D-bifunctional protein is involved in the peroxisomal β-oxidation of very long chain fatty acids, branched chain fatty acids and bile acid intermediates. In line with the central role of D-bifunctional protein in the β-oxidation of these three types of fatty acids, all patients with D-bifunctional protein deficiency so far reported in the literature show elevated levels of very long chain fatty acids, branched chain fatty acids and bile acid intermediates. In contrast, we now report two novel patients with D-bifunctional protein deficiency who both have normal levels of bile acid intermediates. Complementation analysis and D-bifunctional protein activity measurements revealed that both patients had an isolated defect in the enoyl-CoA hydratase domain of D-bifunctional protein. Subsequent mutation analysis showed that both patients are homozygous for a missense mutation (N457Y), which is located in the enoyl-CoA hydratase coding part of the D-bifunctional protein gene. Expression of the mutant protein in the yeast Saccharomyces cerevisiae confirmed that the N457Y mutation s the disease-causing mutation. Immunoblot analysis of patient fibroblast homogenates showed that the protein levels of full-length D-bifunctional protein were strongly reduced while the enoyl-CoA hydratase component produced after processing within the peroxisome was undetectable, which indicates that the mutation leads to an unstable protein.
Peroxisomal β-oxidation of fatty acids in mammals is catalyzed by two distinct pathways (for a review see 1). The first pathway catalyzes the β-oxidation of very long chain fatty acids, like C26:0, and involves the following enzymes: straight chain acyl-CoA oxidase (2,3), enoyl-CoA hydratase/L-3-hydroxyacyl-CoA dehydrogenase (L-bifunctional protein or L-BP) (4–12) and 3-ketoacyl-CoA thiolase (13,14). The second pathway catalyzes the β-oxidation of branched chain fatty acids, like pristanic acid, and bile acid intermediates, like diand trihydroxycholestanoic acid (DHCA and THCA, respectively), and involves the branched chain acyl-CoA oxidase (3,15), enoyl-CoA hydratase/D-3-hydroxyacyl-CoA dehydrogenase (D-bifunctional protein or D-BP) (5–12,16) and sterol carrier protein X (SCPx) (14,17–19). Although D-BP and SCPx are involved in the β-oxidation of branched chain fatty acids and bile acid intermediates, they are also able to react with straight chain fatty acids (5–7,10,11,19).
At present, several patients have been described with a defect in peroxisomal β-oxidation caused by a single enzyme defect, including straight chain acyl-CoA oxidase deficiency (20), 3-ketoacyl-CoA thiolase deficiency (21,22) and D-BP deficiency (23–25). Many of the patients with D-BP deficiency were initially thought to be L-BP-deficient because the patients' fibroblasts did not complement the fibroblasts of a patient diagnosed as being L-BP-deficient (26–34). However, we recently discovered that this latter patient is actually D-BP-deficient and not L-BP-deficient (25). Furthermore, after analysis of nine patients with presumed L-BP deficiency, we identified mutations in the gene coding for D-BP (25).
All patients with D-BP deficiency described so far are characterized by the accumulation of very long chain fatty acids, pristanic acid and bile acid intermediates (23–25). In this paper, we now describe two patients with isolated defects in the enoyl-CoA hydratase component of D-BP. Both patients had elevated levels of very long chain fatty acids and pristanic acid, but the concentrations of bile intermediates were normal.
Clinical and biochemical characterization of the patients
The two patients described in this paper showed a number of abnormalities suggestive of a peroxisomal disorder, including neurological abnormalities (hypotonia, absence of suck reflexes and convulsions). Craniofacial dysmorphia was present in patient 2 but not in patient 1. As shown in Table 1, very long chain fatty acid levels were elevated in plasma from both patients, suggesting a defect in peroxisomal fatty acid β-oxidation. In addition, pristanic acid levels were elevated in both patients, whereas phytanic acid levels were normal. Furthermore, normal levels of DHCA and THCA were found in plasma (Table 1).
In order to identify the extent of peroxisomal dysfunction in the two patients, detailed studies were performed in cultured skin fibroblasts (Table 1). Very long chain fatty acid levels were elevated and oxidation of both C26:0 and pristanic acid was deficient in the patients' fibroblasts. Plasmalogen levels, de novo plasmalogen synthesis and the activity of dihydroxy-acetonephosphate acyltransferase (DHAPAT), which catalyzes the first step in plasmalogen biosynthesis, were all normal.
Immunofluorescence microscopy studies of the patients' fibroblasts using an antibody against human catalase revealed normal appearing peroxisomes in ∼80% of the fibroblasts. However, in ∼20% of the fibroblasts a decreased number of peroxisomes was found, which were much larger than peroxisomes in control fibroblasts (not shown). Similar results were obtained when an antibody against D-BP was used (not shown).
Taken together, these data exclude a generalized peroxisomal defect and indicate a selective defect in the peroxisomal fatty acid β-oxidation pathway.
Complementation analysis and D-BP activity measurements
Since the clinical and biochemical findings in the two patients were reminiscent of the abnormalities found in patients with a deficiency of acyl-CoA oxidase (20) (Table 1), we first focused on this enzyme. However, enzyme activity measurements revealed normal acyl-CoA oxidase activities, suggesting a different enzyme defect (not shown). Subsequently, we studied whether D-BP, the next enzyme in the β-oxidation pathway, was deficient in the two patients. To this end, we performed complementation analysis making use of fibroblasts from a patient with an established D-BP deficiency (25). No restoration of pristanic acid β-oxidation activity was observed (Table 2). This points to a functional deficiency of D-BP in the two patients. Surprisingly, when the patients' fibroblasts were subsequently fused with fibroblasts from a patient with a specific deficiency of the D-3-hydroxyacyl-CoA dehydrogenase activity of D-BP (24), pristanic acid β-oxidation was restored (Table 2). This strongly suggested a selective deficiency of the enoyl-CoA hydratase activity of D-BP with normal activity of the 3-hydroxyacyl-CoA dehydrogenase domain.
To confirm that both patients have a deficient enoyl-CoA hydratase activity, D-BP activity was measured (Table 3). To this end, fibroblast homogenates were incubated with the enoyl-CoA ester of THCA and the formation of 24-hydroxy-THC-CoA and 24-keto-THC-CoA was measured by HPLC. In both patients no 24-hydroxy-THC-CoA and no 24-keto-THC-CoA could be measured, which is in agreement with a deficient enoyl-CoA hydratase domain of D-BP.
Sequence analysis of the D-BP cDNA revealed that both patients were homozygous for two different mutations. The first mutation, 1369A→T, changes the Asn457 into a Tyr (N457Y), whereas the second mutation, 1531T→C, changes the Trp511 into an Arg (W511R). Both mutations are located in the part coding for the enoyl-CoA hydratase domain of D-BP.
To determine whether both patients were also homozygous for these mutations at the genomic level, we took advantage of the fact that the latter mutation eliminates a BsrI restriction site. We therefore amplified exon 18 of the D-BP gene by PCR using primers containing a BsrI restriction site which serves as an internal control for BsrI digestion. Digestion with BsrI of the amplified 290 bp fragment revealed three fragments in a control subject: a 23 bp fragment, a 116 bp fragment and a 130 bp fragment. In both patients, however, only two fragments were found: a 23 bp fragment and a 246 bp fragment, indicative of the disappearance of the internal BsrI site, which demonstrates that both alleles contain the 1531T→C mutation (not shown). Unfortunately, no material from the patients' parents was available, which makes it difficult to determine whether the patients were truly homozygous or heterozygous, with the other allele being the null allele. However, the fact that the parents of both patients are consanguineous makes it very likely that both patients are homozygous for both the N457Y and the W511R mutation.
Sequence analysis of D-BP cDNA of 15 control subjects of Caucasian origin did not identify the N457Y mutation, but two control subjects were heterozygous for the W511R mutation. Since this suggested that the W511R mutation may be a polymorphism and not a disease-causing mutation, we analyzed exon 18 in DNA of 68 additional control subjects of Caucasian origin by restriction analysis with BsrI. This led to the identification of 14 additional control subjects that were heterozygous for this mutation, which makes the overall allele frequency of the W511R mutation ∼10% (16/166).
To study the effect of the mutations at the protein level, we performed immunoblot analysis using an antibody raised against D-BP (Fig. 1). In control fibroblasts, three major bands could be detected: the full-length protein of 79 kDa, the 45 kDa product corresponding to the enoyl-CoA hydratase component of D-BP and the 35 kDa product corresponding to the D-3-hydroxyacyl-CoA dehydrogenase component of D-BP. In contrast, in fibroblasts of both patients, reduced amounts of the full-length protein were observed but the 45 kDa product containing the hydratase domain was never detected. In patient 1, the 35 kDa product containing the dehydrogenase domain was always present at near normal amounts, whereas in patient 2, the 35 kDa product was definitely present, although invariable amounts. The underlying basis for this variability in two patients having the same mutation is unclear.
For comparison, we also analyzed fibroblasts from two patients with selective defects in the PTS1 receptor (Pex5p) (patient PBD018; 35) and PTS2 receptor (Pex7p) (patient RCDP; 36), respectively. In fibroblasts from the PTS1 receptor-deficient patient only the 79 kDa full-length product was found, whereas in fibroblasts of the PTS2 receptor-deficient patient a normal pattern of the 79, 45 and 35 kDa products was observed.
To determine whether the mutations affect the expression and/or activity of D-BP, expression studies were performed in the yeast S.cerevisiae (Fig. 2). When yeast was transformed with the full-length wild-type cDNA of D-BP (pDBP-WT), high enoyl-CoA hydratase activity could be measured, whereas in yeast transformed with the empty expression plasmid (pEL26) no enoyl-CoA hydratase activity could be detected (Fig. 2A). Yeast transformed with the W511R cDNA (pDBP-W511R) displayed wild-type activity, indicating that this mutation has no effect on the enoyl-CoA hydratase activity and that this mutation is indeed a polymorphism, as suggested by the frequency of this mutation in the normal population. Unexpectedly, yeast transformed with the N457Y+W511R cDNA (pDBP-N457Y+W511R) or the N457Y cDNA (pDBP-N457Y) displayed enoyl-CoA hydratase activity of ∼60% compared with the wild-type, while in fibroblasts from the patients no activity was detected (Table 3). An explanation for this phenomenon could be that in yeast the full-length D-BP is not processed (not shown), whereas in man the full-length protein is processed into an enoyl-CoA hydratase component (45 kDa) and a D-3-hydroxyacyl-CoA dehydrogenase component (35 kDa). In both patients, the enoyl-CoA hydratase component was found to be deficient upon immunoblot analysis (Fig. 1). We therefore decided to examine the effect of the mutation(s) on the activity of the enoyl-CoA hydratase component alone. To this end, we expressed a modified cDNA encoding only the enoyl-CoA hydratase domain plus the SCP2-like domain of the wild-type and the two mutant D-BPs in yeast (Fig. 2B). Yeast transformed with the modified wild-type cDNA (pHY-WT) displayed high enoyl-CoA hydratase activity. However, yeast transformed with the modified N457Y+W511R cDNA (pHY-N457Y+W511R) or the modified N457Y cDNA (pHY-N457Y) did not show any residual activity. As was the case for full-length D-BP, the W511R mutation did not affect D-BP activity.
Combined, these data prove that the N457Y mutation is responsible for the loss of enoyl-CoA hydratase activity and that the W511R mutation is a polymorphism.
The two patients described in this paper showed a peculiar combination of abnormalities in plasma suggestive of a peroxisomal defect, including elevated levels of very long chain fatty acids and pristanic acid but normal levels of the bile acid intermediates DHCA and THCA. Our results show that this is caused by a specific deficiency of the enoyl-CoA hydratase component of D-BP in both patients. This is clear from the following observations. Studies in fibroblasts revealed an isolated defect in the peroxisomal β-oxidation system. The availability of a mutant cell line with a specific defect in D-BP (25) allowed complementation analysis which indicated that the two patients belong to the D-BP deficiency group (Table 2). Additional complementation analysis using cells from a patient with an established defect in the D-3-hydroxyacyl-CoA dehydrogenase component of D-BP (24) strongly suggested that the primary defect in the two patients had to be in the enoyl-CoA hydratase domain of D-BP. Enzyme activity measurements indicated that this was true (Table 3). Molecular analysis showed that both patients were homozygous for an N457Y mutation which is located in the enoyl-CoA hydratase domain of D-BP. Expression studies in yeast revealed that the mutant protein displayed abundant enoyl-CoA hydratase activity (60% of normal; Fig. 2A), suggesting that the mutations did not so much affect the catalytic activity of the protein. Rather, the results of immunoblot experiments suggested that the N457Y mutation has a drastic effect on the stability of D-BP. This is not only true for the full-length 79 kDa form, which was present in strongly reduced amounts in the patients' fibroblasts, but also for the 45 kDa form, which harbors the D-enoyl-CoA hydratase activity. In fibroblasts from both patients the 45 kDa band was completely missing, which implies that the 45 kDa mutant protein is very unstable. Expression studies in yeast show that the mutated enoyl-CoA hydratase component does not possess any enoyl-CoA hydratase activity (Fig. 2B). Taken together, these data suggest that the mutant 45 kDa protein per se is very unstable, but less unstable when it is part of the 79 kDa protein and covalently attached to the D-3-hydroxy-acyl-CoA dehydrogenase component of D-BP. The reason that in both patients no enoyl-CoA hydratase activity could be measured despite the presence of low amounts of full-length D-BP is probably because the activity was below the detection limits of the activity assay.
In addition to the disease-causing mutation, N457Y, both patients also had a W511R mutation, which appears to be a polymorphism. This mutation was also found in control subjects of Caucasian origin with a frequency of ∼10%. Furthermore, expression studies in yeast revealed that the W511R mutation does not have any effect on the enoyl-CoA hydratase activity (Fig. 2).
To date, all patients with D-BP deficiency reported in the literature (23–25), including the patients described here, have elevated levels of very long chain fatty acids and deficient C26:0 β-oxidation. This strongly suggests that D-BP is the major enzyme in C26:0 β-oxidation and not L-BP, which until now was generally considered to be the main enzyme involved in the β-oxidation of very long chain fatty acids.
In contrast to the previously published D-BP deficient patients (23–25), normal levels of the bile acid intermediates DHCA and THCA were found in plasma from the patients described in this paper. This difference is difficult to explain but it may be that patients 1 and 2 have sufficient residual amounts of active full-length D-BP to β-oxidize bile acid intermediates, in contrast to the situation in the previously published patients (23–25) with no active D-BP at all. Although the amounts of active, full-length D-BP seem to be enough to β-oxidize DHCA and THCA in vivo, it is apparently not sufficient for β-oxidation of very long chain fatty acids and pristanic acid, since both patients had elevated levels of these substrates in plasma and fibroblasts. Future experiments will have to resolve this problem.
The results presented in this paper indicate that the flowchart which has been introduced for the diagnostic work-up of patients with a peroxisomal defect (37) needs modification. According to this flowchart, a patient suspected of suffering from a peroxisomal β-oxidation defect and with normal plasma bile acids probably suffers from acyl-CoA oxidase deficiency. However, our results show that such a patient could also suffer from D-BP deficiency.
It has been shown that D-BP is cleaved into the enoyl-CoA hydratase component (45 kDa) and the D-3-hydroxyacyl-CoA dehydrogenase component (35 kDa) (5,8,10,11,38,39). Our immunoblot data now show that D-BP is not processed into 45 and 35 kDa products in fibroblasts of a patient with a defective PTS1 receptor. This strongly suggests that this processing event occurs in vivo but that D-BP has to be imported into the peroxisome for processing to occur. Furthermore, D-BP is normally processed in fibroblasts of a patient with a defective PTS2 receptor. This indicates that the processing enzyme, which catalyzes the cleavage, does not contain a PTS2 signal for import into the peroxisome. Although D-BP appears to be processed under in vivo conditions, cleavage does not seem to be important for the activity of the protein, since in fibroblast cell lines of a patient with a defective PTS1 receptor, the activity of D-BP is normal (unpublished data), which is in line with the results of our expression studies in yeast.
Materials and Methods
Patient 1 was born at term to consanguineous Algerian parents after a normal pregnancy. The patient had three healthy siblings (one brother and two sisters). At birth, his weight was 3700 g and the head circumference was 36 cm, which is within the normal range. The patient showed no facial dysmorphia. From birth, the boy was hypotonic and showed no sucking reflex. Two hours after birth he developed a moderate respiratory distress with stridor. Convulsions (partial clonic seizures over 5–10 min, 5–10 per day) appeared at 10 h of life and were initially refractory to anticonvulsant therapy. His EEG was slow with several epileptic discharges. At 11/2 months of age, the patient remained severely hypotonic with no eye contact and no reaction to noise. There was no facial dysmorphia except for the presence of shallow orbital ridges. Furthermore, there was no hepatomegaly and hepatic functions were normal. At 3 months the patient required tube feeding and showed moderate amyotrophia. The stridor was still persistent. At 8 months, the patient developed intracranial hypertension. MRI showed hydrocephalus with ventricular dilation and a mild delay in myelination. He was treated with Diamox and CSF punctures which initially allowed control of the intracranial signs. At 10 months, he was hospitalized again for signs of intracranial hypertension which required a ventricular shunt. After this neurosurgical procedure, the patient showed no progress in psychomotor functioning. The patient died at 1 year from cardio-respiratory failure.
Patient 2 was born at term to consanguineous Moroccan parents after a normal pregnancy. At birth the boy was hypotonic and showed no sucking reflex. He was macrocephalic (head circumference 39 cm) and showed facial dysmorphia typical of Zellweger syndrome. At 2 h of life he had convulsions which were initially refractory to anticonvulsant drugs. The patient was hypotonic with no ocular tracking and no reaction to noise. Hepatomegaly was noted at 2 weeks of life. Ultrasonography of the brain showed moderate ventricular dilation and cardiac ultrasonography showed moderate cardiomyopathy. At 3 months, the patient developed clinical signs of intracranial hypertension (head circumference 45 cm). Upon decompensation of the hydrocephalus (head circumference 50 cm), it was decided to perform a ventricular shunt, but the patient died (at 4 months) before the procedure was performed. The patient had two healthy siblings. Three siblings died from unknown causes in Morocco and two brothers died with a similar clinical presentation.
Cell culture conditions
Skin fibroblasts were cultured in HAM-F10 medium (Gibco BRL, Paisley, UK) supplemented with 10% fetal calf serum (Bio-Whittaker, Verviers, Belgium).
Determination of plasma metabolites and erythrocyte plasmalogens
Very long chain fatty acids, bile acids, phytanic acid and pristanic acid were measured in plasma using gas chromatographic procedures as described in detail previously (40). Erythrocyte plasmalogen levels were measured as described by Björkhem et al. (41).
Peroxisomal functions in cultured skin fibroblasts
De novo plasmalogen synthesis, peroxisomal C26:0 and pristanic acid β-oxidation, phytanic α-oxidation, DHAPAT activity and very long chain fatty acids were measured as described previously (40). Catalase immunofluorescence, immunoblot analysis and the measurement of D-BP activity expressed as combined enoyl-CoA hydratase activity and D-3-hydroxyacyl-CoA dehydrogenase activity were performed as described by Van Grunsven et al. (25). Complementation analysis was performed as described previously (33).
PCR and sequencing
Total RNA was isolated from cultured skin fibroblasts using the acid guanidinium thiocyanate-phenol-chloroform extraction procedure described by Chomczynski and Sacchi (42) and subsequently used to prepare cDNA (43).
For sequence analysis, the cDNA encoding D-BP was amplified by PCR in three overlapping fragments by means of three primer sets tagged with either −21M13 (5′-tgt aaa acg acg gcc agt-3′) or universal M13rev (5′-cag gaa aca gct atg acc-3′) extensions. The first fragment (bases –48 to 806) was amplified with primers -21MDBP-48 (5′-[–21M13]-GGC CAG CGC GTC TGC TTG TTC-3′) and M13RDBP806 (5′- [M13rev]-ACT GCC TCA GGA GTC ATT GG-3′), the second fragment (bases 675 to 1543) was amplified with primers -21MDBP675 (5′-[−21M13]-TTG TCA CGA GAG TTG TGA GG-3′) and M13RDBP1543 (5′-[M13rev]-GTA AGG GAT TCC AGT CTC CAC-3′) and the third fragment (bases 1489 to 2313) was amplified with primers -21MDBP1489 (5′-[-21M13]-ACC TCT CTT AAT CAG GCT GC-3′) and M13RDBP2313 (5′-[M13rev]-CCC TGC ATC TTA GTT CTA ATC AC-3′).
PCR fragments were sequenced in both directions by means of −21M13 and M13rev fluorescent primers on an Applied Biosystems 377A automated DNA sequencer according to the manufacturer's protocol (Perkin Elmer, Foster City, CA).
Genomic DNA was isolated using diatoms matrix as described by Boom et al. (44).
For rapid screening of the 1531T→C mutation, exon 18 (45) was amplified by PCR using the sense primer DBP-exon18-f (5′-TGT AAA ACG ACG GCC AGT ACT GG C CAA TAA CCA GCC ATG TTT CC-3′; positions −85 to −66 in intron 17 starting from the acceptor site) and the antisense primer DBP-exon18-r (5′-CAG GAA ACA GCT ATG ACC ACT GGC AGT CAT AAG TAG TAG TGT C-3′; positions +68 to +87 in intron 18 starting from the donor site). As an internal control for BsrI digestion, both primers introduce a BsrI restriction site (underlined) in the PCR fragment. Subsequently, the PCR products were digested with endonuclease BsrI and were analyzed by agarose gel (2% w/v) electrophoresis.
Construction of expression plasmids
Eight different expression plasmids were constructed for the expression of full-length wild-type D-BP and the mutants N457Y, W511R and N457Y+W511R in Saccharomyces cerevisiae. In addition, plasmids were constructed containing the coding sequence of only the enoyl-CoA hydratase and SCP-related domains of D-BP.
To construct the plasmids containing the wild-type D-BP (pDBP-WT) or N457Y+W511R mutations (pDBP-N457Y+W511R), full-length cDNA encoding D-BP was amplified by PCR from either control fibroblasts or fibroblasts of patient 1, who was homozygous for the N457Y+W511R mutations, using primers xbaDBP (5′-ttt tct aga ATG GGC TCA CCG CTG AGG TTC-3′; positions 1–21 of the cDNA sequence; XbaI extension) and pstDBP (5′-ttt tct gca gCT TCA GAG CTT GGC GTA GTC-3′; positions 2213–2194; PstI extension).
To construct the plasmids containing either the N457Y mutation (pDBP-N457Y) or the W511R mutation (pDBP-W511R), the megaprimer method was used (46). For construction of pDBP-N457Y, part of pDBP-WT was first amplified by PCR using the sense primer xbaDBP and the antisense primer DBP1369T (5′-GAT AGT GGC ATA TAA GTT CC-3′; positions 1372–1353; mutation underlined) containing the 1369A→T mutation. This PCR product was subsequently used as the sense primer in a second round PCR using pstDBP as the antisense primer.
For construction of pDBP-W511R, part of pDBP-WT was first amplified by PCR using the sense primer DBP1531C (5′-CAG TGG AGA CCG GAA TCC CTT ACA C-3′; positions 1521–1545; mutation underlined) and the antisense primer pst-DBP. This PCR product was subsequently used as the sense primer in a second round PCR using xbaDBP as the sense primer.
To obtain the plasmids containing only the C-terminal part (bp 955–2213) of D-BP, including the enoyl-CoA hydratase domain and the SCP-related domain (pHY-WT, pHY-N457Y+W511R, pHY-N457Y and pHY-W511R), part of the plasmids containing full-length D-BP were amplifed by PCR using the primers xbaDBP-955f (5′-ttt tct aga ATG ACA GCA ACA TCA GGA TTT GCT GG-3′; positions 955–977; XbaI extension) and pstDBP.
All PCR reactions were performed with a low error rate mixture of DNA polymerases (Expand High Fidelity; Boehringer Mannheim, Mannheim, Germany).
The PCR products, containing an XbaI restriction site at the 5′ end and a PstI restriction site at the 3′ end, were digested with XbaI and PstI and cloned downstream of the S.cerevisiae CTA1 promoter into the XbaI and PstI restriction sites of pEL 26 (47).
All PCR fragments were sequenced entirely to exclude Taq polymerase introduced errors.
Yeast culture and expression
The yeast strain used in this study, S.cerevisiae BJ1991 (MATα, leu2, ura3–251, prb1–1122, pep4–3), was transformed with pEL26, pDBP-WT, pDBP-N457Y+W511R, pDBP-N457Y, pDBP-W511R, pHY-WT, pHY-N457Y+W511R, pHY-N457Y or pHY-W511R and grown on minimal essential medium containing 0.67% yeast nitrogen base without amino acids (Difco, Detroit, MI), 0.3% glucose and the appropriate amino acids at 30°C. To induce expression, the cells were shifted to rich oleic acid medium containing 0.5% potassium phosphate buffer, pH 6.0 0.3% yeast extract, 0.12% oleic acid, 0.2% Tween-40 and 0.5% peptone with a starting OD600 of 0.1 and incubated for 17 h at 30°C.
For measurement of the enoyl-CoA hydratase and D-3-hydroxyacyl-CoA dehydrogenase activities of D-BP, cells were resuspended in phosphate-buffered saline (pH 7.4) containing 1 µg/ml leupeptin, 1 mM EDTA and 1 mg/ml Pefabloc (Merck, Darmstadt, Germany) and disrupted by agitation at 4°C for 15 min on a vortex mixer in the presence of glass beads (Ø = 0.45 mm). The homogenates were centrifuged at 2000 g at 4°C for 2 min and the supernatants were used for enzyme activity measurements. The enoyl-CoA hydratase activity and D-3-hydroxyacyl-CoA dehydrogenase activity were corrected for the amount of D-BP as quantified by immunoblot analysis.
Induction of expression was checked by measuring the catalase activity in the yeast homogenates. To this end, yeast homogenate was incubated with 0.06% H2O2 in 40 mM KPi pH 7.5 and the disappearance of H2O2 was measured at 240 nm.
We are grateful to Lodewijk Ijlst and Hans Waterham for support and suggestions throughout this study and for critical reading of the manuscript. This work was supported in part by the Princess Beatrix Fund (The Hague, The Netherlands).