation flux of 3 – 12% of controls (3,5).
There is no known treatment for S-MADD patients. In contrast, early dietary treatment (protein and fatty acid restriction)
combined with various diet supplements like riboflavin, carnitine and/or co-enzyme Q10 has been tried with varying
success in patients with M-MADD (6 – 13). A sub-group of
M-MADD patients show a significant response to treatment
with riboflavin, resulting in near-normalized biochemical and
clinical parameters after high doses (100 – 400 mg/day) of
oral riboflavin supplementation. Such patients are referred to
as riboflavin-responsive MADD (RR-MADD) patients
(14,15). Fibroblasts from RR-MADD patients show fatty
acid oxidation flux of 32– 129% of controls when cultured
in standard riboflavin-supplemented media, and disease prognosis is good when the patients are diagnosed and treated early
(8,13). Therefore, it seems that S-MADD can be categorized
as riboflavin-non-responsive, whereas M-MADD is either partially responsive or responsive, the latter with the label
RR-MADD. To substantiate this notion, there is a clinical
and scientific interest in elucidating the molecular mechanisms
of the response to riboflavin observed in MADD patients.
For several years, it was believed that the phenotype of
RR-MADD patients was due to a defect in the formation of
the co-factor flavin adenine dinucleotide (FAD) from riboflavin (vitamin B2) (9,14,16,17). Gene defects of riboflavin metabolism or transport were only recently identified in a few
RR-MADD patients (18– 21). Rather, most RR-MADD
patients reported to date have variations in the ETFDH gene
encoding ETF-QO. This was revealed in 2007, when we and
others reported ETFDH variations in 16 unrelated
RR-MADD patients (8,13). Since then an increasing number
of reports of ETFDH variations in RR-MADD have been published (7,22– 28).
ETF-QO, which is a 64 kDa monomeric protein, is synthesized as a 67 kDa precursor protein. It is processed to the
mature form in the course of translocation into the mitochondria (29). The mature form of ETF-QO contains one FAD
cluster and one iron– sulfur ([4Fe4S]+2 + 1) cluster (2). Electrons derived from the various dehydrogenation reactions are
transferred from ETF to ETF-QO at the iron– sulfur cluster
and subsequently to the FAD co-factor. FAD is the electron
donor to ubiquinone (UQ) (Q10) in the inner mitochondrial
membrane (30 –32). The ETF-QO protein has a three-domain
structure, with the FAD-binding domain and the iron – sulfur
cluster domain facing the mitochondrial matrix and the
UQ-binding domain anchoring the protein to the inner mitochondrial membrane (31).
The mutational spectrum of RR-MADD is large with more
than 50 different ETFDH variations reported to date
(7,8,13,22– 28). All published RR-MADD patients have at
least one missense variation. Some of these missense variations (p.Ser82Pro, p.Ala84Thr, p.Tyr257Cys, p.Leu377Pro,
p.Leu409Phe, p.Val451Leu, p.Pro456Leu and p.Pro483Leu)
have repeatedly been identified in RR-MADD patients
(8,13,22,24), suggesting that the nature of the variations is responsible for riboflavin responsiveness. It is well known that
missense variations often lead to misfolding of the variant
protein with consequently increased proteolytic degradation
and decreased stability (33). FAD binding has been demonstrated to promote chaperone-mediated folding, oligomerization and/or conformational stabilization of a number of
flavoenzymes (34– 38), and FAD deficiency in rat liver mitochondria results in the instability and rapid degradation of ETF
and a number of acyl-CoA dehydrogenases (39). In this
context, it could be speculated that increasing the availability
of riboflavin, which is converted to the FAD, promotes folding
and/or stabilizes the native conformation of certain types of
ETF-QO variant proteins. Such a mechanism of action was recently indicated for the p.Asp128Asn ETF-b variant protein,
where the presence of FAD improved conformational and proteolytic stability as well as enzyme activity (40).
In the present study, we aim to test this hypothesis by using
two different expression systems to study the in vitro influence
of riboflavin and temperature on the folding/stability and catalytic activities of selected variant ETF-QO proteins. We relate
our experimental results to structural data and knowledge
about the influence of the variant proteins on clinical severity
and riboflavin responsiveness in vivo to propose a model for
FAD-assisted folding of ETF-QO in mitochondria.
RESULTS
The effect of riboflavin on the steady-state level
and the activity of variant ETF-QO proteins
Missense variations associated with RR-MADD have been
assigned to the FAD-binding domain or the UQ-binding
domain of ETF-QO, although none of the affected amino
acids makes direct contact with the FAD co-factor (41).
ETF-QO variant proteins with amino acid changes located in
the FAD-binding domain (c.1285G.C/p.Gly429Arg) or in
the UQ-binding domain (c.1367C.T/p.Pro456Leu, c.1448C
.T/p.Pro483Leu) (13) were selected for the current study
and the underlying molecular mechanisms of riboflavin
responsiveness were studied in relation to one ETF-QO
variant protein (c.413T.G/p.Leu138Arg) identified in patients
with M-MADD (6) and two variant ETF-QO proteins
(c.1001T.C/p.Leu334Pro and c.1414G.A/p.Gly472Arg)
identified in S-MADD patients (3,42).
To test, in a cellular system, whether the steady-state level
and the activity of the selected variant ETF-QO proteins
respond to variations in cellular riboflavin concentrations
and correlate to phenotype, we expressed human wild-type
ETF-QO protein and the six different variant ETF-QO proteins
in HEK-293 cells cultured in media containing 16, 28.5 or
530 nmol/l of riboflavin (see Materials and Methods).
ETF-QO activity was determined on extracted membrane
fractions, using an assay which measures the integrity of
both redox prosthetic groups ([4Fe4S]+2 + 1 and FAD), to be
able to measure the whole electron pathway (Fig. 1A). The
ETF-QO protein amount was detected by western blot analysis
of whole cell lysates (Fig. 1B). The identity of the mature
Human Molecular Genetics, 2012, Vol. 21, No. 15
3437
Figure 1. Expression of wild-type and variant ETF-QO proteins in transient transfected HEK-293 cells cultured at 378C at supplemented (530 nmol/l), moderate
deficient (28.5 nmol/l) or severe deficient (16 nmol/l) concentrations of riboflavin. (A) The activity of wild-type and variant ETF-QO proteins was measured as
described in Materials and Methods. The values are expressed relative to wild-type cultivated in the presence of 530 nmol/l riboflavin. (B) Steady-state levels of
wild-type and variant ETF-QO proteins were analyzed by western blot on SDS– PAGE gels immunostained with ETF-QO and VDAC antibodies. The amount of
ETF-QO protein was quantified relative to VDAC, using the Vision WorksLS Image Acquistion software. All values are expressed relative to wild-type cultured
in the presence of 530 nmol/l riboflavin. Bars in (A) and (B) represent mean values of three independent experiments +SEM. A filled triangle represents all
samples, P , 0.05; two filled triangles represent samples between 530 and 16 nmol/l, P , 0.05; three filled triangles represent samples between 530 and
28.5 nmol/l and between 530 and 16 nmol/l, P , 0.05.
64 kDa gel band as human ETF-QO was confirmed by mass
spectrometry analysis (see Materials and Methods).
All ETFDH variations tested compromise the ETF-QO
enzyme amount and activity to various extent and show different dependency on riboflavin availability. At supplemented
riboflavin concentrations (530 nmol/l), at which wild-type
ETF-QO is expected to be fully flavinylated with 100% activity, the activities of variant ETF-QO proteins associated with
an RR-MADD phenotype (p.Pro456Leu, p.Pro483Leu and
p.Gly429Arg) were 70 – 80% of those of wild-type. When
riboflavin was moderately (28.5 nmol/l) or severely
(16 nmol/l) depleted, activities of these variant ETF-QO
proteins were decreased to 40 and 30%, respectively, of
those of fully flavinylated and active wild-type protein
(Fig. 1A). In contrast to this, activities of variant ETF-QO proteins, identified in patients with S-MADD (p.Gly472Arg and
p.Leu334Pro), did not respond significantly to changes in riboflavin concentrations, and they reached a maximum enzyme
activity of only 5 – 20% of that of wild-type when cultured
at supplemented riboflavin concentrations (Fig. 1A). Activities
of the variant p.Leu138Arg protein (M-MADD) showed a
slight but significant (P , 0.05) response to changes in riboflavin concentrations, but surprisingly enzyme activity could be
rescued to only 15% of that of wild-type levels at
3438
Human Molecular Genetics, 2012, Vol. 21, No. 15
Figure 2. Steady-state levels of wild-type and variant ETF-QO proteins expressed in HEK-293 cells cultivated at 408C at supplemented (530 nmol/l), moderate
deficient (28.5 nmol/l) or severe deficient (16 nmol/l) concentrations of riboflavin. Steady-state levels of the ETF-QO protein were analyzed relative to the expression of VDAC by western blot analysis as explained in Figure 1. Bars represent the means of three independent experiments +SEM and all values are
measured relative to wild-type, cultivated in the presence of 530 nmol/l riboflavin.
supplemented riboflavin concentrations. Wild-type ETF-QO
activity was also affected by riboflavin depletion, but to a
lesser extent than the variant ETF-QO proteins.
Four ETF-QO variant proteins (p.Pro483Leu, p.Gly429Arg,
p.Leu138Arg and p.Leu334Pro) showed decreased
steady-state protein levels with proportional reduced enzyme
activities when compared with wild-type, indicating that the
variants mainly cause decreased proteolytic stability with
only minor or no alteration of specific enzyme activity
(Fig. 1A and B). In contrast, the p.Pro456Leu and
p.Gly472Arg variant proteins showed disproportionately less
residual enzyme activity (30 and 70%, respectively) than the
protein amount (60 and 90%, respectively) when compared
with that of wild-type protein at supplemented riboflavin concentrations. This indicates that high riboflavin concentration
increases the proteolytic stability of p.Pro456Leu and
p.Gly472Arg variant proteins. Despite this increase, none of
the two variant proteins reach full activity.
Riboflavin dependency during heat stress
Since proteins unfold at elevated temperatures, the proteolytic
stability and enzymatic function of variant proteins with
impaired folding or stability properties are often more compromised at elevated temperatures in vitro with consequent
increased risk of metabolic decompensation during fever in
vivo (33). Patients with milder forms of MADD (M-MADD
and RR-MADD) often get symptoms in connection with feverish infection (3,13). Therefore, we investigated the relative
steady-state level of wild-type and variant ETF-QO proteins
in HEK-293 cells at an increased temperature (408C), and at
the three different levels of riboflavin. As shown in
Figure 2, the relative amount of all variant ETF-QO proteins
were clearly affected by heat stress. For most ETF-QO
variant proteins, heat stress at 408C seems to cause too
severe folding/stability defects to allow any significant modulation of the protein amount by changes in the riboflavin
levels. Only p.Pro456Leu and p.Gly429Arg showed significant
responses to riboflavin during heat stress. However, they could
be rescued to only 50 and 35% of that of wild-type levels,
respectively. In contrast, at 378C, the same two variant proteins could be rescued to 90 and 70% of wild-type levels, respectively, when cultured at supplemented riboflavin
concentrations (Figs 1C and 2). Since the increased temperature (408C) was so detrimental to the protein, the activity
was not relevant to measure under these conditions.
Overall, our findings that the six ETFDH variations result in
decreased steady-state level of the ETF-QO protein amount to
various extents in response to riboflavin depletion and elevated
temperature show that the variations mainly affect protein biogenesis and are consistent with the hypothesis that FAD
co-factors can act as chemical chaperones, resulting in
increased stability and/or reduced turnover of ETF-QO proteins. The intrinsic capabilities of the variant proteins to fold
and reach a stable and active conformation determine the
extent to which increased riboflavin (FAD) levels and
decreased temperature can increase the protein amount and
rescue the enzyme function. Variant p.Pro456Leu,
p.Pro483Leu and p.Gly429Arg ETF-QO proteins show
clearly a milder impact on overall protein folding/stability
than the p.Leu138Arg, p.Leu334Pro and p.Gly472Arg
ETF-QO proteins.
Thermal stability of mature ETF-QO proteins
For nuclear encoded mitochondrial membrane proteins, like
ETF-QO proteins, loss of steady-state protein amount and activity can occur during ETFDH cDNA expression/translation,
mitochondrial import and maturation to an active membranebound enzyme, or as a consequence of a destabilized native
enzyme once it is formed. To gain further insights into the
mechanisms by which ETFDH variations cause protein dysfunction and RR-MADD, the catalytic stabilities of extracted
membrane fractions of p.Pro456Leu, p.Pro483Leu and
p.Gly429Arg variant ETF-QO proteins, cultured at 378C and
at supplemented riboflavin concentrations (530 nmol/l), were
determined and compared with that of wild-type ETF-QO by
monitoring the loss of activity at 428C as a function of time
Human Molecular Genetics, 2012, Vol. 21, No. 15
3439
Figure 3. The thermal catalytic stability of wild-type and variant ETF-QO proteins. The thermal catalytic stability of wild-type and variant ETF-QO proteins was
monitored by extracting the cellular membranes from cells cultured at 378C at supplemented riboflavin concentrations. The solubilized proteins were incubated at
428C. Aliquots were withdrawn at 0, 10, 20, 30, 60 and 120 min, and the ETF-QO activity was measured.
(Fig. 3). Initial experiments were carried out at 408C.
However, the activities of variant proteins were unchanged
at this temperature (results not shown). As no new ETF-QO
protein is synthesized during the experiment, the experiment
allows us to evaluate whether the variations affect overall conformational stability and activity once the native monomer has
been formed in the membrane. As seen in Figure 3, wild-type
ETF-QO, which is expected to be fully flavinylated at supplemented riboflavin concentrations, was very stable for at least
2 h at 428C. In contrast, the activities of the variant proteins
were severely affected at these conditions. As soon as after
20 min of incubation, the p.Pro456Leu variant protein was
reduced to 50% of its initial activity. The p.Pro483Leu
and p.Gly429Arg variant proteins had lost 50% of their
initial activities after about 60 min. The protein concentrations
in the extracts were 1 mg/ml, so the differences observed in
heat stability do not reflect altered stability due to differences
in protein concentrations. These data show that for the proportion of the variant proteins, which acquires the native
membrane-bound conformation, which is more temperature
sensitive, the catalytic stability is compromised when compared with that of wild-type. Thus, although FAD allows the
variant proteins to fold and reach their mature membrane
forms, FAD cannot correct native structure. Western blot
analysis revealed that wild-type and variant proteins were
stable during incubation at 428C for 2 h (data not shown).
This does not exclude that increased degradation of variant
proteins does take place in vivo as most proteases most
likely are inactive in the extracted membrane fractions.
H2O2 production in HEK-293 cells expressing variant
ETF-QO proteins
Conformational changes in the final structures may allow an
increased flow of electrons from the variant ETF-QO enzymes
to oxygen and thereby give rise to reactive oxygen species
(ROS), such as superoxide. Superoxide produced from mitochondrial oxygen is rapidly converted to H2O2 by mitochondrial
superoxide dismutase (mnSOD or SOD2). The peroxide crosses
the mitochondrial membranes and moves to the cytosol (43). To
investigate whether misfolded ETF-QO proteins may leak electrons and thus give rise to increased cellular peroxide production, HEK-293 cells expressing the p.Pro456Leu variant were
cultured at 378C in supplemented (530 nmol/l) or in severely
depleted (16 nmol/l) riboflavin media and treated with
the peroxide-sensitive fluorescent probe CM-H2DCFDA,
followed by quantification of fluorescence signals by flow
cytometry. For comparison, H2O2 production was analyzed
also in HEK-293 cells expressing a severe stability variant
(p.Leu138Arg) and in HEK-293 cells expressing a relatively
stable variant, but with severely decreased activity
(p.Gly472Arg). When cultured in supplemented riboflavin
medium, HEK-293 cells expressing p.Pro456Leu protein had
increased peroxide production when compared with HEK-293
cells expressing wild-type ETF-QO. Peroxide production in
HEK-293 cells expressing p.Leu138Arg or p.Gly472Arg was
less pronounced than in p.Pro456Leu, relative to HEK-293
cells expressing wild-type ETF-QO (Fig. 4). At these
conditions, the steady-state amount and the activity of the
variant proteins relative to those of wild-type protein are
expected to be 90 and 70%, respectively, for p.Pro456Leu, 15
and 10%, respectively, for p.Leu138Arg and 60 and 30%, respectively, for p.Gly472Arg (Fig. 1A and C). The capacity of
the same cells to produce H2O2 did not change significantly
when the cells were cultured in depleted riboflavin medium
(Fig. 4), even though the steady-state amount of active variant
proteins are expected to be significantly affected during riboflavin deficiency, where the amount and the activity of the variant
proteins relative to those of wild-type protein are 30 and 30%,
respectively, for p.Pro456Leu, 10 and 0%, respectively, for
p.Leu138Arg and 20 and 20%, respectively, for p.Gly472Arg
(Fig. 1A and C). This most likely reflects that the capacity of
the HEK-293 cells to produce H2O2 is saturated with the low
amount of expressed variant ETF-QO proteins produced
during riboflavin deficiency. H2O2 production in HEK-293
cells transfected with wild-type ETFDH was comparable with
3440
Human Molecular Genetics, 2012, Vol. 21, No. 15
Figure 4. H2O2 production in HEK-293 cells expressing wild-type or variant ETF-QO proteins. Transfected HEK-293 cells expressing wild-type or variant
ETF-QO proteins were cultured at steady state in riboflavin-supplemented (530 nmol/l) (+) or depleted (16 nmol/l) (2) media, incubated with CM-H2DCFDA
and analyzed by FACS as described in Materials and Methods. The results are presented as percent increased CM-H2DCFDA fluorescence relative to wild-type.
Results are representative of three different experiments.
that of HEK-293 cells transfected with vector alone, reflecting
that the production of variant ETF-QO proteins and not transfection is responsible for the increased ROS production (data not
shown). The results clearly shows an increase in the ROS
level when HEK-293 cells were transfected with ETF-QO
folding variant proteins compared with ETF-QO activity
variant proteins and wild-type protein.
Mitochondrial import and maturation of the p.Pro456Leu
variant protein
Expression analysis of the ETF-QO variant protein in
HEK-293 cells demonstrated that RR-MADD variant
ETF-QO proteins have decreased thermal stability when
located in the inner mitochondrial membrane. To further elucidate the molecular mechanisms by which RR-MADD
ETFDH gene variations cause decreased steady-state level of
the ETF-QO enzyme, it was of interest to determine whether
the intra-mitochondrial folding of the variant proteins is also
affected. An in vitro mitochondrial system allowed the characterization of the import pathway as well as the investigation of
possible chaperone interactions during import. The import
experiments were performed in mitochondria from rats fed a
normal diet.
As a proof of principle, we investigated the import of the in
vitro-translated ETF-QO variant protein, p.Pro456Leu, into
isolated rat mitochondria for up to 180 min at 26 and 378C.
The import pathway of ETF-QO has not yet been characterized. Therefore, we included an analysis of the mitochondrial
matrix to investigate whether the ETF-QO protein is imported
through the matrix rather than being directly inserted into the
inner mitochondrial membrane. Aliquots withdrawn at different time intervals were divided into three fractions (for details,
see Materials and Methods): a membrane fraction, a matrix
fraction and an insoluble fraction. All fractions were analyzed
by SDS – PAGE. As seen from Figure 5A, wild-type ETF-QO
protein and p.Pro456Leu variant protein behaved similarly
during import at 268C, with a continuous import into the membrane and a continual increase of the insoluble fraction over
time. A difference is observed in the matrix fraction, where
an increase is present in the p.Pro456Leu variant protein compared with the wild-type (after 60 min import, the matrix fraction of wild-type had increased less than 2-fold, compared
with a 3-fold increase in the matrix fraction of
p.Pro456Leu), which suggests that the variant protein is
more difficult to fold, and thus is associated for a longer
period of time with chaperones before insertion into the
membrane.
When the temperature was increased to 378C (Fig. 5B), the
apparent rate of import was faster for both wild-type and
p.Pro456Leu, where maximum import after 30 and 60 min, respectively, is detected, and thereafter there is a turnover of the
membrane fraction. A clear difference is seen in the amount of
wild-type and p.Pro456Leu proteins in the membrane fractions, with almost 25% more wild-type protein imported into
the membrane after 15 min. Moreover, the p.Pro456Leu
variant protein reaches the turnover point after 60 min compared with 30 min for the wild-type protein. The insoluble
fraction for the wild-type and the p.Pro456Leu is similar and
has a more than 4-fold increase over the import period. As
also seen at 268C, a larger fraction of protein in the matrix
fraction for the variant protein compared with the wild-type
protein was identified.
In order to investigate whether the ETF-QO protein found in
the matrix fraction is associated with other proteins, such as
chaperones, the matrix fraction of protein was loaded onto
native gels (Fig. 5C). Variant matrix proteins, having difficulties with folding, are often found in association with molecular
chaperones like Hsp60 (44,45). As seen in Figure 5C, the gel
showed two bands of larger size (120 and 480 kDa) than
native ETF-QO (64 kDa). The intensity of the 120 kDa band
is constant over time, whereas the band of 480 kDa for the
wild-type increased in intensity up to 30 min and then
decreased. For the variant p.Pro456Leu protein, there is a
Human Molecular Genetics, 2012, Vol. 21, No. 15
3441
Figure 5. Import and maturation of the p.Pro456Leu ETF-QO precursor into isolated rat liver mitochondria. (A) Experiments were performed at 268C (A) and
378C (B). The wild-type protein is imported in parallel with the variant protein. Samples are collected at four different times: 15, 30, 60 and 180 min. As
described in detail in the Materials and Methods section, the samples are separated into three fractions: a membrane fraction, a matrix fraction and an insoluble
fraction. The samples were analyzed by SDS–PAGE, followed by radiography. Each sample (wild-type or variant) was normalized to its respective wild-type
sample taken at 30 min. Data are representative of three independent sets of experiments. (C) Native gel electrophoresis of the matrix fraction. The matrix fraction loaded on the native gel is the fraction dissolved in lysis buffer without a detergent. Samples were taken over a time period from 0 to 240 min. The gel shows
a strong band with a molecular mass of 480 kDa and a weaker band with a molecular mass of 120 kDa. No bands corresponding to 64 kDa could be detected.
constant and more pronounced increase in amounts of this
480 kDa complex over time until 180 min, followed by a
small decrease at 240 min. The 120 kDa band might be a
dimer of ETF-QO as the electrophoresis was performed in
the absence of a detergent. A band of similar size was seen
when human ETF-QO proteins were overexpressed in Sf9
3442
Human Molecular Genetics, 2012, Vol. 21, No. 15
Figure 6. Location of the investigated ETF-QO variations and structures surrounding the variations. The ribbon representation of ETF-QO is based on the crystal
structure of porcine ETF-QO (31). The FAD-binding domain is in blue; the 4Fe4S cluster domain is in red; the UQ-binding domain is in dark green; and the
membrane spanning part of the UQ-binding domain is in light green. The three redox centers are shown as stick models: FAD in orange, 4Fe4S in pink and UQ in
purple. The mutated amino acids investigated in the present study are shown in yellow space filling. Magnifications of the ETF-QO structural regions in which
the variations are located are shown in gray overlaying the variation (in red) and the original residues (in green). Structures were prepared using PyMOL.
insect cells and analyzed in their native structure, using
perfluoro-octanoicacid PAGE (46). The 480 kDa band is
very interesting since it is dynamic over time and the size of
the band correlates with ETF-QO in complex with an Hsp60
chaperone complex. Heptameric, functional Hsp60 migrates
in native PAGE as a 420 kDa complex corresponding to a heptamer, as previously documented for another misfolded protein
(44). If ETF-QO forms a complex with the Hsp60 heptamer,
this would give an expected band size of 484 kDa, which
correlated with the band we identify at 480 kDa on the
native gel. All together, the mitochondrial import experiments
demonstrate that the variant p.Pro456Leu is a mild folding
variant with increased chaperone-mediated folding and
holding in the matrix. Finally, the experiments show, for the
first time, that the ETF-QO protein is imported through the
matrix and then inserted into the inner membrane corresponding to a heptamer (44).
DISCUSSION
In this paper, we demonstrate that three variant ETF-QO proteins (p.Pro456Leu, p.Pro483Leu and p.Gly429Arg) identified
in RR-MADD patients respond to increasing concentrations of
riboflavin in the cultivate medium when expressed in
HEK-293 cells. In contrast, ETF-QO variant proteins
(p.Leu334Pro, p.Gly472Arg and p.Leu138Arg) identified in
patients who do not respond, or respond poorly, to riboflavin
therapy in vivo do not respond in the cellular system. Thus,
a clear genotype – phenotype relationship for the observed
riboflavin responsiveness in patients with genetic defects of
ETF-QO is emerging. Our data are also consistent with the hypothesis that the riboflavin response results from the ability of
FAD to promote the folding of the variant proteins or otherwise stabilize their mature proteins or folding intermediates.
This was inferred from (i) the negative effect of temperature
on the steady-state level of ETF-QO in cells, which must
process and translocate the polypeptides, incorporate two
redox centers and insert the proteins into the inner mitochondrial membrane and (ii) the thermal stability of the partially
purified riboflavin-responsive proteins, which suggests that
these variant proteins can be stabilized by riboflavin in spite
of structural differences from the wild-type protein.
The effect of riboflavin on expression of ETF-QO proteins
with variations in the FAD domain, like the p.Gly429Arg,
can be reasonably explained by FAD binding, which affects
the kinetics and/or thermodynamics of the folding of that
FAD domain, limiting the number of favorable conformations
leading to a near-native protein structure, as shown by Er et al.
(47). Gly429 is located in a loop between the two helices a6
and a7, but far from the active FAD-binding site (Fig. 6). In
the early folding pathway, glycine residues play a critical
role in loop structure by maintaining flexibility (48). Substitution to arginine with a large cationic side chain into this 14
Human Molecular Genetics, 2012, Vol. 21, No. 15
amino acid loop that emerges from the surface of ETF-QO is
an obstacle to maintaining flexibility and likely presents
barrier, which may be partially overcome by binding FAD
to an intermediate in the folding pathway, decreasing the ensemble of pathways open to the final structure. The other
two RR-MADD amino acid variations investigated here, and
several identified by others (8,13,22), are not located in the
flavin domain, but are located in the UQ-binding domain
(Fig. 6). The FAD and UQ structural domains are not isolated
in the primary structure of ETF-QO, although the functional
domains are closely packed and share structural elements.
The interspersed, non-contiguous elements of the FAD and
UQ domains in the primary structure (see Fig. 6 and reference
31) suggest that a variation in a UQ functional domain may
well be influenced by FAD interaction with the folding of
the protein in the flavin domain. The p.Pro456Leu amino
acid variation is located in the UQ domain in a loop
between helix a8 in the FAD domain and a9 in the membrane
domain of the tertiary structure (31). Loop and turn formation
is critical early in the folding pathway because these elements
permit sampling of more energetically favorable structures
that lead to chain compaction (48). Proline imparts rigidity
to a polypeptide and therefore, the entropy of the unfolded
protein would increase in the p.Pro456Leu and p.Pro483Leu
variants by substitution of a leucine residue. p.Pro483Leu,
which is located four residues to the C terminus of the a9
helix and at the C terminus of the membrane-associated
region of the protein in the middle of a rather structure-less
region (Fig. 6) so that p.Pro483 would also limit the number
of conformations of this region as the protein folds (31).
FAD binding to folding intermediates could counteract the
resulting increase in entropy and increase the ensemble of
unfolded structures. Previous work by Wittung-Stafshade
(38) on the in vitro unfolding and refolding of desulfovibrio
desulfuricans flavodoxin, a small flavoprotein, showed that
the flavin co-factor, FMN, can bind to the unfolded protein
and accelerate its folding . It is conceivable that an increase
in the concentration of intra-mitochondrial FAD in a similar
way may affect both the kinetics and thermodynamics of the
folding of the variant p.Pro456Leu and p.Pro483Leu
ETF-QO proteins. Another indication of early defective
folding associated with these ETF-QO variants was the
result from native gel analyses of in vitro-translated and
mitochondria-imported ETF-QO proteins (Fig. 5C), where
we found indications of prolonged association of the
p.Pro456Leu variant protein with the Hsp60 chaperonin
system in the mitochondrial matrix compared with wild-type
protein. Although FAD can assist in overcoming misfolding
and allow the RR-MADD variant ETF-QO proteins to be
inserted into the inner mitochondrial membrane, it does not
correct the mutation-induced structural defects as reflected
by the decreased thermal catalytic stability of the partially
purified variant proteins (Fig. 3) along with the increased tendency of the p.Pro456Leu variant to produce ROS when
expressed at supplemented riboflavin concentrations (Fig. 4).
The p.Leu138Arg variant has profoundly reduced the
protein amount and activity in the cell system, with a slight
but significant response to increasing concentrations of riboflavin in the culture media (Fig. 1A and B). A patient, compound
heterozygous for the p.Leu138Arg variation and a splice
3443
variation (IVS3 +3A.T) causing degradation of the corresponding ETFDH transcripts, showed a partial clinical response to treatment with riboflavin and a remarkably mild
fatty acid oxidation flux in fibroblasts (24 – 52% of those of
controls) when cultured in riboflavin-supplemented media.
The patient died at 14 years of age despite intensive treatment.
Other patients reported with this variation have died, probably
resulting from not as intensive care as the patient reported by
Olsen et al. (6). Leu138 is buried in the structure of the protein
between the a2 helix in the flavin domain and the b3 sheet of
the UQ domain (Fig. 6). Substitution of the cationic arginine
residue in the p.Leu138Arg variation would be energetically
unfavorable and likely too structurally distorted to be overcome by riboflavin.
The p.Leu334Pro and p.Gly472Arg variants have been
reported in the homozygous state in S-MADD patients with
neonatal death and very impaired the ETF-QO enzyme
function in fibroblasts cultured in riboflavin-supplemented
media (3,42,49,50). Expression studies showed that the
p.Leu334Pro and the p.Gly472Arg variant ETF-QO proteins
profoundly reduced the protein amount and activity, with no
significant response to increasing concentrations of riboflavin
in the culture media (Fig. 1A and C). The p.Leu334Pro
variation is located at the N + 1 position relative to the a5
helix which is located in the UQ domain (Fig. 6). Substitution
of a proline residue could twist the helix, altering the orientation of the amino acid side chains in the helix. Gly472 is
located in the helical membrane domain (a9) that forms part
of the UQ entry channel and it is not unreasonable that the
p.Gly472Arg variant would be unstable with less activity
because the cationic arginine side chain would extend
toward the lipid phase. These mutations likely cause too
severe structural distortion to be stimulated by riboflavin.
In summary, all mutations investigated in this study showed
significant abnormalities in molecular phenotype ranging from
mild to severe and correlated with clinical phenotype.
Decreased proteolytic stability seems to be their major molecular mechanism of action. All variant proteins studied
were affected by increased temperature, but only milder variants responded to changes in riboflavin concentrations and
only those associated with RR-MADD could be rescued to significant levels when expressed at supplemented concentrations
of riboflavin. The results from experimental analyses and presumed structural effects of the variants are consistent with the
hypothesis that the riboflavin response results from the ability
of FAD to act as a chemical chaperone that can promote
folding of secondary elements already early in the folding
pathway of ETF-QO or at later steps by stabilizing folding
intermediates or mature membrane-inserted proteins. The
ability of riboflavin and temperature to modulate the molecular and clinical phenotypes depends on the inherited folding
properties of the variant ETF-QO proteins as illustrated in
our proposed model for the mitochondrial folding pathway
of ETF-QO variants (Fig. 7). According to this model,
increased cellular temperature (such as fever) or genetic
factors or physiologic stressors like pregnancy, restricted
diet and viral infections that are known to decrease cellular
FAD content (51– 53) may push milder variant ETF-QO proteins, like the p.Gly429Arg, p.Pro456Leu and p.Pro483Leu
proteins or even wild-type ETF-QO, toward a non-productive
3444
Human Molecular Genetics, 2012, Vol. 21, No. 15
Figure 7. Proposed model for the effect of riboflavin/FAD and temperature on mitochondrial folding of wild-type and variant ETF-QO proteins. Newly imported
ETF-QO polypeptides interact with mitochondrial heat shock protein 70 (mhsp70). ETF-QO polypeptides carrying severe variations will be recognized by mitochondrial proteases and degraded. Some mild variations will also be recognized for degradation in this early step and here we hypothesize that FAD might have
an early effect in stabilizing the polypeptide so that it will not be recognized for degradation, but go on for further folding in interaction with the hsp60/hsp10
chaperonin folding machinery, as what would be expected for the wild-type. The polypeptide may have several rounds of folding in interaction with the chaperonin system, especially if the polypeptide carries a mild variation that makes it difficult to fold into the right conformation. Here, there may be several factors
that can influence the outcome; if FAD is high (in patients treated with riboflavin), it might help stabilize the polypeptide and thereby boost the refolding/folding
of the polypeptide, which will now form a more stabile intermediate that can fold into a native ETF-QO protein. If FAD is low (patients before the riboflavin
treatment), the polypeptide may be too unstable for the chaperonin system to fold and it will be degraded after one or several rounds of folding with the chaperonin system. Another factor that may influence the final destination of the polypeptide is temperature. If there is increased temperature (in the form of fever) the
polypeptide is expected to be more unstable and the chaperonin system will probably not be able to fold it; it will therefore be degraded. The temperature might
also have a destabilizing effect on the stably folded intermediate or the native membrane-inserted ETF-QO proteins, which will be more susceptible to proteolytic
degradation. Our experiments show that wild-type ETF-QO also can be shifted toward the unproductive pathway (red) by changes in the cellular environment
(FAD and temperature).
folding pathway (red arrows), giving rise to significant accumulation of MADD metabolites and clinical disease.
These gene – environmental interactions present a challenge
for the diagnostic evaluation of RR-MADD patients, and compiled knowledge gained by documentation of riboflavin
responsiveness in relation to the clinical and biochemical
course and precise knowledge on the pathogenic nature of
the specific genotypes are needed to further improve our
understanding of genotype – phenotype relations, and gene –
environmental interaction in MADD.
Human Molecular Genetics, 2012, Vol. 21, No. 15
MATERIALS AND METHODS
Plasmid constructs
The region from position 229 to ∗ 40 in the 3′ UTR of a human
wild-type ETFDH cDNA was cloned into the Srf I cloning
site of the pPCR-ScriptTM Amp SK(+) expression vector
(Stratagene), using the pPCR-ScriptTM Amp SK(+) Cloning
Kit (Stratagene). The c.1367C.T/p.Pro456Leu, c.1448C.
T/p.Pro483Leu, c.1285G.C/p.Gly429Arg, c.413T.G/p.
Leu138Arg, c.1414G.A/p.Gly472Arg and c.1001T.C/
p.Leu334Pro variations were each introduced into the
pPCR-Script-ETFDH wild-type plasmid by use of the QuikChange Site-Directed Mutagenesis Kit (Stratagene). The
sequences of the mutagenic inserts were confirmed by DNA
sequencing and subcloned into the HindIII and NotI cloning
sites of the pcDNA3.1(+) expression vector (Invitrogen).
Correct orientation and sequences of the final expression
vectors were confirmed by DNA sequencing. All references
to nucleotides or amino acids in the text are based on
the cDNA sequence of human ETFDH (NM_004453). The
initiating ATG codon is numbered as bp 1_3, and the initiator
methionine is numbered as amino acid 1.
Cell culture
Unless otherwise stated, human embryonic kidney (HEK-293)
cells (cat. no. CRL-1573, American Type Culture Collection)
were maintained in RPMI-1640 medium (Lonza) with 10%
fetal bovine serum (Sigma-Aldrich), 0.06 mg/ml penicillin
(FarmaPlus), 0.1 mg/ml streptomycin and 0.29 mg/ml glutamine (both Sigma-Aldrich). Cells were cultivated in tissue
culture flasks or plates (TPP Techno Plastic Products AG) at
378C with 5% (v/v) CO2 as previously described (54). Transient transfection of cells was performed using FuGENE 6 transfection reagent (Roche Diagnostics) according to Corydon
et al. (54). To test differences in riboflavin responsiveness
among various variant ETF-QO proteins, transfected
HEK-293 cells were cultured in RPMI-1640 media containing
16, 28.5 or 530 nmol/l of riboflavin. These riboflavin concentrations were established in a pilot study where changes in the
steady-state amount of expressed wild-type and variant
p.Pro456Leu ETF-QO protein were measured by western
blot analysis in response to changes in riboflavin concentrations of culture media (data not shown): 16 nmol/l (severe deficient media) is the riboflavin concentration, at which the
amount of both variant and wild-type ETF-QO protein is profoundly decreased and to a similar level; 28.5 nmol/l (moderate deficient media) is the riboflavin concentration, at which
the amount of p.Pro456Leu variant ETF-QO protein is significantly more decreased than that of wild-type ETF-QO;
530 nmol/l (supplemented) is the riboflavin concentration in
standard RPMI-1640 media, at which no significant changes
in the wild-type and variant ETF-QO protein amounts are
observed. Deficient media were made by adding the appropriate amount of riboflavin (Sigma) to riboflavin-depleted
RPMI-1640 media (custom-made by PromoCell) supplemented with 10% dialyzed fetal calf serum (Sigma-Aldrich). For
the experiments conducted at 378C, the cells were cultivated
for 6 days in media with the different concentrations of riboflavin before harvest. For the experiments performed at
3445
408C, the cells were initially cultivated at 378C. Twelve
hours after transfection, the cells were transferred to 408C
and maintained at this temperature until harvest.
Western blot analysis
Protein was extracted in a lysis buffer [50 mM Tris – HCL, pH
7.8, 5 mM EDTA, pH 8.0, 1 mM dithiothreitol (DTT), 10 mg/
ml aprotinin (Sigma-Aldrich), 1 mg/ml trypsin inhibitor (Bie
& Berntsen), one tablet of protease inhibitors (Roche) in
10 ml and 1% Triton X-100] according to Schmidt et al.
(55). Following centrifugation, 25 mg of the lysate was analyzed by SDS – PAGE on 12.5% Tris – HCl Criterion Gels
(Bio-Rad). Western blot analysis was performed according
to Schmidt et al. (55) with polyclonal porcine anti-ETFQO,
a kind gift from Frank Frerman (29), and VDAC (Abcam)
antibodies, followed by incubation with a secondary goat
anti-rabbit-HRP antibody (Dako), using the ECL plus
Western Blotting Detection System (Amersham Biosciences).
Detection was done using the ChemiDoc-Itw Imaging System
(UVP). The intensities of bands were quantified using VisionWorksLS Image Acquisition (UVP).
Mass spectrometry analysis to confirm the identity
of the 64 kDa gel band as human ETF-QO
Twenty-five micrograms of proteins from HEK-293 cells,
expressing wild-type ETF-QO, were extracted and separated
by SDS – PAGE on a 12.5% Tris– HCl Criterion Gel
(Bio-Rad) as described above. The gel was stained with Coomassie Blue, and a gel slice containing 45 –75 kDa proteins
was removed and de-stained. Subsequently, proteins in the
gel were digested by trypsin, and the resulting peptides were
extracted and purified essentially as described in Pedersen
et al. (56). The peptide mixture was analyzed by nano-liquid
chromatography tandem mass spectrometry (nanoLC-MS/
MS) on an Easy nLC (Proxeon, Odense, Denmark) coupled
to an LTQ-Orbitrap mass spectrometer (Thermo Fisher,
Waltham, MA, USA) (57). For the identification of peptides
and proteins, the MS data were searched against IPI human
database version 3.45 (71 983 sequences, released 6/10/
2008) using the MudPIT scoring algorithm in the Mascot software (Matrix Science, version 2.2.04, London, UK). Peptide
mass tolerance and fragment mass tolerance were set to
5 p.p.m. and 0.5 Da, respectively. Eleven peptides were identified (P , 0.006) as peptides originating from human
ETF-QO protein (UniProt #Q16134). The eleven peptides
covered 25% of the protein sequence.
ETF-QO enzymatic assay
Cell pellets were suspended in 0.7 ml of 10 mM Tris – HCl, pH
7.4, and sonicated in an ice-salt bath using a Branson 450 sonifier at 30% duty and a power setting of 2. The cells were sonicated with eight cycles of five bursts and were kept on ice for
1 min between cycles. The disrupted cells were centrifuged at
100 000g for 30 min at 48C and the sedimented membranes
were suspended in 0.15 ml of 10 mM Tris –HCl, pH 7.4, containing 1% dodecyl maltoside and homogenized with 10
strokes in a glass homogenizer. The preparations were then
3446
Human Molecular Genetics, 2012, Vol. 21, No. 15
incubated on ice for 1 h, homogenized and centrifuged as
described above. The resulting supernatant was assayed for
ETF-QO activity. ETF-QO activity was assayed spectrophotometrically at 308C in reaction mixtures containing 20 mM
Hepes (K+), pH 7.4, 30 mM glutaryl-CoA, 0.5 mM glutarylCoA dehydrogenase, 1.5 mM ETF, 0.01% dodecyl maltoside
and 100 mM UQ homolog, Q1 (46,58). Q1 reduction was monitored at 275 nm, using D1 ¼ 7.5 mM21 cm21. Protein concentrations were determined with the Bradford reagent kit
(Bio-Rad). Human glutaryl-CoA dehydrogenase and human
ETF were expressed in Escherichia coli and purified as previously described (59,60). Q1 was obtained from Sigma and
dodecyl maltoside was obtained from Anatrace.
The thermal stability of the ETF-QO variant proteins and
the wild type protein was assayed by measuring ETF-QO enzymatic activity after incubation at an elevated temperature
(428C). Cellular membranes were extracted from cells cultured at 378C, and at supplemented riboflavin concentrations.
The solubilized proteins were incubated at 428C. Aliquots
were removed at 0, 10, 20, 30, 60 and 120 min and the activity
measured as described above.
4– 15% Tris– HCL Criterion gels (Bio-Rad) and by SDS –
PAGE on 12.5% Tris – HCL Criterion gels (Bio-Rad). The
pellet was resuspended in the lysis buffer containing 1%
Triton X-100 (see western blot analysis) and centrifuged at
14 800g for 20 min at 48C. The supernatant (membrane fraction) was analyzed by SDS– PAGE on 12.5% Tris – HCL Criterion gels (Bio-Rad). The final pellet (insoluble fraction) was
resuspended in SDS sample buffer (350 mmol/l Tris– HCL,
6 mmol/l DTT, 10% SDS, 30% glycerol, 0.12% bromphenol
blue) and analyzed on 12.5% Tris– HCL Criterion gels
(Bio-Rad). The gels were dried on a slab gel dryer (DrygelSr,
model SE1160, Hoefer Scientific Instruments, CA, USA),
and radio-labeled ETF-QO proteins were visualized on an
Amersham Biosciences PhosphorImager (STORM 840) and
quantified using the ImageQuantw software.
Statistical analysis
Indicated data are expressed as a mean + SEM of triplicate
experiments. Two-tailed Student’s t-test was used to
compare the mean between groups.
Measurements of intra-cellular hydrogen peroxide
production
ACKNOWLEDGEMENTS
HEK-293 cells expressing wild-type or variant ETF-QO proteins were cultured in standard RPMI-1640 media at 378C as
described above. To measure H2O2, 80% confluent cells
were incubated for 20 min at 378C in phosphate buffered
saline (PBS) containing 10 mmol/l of the redox-sensitive dye
5-(and 6-)chloromethyl-2′ ,7′ -dichlorodihydrofluorescein diacetate (CM-H2DCFDA) (Invitrogen). The cells were washed
three times in PBS, harvested and resuspended in
RPMI-1640 media before fluorescence-activated cell sorting
(FACS) analysis. A minimum of 1 × 104 detached cells
were counted per sample and analyzed on a FACS Aria II
flow cytometer with the FACS Diva software v.5.0.3
(Beckton-Dickinson). Propidium iodide staining was used for
the exclusion of dead cells.
We thank technicians Tina F. Hindkær and Hanne Jacobsen,
Department of Biomedicine, Aarhus University, Aarhus,
Denmark for excellent technical assistance concerning cell
work and FACS analysis, respectively. We also thank technician Marissa Hawkins, Department of Pediatrics, University of
Colorado –Denver and Colorado Intellectual and Developmental Disabilities Research Center, Aurora, CO, USA for excellent assistance during measurement of the ETF-QO enzymatic
activity.
In vitro rat mitochondria import assay
In vitro transcription and translation of wild-type or
c.1367C.T ETFDH cDNA inserted in pcDNA3.1(+) expression vectors (see above) were performed in the presence of
[35S]methionine, and the translation products were imported
to isolated rat liver mitochondria as described in Pedersen
et al. (44). Experiments were carried out at 26 and 378C,
and intra-mitochondrial biogenesis of ETF-QO protein was
followed by withdrawing samples after 15, 30, 60 and
180 min of import. The samples were treated with trypsin to
remove non-imported protein, and mitochondrial pellets
were isolated as described in Pedersen et al. (44). The
pellets were kept at 2808C and resuspended in 20 ml of
lysis buffer [50 mmol/l Tris – HCL, pH 7.8, 5 mmol/l EDTA,
1 mmol/l DTT, 10 mg/ml aprotinin, 1 mg/ml soybean-trypsin
inhibitor, 250 mmol/l sucrose]. Samples were sonicated for
20 s followed by centrifugation at 14 800g for 20 min at
48C. The supernatant (matrix and inter-membrane space)
was carefully removed and analyzed by native PAGE on
Conflict of Interest statement. None declared.
FUNDING
This work was supported by grants from the Danish Council
for Independent Research (271-08-0120 to R.K.J.O.) and by
the Lundbeck Foundation (R44-A4454 to R.K.J.O.).
REFERENCES
1. Ghisla, S. and Thorpe, C. (2004) Acyl-CoA dehydrogenases. A
mechanistic overview. Eur. J. Biochem., 271, 494–508.
2. Frerman, F.E. and Goodman, S.I. (2001) Defects of electron transfer
flavoprotein and electron transfer flavoprotein-ubiquinone oxidoreductase.
In Scriver, C.R., Beaudet, A.L., Sly, W.S. and Valle, D. (eds), The
Metabolic and Molecular Bases of Inherited Diseases. McGraw-Hill,
New York, NY, pp. 2357– 2365.
3. Olsen, R.K., Andresen, B.S., Christensen, E., Bross, P., Skovby, F. and
Gregersen, N. (2003) Clear relationship between ETF/ETFDH genotype
and phenotype in patients with multiple acyl-CoA dehydrogenation
deficiency. Hum. Mutat., 22, 12–23.
4. Curcoy, A., Olsen, R.K., Ribes, A., Trenchs, V., Vilaseca, M.A.,
Campistol, J., Osorio, J.H., Andresen, B.S. and Gregersen, N. (2003)
Late-onset form of beta-electron transfer flavoprotein deficiency. Mol.
Genet. Metab., 78, 247–249.
5. Purevjav, E., Kimura, M., Takusa, Y., Ohura, T., Tsuchiya, M., Hara, N.,
Fukao, T. and Yamaguchi, S. (2002) Molecular study of electron transfer
flavoprotein alpha-subunit deficiency in two Japanese children with
Human Molecular Genetics, 2012, Vol. 21, No. 15
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
different phenotypes of glutaric acidemia type II. Eur. J. Clin. Invest., 32,
707–712.
Olsen, R.K., Pourfarzam, M., Morris, A.A., Dias, R.C., Knudsen, I.,
Andresen, B.S., Gregersen, N. and Olpin, S.E. (2004) Lipid-storage
myopathy and respiratory insufficiency due to ETFQO mutations in a
patient with late-onset multiple acyl-CoA dehydrogenation deficiency.
J. Inherit. Metab. Dis., 27, 671–678.
Liang, W.C., Ohkuma, A., Hayashi, Y.K., Lopez, L.C., Hirano, M.,
Nonaka, I., Noguchi, S., Chen, L.H., Jong, Y.J. and Nishino, I. (2009)
ETFDH mutations, CoQ10 levels, and respiratory chain activities in
patients with riboflavin-responsive multiple acyl-CoA dehydrogenase
deficiency. Neuromuscul. Disord., 19, 212–216.
Gempel, K., Topaloglu, H., Talim, B., Schneiderat, P., Schoser, B.G.H.,
Hans, V.H., Palmafy, B., Kale, G., Tokatli, A., Quinzii, C. et al. (2007)
The myopathic form of coenzyme Q10 deficiency is caused by mutations
in the electron-transferring-flavoprotein dehydrogenase (ETFDH) gene.
Brain, 130, 2037– 2044.
Rhead, W., Roettger, V., Marshall, T. and Amendt, B. (1993) Multiple
acyl-coenzyme A dehydrogenation disorder responsive to riboflavin:
substrate oxidation, flavin metabolism, and flavoenzyme activities in
fibroblasts. Pediatr. Res., 33, 129–135.
Di, D.S., Frerman, F.E., Rimoldi, M., Rinaldo, P., Taroni, F. and
Wiesmann, U.N. (1986) Systemic carnitine deficiency due to lack of
electron transfer flavoprotein:ubiquinone oxidoreductase. Neurology, 36,
957–963.
Beresford, M.W., Pourfarzam, M. and Davidson, J.E. (2006) So doctor,
what exactly is wrong with my muscles? Glutaric aciduria type II
presenting in a teenager. Neuromuscul. Disord., 16, 613.
de Visser, M., Scholte, H.R., Schutgens, R.B., Bolhuis, P.A.,
Luyt-Houwen, I.E., Vaandrager-Verduin, M.H., Veder, H.A. and Oey,
P.L. (1986) Riboflavin-responsive lipid-storage myopathy and glutaric
aciduria type II of early adult onset. Neurology, 36, 367– 372.
Olsen, R.K., Olpin, S.E., Andresen, B.S., Miedzybrodzka, Z.H.,
Pourfarzam, M., Merinero, B., Frerman, F.E., Beresford, M.W., Dean,
J.C., Cornelius, N. et al. (2007) ETFDH mutations as a major cause of
riboflavin-responsive multiple acyl-CoA dehydrogenation deficiency.
Brain, 130, 2045– 2054.
Gregersen, N., Wintzensen, H., Christensen, S.K., Christensen, M.F.,
Brandt, N.J. and Rasmussen, K. (1982) C6-C10-dicarboxylic aciduria:
investigations of a patient with riboflavin responsive multiple acyl-CoA
dehydrogenation defects. Pediatr. Res., 16, 861–868.
Gregersen, N., Rhead, W. and Christensen, E. (1990) Riboflavin
responsive glutaric aciduria type II. Prog. Clin. Biol. Res., 321, 477–494.
Gianazza, E., Vergani, L., Wait, R., Brizio, C., Brambilla, D., Begum, S.,
Giancaspero, T.A., Conserva, F., Eberini, I., Bufano, D. et al. (2006)
Coordinated and reversible reduction of enzymes involved in terminal
oxidative metabolism in skeletal muscle mitochondria from a
riboflavin-responsive, multiple acyl-CoA dehydrogenase deficiency
patient. Electrophoresis, 27, 1182– 1198.
Vergani, L., Barile, M., Angelini, C., Burlina, A.B., Nijtmans, L., Freda,
M.P., Brizio, C., Zerbetto, E. and Dabbeni-Sala, F. (1999) Riboflavin
therapy. Biochemical heterogeneity in two adult lipid storage myopathies.
Brain, 122, 2401– 2411.
Bosch, A.M., Abeling, N.G., Ijlst, L., Knoester, H., van der Pol, W.L.,
Stroomer, A.E., Wanders, R.J., Visser, G., Wijburg, F.A., Duran, M. and
Waterham, H.R. (2011) Brown-Vialetto-Van Laere and Fazio Londe
syndrome is associated with a riboflavin transporter defect mimicking
mild MADD: a new inborn error of metabolism with potential treatment.
J. Inherit. Metab. Dis., 34, 159–164.
Ho, G., Yonezawa, A., Masuda, S., Inui, K., Sim, K.G., Carpenter, K.,
Olsen, R.K., Mitchell, J.J., Rhead, W.J., Peters, G. and Christodoulou, J.
(2011) Maternal riboflavin deficiency, resulting in transient neonatal-onset
glutaric aciduria Type 2, is caused by a microdeletion in the riboflavin
transporter gene GPR172B. Hum. Mutat., 32, E1976–E1984.
Johnson, J.O., Gibbs, J.R., Van, M.L., Houlden, H. and Singleton, A.B.
(2010) Exome sequencing in Brown-Vialetto-van Laere syndrome.
Am. J. Hum. Genet., 87, 567–569.
Green, P., Wiseman, M., Crow, Y.J., Houlden, H., Riphagen, S., Lin, J.P.,
Raymond, F.L., Childs, A.M., Sheridan, E., Edwards, S. and Josifova, D.J.
(2010) Brown-Vialetto-Van Laere syndrome, a ponto-bulbar palsy with
deafness, is caused by mutations in c20orf54. Am. J. Hum. Genet., 86,
485–489.
3447
22. Wen, B., Dai, T., Li, W., Zhao, Y., Liu, S., Zhang, C., Li, H., Wu, J., Li,
D. and Yan, C. (2010) Riboflavin-responsive lipid-storage myopathy
caused by ETFDH gene mutations. J. Neurol. Neurosurg. Psychiatry, 81,
231– 236.
23. Lammer, A.B., Rolinski, B., Ahting, U. and Heuss, D. (2011) Multiple
acyl-CoA-dehydrogenase deficiency (MADD)—a novel mutation of
electron-transferring-flavoprotein dehydrogenase ETFDH. J. Neurol. Sci.,
307, 166–167.
24. Wang, Z.Q., Chen, X.J., Murong, S.X., Wang, N. and Wu, Z.Y. (2011)
Molecular analysis of 51 unrelated pedigrees with late-onset multiple
acyl-CoA dehydrogenation deficiency (MADD) in southern China
confirmed the most common ETFDH mutation and high carrier frequency
of c.250G.A. J. Mol. Med. (Berl.), 89, 569– 576.
25. Wolfe, L.A., He, M., Vockley, J., Payne, N., Rhead, W., Hoppel, C.,
Spector, E., Gernert, K. and Gibson, K.M. (2010) Novel ETF
dehydrogenase mutations in a patient with mild glutaric aciduria
type II and complex II-III deficiency in liver and muscle. J. Inherit.
Metab. Dis. [Epub ahead of print].
26. Lan, M.Y., Fu, M.H., Liu, Y.F., Huang, C.C., Chang, Y.Y., Liu, J.S.,
Peng, C.H. and Chen, S.S. (2010) High frequency of ETFDH c.250G.A
mutation in Taiwanese patients with late-onset lipid storage myopathy.
Clin. Genet., 78, 565– 569.
27. Er, T.K., Liang, W.C., Chang, J.G. and Jong, Y.J. (2010) High resolution
melting analysis facilitates mutation screening of ETFDH gene:
applications in riboflavin-responsive multiple acyl-CoA dehydrogenase
deficiency. Clin. Chim. Acta, 411, 690–699.
28. Law, L.K., Tang, N.L., Hui, J., Fung, S.L., Ruiter, J., Wanders, R.J., Fok,
T.F. and Lam, C.W. (2009) Novel mutations in ETFDH gene in Chinese
patients with riboflavin-responsive multiple acyl-CoA dehydrogenase
deficiency. Clin. Chim. Acta, 404, 95–99.
29. Goodman, S.I., Axtell, K.M., Bindoff, L.A., Beard, S.E., Gill, R.E. and
Frerman, F.E. (1994) Molecular cloning and expression of a cDNA
encoding human electron transfer flavoprotein-ubiquinone
oxidoreductase. Eur. J. Biochem., 219, 277–286.
30. Usselman, R.J., Fielding, A.J., Frerman, F.E., Watmough, N.J., Eaton,
G.R. and Eaton, S.S. (2008) Impact of mutations on the midpoint potential
of the [4Fe-4S]+1,+2 cluster and on catalytic activity in electron transfer
flavoprotein-ubiquinone oxidoreductase (ETF-QO). Biochemistry, 47,
92– 100.
31. Zhang, J., Frerman, F.E. and Kim, J.J. (2006) Structure of electron transfer
flavoprotein-ubiquinone oxidoreductase and electron transfer to the
mitochondrial ubiquinone pool. Proc. Natl Acad. Sci. USA, 103,
16212– 16217.
32. Swanson, M.A., Usselman, R.J., Frerman, F.E., Eaton, G.R. and Eaton,
S.S. (2008) The iron-sulfur cluster of electron transfer
flavoprotein-ubiquinone oxidoreductase is the electron acceptor for
electron transfer flavoprotein. Biochemistry, 47, 8894– 8901.
33. Gregersen, N., Bross, P., Vang, S. and Christensen, J.H. (2006) Protein
misfolding and human disease. Annu. Rev. Genomics Hum. Genet., 7,
103– 124.
34. Henriques, B.J., Olsen, R.K., Bross, P. and Gomes, C.M. (2010) Emerging
roles for riboflavin in functional rescue of mitochondrial beta-oxidation
flavoenzymes. Curr. Med. Chem., 17, 3842–3854.
35. Otto, A., Stoltz, M., Sailer, H.P. and Brandsch, R. (1996) Biogenesis of
the covalently flavinylated mitochondrial enzyme dimethylglycine
dehydrogenase. J. Biol. Chem., 271, 9823–9829.
36. Tahallah, N., Van Den Heuvel, R.H., Van Den Berg, W.A., Maier, C.S.,
Van Berkel, W.J. and Heck, A.J. (2002) Cofactor-dependent assembly
of the flavoenzyme vanillyl-alcohol oxidase. J. Biol. Chem., 277,
36425– 36432.
37. Sato, K., Nishina, Y. and Shiga, K. (1996) In vitro refolding and unfolding
of subunits of electron-transferring flavoprotein: characterization of the
folding intermediates and the effects of FAD and AMP on the folding
reaction. J. Biochem., 120, 276–285.
38. Wittung-Stafshede, P. (2002) Role of cofactors in protein folding. Acc.
Chem. Res., 35, 201–208.
39. Nagao, M. and Tanaka, K. (1992) FAD-dependent regulation of
transcription, translation, post-translational processing, and
post-processing stability of various mitochondrial acyl-CoA
dehydrogenases and of electron transfer flavoprotein and the site of
holoenzyme formation. J. Biol. Chem., 267, 17925– 17932.
40. Henriques, B.J., Rodrigues, J.V., Olsen, R.K., Bross, P. and Gomes, C.M.
(2009) Role of flavinylation in a mild variant of multiple acyl-CoA
3448
41.
42.
43.
44.
45.
46.
47.
48.
49.
Human Molecular Genetics, 2012, Vol. 21, No. 15
dehydrogenation deficiency: a molecular rationale for the effects of
riboflavin supplementation. J. Biol. Chem., 284, 4222– 4229.
Gregersen, N., Andresen, B.S., Pedersen, C.B., Olsen, R.K., Corydon, T.J.
and Bross, P. (2008) Mitochondrial fatty acid oxidation defects—
remaining challenges. J. Inherit. Metab. Dis., 31, 643–657.
Goodman, S.I., Binard, R.J., Woontner, M.R. and Frerman, F.E. (2002)
Glutaric acidemia type II: gene structure and mutations of the electron
transfer flavoprotein:ubiquinone oxidoreductase (ETF:QO) gene. Mol.
Genet. Metab., 77, 86–90.
Zeevalk, G.D., Bernard, L.P., Song, C., Gluck, M. and Ehrhart, J. (2005)
Mitochondrial inhibition and oxidative stress: reciprocating players in
neurodegeneration. Antioxid. Redox. Signal., 7, 1117– 1139.
Pedersen, C.B., Bross, P., Winter, V.S., Corydon, T.J., Bolund, L.,
Bartlett, K., Vockley, J. and Gregersen, N. (2003) Misfolding,
degradation, and aggregation of variant proteins. The molecular
pathogenesis of short chain acyl-CoA dehydrogenase (SCAD) deficiency.
J. Biol. Chem., 278, 47449–47458.
Pedersen, C.B., Kolvraa, S., Kolvraa, A., Stenbroen, V., Kjeldsen, M.,
Ensenauer, R., Tein, I., Matern, D., Rinaldo, P., Vianey-Saban, C. et al.
(2008) The ACADS gene variation spectrum in 114 patients with
short-chain acyl-CoA dehydrogenase (SCAD) deficiency is dominated by
missense variations leading to protein misfolding at the cellular level.
Hum. Genet., 124, 43– 56.
Simkovic, M., Degala, G.D., Eaton, S.S. and Frerman, F.E. (2002)
Expression of human electron transfer flavoprotein-ubiquinone
oxidoreductase from a baculovirus vector: kinetic and spectral
characterization of the human protein. Biochem. J., 364,
659– 667.
Er, T.K., Chen, C.C., Liu, Y.Y., Chang, H.C., Chien, Y.H., Chang, J.G.,
Hwang, J.K. and Jong, Y.J. (2011) Computational analysis of a novel
mutation in ETFDH gene highlights its long-range effects on the
FAD-binding motif. BMC Struct. Biol., 11, 43.
Krieger, F., Moglich, A. and Kiefhaber, T. (2005) Effect of proline and
glycine residues on dynamics and barriers of loop formation in
polypeptide chains. J. Am. Chem. Soc., 127, 3346– 3352.
Frerman, F.E. and Goodman, S.I. (1985) Deficiency of electron transfer
flavoprotein or electron transfer flavoprotein:ubiquinone oxidoreductase
in glutaric acidemia type II fibroblasts. Proc. Natl Acad. Sci. USA, 82,
4517– 4520.
50. Loehr, J.P., Goodman, S.I. and Frerman, F.E. (1990) Glutaric acidemia
type II: heterogeneity of clinical and biochemical phenotypes. Pediatr.
Res., 27, 311–315.
51. Lakshmi, A.V. (1998) Riboflavin metabolism—relevance to human
nutrition. Indian J. Med. Res., 108, 182– 190.
52. Powers, H.J. (2003) Riboflavin (vitamin B-2) and health. Am. J. Clin.
Nutr., 77, 1352– 1360.
53. Windmueller, H.G., Anderson, A.A. and Mickelsen, O. (1964) Elevated
riboflavin levels in urine of fasting human subjects. Am. J. Clin. Nutr.,
15, 73– 76.
54. Corydon, T.J., Bross, P., Holst, H.U., Neve, S., Kristiansen, K., Gregersen,
N. and Bolund, L. (1998) A human homologue of Escherichia coli ClpP
caseinolytic protease: recombinant expression, intracellular processing
and subcellular localization. Biochem. J., 331, 309–316.
55. Schmidt, S.P., Corydon, T.J., Pedersen, C.B., Bross, P. and Gregersen, N.
(2010) Misfolding of short-chain acyl-CoA dehydrogenase leads to
mitochondrial fission and oxidative stress. Mol. Genet. Metab., 100,
155–162.
56. Pedersen, C.B., Zolkipli, Z., Vang, S., Palmfeldt, J., Kjeldsen, M.,
Stenbroen, V., Schmidt, S.P., Wanders, R.J., Ruiter, J.P., Wibrand, F.
et al. (2010) Antioxidant dysfunction: potential risk for neurotoxicity
in ethylmalonic aciduria. J. Inherit. Metab. Dis., 33,
211–222.
57. Palmfeldt, J., Vang, S., Stenbroen, V., Pavlou, E., Baycheva, M., Buchal,
G., Monavari, A.A., Augoustides-Savvopoulou, P., Mandel, H. and
Gregersen, N. (2011) Proteomics reveals that redox regulation is disrupted
in patients with ethylmalonic encephalopathy. J. Proteome. Res., 10,
2389– 2396.
58. Ramsay, R.R., Steenkamp, D.J. and Husain, M. (1987) Reactions of
electron-transfer flavoprotein and electron-transfer flavoprotein:
ubiquinone oxidoreductase. Biochem. J., 241, 883– 892.
59. Dwyer, T.M., Rao, K.S., Goodman, S.I. and Frerman, F.E. (2000) Proton
abstraction reaction, steady-state kinetics, and oxidation-reduction
potential of human glutaryl-CoA dehydrogenase. Biochemistry, 39,
11488– 11499.
60. Griffin, K.J., Degala, G.D., Eisenreich, W., Muller, F., Bacher, A. and
Frerman, F.E. (1998) 31P-NMR spectroscopy of human and Paracoccus
denitrificans electron transfer flavoproteins, and 13C- and 15N-NMR
spectroscopy of human electron transfer flavoprotein in the oxidised and
reduced states. Eur. J. Biochem., 255, 125– 132.