Main

Elevated urinary excretion of EMA can be observed in a number of inherited deficiencies of mitochondrial energy metabolism. It does not reflect a distinct disease entity, but rather is an unspecific biochemical marker for metabolic disturbances of a variety of etiologies affecting the activity of SCAD (EC 1.3.99.2) in the β-oxidation of fatty acids. Primary or secondary deficiency of SCAD results in intracellular accumulation of butyryl-CoA, which can be converted to EMA by mitochondrial propionyl-CoA carboxylase and subsequent hydrolysis of ethylmalonyl-CoA(1).

Secondary deficiencies of SCAD are the best known. It is observed in multiple acyl-CoA dehydrogenation defects, where SCAD activity is affected by deficiencies of either electron transfer flavoprotein(2), electron transfer flavoproteinubiquinone oxidoreductase(3, 4), or intramitochondrial flavin adenine dinucleotide(5, 6). Ethylmalonic aciduria, probably due to secondary SCAD deficiency, has also been observed in patients with predominantly neuromuscular symptoms and respiratory chain deficiencies(7–9) in addition to patients with ethylmalonic encephalopathy syndrome(10, 11).

Primary deficiency of SCAD has been documented at the molecular level in only a single patient(12, 13). Other patients have been described either with low SCAD catalytic activity in cultured skin fibroblasts(14–16) or with low activity in skeletal muscle but normal activity in fibroblasts(17, 18).

In addition to the above-mentioned patients with ethylmalonic aciduria, there are a large number of still uncharacterized patients with persistent or sporadic elevated urinary EMA. To our knowledge, such patients have been recognized in many metabolic centers worldwide, leading to the definition of a broad spectrum of clinical, biochemical, and molecular findings in ethylmalonic aciduria. Taking into consideration the difficulties of collecting sufficient amounts of critical specimens, muscle and liver in particular, and the need to perform many different assays which seldom are accessible in a single laboratory, only a small fraction of these patients have been investigated to the full extent with available methods. Therefore, we believe that association studies with candidate genetic markers could provide an effective first approach toward a stratification of genetic subpopulations among these patients. Subsequently the various subgroups would be candidates for more specific molecular and biochemical investigations.

Because SCAD activity is suspected to be implicated in the etiology of many cases of ethylmalonic aciduria, we postulated that SCAD gene polymorphisms are candidate markers to be tested in such studies. We have recently characterized a polymorphism of the SCAD gene, with an adenine at position 625 of the SCAD cDNA instead of a guanine, replacing a glycine at position 209 of the precursor SCAD protein with a serine (G209S)(19, 20). In 90 randomly chosen individuals we found that the variant allele(A625) was present in 20% of all SCAD alleles, corresponding to a population distribution of 4% homozygous A625/A625, 31% heterozygous A625/G625 carriers, and 65% homozygous for the wild-type allele(20).

To investigate whether or not this variant allele (A625) is implicated in ethylmalonic aciduria, we have determined the allelic nucleotide configuration at position 625 in the SCAD gene in 135 patients from four European countries and the United States chosen because of elevated ethylmalonic aciduria. We also studied the effect of the variant allele on the biogenesis of SCAD protein, its catalytic activity, and enzyme stability, after overexpression of the variant enzyme in both eukaryotic (COS) and prokaryotic(Escherichia coli) cells.

METHODS

Subjects

The EMA group comprised 135 selected patients from Germany, Denmark, the Czech Republic, Spain, and the United States who had a history of elevated EMA excretion (18-1185 mmol/mol of creatinine). EMA excretion in these patients was measured either by standard gas chromatographic/mass spectrometric methods(21) or by selected ion monitoring/stable isotope dilution methods(4, 22). The EMA values in urine from patients were obtained prospectively from specimens analyzed for organic acids for diagnostic purposes. In about half of the cases more than one urine sample was analyzed, and the highest determined EMA value (peak value) was used for the histogram in Figure 1. Besides ethylmalonic aciduria these patients exhibited a variety of clinical symptoms: neuromuscular symptoms with hypotonia, convulsions, and psychomotor retardation were prevalent, but classical symptoms pointing to a fatty acid oxidation defect, such as hypoglycemia and lethargy, were also seen. No obvious association between the EMA value and the described clinical symptoms was seen. Control genomic DNA and dried blood spots were from 90 healthy Danish individuals and 197 newborn babies (102 German, 95 Spanish), respectively.

Figure 1
figure 1

Correlation between EMA excretion (peak values) and SCAD A625 polymorphism. Patients with EMA excretion in the range from 18 mmol/mol of creatinine to 107 were arranged into nine subgroups (122 patients). Patients with excretion above 107 were arranged in one group (13 patients). In this group patients with A625 in homozygous and heterozygous form excreted EMA in the range of 120 to 136 mmol/mol of creatinine and 114-1185 mmol/mol of creatinine, respectively. For illustration, the A625 allele distribution in the control individuals was included.

Full size image

Measurements

A625 SCAD gene polymorphism in genomic DNA from EMA patients and controls. Genomic DNA was extracted from blood lymphocytes, cultured fibroblasts, and dried blood spots according to established isolation methods(23–25). The nucleotide configuration at position 625 in the SCAD gene was determined by a SSCP assay as previously described(20). The specificity of the SSCP assay was verified by sequencing the PCR products from DNA of one homozygous A625/A625 and 10 heterozygous A625/G625 patients(20). In addition, the configuration at position 625 was verified in 40 control individuals with a restriction enzyme cleavage assay for A625 (see below).

DNA sequencing of SCAD gene fragments with the 625 A/G polymorphism. PCR was carried out in an automated thermal cycler(Perkin-Elmer, Norwalk, CT, DNA Thermal Cycler 480) with 1 μg of genomic DNA as template in a 100-μL reaction volume with 150 ng of each primer(pri625s and pri745as, Table 1), 2 U of recombinantTaq polymerase (Perkin-Elmer), 1 × PCR buffer (Perkin-Elmer), and the following program for 30 cycles: 95, 58, and 74°C for 1, 1, and 4 min, respectively. The PCR product (one of the primers is biotin-labeled) was purified using magnetic beads (DYNABEADS M-280, Dynal AS, Oslo, Norway), as described by the manufacturer. Sequenase sequencing with dye terminators(Applied Biosystems Inc., Foster City, CA; Perkin-Elmer, Norwalk, CT) was performed on a model 373A automated DNA sequencer from Applied Biosystems, exactly as described by the manufacturer.

Table 1 Primers used for PCR (restriction sites are underlined)
Full size table

Restriction enzyme cleavage assay for A/G 625 SCAD gene polymorphism. The assay is based on two nested PCR procedures.

  • PCR 1 (141-bp). PCR amplification with 1 μg of genomic DNA was performed as described above for 35 cycles with 75 ng of each primer(priSIP-s and pri745as, Table 1) and the following program: 95, 61, and 74°C for 1, 1, and 4 min, respectively. The band of expected length was purified from the agarose gel (GENECLEAN, Bio 101 Inc., La Jolla, CA).

  • PCR 2 (47-bp). The purified PCR product (141 bp) was used as template in a second PCR with priSIP-s as sense primer. A BstNI restriction site (CCAGG) is produced by PCR when the wild type is copied, but not when the mutant sequence (CCAGA) is amplified. The antisense primer, pri625as (Table 1), contains a BstNI restriction site, which functions as an internal control of the enzyme cleavage efficiency. PCR was performed as described above, except that the polymerization time was 2 min.

Solid phase direct sequencing of SCAD cDNA from patients with EMA excretion. Sequencing of PCR amplified patient cDNA with SCAD-specific primers was performed according to the method mentioned above. PCR was carried out in an automated thermal cycler (Perkin-Elmer, Geneamp PCR system 9600) with cDNA as template in a 50-μL reaction volume with two primer pairs(pri-29s/pri745as and pri645s/pri1285as, Table 1), 2 U of recombinant Taq polymerase (Perkin-Elmer), 5 μL of 10 × DMSO buffer (670 mM Tris-HCl, pH 8.8, 166 mM ammonium sulfate, 100 mM mercaptoethanol), 4 μL of DMSO, 1 μL of MgCl2 (50 mM) and the following program for 35 cycles: 94, 64, and 72°C for 30, 15, and 60 s, respectively. The Taq polymerase was added after 5 min of initial denaturation at 94°C. The two PCR fragments have overlapping sequences and cover the entire full-length coding region of SCAD cDNA. cDNA was synthesized from 1 μg of fibroblast total RNA, using a cDNA synthesis kit (Clontech Laboratories, Inc., Palo Alto, CA) and an oligo(dT) primer.

Allele-specific analysis of SCAD mRNA from cultured fibroblasts. To measure the relative amounts of SCAD mRNA from the A625 and G625 SCAD alleles, we performed a PCR-based assay on SCAD cDNA with pri625s and pri625as (Table 1) surrounding position 625. cDNA was synthesized from 8 μg of fibroblast total RNA using a cDNA synthesis kit(Invitrogen, San Diego, CA) and an oligo(dT) primer. The sense primer, pri625s(Table 1), covers nucleotides 603 to 624, with C instead of A at position 620, and creates a StyI site (CCAAGG) in the amplification product from the normal alleles but not in the product from alleles harboring the A625 variant (CCAAGA). The PCR was carried out as described above using SCAD cDNA as template with 75 ng of each primer and the following program for 35 cycles: 94, 58, and 74°C for 1, 1, and 2 min, respectively. After PCR, DNA was digested with StyI restriction endonuclease and analyzed by 16% PAGE.

Construction of expression vectors containing full-length SCAD cDNA and “signal peptide”-deficient SCAD cDNA. Full-length cDNA was amplified with primer pri-30s (Table 1) and primer pri1300as. PCR amplification was performed for 40 cycles with first strand SCAD cDNA (produced as described above) and 300 ng of each primer and the following program: 94, 59, and 74°C for 1, 2, and 6 min, respectively. In the last cycle the polymeration was extended to 10 min at 74°C. Two additional units of Taq polymerase were added after the 20th and 35th cycles. The PCR product was gel-purified as described above and ligated into pCRII cloning vector (Invitrogen) as recommended by the supplier. The recombinant plasmid was designated pCpre (pCpreWt or pCpre625).

Signal peptide-deficient SCAD cDNA was synthesized from plasmid pCpreWt(pCRII cloning vector harboring full-length wild-type cDNA; see above) using the primers: pri60s which introduces a NdeI restriction site and primer pri1300as. Fifty nanograms of the plasmid were mixed with 300 ng of each primer, and the following program was run for 20 cycles: 94, 59, and 74°C for 1, 2, and 6 min, respectively. The PCR product was gel-purified and ligated into pCRII cloning vector as described above. The recombinant plasmid containing the signal peptide-deficient SCAD cDNA was designated pCmatWt. The signal peptide-deficient SCAD cDNA fragment from this vector was subcloned into the pBMCK2-(26), replacing the fragment encoding mature wild-type human MCAD. This vector is under control of the lac promoter and was designated pBmatWt. Another vector(pBmat625) containing signal peptide-deficient A625 variant cDNA was constructed in a similar way.

For expression in the eukaryotic expression system the full-length constructs were used (pCpreWt and pCpre625). A 1309-bp long SCAD cDNA fragment was cut out from these plasmids with SalI and ligated into the unique XhoI cloning site of the Epstein-Barr virus-based expression vector pTPS(27). The expression vectors were designated pTPSWt and pTPS625, respectively. The pTPS plasmids were transfected into COS cells by a calcium phosphate coprecipitation method as described elsewhere(27).

For expression in a prokaryotic system, the two plasmids (pBmatWt and pBmat625) with the signal peptide-deficient cDNA were used together with plasmid pGroESL described by Goloubinoff et al.(28). pGroESL carries the GroEl and GroES genes under control of the lac promoter and carries a chloramphenicol resistance marker. To check the constructs, the plasmids were sequenced according to the method mentioned above except that Taq polymerase (cycle sequencing) was used instead of T7 DNA polymerase.

Expression of recombinant SCAD. Transformed E. coli cells were grown at 31°C in dYT medium (16 g of Bactotryptone, 10 g of Bacto-yeast extract, 5 g of NaCl per liter)(29) supplemented with ampicillin and chloramphenicol to a cell density of approximately 1 × 109 cells/mL, and plasmid expression was induced with 1 mM isopropyl-β-D-thiogalactopyranoside for 3 h.

Expression in eucaryotic COS-7 cells was performed as previously described(27). Total cellular RNA was isolated from COS cells according to Chomczynski and Sacchi(30), and Northern blot analysis was performed as described by Fourney et al.(31).

Protein extracts for Western blot analysis were obtained from COS cells harboring expression vectors encoding wildtype and A625 variant (G209S) SCAD. The cells were harvested and disrupted as described previously(27). Electrophoretic analysis of SCAD was performed by SDS-PAGE. The proteins were blotted onto Immobilon membranes (Millipore) by semidry electroblotting (Pharmacia) as described by Blake et al.(32). Rabbit anti-human SCAD antibodies (see below) were used and visualized with alkaline phosphatase conjugated goat anti-rabbit secondary antibodies (Dako Corp., Carpinteria, CA) and chemiluminescence detection using AMPPD (Tropix Inc., Bedford, MA). SCAD enzyme activity in extracts from transformed E. coli or transfected COS cells were determined with the ferricenium ion assay using n-butyryl-CoA as substrate(33).

Rabbit anti-human SCAD antibodies. A 717-bp PCR fragment covering position 584 to 1300 in the cDNA (215 amino acid carboxyl-terminal end of SCAD) was cloned into the pEX2 vector(34). Expression in E. coli strain POP 2136 results in a fusion protein(≈ 141 kD) of cro-β-galactosidase and the SCAD carboxyl terminus. Purification of the fusion protein was performed by preparative gel electrophoresis (ProSieve, FMC Corp. BioProducts, Rockland, ME), and a polyclonal antibody against this protein was raised in rabbits. The antiserum was adsorbed against a Sepharose column coated with a soluble extract ofE. coli protein and used for immunoblotting.

RESULTS

Analysis for A625 in genomic DNA from patients. The SSCP method used to determine the A/G polymorphism at cDNA position 625 in the SCAD gene proved to be reliable and specific, suitable for analysis of large series of samples. The specificity of the assay was checked by means of sequencing the SSCP fragments from 11 individuals(20) and by a restriction enzyme cleavage assay for the A/G 625 polymorphism in 40 individuals (results not shown). In all cases we found complete agreement between the three methods. Figure 2 shows examples of the analyses where the two alleles are clearly separated.

Figure 2
figure 2

Diagnosis of the A625 SCAD gene variant. Autoradiogram of a polyacrylamide gel showing SSCP analysis of the 141-bp genomic fragment of the SCAD gene comprising the A/G 625 polymorphism. Lanes 1-3, isolated genomic DNA from control individual A/G heterozygous, control individual A/A homozygous and control individual G/G homozygous; lanes 4-11, DNA from blood spots of patients. Analytical procedure; see“Methods,” and Kristensen et al.(20).

Full size image

The SSCP assay was used in the prospective study, with blood spots obtained from 135 patients with urinary EMA excretion above 18 mmol/mol creatinine(control <18 mmol/mol of creatinine)(16, 35), referred to the collaborative metabolic centers in Germany, Denmark, the Czech Republic, Spain, and the United States for investigation of inborn errors of metabolism. The criteria for inclusion was solely an elevated urinary excretion of EMA. The A625 allele frequency in the EMA group of patients compared with controls is shown in Table 2.

Table 2 Nucleotide configuration at position 625 in the SCAD cDNA in 135 selected patients with ethylmalonic aciduria (EMA group)
Full size table

The expected frequencies in the group of controls were calculated from the combined estimate of the allele frequencies and Hardy-Weinberg law. The overrepresentation of A625 alleles in the EMA group (202 A625 alleles out of 270) compared with the control distribution (140 A625 alleles out of 574) is striking and highly significant with p < 0.001 (twosided). Interestingly, not only the A625 homozygous individuals are overrepresented in the EMA group compared with G625/G625 individuals (p < 0.001), but also the heterozygous A625/G625 individuals are overrepresented(p < 0.001).

Stratification of the A/G 625 genotype distribution according to the EMA excretion in the range 18-107 mmol/mol of creatinine is shown in Figure 1. The overrepresentation of the A allele is seen at all levels of EMA excretion, but it is noteworthy that, whereas the number of patients with the A/G and G/G configuration decrease gradually to a few individuals with EMA excretion above 50-70 mmol/mol of creatinine, the number of patients with the A/A configuration exhibit a second peak in this interval (Fig. 1). This distribution is also observed when the EMA excretion intervals are chosen differently from those shown in Figure 1.

This distribution indicates an effect of the A625 alleles in addition to other genetic and/or environmental factors, suggesting that the A625 is a susceptibility allele. In the high range (>70 mmol/mol of creatinine) of EMA excretion the overrepresentation of A/A homozygous is less pronounced than in the lower range indicating that the confounding effect from other factors are decisive for the EMA excretion (Fig. 1). Two of these patients, one from Spain and one from the United States(11, 36), with the A625 polymorphism in heterozygous form and repeatedly high EMA excretion (>100 mmol/mol of creatinine) suffer from the newly defined disease entity, ethylmalonic encephalopathy syndrome(10, 11). The etiology of this disease has not yet been elucidated.

The first approach to define the molecular mechanism of increased urinary EMA excretion was to investigate the molecular properties of the A625 allele, both at the mRNA and protein level. To ensure that there exist patients where A625 is the only substitution in the coding region of the SCAD gene, we sequenced the entire coding region in three A625/A625 homozygous patients with EMA excretions in the low and middle range (26, 35, and 65 mmol/mol of creatinine). cDNA sequencing of fibroblasts from the three patients revealed only the A625 in addition to three silent polymorphisms (T321C, C990T, and G1260C), which have been identified in the A625 allele previously (unpublished results). The three silent polymorphisms were present in homozygous form in all three patients.

Analysis of the relative expression of SCAD mRNA from A625 and G625 alleles. Because the A625 polymorphism is located as the first nucleotide of an exon(20) it might affect the splicing of SCAD mRNA. To measure the relative amount of SCAD mRNA from A625 and G625 alleles in heterozygous individuals we performed allele specific analysis of SCAD mRNA in three normal A625 carrier subjects.

PAGE analysis of the PCR-amplified cDNA from these individuals revealed a specific band (49 bp) with an intensity comparable to the bands produced from normal individuals homozygous for G625 (results not shown). The 49-bp fragment was digested with StyI and analyzed by 16% PAGE. The analysis revealed in all three heterozygous individuals three bands [49-bp band (A625), 31-bp band (G625), and 18-bp band (G625); Fig. 3]. The intensity of the 49-bp band is stronger than the cleaved bands due to the length dependent ethidium bromide staining. To demonstrate this discrepancy a mix of equal amounts of StyI cleaved G/G control and uncleaved G/G control was loaded in lane 4, Figure 3. The results show that the amounts of mRNA from the variant allele and the wild-type allele are comparable, suggesting that the A625 allele is spliced with a similar efficiency as the G625 allele.

Figure 3
figure 3

PCR-based assay for the A625 variant in cDNA. cDNA from three normal controls, heterozygous for A625, was used as template. PCR product is 49 bp. StyI digestion of the product with normal sequence(G625) results in two fragments of 31 and 18 bp, respectively. A625-bearing fragments are not cleaved. The intensity of the 49-bp band is stronger than the cleaved bands due to the length dependent ethidium bromide staining.Lanes 1-3, heterozygous A/G control individuals cleaved withSty I; lane 4, mix of equal amounts ofSty I-cleaved G/G control and uncleaved G/G control; lane 5, G/G control cleaved with StyI.

Full size image

Expression of variant and wild-type SCAD . Eucaryotic cells (COS cells). Northern blot analysis of SCAD mRNA from COS cells transfected with the expression vector containing wild-type or A625 variant(G209S) SCAD cDNA showed that the amounts of G209S SCAD mRNAs did not differ significantly from that expressed from wild-type cDNA (results not shown). Western blot analysis showed that the wild-type and the A625 variant vector produced mature SCAD protein of correct size and that the amounts of SCAD protein were comparable (Fig. 4). SCAD activity measurement in the COS cell extracts showed that the activity of the G209S variant enzyme was in the same range as that of the wild type(Table 3). These results indicate that the G209S variant protein, like the wild-type protein, is able to fold into a stable and active conformation.

Figure 4
figure 4

Expression of wild-type and variant SCAD in COS cells. Western blot of extracts from COS cells harboring expression vectors encoding wild-type and A625 variant (G209S) SCAD protein. Approximately 0.20 μg of protein was loaded per lane. Purified bovine SCAD was coelectrophoresed as a marker. Lane 1, marker; lane 2, cells transfected with A625 variant SCAD cDNA; lane 3, cells transfected with wild-type SCAD cDNA; lane 4, cells transfected with vector without the SCAD gene.

Full size image
Table 3 SCAD enzyme activity in extracts from transfected COS cells
Full size table

Prokaryotic cells (E. coli). The GroE chaperonin complex, consisting of the proteins GroES and GroEL, has been shown to facilitate the refolding of several proteins in vitro(37). The proposed role of the GroE complex in vivo includes protein folding, transport, and degradation(38–40). We have demonstrated that coexpression of MCAD mutants in E. coli together with GroES and GroEL mimics expression in eukaryotic COS cells(41), and that expression in E. coli without GroES and GroEL represents a situation where folding defects of MCAD mutants can be revealed(42). To investigate the influence of the G625 to A transition on folding of the SCAD variant, wild-type, and variant SCAD proteins were expressed with and without co-overexpression of GroESL. Expression of G209S variant SCAD without GroESL co-overexpression resulted in one-fifth to one-third of the wild-type activity in repeated independent expression experiments (Table 4), thus indicating impaired biogenesis of the G209S variant. Co-overexpression of the GroESL chaperonins resulted in a dramatically increased SCAD enzyme activity levels both for wild-type and G209S variant SCAD (Table 4). These results indicate that G209S can fold into a correct conformation as efficiently as the wild-type protein, but only in the presence of excessive amounts of the chaperonins. The reproducible production of active SCAD wild-type and variant enzyme in E. coli have been unambiguously confirmed through repeated expression experiments.

Table 4 SCAD enzyme activity in extracts from transformed E. coli
Full size table

Investigation of A625 variant SCAD. To further investigate the effect of the G209S substitution on SCAD enzyme function we analyzed the kinetic parameters of the G209S variant and the wild type. In protein extracts from E. coli cells co-overexpressing the GroESL chaperonins and the respective SCAD variant, we observed that the apparent Km andVmax of the G209S variant enzyme did not differ significantly from those of the wild type (G209S: 25 μM and 19.2 μmol/mg/h, respectively; wild type: 21 and 19, respectively)(43). This indicates that the G209S mutation does not significantly affect substrate binding and turnover of the enzyme.

To investigate the temperature stability of the enzyme, the protein extracts from E. coli cells coexpressing GroESL and the respective SCAD variant were incubated at different temperatures ranging from 0°C to 62°C for 10 min and the SCAD enzyme activity was determined subsequently. The thermal inactivation profiles were performed two times with different protein extracts from two different expression experiments. The profiles demonstrate that the half-denaturation temperature of the G209S variant is 2.5°C lower than that of the wild type (Fig. 5), indicating that the change from glycine to serine compromises the stability of the enzyme.

Figure 5
figure 5

Temperature inactivation of wild-type and G209S variant enzyme. Thermal inactivation curves showing the percentage of protein remaining undenatured and active after a 10-min incubation. The protein extracts from E. coli cells coexpressing GroESL and the respective variant SCAD enzyme were incubated at different temperatures ranging from 0°C to 62°C for 10 min, and the enzyme activities were then determined with the ferricenium ion assay. The half-denaturation temperatures of wild-type and the G209S variant were 49.5 and 47°C, respectively.

Full size image

DISCUSSION

The principal result of this study is the highly significant association between elevated EMA excretion and the A625 SCAD allele. About 90% of the patients in the EMA group are either homozygote or heterozygote individuals, a proportion significantly higher than the control group (42%). This demonstrates that association studies with a candidate genetic marker can provide an approach toward a stratification of genetic subpopulations among the EMA patients. Because of the direct metabolic relationship between the SCAD enzyme and EMA through butyryl-CoA(1), it is reasonable to assume that the overrepresentation of the A625 allele in the patient group either is due to linkage disequilibrium with a disease-associated mutation in the SCAD gene, or that the A625 gene variant in itself is causally related to the abnormal organic aciduria, possibly with other-yet unknown-genetic and/or environmental factors interacting with it. Both possibilities exist, and as indicated from the genotype distribution over the range of EMA excretions in the patients, there seem to exist several subgroups, which need further study by clinical, biochemical, and molecular methods. However, the total sequencing of the coding region of the SCAD gene from three representative cases with EMA excretion of 26, 35, and 65 mmol/mol of creatinine, respectively, shows that there exist patients in whom A625 is the only amino acid variation compared with the control SCAD gene. Thus, the results indicate that the A625 allele in itself is implicated in the ethylmalonic aciduria (Fig. 1) and that other environmental or genetic factors likely contribute with the A625 variant to cause elevation of EMA excretion.

This first approach toward an elucidation of the disease-causing molecular mechanism was therefore focused on the molecular properties of the SCAD A625 variant allele and protein. To address the hypothesis that the A625 allele is a susceptibility allele for the development of ethylmalonic aciduria, we have performed a number of structural and functional studies. First, the possibility that this one-nucleotide change located at an intron-exon junction may have an effect on the splicing process was excluded by our finding that the steady state level of mRNA from the two alleles is similar (Fig. 3). Therefore, it is unlikely that A625 affects the splicing of SCAD mRNA. Alternatively, the change at position 209 (equivalent to nucleotide position 625) of the SCAD protein from glycine to serine, a large residue with a polar hydroxy group, might have a negative effect on substrate binding, because the glycine 209 could be located at a critical position (based on homology to the known MCAD structure)(44) in the proximity of the substrate binding site. Although serine 209 could distort the conformation of the domain, our analysis of the apparent kinetic parameters of wild-type SCAD and the G209S variant indicates that the serine-containing variant has kinetic characteristics similar to those of the wild type.

Because expression studies in E. coli have shown that several MCAD mutants have a primary effect on folding and/or oligomer assembly(41, 42, 45, 46) we investigated the G209S SCAD variant in this system. Accordingly, expression of the G209S variant in E. coli without GroESL co-overexpression shows that the level of active serine 209 variant is decreased compared with wild-type(Table 4), presumably due to slightly impaired folding. In addition to impaired folding the stability of the tetrameric enzyme is decisive for the steady-state amount of active enzyme. Therefore we measured the protein stability of wild-type protein and the G209S variant. The thermal inactivation experiment (Fig. 5) revealed that the half-denaturation temperature of the G209S variant was 2.5°C lower than the wild type, showing that the variant enzyme has a reduced stability reflecting a less tight structure. Although the G209S does not appear to affect the structures directly involved in catalysis, this observation suggests that it does perturb the enzyme stability, as is known for variants of β-glucuronidase(47).

The observed effect on folding of the G209S variant in the E. coli expression system could not be detected in the COS cells. However, we have previously shown that the sensitivity of the COS expression system is insufficient to reveal certain disease-causing folding mutants compared with the E. coli system. Actually, the disease-causing R28C MCAD mutant exhibits 51-110% of the wild-type enzyme activity in COS cells(41), whereas E. coli expression revealed less that 5% of the wild-type activity.

The physiologic relevance of the slightly impaired folding and the decreased heat stability of the G209S SCAD protein is not clear at present. However, it is reasonable to suspect that a less tight structure of the protein could make it more susceptible to degradation in some tissues under certain conditions. Thus, a slightly defective folding and/or a decreased stability of the tetrameric variant enzyme may render the steady state amount of SCAD enzyme rate-limiting in the fatty acid oxidation pathway, resulting in butyryl-CoA accumulation and EMA excretion.