Andresen1,2 1 Research Unit for Molecular Medicine, Aarhus University Hospital and Faculty of Health Sciences and2Institute of Human Genetics, Aarhus University, Aarhus, Denmark Mitochon
Trang 1M I N I R E V I E W
Genetic defects in fatty acid b-oxidation and acyl-CoA dehydrogenases Molecular pathogenesis and genotype–phenotype relationships
Niels Gregersen1, Peter Bross1and Brage S Andresen1,2
1
Research Unit for Molecular Medicine, Aarhus University Hospital and Faculty of Health Sciences and2Institute of Human Genetics, Aarhus University, Aarhus, Denmark
Mitochondrial fatty acid oxidation deficiencies are due to
genetic defects in enzymes of fatty acid b-oxidation and
transport proteins.Genetic defects have been identified in
most of the genes where nearly all types of sequence
vari-ations (mutation types) have been associated with disease.In
this paper, we will discuss the effects of the various types of
sequence variations encountered and review current
know-ledge regarding the genotype–phenotype relationship,
espe-cially in patients with acyl-CoA dehydrogenase deficiencies
where sufficient material exists for a meaningful discussion
Because mis-sense sequence variations are prevalent in
these diseases, we will discuss the implications of these types
of sequence variations on the processing and folding of
mis-sense variant proteins.As the prevalent mis-sense
vari-ant K304E MCAD protein has been studied intensively, the investigations on biogenesis, stability and kinetic properties for this variant enzyme will be discussed in detail and used as
a paradigm for the study of other mis-sense variant proteins
We conclude that the total effect of mis-sense sequence variations may comprise an invariable – sequence variation specific – effect on the catalytic parameters and a conditional effect, which is dependent on cellular, physiological and genetic factors other than the sequence variation itself Keywords: fatty acid b-oxidation; acyl-CoA dehydrogenase; VLCAD; MCAD; SCAD; mutation type; protein quality control system; molecular chaperones; intracellular proteases; genotype–phenotype
Introduction
During the last 25 years, the number of known
mitochond-rial fatty acid oxidation defects, as well as the number of
patients with associated disease states, has been increasing
steadily [1,2].Since the first descriptions of muscle carnitine
palmitoyltransferase (carnitine palmitoyl-CoA transferase
II; CPTII) deficiency [3]; systemic carnitine (carnitine
transporter; CAT) deficiency [4] and nonketotic
dicarboxy-lic aciduria [medium-chain acyl-CoA dehydrogenase (MCAD) deficiency] in the 1970s [5], defects in many enzymes and transport proteins involved in the oxidation of fatty acids have been discovered (Table 1)
The clinical features in patients with different defects, and among patients with deficiencies of the same transport protein/enzyme, are very diverse but the most prevalent symptoms are always related to heart, liver and/or the neuromuscular systems
Deficiencies in the transporters and enzymes involved in the oxidation of long-chain fatty acids are generally severe and may cause death and severe morbidity early in life.In contrast, the most common features of disorders of enzymes involved in the metabolism of medium-chain fatty acids are episodic hypoglycaemia and liver-associated disturbances
of consciousness, which – if untreated – may lead to coma and death.These severe, acute life-threatening episodes are rarely seen in the defects of short-chain fatty acid oxidation, where the most common symptoms are neuromuscular Despite the fact that defects of the long-chain fatty acid metabolism often cause severe fatal disease, it has become evident that the whole range of clinical symptoms, from fatal heart or liver failure to mild muscular disabilities, has been observed in patients with these diseases.An exception
is in CPTI deficiency, where liver symptoms predominate
On the other hand, it is unusual to observe heart and liver pathologies in patients with deficiencies of short-chain fatty acid metabolism
Furthermore, in patients with very-long-chain acyl-CoA dehydrogenase (VLCAD), CPTII and electron transfer flavoprotein (ETF)/ETF dehydrogenase (ETFDH) defects
Correspondence to N.Gregersen, Research Unit for Molecular
Medicine, Skejby Sygehus, 8200 Aarhus N, Denmark.
Fax: + 45 89496018, Tel.: + 45 89495140, E-mail: nig@mmf.au.dk
Abbreviations: CPTI (II), carnitine palmitoyl-CoA transferase I (or II);
ETF, electron transfer flavoprotein; ETFDH, ETF dehydrogenase;
Hsp, heat shock protein; HGMD, Human Gene Mutation Database;
MCAD, medium-chain acyl-CoA dehydrogenase; NCBI, National
Centre for Biotechnology Information; PKU, phenylketonuria; PTC,
premature termination codon; SNP, single nucleotide polymorphism;
VLCAD, very-long-chain acyl-CoA dehydrogenase.
Definitions: Sequence variation designates all types of gene sequence
changes, including conventional disease-causing mutations and
null-mutations as well as neutral and susceptibility polymorphisms, as
recommended by The Human Genome Variation Society [den
Dun-nen, J.T & Antonarakis, S.E (2001) Hum Genet 109, 121–124].
Where not featured in the abbreviations list, enzyme and transport
protein abbreviations are defined in Table 1.
Note: A web site is available at http://www.auh.dk/sks/afd/mmf.dk
(Received 17 July 2003, revised 13 October 2003,
accepted 23 October 2003)
Trang 2there is a clear correlation between the degree of deficiency
and the clinical phenotype ([1] and references therein).Severe
deficiencies generally result in fatal or severe disabilities,
while milder defects are associated with mainly muscular
symptoms.Such a correlation is not seen in patients with
medium- and short-chain defects.In these diseases mild
defects may not be associated with detectable disease
The realization of these associations – and lack of
connections – between enzymatic phenotypes and clinical
phenotypes has emerged through careful studies of many
patients over many years.However, the cloning and
elucidation of the genes and genomic structures for nearly
all clinically relevant enzymes and transport proteins of fatty
acid oxidation has stimulated our knowledge considerably,
both with respect to the possibility of specific molecular
genetic diagnostics – which are insensitive to disturbances in
the biochemical and cellular factors – and because this
knowledge has made genotype–phenotype investigations
possible
In the following we will summarize the current knowledge
regarding the genes that code for clinically relevant
trans-port proteins and the enzymes of mitochondrial fatty acid
oxidation, as available in publicly accessible databases
developed and maintained at the National Center for
Biotechnology Information (NCBI; http://www.ncbi.nlm
nih.gov/genome/guide/human/)
Despite the fact that the annotated genomic structures
and cDNAs may not be exact, the information is sufficiently
accurate for the purpose of the present discussion and the
databases are an extremely valuable resource with links to
existing original literature.For the discussion concerning
the effects of the various types of sequence variations we
have used the information in the Human Gene Mutation
Database [6] (HGMD, http://www.hgmd.org/; Cardiff,
UK), which remains the most comprehensive database containing published disease-associated sequence variations
in fatty acid oxidation genes
Lastly, to give the descriptions of genes and sequence variations biological significance, we will review the current knowledge concerning genotype–phenotype relationships in acyl-CoA dehydrogenase deficiencies, which will illuminate considerations and ideas that are applicable to the other fatty acid oxidation deficiencies and many other genetic disorders
Genomic structures and disease-associated sequence variations in genes encoding enzymes of fatty acid oxidation
The draft sequence of the human genome was published in
2001 [7,8] and the assembly of large contigs and the annotation of genes makes it possible to find gene and genome structures for all genes that encode the enzymes and transport proteins of mitochondrial fatty acid oxida-tion (except for carnitine/acylcarnitine translocase) in the NCBI databases (Table 2).The information extracted includes: chromosome localization; gene length (total sequence) and the number of exons in the gene and nucleotides in the coding region of each gene.In addition, the types of sequence alterations identified in patients with fatty acid oxidation defects, as extracted from the HGMD
in Cardiff, are also summarized in Table 2.The sequence variations are categorized into those that probably result
in no enzyme protein (null-mutations) and those for which the effect is more unpredictable.This is a little different from the categorization in the database.In Table 2 we have on one hand counted large deletions, small out-of-frame deletions/insertions, stop-codon introductions and
Table 1 Transporter proteins and enzymes involved in the mitochondrial saturated fatty acid oxidation FATP, fatty acid transport protein; CAT, carnitine transporter; CACT, carnitine/acylcarnitine translocase; CPT I, carnitine palmitoyltransferase I (liver); CPT II, carnitine palmitoyl-transferase II; ETF/ETFDH, electron transport flavoprotein/electron transport flavoprotein dehydrogenase; VLCAD, very-long-chain acyl-CoA dehydrogenase; MTP, mitochondrial trifunctional protein (including long-chain enoyl-CoA hydratase, long-chain 3-hydroxyacyl-CoA dehy-drogenase and long-chain 3-oxoacyl-CoA thiolase); MCAD, medium-chain acyl-CoA dehydehy-drogenase; SCHAD, short-chain 3-hydroxyacyl-CoA dehydrogenase; SCKAT, short-chain 3-oxoacyl-CoA thiolase; SCAD, short-chain acyl-CoA dehydrogenase.
Year disease discovered Typical organ involvement Recent updates/references Transporters
CAT (plasma membrane) 1975 [4] Liver, heart, muscle [32]
CACT (mitochondrial membrane) 1992 [72] Heart, liver, muscle [35]
Enzymes (mitochondrial membrane)
MTP
Enzymes (mitochondrial matrix)
Trang 3consensus splice site changes, and on the other hand, mis-sense variations, small in-frame deletions/insertions and nonconsensus splice site changes.We will discuss the various types of sequence variations below.In Fig.1 the genes, including information on the HGMD accessible disease associated gene defects, are depicted for VLCAD, MCAD and short-chain acyl-CoA dehydrogenase (SCAD)
Types of sequence variations in fatty acid oxidation genes
The first level of analysis of the genotype–phenotype relationship in fatty acid oxidation deficiencies is a discus-sion of the various types of sequence variations identified and associated to the disease in patients.As large deletions – where whole parts of the genes are missing – are rare and because the description in the database is restricted to the cDNA level, we do not discuss this type of gene defect further, but concentrate on the other types
Small out-of-frame deletions/insertions, including stop-codon introductions
These have been encountered in nine of the 12 fatty acid oxidation defects where sequence variations have been identified in the corresponding genes (Table 2).The change
in reading frame resulting from this type of sequence variation leads to the introduction of a premature termin-ation codon (PTC) shortly downstream of the deletion/ insertion.A PTC may also be created by changing an amino acid codon to a stop-codon.By means of a number of poorly understood mechanisms, the PTC – if it is present more than 50 nucleotides upstream of the last intron in the gene – will be recognized by a RNA surveillance mechanism [9,10].This mechanism is mediated by a general mRNA quality control system, which targets mRNA species containing PTCs to the so-called nonsense mediated decay (NMD) pathway.The consequence is that the mRNA is degraded and no polypeptide is synthesized.If small amounts of PTC-containing mRNA should escape the NMD system, it is most probable that the encoded truncated polypeptide will be rapidly degraded by intracel-lular proteases, which are part of the protein quality control system, which will be discussed below.Thus, these types of sequence variations will, as a rule, result in null-mutations, characterized by negligible amounts of variant protein product formed
Splice site changes
A number of different splice site sequence variations have been encountered in genes resulting in fatty acid oxidation deficiencies.Depending on the position in relation to the intron–exon border the effect may vary.Variations in 100% conserved AG and GT dinucleotides immediately before and after an exon may result in exon skipping, intron retention or activation of cryptic splice sites [11], usually resulting in a change of reading frame and consequently degradation of mRNA.In cases where the reading frame is unchanged, the truncated protein is most probably rapidly degraded due to misfolding (see below)
Gene (kbp)
Stop- codons
Trang 4Splice site variations located further from the exons,
as seen in carnitine/acylcarnitine translocase (CACT),
VLCAD and long-chain 3-hydroxy acyl-CoA
dehydro-genase (LCHAD) deficiencies, may or may not result in
complete abolition of the active enzyme [12].Thus, the effect
may range from severe to mild as discussed for mis-sense
variations below
Small in-frame deletions/insertions
These sequence defects delete or insert one or more
amino acid codons in the mRNA.Usually sequence
variations of this type have no consequences for the
sta-bility and processing of the mRNA and a truncated or
elongated polypeptide will be produced.This is the case
in several of the fatty acid oxidation deficiencies (Table 2)
but the consequences are difficult to predict.However, as
small insertions or deletions will affect the structural
stability more severely if located in a-helixes or b-sheets
than in structural loops, some idea of the effect can be
predicted if the crystal structure of the protein in question
is known
In general, the polypeptide is synthesized, but it may have
difficulties in achieving the correct active structure and will
most often be degraded by the protein quality control
system, which is dependent on the nature of the sequence
variation at the protein level and on the cellular conditions,
as discussed below
Mis-sense sequence variations About two thirds of all disease-associated sequence varia-tions in patients with fatty acid oxidation deficiencies are of the mis-sense type (Table 2), which changes a codon from one amino acid into another.Usually such sequence variations result in normal mRNA production and pro-cessing and normal translation to the corresponding variant polypeptide.By inspecting the available crystal structures of wild-type protein it is seen that the vast majority of such changes are located distant from the active centres.Only a few seem to be involved in the catalytic mechanism.The rest perturb folding, resulting in either impaired production of a correctly folded active enzyme, or in an unstable active enzyme [13].Although there have been several attempts, it is only possible to predict the effect of the mutation from the nature and position of the altered amino acid [14–16] in a minority of cases.In certain cases, some rationalization – mostly post hoc – may be possible.However, the general conclusion seems to be that predictions on the severity of a given mis-sense variation are still very uncertain.Despite the fact that a certain correlation exists between the molecular interactions in the structured active protein and the
Fig 1 The gene structures of the ACADVL (VLCAD), ACADM (MCAD) and ACADS (SCAD) genes The number and approximate size of all coding regions are shown and the 5¢-UTR (untranslated region) as well as the 3¢-UTR are indicated.The information used for the constructions are: VLCAD [82,83] and NCBI nucleotide database gi: 3273227; MCAD [84] and NCBI nucleotide database gi: 187432 and SCAD [85] and NCBI nucleotide database gi: 2995253; 2821943.Sequence variations are designated according to the position relative to the first nucleotide in the start-codon ATG, and they are taken from the Human Gene Mutation Database (HGMD) in Cardiff (http://www.hgmd.org/).
Trang 5interactions that are used during the folding process, the
folding pathway and the molecular forces along this (or
these) path(s) cannot presently be modelled for molecules of
more than 15 kDa [17].Mis-sense sequence variations may
therefore affect the folding of the enzyme protein severely or
they may only perturb it slightly.The folding process is
monitored by the protein quality control systems,
compri-sing molecular chaperones, assisting the folding, and
intracellular proteases, which eliminate misfolded proteins
[13].As the efficiency of these systems is dependent on the
cellular conditions, e.g the temperature and energy level,
and probably also on genetic differences between
individ-uals, the effect of mis-sense sequence variations cannot, in
general, be predicted [18].As will be discussed in the next
section, experimental evaluation can and should be
performed
Recently it has been demonstrated that mis-sense sequence
variations, in addition to influencing protein biogenesis, also
may affect the splicing efficiency by interfering with binding
sites for splice modulating factors [19].Although the effect of
mis-sense variations on splicing has not yet been published in
relation to fatty acid oxidation defects, it has been identified
in relation to isovaleryl CoA dehydrogenase [20] as well as
2-methyl butyryl-CoA dehydrogenase deficiencies [21].This
fascinating phenomenon is in the process of being
charac-terized in the MCAD and VLCAD genes (K.B.Nielsen,
T.Sinnathamby, T.J.Corydon, L.Cartegni, A.R.Krainer,
O.N.Elpeleg, N.Gregersen, J.Kjems & B.S.Andresen,
unpublished observation)
In conclusion, it is only possible to predict the effect for
one third of the disease associated sequence variations in the
fatty acid oxidation genes, i.e for large deletions,
stop-codon (PTC) introductions, consensus splice site changes
and small out-of-frame deletions/insertions.These are
sequence variations preventing formation of functional
protein (putative null-mutations).The rest, i.e.in-frame
deletions/insertions, nonconsensus splice site changes and
mis-sense sequence variations, may show an a priori
unpre-dictable effect, which should be studied experimentally
With the exception of the few variations directly affecting
the catalytic sites, these types of defects represent sequence
variations with potential variable effects.However, even if
the effect of the sequence variation can be elucidated in vitro,
the in vivo effect may be modulated by cellular and genetic
factors.In spite of these reservations concerning the
predictive value of knowing the disease associated sequence
variations in a given patient, this knowledge has obvious
diagnostic implications, which have been discussed in detail
elsewhere [1,22].Furthermore, by performing careful
genotype–phenotype studies the relative importance of the
genetic predisposition and possible cellular and metabolic
disturbances, which are often determinants for the
precipitation of the fatty acid oxidation deficiencies, may
be assessed
Genotype–phenotype relation in fatty acid oxidation deficiencies
A second level of analysis of the genotype–phenotype relation in fatty acid oxidation deficiencies is the investiga-tion of possible associainvestiga-tions between the type of sequence variations and the clinical phenotype.As mentioned in the Introduction, such associations seem to exist in patients with certain of the long-chain defects but not – or at least to
a lesser extent – in patients with medium- or short-chain defects.As our research has focussed on the acyl-CoA dehydrogenase deficiencies, we will use VLCAD, MCAD and SCAD deficiencies as examples and try to extrapolate conclusions drawn from these diseases to the other fatty acid oxidation defects
Very-long-chain acyl-CoA dehydrogenase (VLCAD) deficiency
The clinical spectrum seen in patients with VLCAD deficiency is a prototype for other long-chain defects.As discussed in more detail elsewhere [23], it is possible to distinguish three phenotypes.The first comprises very young infants who die from cardiac and liver disease within the first year of life.The second group comprises older children who do not have cardiac symptoms but show hypoketotic hypoglycemia and hepatomegaly, symptoms which are MCAD deficiency-like (see below).The third group are composed of adolescents and adults who do not show cardiac and hepatic symptoms but who suffer from muscle weakness, which may develop to degenerative disability [24–27].A large number of patients with VLCAD deficiency have been genotyped and Table 3 shows the distribution of the null-mutation/null-mutation and poten-tial variable/potenpoten-tial variable genotypes in the three clinical groups
The most striking result is that homozygosity for sequence variations, which lead to mRNA/protein elimin-ation (null-mutelimin-ations), is exclusively present in the patient group with severe symptoms.There is little doubt that this is
a reflection of a severe enzyme deficiency, which is also reflected in the profile of acyl-carnitines in blood and in patient cells metabolizing long-chain fatty acids [27].Not surprisingly, the severe metabolic block results in profound energy deficiency and corresponding severe clinical symp-toms.To what degree the accumulated long-chain fatty acids and their derivatives, especially the acylcarnitines, may contribute to the clinical phenotype is not known with certainty but these species may disturb membrane function,
Table 3 Distribution of VLCADgenotypes among three clinical subtypes of VLCADdeficiency Data from [23].
Group1 Number of patients with severe childhood form
Group 2 Number of patients with mild childhood form
Group 3 Number of patients with adult form
Potential variable genotype/
potential variable genotype
Trang 6including ion-channels, and thereby perhaps, promote
arrhythmia ([28] and references therein)
It is also noteworthy that the sequence variations with
potentially varying effects (potentially variable genotypes)
are distributed to all three groups.This is most probably a
reflection of the fact that the effect of these sequence
variations may be total inactivation of VLCAD function or
it may be mild, leaving sufficient residual activity to avoid
severe energy deficiency.However, the deficiency is
suffi-cient to promote long-term muscle damage [24,25].Whether
this damage is a result of energy deficiency or the long-term
effect of toxic long-chain fatty acids and their derivatives is
impossible to judge at present
As far as we can determine, the same situation and
arguments apply to CPTII [29] and to ETF/ETFDH
deficiencies [30,31], where the number of patients and
sequence alterations are sufficiently large to perform a
similar analysis
In contrast to these diseases, the other diseases related to
long-chain fatty acid metabolism may not show any
significant association between the severity of the defect/
type of sequence variation and the clinical phenotype, i.e
CAT deficiency [32,33], LCHAD/mitochondrial
trifunc-tional protein (MTP) deficiencies [34], carnitine
acylcarni-tine translocase (CACT) deficiency [35] and CPTI deficiency
[36].However, despite the fact that the number of patients
as well as the number of known disease-associated sequence
variations is still too small to provide a clear picture of the
genotype–phenotype in these diseases, liver-related
patho-logies are most often encountered.That other pathopatho-logies,
such as cardiac dysrhythmia in CPTI and cardiomyopathy
in LCHAD deficiencies, are observed emphasizes the notion
that factors other than the gene defect itself may be decisive,
as is also the case in MCAD deficiency, as discussed below
Medium-chain acyl-CoA dehydrogenase (MCAD)
deficiency
The situation in MCAD deficiency is different from that in
VLCAD deficiency as a single prevalent sequence variation
(985AfiG), resulting in a mis-sense variant protein
(K304E), is present in homozygous form in 80% of all
patients diagnosed with MCAD deficiency.Eighteen
per-cent of patients are compound heterozygous with 985AfiG
on one allele and a rare disease associated sequence
variation on the other, and only about 2% carry other
(rare) sequence variations on both alleles [37].Thus, studies
of the clinical impact of different types of sequence
variations are hampered by the fact that nearly all patients
carry one or two copies of the K304E variant MCAD
enzyme.This amino acid change exerts its effect primarily
by compromising the folding [38–40], but the variant
protein is also unstable [41] and the function is impaired
[42].At least the misfolding and instability are influenced by
cellular and probably also by genetic factors, thus, resulting
in an effect that is totally unpredictable without
experimen-tal approaches, which will be discussed in detail below
Suffice to say that the 985AfiG sequence variation may
result in varying effects and may blur an analysis similar to
the one described for VLCAD above
The age of MCAD deficiency at presentation may vary
from birth to middle age.Clinically, the severity ranges from
fatal, through treatable acute symptoms, mild disabilities, to asymptomatic throughout life.The features are, however, rather uniform; episodic attacks of hypoketotic hypogly-caemia accompanied by lethargy and vomiting, that may develop into hepatic coma and death if not treated by administration of carbohydrate.Usually, MCAD-deficient patients do not experience life-threatening heart-related symptoms, such as arrhythmia [28,43].This indicates that the energy deficiency is milder in MCAD than in VLCAD deficiency or, alternatively, that medium chain-fatty acids and their derivatives are less toxic to cardiac function than long-chain fatty acids and their derivatives.However, it is interesting to note that early investigations of the toxicity of medium-chain fatty acids, such as octanoic acid, showed narcotic properties that may contribute significantly to the lethargy and hepatic coma observed in patients during periods of metabolic decomposition [44,45]
With respect to the energy deficiency, the milder mani-festation is probably caused by a combination of the fact that several cycles of oxidation can proceed before the pathway is blocked and that some enzyme activity may arise from long-chain acyl-CoA dehydrogenase (LCAD) and SCAD because of their overlapping substrate activity with that of MCAD [46]
Although it is known that physiological factors, i.e metabolic stress in connection with fasting and fevers, are important factors for the expression of the disease, it is still
an open question whether there exist other cellular and genetic factors that contribute to the susceptibility in some, but not in all, individuals [18]
Short-chain acyl-CoA dehydrogenase (SCAD) deficiency SCAD deficiency remains difficult to analyse.Compared to long-chain and medium-chain defects, there are at least three peculiarities that are noteworthy with respect to the relationship between the type of sequence variations and clinical features in patients.The first is that nearly all sequence variations identified so far are of the mis-sense type, which in all cases may result in production of some variant protein possessing some residual activity.Among the first 13 disease-associated sequence variations identified
in SCAD-deficient patients, 12 are mis-sense sequence variations and one is an in-frame deletion of three nucleo-tides (Table 2).This trend has continued in more than 100 unpublished cases (N.Gregersen, A.Kølvraa, J.Vockley,
D Matern, I Tein, R Ensenauer, C Vianey-Saban, M.Kjeldsen, V.S.Winter, C.B.Petersen & S.Kølvraa, unpublished observations).Although about half of the mis-sense sequence variations in the SCAD gene are cytidine to thymine substitutions, which usually arises at CpG dinu-cleotides, the overrepresentation is still striking and could be
a reflection of negative selection of germ cells with sequence variations abolishing all enzyme activity.This explanation has been proposed for the similar phenomenon in fumarate hydratase deficiency, where the absence of such sequence variations also is striking [47]
The second point of note is the spectrum of sequence variations observed in patients with SCAD deficiency.In a minority of cases, rare inactivating mis-sense sequence variations are present in homozygous or compound hetero-zygous form, whereas such variations in many cases are
Trang 7present in compound heterozygous form together with one
of two common susceptibility mis-sense sequence variations,
625GfiA and 511CfiT.These variations are present in
homozygous or compound heterozygous form in 14% of the
general population [48].Further, a significant fraction of
patients with SCAD deficiency carry this genotype, in the
absence of any rare inactivating variations [49]
The above situation is apparently different from that
encountered in MCAD deficiency, where a prevalent
sequence variation likewise is present in nearly all patients
but where all individuals carrying the prevalent 985Adefi
Gua on both chromosomes, or in one chromosome with a
rare sequence variation in the other, are at risk of developing
disease.Although preliminary results have shown that the
spectrum of clinical symptoms in patients who are
homo-zygous or compound heterohomo-zygous for 625GuafiAde
or/and 511CytfiThy, are indistinguishable from patients
harbouring rare inactivating sequence variation [50], it must
be assumed that only a minority of individuals, who carry
the two susceptibility variations, are at risk of developing
disease.From this, it follows that there must be other
factors, physiological as well as cellular and/or genetic, that
are implicated in the expression of the disease
This leads to the third point, namely the clinical features
in patients with SCAD deficiency.Only a few patients have
presented with symptoms related to energy deficiency, such
as cardiac symptoms or hypoglycaemia ([49]; N.Gregersen,
A.Kølvraa, J.Vockley, D.Matern, I.Tein, R.Ensenauer,
C Vianey-Saban, M Kjeldsen, V S Winter, C B Petersen
& S.Kølvraa, unpublished observation).The most probable
explanation for this is that SCAD deficiency only blocks the
last cycle of the pathway and that MCAD activity overlaps
with that of SCAD [46], thus, it is probable that near normal
amounts of reducing equivalents are generated.A further
reason is that butyric acid and its derivatives are neither
heart nor liver toxic.On the other hand, butyric acid is
known to exert severe cell toxicity by promotion of cell
differentiation, inhibition of the cell cycle and induction of
apoptosis [51,52].This may be the reason why the
predomi-nant clinical symptoms are neuromuscular
Without going into a detailed discussion about this issue,
there are two important remaining questions.First, how is it
possible that two genetic variations, which are implicated in
severe neuromuscular disease, can achieve such high
frequencies in the general population, and second, what
are the genetic/cellular/biochemical mechanisms which
renders some of the individuals carrying the variations in
homozygous or compound heterozygous form at risk of
developing clinically relevant disease.The first question can
not be answered at this time but we will attempt to answer
the second in the next section
The only other disease affecting the metabolism of
short-chain fatty acids, where sequence variations have been
associated to an enzyme deficiency, is short-chain 3-hydroxy
acyl-CoA dehydrogenase (SCHAD) deficiency.Although
the deficiency has been known in a number of patients for
more than 10 years [53,54] only one patient with
disease-associated sequence variations in both alleles of the SCHAD
gene has been published [55].The clinical symptoms in this
patient, who was shown to be homozygous for a mis-sense
sequence variation in the SCHAD gene (773CytfiThy;
P258L), are quite different from those found in patients with
SCAD deficiency, and include hyperinsulinism.Only time will show the degree of clinical and genetic heterogeneity in this rare disease
Molecular effect of sequence variations with
an a priori unpredictable effect
A third level of analysis of the genotype–phenotype correlation in fatty acid oxidation deficiencies may be the experimental dissection of the molecular effects of sequence variations with a priori unpredictable effects
Traditionally, and long before the genes and protein structures were elucidated, enzymatic diagnosis of most fatty acid oxidation deficiencies was possible and, furthermore, practised in many laboratories.The enzymatic analyses could correctly determine the residual activity in patient cells but the question remained whether the enzyme protein was present with reduced activity, or it was present in diminished amounts.When antibodies against several of the fatty acid oxidation enzyme proteins became available, it was possible
to approach this question.It was soon realized that decreased amounts of protein are the rule rather than the exception However, it was only after gene cloning and synthesis of proteins by recombinant techniques, that it became possible
to resolve the molecular pathogenesis of the increasing number of identified disease associated sequence variations Although it might seem unnecessary to use large resources
to investigate molecular mechanisms, particularly as the diagnoses can be made by direct enzyme activity measure-ment in patient cells, there are at least three good reasons for doing so.The first is to corroborate that an identified sequence variation is associated to the clinical phenotype through a functional effect on the variant protein.This is a very practical and important goal that should be achieved every time a new sequence variation is encountered The second reason is related to the first but extends the purpose to future diagnostic procedures.The ongoing genotype–phenotype studies data will alter the content of sequence variation databases, which, in addition to raw variation data should also contain information about the effects on protein and cellular metabolism.For many diseases, including the fatty acid oxidation deficiencies, it will thus be possible to replace the laborious and expensive enzymatic analysis by gene-based in vitro and in silico methods
The third reason for elucidating the molecular patho-genesis, at least for some model variant proteins, is that knowledge gained from detailed investigation of such proteins may be generalized to other proteins in other diseases.As a paradigm – with possible implications for future treatment of patients – we will discuss the careful investigation of the biogenesis, stability and function of the disease-associated K304E MCAD enzyme protein, and the application of the gained knowledge and methodological approaches to define the role of the two common suscep-tibility variations in the SCAD gene
Molecular effects of the 985AfiG MCAD sequence variation
Surprisingly, at least at the time of discovery, the amino acid lysine, which is replaced by glutamic acid by the
Trang 8985AdefiGua sequence variation at position 304 of the
MCAD protein, is located far from the active centre
However, Western blot analysis of soluble variant protein
after expression in Escherichia coli cells showed diminished
amounts of the K304E MCAD protein compared to the
normal (wild-type) MCAD protein [56].This result
stimulated further investigations, some of which have
provided the basis for the emerging concept that the
effects of mis-sense sequence variations are not only
dependent on the nature and location of the particular
variation but influenced to varying extents by cellular
factors related to the protein quality control systems
[18,38,41,57]
The studies, which will be summarized below, focused
on one hand on the biogenesis and stability of the
variant K304E MCAD protein, coded from the
985AdefiGua allele, and on the other hand on
the activity and chain-length selectivity of active K304E
MCAD enzyme
Investigations of biogenesis and stability
In the late 1980s and early 1990s it was realized that the
expression and folding of many cellular proteins, including
mitochondrial proteins, are dependent on assistance by
molecular chaperones [58,59].As the location was distant
from the active centre, it was hypothesized that the glutamic
acid in position 304 of the K304E MCAD protein distorted
the normal folding.This hypothesis was corroborated by
experiments performed in K.Tanaka’s group [38,39,60] and
in our own [40,41]
Tanaka’s group showed that both wild-type and variant
MCAD proteins were assisted in their intramitochondrial
folding by, first the chaperone heat shock protein 70
(Hsp70) and, subsequently, by chaperonin Hsp60, and
that the K304E protein was retained longer in association
with Hsp60 than was the wild-type protein.These elegant
experiments, which were performed by using rat liver
mitochondria, clearly indicated that MCAD deficiency is
at least in part due to compromised folding of the variant
enzyme protein.In parallel with these studies, we
inves-tigated the effect of co-overexpression of the bacterial
groELS (homologous to human Hsp60/10) in E coli cells,
which over-expressed K304E or wild-type MCAD protein
We found that it was possible to increase the yield of
active variant enzyme considerably but further studies also
showed that it was not possible to rescue more than
40–50% of wild-type activity.These experiments clearly
showed that the folding of the variant protein is
compromised
Inspection of the molecular structure surrounding
posi-tion 304 in the mature MCAD protein [41] indicated that
the lysine at position 304 is in close vicinity to two
opposite-charged aspartate residues at positions 300 and 346,
respectively (Fig.2).Second site mutations at, respectively,
position 300 and 346, indicated that the presence of lysine
at position 304 is important for efficient folding of the
monomer and that the charge interaction between lysine 304
and aspartate 346 is important for tetramer assembly and,
therefore, for the stability of the assembled enzyme protein
[41].These effects may be decisive for the steady-state level
of variant K304E MCAD but these experiments did not
give any data relating to either enzyme activity or substrate selectivity
Investigations of enzyme kinetics Although the main effect of the 985AdefiGua sequence variation is most probably due to distortion of folding and tetramer assembly/stability, small distortions in the conformation at the active site and substrate binding pocket could contribute to the pathogenesis of the 985AdefiGua MCAD gene variation.Kieweg and coworkers [42] addressed this question by determining the kinetic parameters for purified wild-type and variant MCAD protein from over-expressing E coli cells.The authors showed that Vmaxwas similar for wild-type and the variant K304E MCAD proteins (980 vs.970 lmolÆ min)1), whereas Kmwas 3–4 times higher for the variant enzyme, indicating a higher saturation concentration for the optimum substrate octanoyl-CoA compared to the wild-type enzyme.This may have consequences for the amounts of available free CoA for other important cellular processes
Interestingly, the preferred substrate for K304E variant MCAD is dodecanoyl-CoA.At this chain length, both Vmax and Km are similar for wild-type and variant MCAD enzyme
Taken together, these detailed studies on the molecular pathogenesis of the K304E variant enzyme protein have illuminated a number of important aspects of the effects of mis-sense sequence variations in MCAD deficiency in particular but also in fatty acid oxidation deficiencies in general – which will be discussed for SCAD deficiency below – as well as in other genetic diseases, such as phenylketonuria (PKU) [61,62]
Fig 2 An enlarged view of the vicinity of K304 of a monomer of porcine MCAD(PDB accession no 3MDDor 3MDE) Helices H and I are shown in ribbons and side-chain atoms of K304, D346, Q342, D300 and the main chain carbonyl atoms of Q342 are shown as solid balls The side chain of R383 of the neighbouring monomer is represented by open ball-and-stick.Distances between polar atoms in A˚ are shown with dotted lines.Reproduced with permission from Journal of Biological Chemistry [41].
Trang 9The SCAD enigma
As mentioned earlier in this review, the genetic defect in
most patients with SCAD deficiency is not due to rare
inactivating sequence variations but rather to the presence of
one of two (or both) susceptibility gene variations, which are
present in 14% of the general population in configurations
also seen in patients with enzymatically proven SCAD
deficiency [49,63].The goal is to delineate the nature of these
variations, which may help to explain why only certain
individuals carrying these variations develop clinically
relevant disease
The structure of SCAD from rat has been elucidated [64]
From an inspection of the positions of the two variations,
G185S and R147W, it is not obvious how amino acid
changes at these positions could be pathogenic (Fig.3)
Both positions are at the outer surface of the monomeric
structure.In agreement with this location and the fact that
severe defects on enzyme function would not be compatible
with the high frequency in the general population, the
kinetic disturbances were not found to be serious.Purified
R147W protein had kinetic properties similar to the
wild-type, and the kinetic efficiency of G185S protein was about
50% compared to the wild-type enzyme [65].This probably
reflects the change from glycine to serine distorting the
conformation and exact positions of other amino acids
involved in the enzyme mechanism.These results – at least
concerning the G185S variant enzyme – underscore the
predisposing nature and indicate that other factors must be
involved
Early biogenesis and stability studies showed that
wild-type SCAD is more dependent on the chaperonin system
Hsp60/10 (GroELS in E coli) than MCAD [66].While
wild-type MCAD does not need additional assistance by the
chaperonins in E coli at 31C to achieve the active
conformation [40], the yield of functional wild-type SCAD
is increased eightfold by co-overexpression of GroELS at
the same temperature In the same type of experiment,
G185S variant SCAD showed about 30% of wild-type
activity without co-overexpression of GroELS but achieved
wild-type activity after co-overexpression of GroELS.This indicated a greater dependence on chaperonin assistance for the variant protein than for the wild-type but also that when the folding capacity is sufficiently high, the biogenesis of the variant enzyme is as effective as the wild-type enzyme Due to the ineffective folding behaviour of SCAD compared to MCAD in E coli cells, we looked into eukaryotic expression that has been shown to be intrinsic-ally more effective than bacterial expression for a number of MCAD mutant proteins [67].By varying the culture temperature it was possible to detect differences in biogen-esis between wild-type and the two variant proteins, G185S and R147W SCAD.At physiological temperature, 37C, the relative SCAD activities in extracts from transfected COS-7 cells for G185S and R147W were 136 and 45%, respectively.At 41C the relative activities were, respect-ively, 58 and 13% for G185S and R147W SCAD, while they were 183 and 85% at 26C [49].These results support the notion that the variant proteins in their biogenesis at physiological temperatures may achieve sufficient activity to sustain normal fatty acid oxidation but that both variant proteins at higher temperatures, as experienced during fevers, may result in insufficient amounts and activity and thus the development of SCAD deficiency.This conclusion
is supported by further in vitro studies, where the biogenesis
of the two variant SCAD enzyme proteins was shown to be delayed and compromised, especially at higher temperatures [68].Together with the fact that the stability of the active G185S SCAD protein is decreased compared to that of wild-type SCAD [66], these studies further contribute to the notion that especially the G185S SCAD protein may be disease-associated
Whether other perturbations of the cellular homeostasis
in addition to high temperatures, such as alterations of redox state, ATP depletion and pH changes, may show differential effects on the biogenesis and/or stability of the two variant proteins are pressing questions.If this is the case, a number of conditions encountered in other metabolic and endocrine diseases may result in functional SCAD deficiency and add to the clinical features of these other diseases
With the present knowledge levels we still do not know how many of the 14% of the general population are at risk
of developing – perhaps in a mild and unrecognized form – SCAD deficiency.We only know that a small fraction develop clinically relevant disease [50], and we know that this is possible by a combination of high fatty acid oxidation activity and high temperature, which may result in accu-mulation of cytotoxic butyric acid
The challenge is to define further the cellular conditions under which the deficiency occurs and to delineate whether there exist inter-individual genetic differences in susceptibi-lity to develop clinical disease
Generalization and future aspects
Many elements of the above discussion can be generalized
to defects in other fatty acid oxidation enzymes and to variant proteins present in other genetic diseases.To our knowledge only a few other metabolic diseases have been investigated in the same detail as MCAD and SCAD deficiencies, and with PKU as a prominent example [61,62]
Fig 3 Schematic overview of monomeric SCADwith the positions of
the 12 published mutations The figure is based on the coordinates for
rat SCAD (PDB acc.no.1JQI).SCAD protein is shown as a solid
ribbon and the Ca atoms at variant residues are represented as balls.
FAD and butyryl-CoA are shown as sticks.
Trang 10In the future it may be important, in special cases, to do
similar experiments but the real challenge for disease-related
research, in relation to the biological significance of varying
effects of disease-causing and disease associated
susceptibi-lity variations, is to define the cellular conditions and
perturbations as well as genetic factors that may modulate
their effect.This challenge is still unappreciated but with the
identification of single nucleotide polymorphisms (SNPs),
which may associate to clinical features in complex diseases,
there will be a need for biochemical and cellular approaches
to delineate the functional significance of putative
disease-associated SNPs
Furthermore, another challenge, which has not been
addressed in this review, is to describe and characterize
sequence variations that influence splicing by modulation of
the binding of splicing factors [19].Doing this will also – as
has been seen for the mis-sense sequence variations – open
avenues to new questions about the plasticity of the cellular
response to gene variations, and give new insights in
biological mechanisms, which may be the target for
intervention by conventional treatments or future gene
therapeutic treatment
Acknowledgements
The molecular genetic analyses of the VLCAD, MCAD and SCAD
genes have been performed by medical laboratory technologists
Vibeke Winter, Inga Knudsen, Margrethe Kjeldsen and Lisbeth
Schrøder.The investigations of our own group referred to in this
review have been supported by The Danish Medical Research
Council; Danish Human Genome Centre; Karen Elise Jensen
Foundation; Aarhus County Research Initiative; Institute of
Experi-mental Clinical Research, Aarhus University; Institute of Human
Genetics, Aarhus University and Aarhus University Hospital.We
thank colleagues from all over the world for providing genetic and
cell material for the studies and certain of them for inspiring
discussions concerning genotype–phenotype interactions in especially
the acyl-CoA dehydrogenase deficiencies.
References
1 Gregersen, N., Andresen, B.S., Corydon, M.J., Corydon, T.J.,
Olsen, R.K., Bolund, L & Bross, P (2001) Mutation analysis in
mitochondrial fatty acid oxidation defects exemplified by
acyl-CoA dehydrogenase deficiencies, with special focus on genotype–
phenotype relationship Hum Mutat 18, 169–189.
2.Vockley, J.& Whiteman, D A.(2002) Defects of mitochondrial
beta-oxidation: a growing group of disorders Neuromuscul.
Disord 12, 235–246.
3 DiMauro, S.& DiMauro, P.M.(1973) Muscle carnitine
palmi-tyltransferase deficiency and myoglobinuria Science 182, 929–931.
4 Karpati, G , Carpenter, S , Engel, A G , Watters, G , Allen, J ,
Rothman, S., Klassen, G & Mamer, O.A (1975) The syndrome of
systemic carnitine deficiency.Clinical, morphologic, biochemical,
and pathophysiologic features Neurology 25, 16–24.
5.Gregersen, N , Lauritzen, R.& Rasmussen, K.(1976)
Sub-erylglycine excretion in the urine from a patient with dicarboxylic
aciduria Clin Chim Acta 70, 417–425.
6 Krawczak, M , Ball, E V , Fenton, I , Stenson, P D , Abeysinghe,
S., Thomas, N & Cooper, D.N (2000) Human gene mutation
database – a biomedical information and research resource Hum.
Mutat 15, 45–51.
7 Lander, E.S., Linton, L.M., Birren, B., Nusbaum, C., Zody, M.C.,
Baldwin, J , Devon, K , Dewar, K , Doyle, M , FitzHugh, W ,
et al.(2001) Initial sequencing and analysis of the human genome Nature 409, 860–921.
8 Venter, J C , Adams, M D , Myers, E W , Li, P W , Mural, R J , Sutton, G G , Smith, H O , Yandell, M , Evans, C A , Holt, R A ,
et al.(2001) The sequence of the human genome.Science 291, 1304–1351.
9.Culbertson, M R.(1999) RNA surveillance.Unforeseen conse-quences for gene expression, inherited genetic disorders and cancer Trends Genet 15, 74–80.
10 Frischmeyer, P.A & Dietz, H.C (1999) Nonsense-mediated mRNA decay in health and disease Hum Mol Genet 8, 1893– 1900.
11.Cooper, D N.& Krawczak, M.(1993) Human Gene Mutation Bios Scientific Publishers Ltd, Oxford, UK.
12 Nissim-Rafinia, M.& Kerem, B.(2002) Splicing regulation as a potential genetic modifier Trends Genet 18, 123–127.
13 Bross, P., Corydon, T.J., Andresen, B.S., Jørgensen, M.M., Bolund, L.& Gregersen, N.(1999) Protein misfolding and degradation in genetic disease Hum Mutat 14, 186–198.
14 Wang, Z.& Moult, J.(2001) SNPs, protein structure, and disease Hum Mutat 17, 263–270.
15 Sunyaev, S., Ramensky, V., Koch, I., Lathe, W III, Kondrashov, A.S & Bork, P (2001) Prediction of deleterious human alleles Hum Mol Genet 10, 591–597.
16 Terp, B N , Cooper, D N , Christensen, I T , Jorgensen, F S , Bross, P., Gregersen, N & Krawczak, M (2002) Assessing the relative importance of the biophysical properties of amino acid substitutions associated with human genetic disease Hum Mutat.
20, 98–109.
17 Fersht, A.R.& Daggett, V.(2002) Protein folding and unfolding
at atomic resolution Cell 108, 573–582.
18 Gregersen, N., Bross, P., Andresen, B.S., Pedersen, C.B., Cory-don, T J.& Bolund, L.(2001) The role of chaperone-assisted folding and quality control in inborn errors of metabolism: protein folding disorders J Inherit Metab Dis 24, 189–212.
19 Cartegni, L., Chew, S.L & Krainer, A.R (2002) Listening to silence and understanding nonsense: exonic mutations that affect splicing Nat Rev Genet 3, 285–298.
20 Vockley, J , Rogan, P K , Anderson, B D , Willard, J , Seelan, R.S., Smith, D.I & Liu, W (2000) Exon skipping in IVD RNA processing in isovaleric acidemia caused by point mutations in the coding region of the IVD gene Am J Hum Genet 66, 356–367.
21 Andresen, B S , Christensen, E , Corydon, T J , Bross, P , Pilg-aard, B , Wanders, R J , Ruiter, J P , Simonsen, H , Winter, V , Knudsen, I , Schroeder, L D , Gregersen, N & Skovby, F (2000) Isolated 2-methylbutyrylglycinuria caused by short/branched-chain acyl-CoA dehydrogenase deficiency: identification of a new enzyme defect, resolution of its molecular basis, and evidence for distinct acyl-CoA dehydrogenases in isoleucine and valine meta-bolism Am J Hum Genet 67, 1095–1103.
22 Gregersen, N., Andresen, B.S & Bross, P (2000) Prevalent mutations in fatty acid oxidation disorders: diagnostic considera-tions Eur J Pediatr 159, S213–S218.
23 Andresen, B.S., Olpin, S., Poorthuis, B.J., Scholte, H.R., Vianey-Saban, C., Wanders, R., Ijlst, L., Morris, A., Pourfarzam, M., Bartlett, K., Baumgartner, E.R., deKlerk, J.B., Schrøder, L.D., Corydon, T.J., Lund, H., Winter, V., Bross, P., Bolund, L & Gregersen, N.(1999) Clear correlation of genotype with disease phenotype in very- long-chain acyl-CoA dehydrogenase defi-ciency Am J Hum Genet 64, 479–494.
24 Smelt, A.H., Poorthuis, B.J., Onkenhout, W., Scholte, H.R., Andresen, B S , van Duinen, S G , Gregersen, N & Wintzen, A R (1998) Very long chain acyl-coenzyme A dehydrogenase deficiency with adult onset Ann Neurol 43, 540–544.
25 Scholte, H R , Van Coster, R N , de Jonge, P C , Poorthuis, B J , Jeneson, J.A., Andresen, B.S., Gregersen, N., de Klerk, J.B &