Mitochondrial fatty acid oxidation defects (FAOD) are rare inherited diseases with clinical presentation ranging from severe outcome to asymptomatic status. Diagnosis of FAOD requires specific investigations including biochemical and genetic studies. The most common FAOD is represented by a deficiency in the activity of the medium-chain acyl-CoA dehydrogenase (MCAD, EC 3.1.2.20) protein, which is coded by the ACADM gene (chromosome 1p31) 1 2 3 4 . First reported in 1976 5 and biochemically characterized in 1982 6 , MCAD deficiency (MCADD) has since been extensively described. Point mutations in the ACADM gene were reported early 7 8 and many mutations have been documented subsequently, in part as a result of the large development of screening strategies 9 10 11 12 13 14 15 16 17 18 . As a general rule, these mutations lead to either truncated/absent protein synthesis, or replacement of one amino acid by another with a consequent drop in the activity of the mutant protein. The most common ACADM mutation is c.985A>G, which causes the replacement of a lysine by a glutamate at position 329 of the precursor protein (position 304 of the mature protein) 3 17 , being for this reason also referred to as the K329E (or K304E) protein variant. Eighty percent of symptomatic MCADD patients of European origin are homozygous for this mutation, and a further 18% of symptomatic MCADD are associated with the c.985A>G mutation in one allele and a rare mutation in the other allele 1 3 19 20 . Mutations other than c.985A>G may predominate in non-European populations 21 22 .

MCADD is inherited as an autosomal recessive trait 2 3 23 24 . Like other genetic β-oxidation defects, MCADD may cause in early childhood life-threatening events like hypoketotic/hypoglycaemic comas occurring during the course of catabolic states which disrupt the body energetic production/expense balance (infectious episodes, for instance) 2 3 23 24 . In fact, MCADD may present phenotypically as either a classical (severe) or a mild form. In mild MCADD, the residual MCAD activity is intermediate between that observed in classical MCADD (usually less than a few percent) and their heterozygous carriers (around 50%) 25 , the classical form of the disease resulting mainly from homozygosity for the c.985A>G mutation as mentioned above. The residual activity of the ACADM gene product and hence its impact on cell metabolism and patient clinical presentation depends on a series of factors including protein misfolding, MCAD tetramer stability, and thermosensitivity of the mutant protein 3 26 27 28 along with contribution from the overlapping activities of other mitochondrial acyl-CoA dehydrogenases and efficacy of L-carnitine-based detoxification. In turn, a toxic "gain-of-function" pathogenesis has been also proposed for some MCAD variants 3 .

Febrile episodes may negatively impact the disease outcome, especially when the mutant protein is thermosensitive, which is the case for the most common K304E protein variant 28 . The patient has then to face concomitantly not only the increased energy demand but also the sudden lowering of the fatty acid oxidation-derived energy production secondary to a drop in residual MCAD activity. This might explain why one quarter of the patients with classical MCADD do not survive their first life-threatening episode and why this disease is present in the etiology list for sudden infant death syndrome 29 30 31 .

Like patients with classical MCADD, patients with mild MCADD may develop metabolic decompensation in their early childhood. They may have to cope throughout life with neurological sequels or other irreversible lesions caused by early childhood metabolic decompensation. These sequels do not necessarily reflect the MCADD genotype but rather are the consequences of the life threatening episode before medical intervention. Recent work, however, suggests that the link between susceptibility to develop metabolic decompensation and MCADD patient genotype is not straightforward and that this susceptibility is largely dependent on environmental triggers such as starvation stress, infection and/or fever 3 18 . We report here on a patient with mild MCADD having undergone a severe metabolic decompensation and subsequent clinical sequels. Genetic studies revealed a previously unreported mutation (c.145C<G) in one of the ACADM alleles combined with the common c.985A>G mutation in the other allele. Biochemical investigations of the patient are reported along with clinical data from the same time and an assessment of the impact of the new mutant protein variant is discussed.

Our patient undoubtedly suffered from a deficiency of MCAD, presenting with a mild phenotype of the disease as ascertained by relatively high residual activity from lymphocytes and fibroblasts and by a low urinary excretion of hexanoylglycine and suberylgycine during metabolic decompensation. At the time of this urine analysis, acylglycines were determined by GC/MS (with a detection limit of 2 mmol/mol creatine) and could not be discriminated from control values. Better analytical detections of urinary acylglycines are obtained by GC/MS-MS or GC-MS stable-isotope dilution analysis (SID) of methyl esters, with detection limits less than 0.7-1 mmol/mol creatinine, allowing for upper limit of controls a clear cut-off value close to 1.1 mmol/mol 25 41 .

Biochemically, our patient resembled the four mild MCADD patients described by Zschocke et al. 25 who had a residual MCAD activity (10-20%) of the same order of magnitude as in our patient. In fact, these four cases and our patient all show intermediate (between controls and severe MCADD patients) values for both enzyme activities and acylcarnitine levels and ratios. The residual enzyme activity observed in classical MCADD, caused by homozygosity for c.985A>G, is usually below 1%. Two of the four patients previously described by Zschocke et al. 25 were homozygous for ACADM gene mutations (c.734C>T or c.799G>A [exon 9]) distinct from the c.985A>G mutation; the other two patients were compound heterozygous for the c.985A>G mutation in one allele and the c.199T>C (exon 3) mutation in the other ACADM allele. Genotypically, our patient resembled the latter two patients, with the c.985A>G mutation (exon 11) in one allele and a mutation in exon 3 (namely, c.145C>G mutation) in the other allele. This mutation has never been observed in controls. To our knowledge, this is the first report for the c.145 C>G mutation in the ACADM gene.

As a general rule, amino acid changes introduced in a mature protein affect its conformation and this may impact its final stability and function/activity. This has been well documented for the MCAD mutant protein (K304E) resulting from the c.985A>G mutation. The MCAD protein requires assembly of the monomers in a tetramer, which represents the active functional form of MCAD. In the case of monomers resulting from the wild-type ACADM gene, the MCAD tetramer is stable. In the case of K304E monomers, the tetramer dissociates abnormally rapidly 26 , giving rise to a reduced steady-state concentration of the tetrameric form. Another contributing factor is impairment of intramitochondrial processes including altered monomer folding or management 27 . Several studies indicate that protein misfolding is the key event in the loss of MCAD activity from many protein variants, and that overexpressing chaperonins may rescue monomers from aggregation and enhance residual MCAD activity (see 3 and references therein). As a result of protein misfolding, accelerated temperature-driven unfolding may affect MCAD protein variants. This abnormal MCAD loss of function and thermosensitivity first documented for the K304E variant, has been recently extended to a number of MCAD variants 18 42 , questioning the view that some of the missense mutations detected by newborn screening are less life-threatening than those clinically proven to be disease causing. In this respect, the relatively high residual activity found in our mild MCADD patient has not prevented the occurrence of a severe metabolic decompensation which might have been fatal in the absence of adequate medical management. A patient with the R28C mutant protein who had a mutation (c.157C>T), also less severe than the c.985A>G mutation, died unexpectedly in early childhood 43 .

Our mild MCADD patient combines MCAD K304E and Q24E variant monomers. Conformational characteristics of the MCAD K304E variant have been repeatedly described in the literature. The amino acid change is located in the MCAD C terminal α domain which is involved in helix-helix interactions taking place in tetramer assembly (see 18 and references therein). As a result, tetramer assembly is compromised: unemployed monomers aggregate and are subjected to degradation. Experimental prevention of monomer aggregation by overexpressed chaperonins has been shown to result in increased tetramer assembly along with higher residual activity (see 3 and references therein).

The Q24E variant has not been previously described. The amino acid change here occurs in the MCAD N-terminal α domain. Some of the protein variants with mutations in this protein region may present with normal or close to normal enzyme activity 3 18 . Mutations of this portion of the MCAD protein are also shown to induce moderate alterations in tetramer assembly and to disrupt thermal stability. This is the case for the Y42 H mutant protein for which temperature-driven loss of activity is greater than normal though lower than that of common K304E mutant protein 42 . De novo synthesis of acylcarnitines monitored on whole blood samples from our patient was consistent with MCADD. The major acylcarnitine produced from deuterated palmitate was octanoylcarnitine, other medium-chain (hexanoyl- and decanoyl-) carnitine esters being also generated in excess. The C8 to C4 acylcarnitine production rate ratio was 54-fold that of controls. As a whole, profile and ratio of acyl-carnitine production rates in our mild MCADD patient resemble those found in patients with classical MCADD, indicating clearly functional repercussion of the enzyme defect on the disposal of acyl-CoA. The carnitine added to incubation media appeared to mobilize acyl-CoAs unmetabolised by β-oxidation and, by comparing acylcarnitines production rates to control values, revealed the level(s) at which β-oxidation was defective. The acylcarnitine production rates and increased C8/C4 acylcarnitine de novo synthesis ratio of our mild MCADD patient are clearly different from those of the adult carrier of the c.985A>G mutation (figure 2).

The last but not least issue concerns the carnitine status of our patient at the time of metabolic decompensation. The patient was then not receiving L-carnitine therapy and her systemic carnitine levels were below normal (Table 1). This chronic hypocarnitinaemia was concomitant with a low urinary excretion of hexanoylglycine and suberylglycine. Interestingly, no elevation of urinary suberylglycine was reported in the urine from one patient with hypocarnitinaemia in the description of four MCADD asymptomatic siblings (compound heterozygotes for the common c.985A>G and the c.842G>C ACADM gene mutations) provided by Albers and coworkers 44 . Zschocke and coworkers on the other hand discussed elevated acylcarnitine and undetectable 5-hydroxyhexanoic acid and suberylglycine in their four mild MCADD patients 25 . Patients with genotypes representing mild MCADD (high residual enzyme activity and low urinary levels of glycine conjugates) should get the same advice as patients with classical MCADD regarding risks associated with catabolic states and carnitine deficiency.

The authors declare that they have no competing interests.

AFD and MF contributed equally to this work. AFD, MF, AMP, GB and JV designed and performed measurements of de novo acylcarnitine syntheses by whole blood samples. BSA and NG contributed to genetic studies. MB, CVS and RJAW performed the biochemical and enzyme measurements. SNG, DD and KMM contributed to the clinical follow-up of studied patients. DR determined organic acids in body fluids and DSM performed acylcarnitine analysis on dried blood spots. Each author read the text and JV coordinated the writing of the manuscript. All authors read and approved the final manuscript.

Written informed consent was obtained by mother of the patient for publication of this case report.