Main

Holocarboxylase synthetase (HCS: EC 6.3.1.10) is an enzyme that catalyzes biotin incorporation into multiple carboxylases. In humans, there are three mitochondrial carboxylases, namely pyruvate carboxylase, propionyl-CoA carboxylase (PCC), and methylcrotonyl-CoA carboxylase. Acetyl-CoA carboxylase (ACC) is the only known cytosolic carboxylase. Biotin is the prosthetic group of these carboxylases. A deficiency in HCS has been shown to be the cause of the early-onset type of biotin-responsive multiple carboxylase deficiency (McKusick 253270). Most patients with HCS deficiency show acute episodes of metabolic acidosis and characteristic organic aciduria due to the decreased activity of multiple carboxylases before a few months of age (1, 2). Symptoms of the patients include tachypnea, feeding difficulties, and seizures, which may lead to coma or even death. Some patients become symptomatic in the later infantile period, at the age of several months to years (3–6). All HCS-deficient patients so far reported have responded to biotin administration, although in some patients the response was only partial as manifested by continued excretion of abnormal metabolites in the urine (7–9). Developmental abnormalities have also been reported in some cases in spite of high-dose biotin therapy (7, 9).

Burri et al. characterized HCS activity in cultured cells from seven patients with considerable differences in disease severity (10). They showed that the Vmax values of patients' HCS were lower than the normal mean and were similar in all except one patient, and that the Km values for biotin were elevated to various degrees. From their clinical and kinetic observations, they concluded that onset of the disease and its responsiveness to biotin administration is governed by the degree of abnormality in the Km of HCS. Suormala et al. investigated the effects of biotin concentration on carboxylase activities in fibroblasts in five HCS-deficient patients (5). The carboxylase activities of all patients normalized when the biotin concentration was increased to 10 000 nM. In carboxylase reactivation studies, although the kinetics of PCC activation varied greatly, all cells restored their PCC activity to the normal or nearly normal level (>87%). Reactivation of PCC activity in relation to time and biotin concentration correlated well with the severity and age at onset of the illness in four patients. These results were consistent with the original “Km mutant theory” proposed by Burri et al. (10).

Human HCS cDNA has been cloned (11, 12) and has enabled investigation of HCS at the molecular level. The protein encoded by this cDNA is 726 amino acids in length and has a homologous region (aa 448–701) to Bir A, the biotin apo-carboxyl carrier protein ligase of Escherichia coli (11). This portion of human HCS is thought to be the putative biotin-binding region. In a previous report, we characterized two mutations, Val550Met and Leu237Pro (13). The former was found in a patient with mild clinical phenotype and the latter in patients with severe form. An expression study demonstrated that the Km of the Val550Met HCS mutant was higher than, but the Km of the Leu237Pro-mutant was the same as, that of the wild-type enzyme (13). Thus, we questioned whether the “Km mutant theory” could be applied to all patients with HCS deficiency, and proposed that not only the Km of HCS for biotin, but also the Vmax, is an important factor in determining the severity of symptoms and their responsiveness to biotin therapy (13). We tested our hypothesis in the present study by analyzing the kinetics of HCS mutants, including a newly identified mutation. We also examined the relationship between the kinetic characteristics of HCS mutants and the clinical and biochemical features of the HCS deficient patients.

METHODS

Patients.

Nine HCS-deficient patients were examined in the present study. Patient AD was born to healthy consanguineous Turkish parents (5, 14). On the 2nd day of life, she showed metabolic acidosis. The response of symptoms to biotin treatment was good, although small amounts of 3-hydroxyisovalerate were detected in the urine with 20 to 40 mg/day of biotin. Patient HR is a girl of Lebanese descent born to consanguineous parents (15). After the first severe symptoms appeared at the age of 20 mo, she showed recurrent attacks of vomiting associated with tachypnea and lethargy. Diagnosis was achieved at the age of 5 y and biotin administration was started (10 mg/day). The clinical manifestations of patients UW and KT have been described earlier as patients 1 and 4b, respectively (16). Briefly, patient UW developed metabolic acidosis on the 1st day of life and oral biotin administration (10 mg/day) led to remarkable improvement. With 40 mg/day of biotin, she has remained asymptomatic but urinary organic acid analysis revealed slightly elevated excretion of 3-hydroxyisovalerate. The elder sister of KT presented multiple carboxylase deficiency on the 2nd day of life (2). Consequently, the mother was given biotin in the last trimester of pregnancy since a prenatal diagnosis predicted KT to be affected (8). Biotin treatment was implemented immediately after the birth. However, lactate and 3-hydroxyisovalerate in the urine of KT were elevated even with high-dose biotin therapy (40–100 mg/day), and occasionally she showed acidosis associated with infection. Her psychomotor development was retarded (IQ = 64 at 6.5 y old). Patient TM suffered from two life-threatening episodes associated with metabolic acidosis at the age of 13 and 18 mo (6). During biotin therapy (20 mg/day), no further episodes have occurred and urinary excretion of organic acids was normal. Patient KE was born to nonconsanguineous French parents. She became symptomatic at 5 mo of age and HCS deficiency was diagnosed at 8 mo of age. She exhibited a partial clinical and biochemical response to biotin therapy. When she was 8 y old, her psychomotor development corresponded to that of a 3.5 to 4 y old child (5). Patient FE is a German girl and is described previously in brief (17). Patient FE had severe metabolic acidosis, muscular hypotonia, and developmental delay at the age of 11 mo. The carboxylase activities in her lymphocytes remained below the normal range (23–44% of normal mean) with 100 mg/day of biotin therapy. The clinical response was also partial. Her psychomotor development has remained slightly retarded (3–3.5 y at the chronological age of 4 y). Clinical picture of patient VE was described previously in brief (18). Her clinical responsiveness to biotin was good with slight excretion of organic acid in the urine. Clinical information for patient FA was not available. This study was approved by the Ethical Committee of Tohoku University School of Medicine.

Mutation analysis in patients.

Sequence analysis of the HCS cDNA of patient FA was performed using an A.L.F. red DNA sequencer (Pharmacia Biotech, Uppsala, Sweden) as previously described (11). To detect the Arg183Pro mutation by PCR-RFLP, genomic DNA was extracted from cultured fibroblasts using the Sepa Gene Kit (Sanko Junyaku, Tokyo, Japan). PCR reactions were conducted with primers S183 (5′AGGCACCCAACATCCTCCTCTA3′) and AS183 (5′TCAGGTAAGGGCTTCTGGACAT3′) (Fig. 1B). The PCR products were digested with Ppu MI, electrophoresed on a 3% agarose gel and visualized by ethidium bromide staining. To detect the IVS10n+5(g→a) mutation found in a Swedish patient, PCR-RFLP on the genomic DNA from patients FE and KE was performed as described previously (19). In brief, PCR was conducted using a sense (5′CGAGGTCGACGGTATCGAACCCTGTCCCTCCTGTGT3′) and an antisense (5′CAGCAATGATCACAAAAGAT A A3′) primer that contained one mismatch (bold) to create a Tsp 509I site on the mutant allele. The PCR products were digested with Tsp 509I (New England Biolabs, Beverly, MA) and separated on a 6% agarose gel.

Figure 1
figure 1

Mutation analysis of patients. (A) Schematic representation of PCR-RFLP analysis for detection of Arg283Pro mutation. (B) The banding pattern indicates that patient FA is homozygous for an Arg183Pro mutation. (C) PCR-RFLP analysis for detection of IVS10n+5(g→a). The genomic DNA from patient FE and KE was amplified and the PCR products were digested with Tsp 509I. The banding pattern suggests that both patients are heterozygous for IVS10n+5(g→a).

Full size image

Kinetic study of expressed mutant HCS proteins.

The expression vector pCAGGS (20) containing the entire coding region of the wild-type human HCS cDNA was constructed as described previously (16). The Leu237Pro mutation was inserted into the corresponding site of the HCS cDNA in pCAGGS by digestion with Bbs I (16). Similarly, for the delThr610 mutation, a fragment between the Hpa I site and the Afl II site was obtained from the HCS cDNA of patient HR and inserted into the expression vector. The Arg183Pro, Leu216Arg, Val333Glu, and Val363Asp mutations were introduced using the Transformer Site-Directed Mutagenesis Kit (Clontech, Palo Alto, CA). Each plasmid construct was sequenced to confirm the sequence fidelity of the insert and the insert-vector junctions. The plasmids were expressed in SV40-transformed HCS-deficient fibroblasts as described previously (13). Expression of HCS proteins was determined by Western blot analysis as described previously (13). HCS activity was estimated by measuring incorporation of 3H-biotin into apo-carboxyl carrier protein (CCP), a subunit of E. coli ACC, using boiled lysate as a blank (11, 13). To determine kinetic properties, a series of various 3H-biotin concentrations was assessed. The Km values for biotin and Vmax values were determined using Hanes's plots (21). Since we noticed that estimated Vmax values differed from preparation to preparation of CCP (13), we used the same CCP preparation throughout this study.

PCC activity in HCS deficient cells.

To maintain SV40-transformed fibroblasts and EB virus-transformed lymphoblasts, MEM (GIBCO BRL, Life Technologies, Rockville, MD) supplemented with 10% (vol/vol) FCS and AIM-V medium (GIBCO BRL, Life Technologies) were used as basal media. The biotin concentration of these media is about 10 nM. For PCC restoration experiments, cells were split and cultured in media supplied with biotin at various concentrations for 5 days. PCC activity in cell extracts was determined by fixation of [14C] bicarbonate as described (22).

Construction of plasmids for expression of N-terminal deletionmutants of HCS.

λgt10 DNA containing full-length HCS cDNA was subjected to PCR using Pfu DNA polymerase (Stratagene, La Jolla, CA) and three sets of primers (S51-AS9, S83-AS9 and S8-AS9). Primers S51 (5′GAGATTAAGCTCGAGCAGGACG3′: for expression from Met58), S8 (5′AGTCCATTCTCGAGGACCTGTAC3′: for expression from Met234) and AS9 had one mismatch to create a Xho I site. In primer S83 (5′AACTCGAGAACAGCACCAATGGAG3′: for expression from Ile117), two mismatches were introduced, one for a Xho I site, and the other for mutating codon Ile117 to Met. Each PCR product was digested with Xho I and subcloned into pBluescript II KS+. After confirming the nucleotide sequence, Xho I fragments were isolated and inserted into the Xho I site of pCAGGS. Expression and activity of HCS were analyzed as described above.

RESULTS

Genotype of patients.

Patient FA showed a G-to-C substitution at nt 835 (numbering as in Ref. (11), which converts Arg183 to Pro (Arg183Pro) (data not shown). To confirm the mutation in this patient's genomic DNA, PCR-RFLP was conducted as shown in Fig. 1A. The patterns of PCR products digested with Ppu MI suggested that she is homozygous for this novel mutation (Fig. 1B). Patients KE and FE have been shown to have 68 bp deletion on their HCS cDNAs in our previous work (17). We found an intronic mutation, IVS10n+5(g→a), in a Swedish patient and showed that this mutation produced the same 68 bp deletion on his HCS transcript that is a consequence of the exon 10 skipping of the HCS gene (19). Thus, we performed PCR-RFLP analysis for IVS10n+5(g→a) on genomic DNAs of patients KE and FE. The pattern of the digested DNA fragments suggested that both patients had the same mutation in a heterozygous form (Fig. 1C). The IVS10n+5(g→a) mutation was confirmed by direct DNA sequencing of genomic DNA (data not shown). Another mutation of each patient had already been identified in our previous work (17). Thus, the genotype of patient KE was determined as delC2279/IVS10n+5(g→a) and that of patient FE as delT1876/IVS10n+5(g→a). Patients AD, HR, and UW were homozygous for Gly581Ser, delThr610, and Leu237Pro, respectively. Patient KT, TM, and VE were compound heterozygotes of Leu237Pro/delG1067, Val333Glu/?, and Leu216Arg/Val363Asp, respectively. Details of the mutation analysis on patients other than FA have been described elsewhere (11, 13, 17, 19, 23).

Kinetic analysis of expressed wild-type and mutant HCSproteins.

We determined the kinetic properties of the wild-type and seven mutant HCS proteins using apo-CCP as a substrate. Table 1 shows the Km values for biotin and the Vmax values of the expressed HCS proteins. Two mutations, Gly581Ser and delThr610, were within the putative biotin-binding region (aa448–701), and the Km values of these mutant enzymes for biotin were higher than that of the wild type (Gly581Ser: 45 times, delThr610: 3 times). The Km values of the mutant HCS proteins containing Arg183Pro, Leu237Pro, Val333Glu, Leu216Arg, and Val363Asp, all of which were located outside the biotin-binding region, were not elevated. However, the Vmax values of all seven mutant HCS proteins were markedly decreased (percentage of the wild-type value): Arg183Pro 1.7%, Leu216Arg 0.6%, Leu237Pro 1.2%, Val333Glu 1.9%, Val363Asp 3.7%, Gly581Ser 2.4%, delThr610 7.6%. Western blot analysis showed that the expression levels of all mutant HCS proteins were essentially the same as that of the wild type protein (data not shown).

Table 1 Kmfor biotin and Vmaxvalues of expressed mutant HCS proteins The pCAGGS vector containing the wild-type or mutant cDNA was transfected into SV40-transformed fibroblasts from patient KT. Incorporation of biotin into apo-CCP was measured at a series of various biotin concentrations. * Mean and [range of values]. †n, Number of experiments performed.
Full size table

Effect of biotin supplementation of the culture media on PCCactivity.

To determine the effect of biotin concentration on the biotinylation activity of mutant HCS proteins, we assayed PCC activity in cultured cells grown in various biotin concentrations (Fig. 2). PCC activity was very low in the basal media (biotin concentration 10 nM), except in fibroblasts of patient HR who is homozygous for delThr610 mutation. The Km of this mutant HCS determined by the expression study was moderately higher than that of the wild type enzyme and the Vmax was decreased to 7.6% of the wild type value, which is the highest among the mutant HCS proteins analyzed in this study (Table 1). This mutant protein exerted enough HCS activity to keep PCC activity within the normal range under our basal culture conditions. In lymphoblasts of AD who is homozygous for Gly581Ser mutation, the PCC activity was increased to the control value at high level biotin, although the activity was very low in the basal media. On the other hand, PCC activity could not be normalized in fibroblasts of FA, UW, or KT even at 100 000 nM biotin. This was also the case with the PCC activity in lymphoblasts of UW and KT. The maximal PCC activities in fibroblasts of the patients decreased in the order FA (65% of the control cell value), UW (20%), KT (10%). The maximal PCC activities in lymphoblasts of UW and KT were about 9% and 7% of the control cell value, respectively. FA and UW are homozygous for Arg183Pro and Leu237Pro, respectively, whereas KT is a compound heterozygote of Leu237Pro and delG1067. The Vmax of the Arg183Pro mutant was higher than that of the Leu237Pro mutant (Table 1), and the delG1067 mutation has been shown to abolish the HCS activity totally (13). Thus, cells that have mutant HCS proteins with lower Vmax showed less maximal PCC activity. The biotin concentrations greater than 1 000 nM failed to increase PCC activity in fibroblasts of FA, UW, and KT and in lymphoblasts of UW and KT, whereas the maximal activity was reached at a biotin concentration of 10 000 nM in the lymphoblasts of AD. This observation is compatible with the data that the Km for biotin of the Gly581Ser mutant found in AD was much higher than that of HCS proteins from FA, UW, and KT.

Figure 2
figure 2

PCC activity in lymphoblasts and fibroblasts cultured in media containing various biotin concentrations. The average activity (233 pmol/min/mg protein in lymphoblasts, 228 pmol/min/mg protein in fibroblasts) in control cells cultured in the basal medium (biotin: 10 nM) was defined as 100%, and each activity was represented by a percentage of this. Error bars depict SD (n= 4). Significance was assessed using t test. ** indicate p< 0.01. No significant differences in the activity were found among the values at biotin concentrations of 1 000 nM, 10 000 nM, or 100 000 nM in FA, UW or KT.

Full size image

Expression of a series of N-terminal deletion mutants of HCS.

To evaluate the importance of sequence lying outside the biotin-binding region of HCS, we expressed HCS deletion mutants lacking the N-terminal region (Table 2).

Table 2 Biotin incorporation activity for N-terminal deletion mutants of HCS proteins Each deletion mutant was expressed using SV40-transformed fibroblasts from patient KT. Incorporation of biotin into apo-CCP was measured at 1500 nM biotin. * Mean and [range of values]. †n, Number of experiments. ‡ ND, not detected.
Full size table

The activity of HCS lacking Met1 to Met58 was less than 40% of the wild-type value, whereas the activity of HCS lacking Met1 to Ile117 was greater than that of wild-type value by 40%. Deletion up to Met234 abolished the majority of the catalytic activity. These data suggest that the amino acid sequence starting between Ile117 and Met234 is essential for HCS to exert catalytic activity and that the N-terminal sequence up to Met117 may be necessary for defining substrate specificity and/or regulation of this enzyme.

DISCUSSION

We have identified several mutations located within and outside the putative biotin-binding site of HCS (11, 13, 16, 17, and this study). The Gly581Ser and delThr610 mutations are located in the biotin-binding region. The Km values for biotin of the Gly581Ser and delThr610 mutant proteins were 45 and 3 times higher, respectively, than the wild-type value. The present and earlier (13) studies suggest that mutations within the putative biotin-binding region result in elevated Km values (Km mutants). The mutations Arg183Pro, Leu216Arg, Leu237Pro, Val333Glu, and Val363Asp were located outside the biotin-binding region. The Km values of these mutants differed only slightly from that of the wild type, whereas the Vmax values were extremely lower than that of the wild type (non-Km mutants). The result suggests that non-Km mutants can be a cause of biotin responsive HCS deficiency, thereby confirming our hypothesis (13). Recently, Dupuis et al. evaluated biotinylation of CCP by expressing plasmids containing either Km mutant or non-Km mutant in an E. coli strain mutated in the Bir A gene. The authors observed that both types of mutant HCS proteins showed biotin responsiveness (18). The results are in agreement with our present study.

PCC activity in the cells with mutant HCS proteins containing Arg183Pro or Leu237Pro was increased to varying degrees when the concentration of biotin in the medium was increased. The mechanism of the apparent biotin responsiveness in this culture system may be as follows. In the case of wild-type HCS, the biotin level in the basal media is still below the level for HCS to exert full activity. However, at this concentration of biotin (10 nM), the velocity of biotinylation is sufficient to maintain normal amounts of holocarboxylases, i.e. all apocarboxylases synthesized are converted to holoenzymes. Thus, no increase in the carboxylase activity can be obtained when the biotin concentration is increased from the basal to a higher level. On the other hand, in cells expressing normal-Km/low-Vmax type HCS proteins, the rate of biotinylation at 10 nM biotin is not high enough to maintain all carboxylases biotinylated (ref.24 and this study). The velocity of biotinylation with the mutant enzymes increased to some extent when we increased the biotin concentration from 10 nM to 1000 nM, and would most likely be maximized in this biotin concentration range because the Kms of these mutants are normal. In fact, biotin concentrations from 1000 nM to 100,000 nM failed to increase PCC activity in the cells from patients FA, UW, or KT. The maximal PCC activity of fibroblasts decreased in the order HR (homozygous for delThr610), FA (homozygous for Arg183Pro), UW (homozygous for Leu237Pro), and KT (heterozygous for Leu237Pro and delG1067). The Vmax value of the mutant enzymes determined in the expression study decreased in the order delThr610, Arg183Pro, Leu237Pro, and delG1067 (Table 1 and Ref.13). Thus, the maximal PCC activity obtained after increasing the biotin concentration in the culture medium was related to the Vmax of the HCS and its gene dosage. The correlation of the Vmax with the extent of restoration of PCC activity suggests that the Vmax of mutant HCS is an important factor for responsiveness of HCS deficient patients to biotin therapy.

Patients KE (delC2279/IVS10n+5(g→a)) and FE (delT1876/IVS10n+5(g→a)) showed relatively later-onset of HCS deficiency (5 and 11 mo, respectively), and partial clinical and biochemical responses to high dose biotin therapy. The IVS10n+5(g→a) mutation causes exon skipping and therefore results in a decreased amount of normal HCS mRNA and protein (18). The expression study has shown that delC2279 and delT1876 mutations abolish HCS activity (17); the biotinylation activity of these patients would be dependent on the normal HCS protein derived from the allele with IVS10n+5(g→a). Since these patients virtually have only normal HCS protein (i.e. normal Km), these patients are the proof that biotin responsiveness can be observed in patients who express the normal-Km/low-Vmax type mutant enzymes.

The physiologic biotin concentration of human serum was estimated to be 1–3 nM and that of liver 0.5–4 nM (25). More recently, the serum biotin level was estimated to be about 0.24 nM (26). These concentrations are lower than the Km values of wild-type HCS that were measured in cultured fibroblasts (8, 10) or lymphoblasts (10), using CCP or PCC as a substrate. Therefore, the mutant HCS proteins harboring the Leu216Arg, Arg183Pro, Leu237Pro, Val333Glu, or Val363Asp would not show full activity in the physiologic state. Biotin treatment increases the biotin concentration in serum up to several hundred nanomolars (5, 10), thereby helping to increase the velocity of biotinylation in the cells.

The importance of the domain outside the biotin-binding region of HCS for enzymatic activity was examined in our expression study with HCS deletion mutants. The N-terminal amino acid region up to Ile117 had a much less effect on the activity than the amino acid sequence starting between Ile117 and Met234. The data are consistent with the results of an immunosupression study (27) and the fact that the mutations identified so far are located beyond amino acid residue Arg183 (11, 13, 14, 17, 23).

The Vmax of the HCS with the delThr610 mutation was the highest among mutant proteins analyzed in this study, and the PCC activity in fibroblasts of patient HR (homozygous for delThr610) was within the normal range in basal medium (Ref. (15) and this study). As pointed out by Suormala et al., data from patient HR reveal the pitfalls of overlooking slight to moderate changes in HCS activity when carboxylase activities were measured in cells cultured under standard conditions (∼10 nM biotin) to evaluate HCS. When cells from a patient such as HR are cultured under standard conditions, carboxylase activity may fall into the normal range, failing to reveal an HCS abnormality. Our direct HCS assay method is useful in such cases.

By comparing the clinical pictures and kinetic properties of mutant HCS, we can speculate possible relationship between a kinetic parameter of mutant enzyme and biotin responsiveness in HCS deficiency. Assuming that both alleles expressed equally, expected Vmax in the cells decreased in the order HR (delThr610 + delThr610), AD (Gly581Ser + Gly581Ser), VE (Leu216Arg + Val363Asp), UW (Leu237Pro+ Leu237Pro), KT (delG1067+ Leu237Pro). Patients HR and AD responded excellently to biotin and the outcome was very fine. On the other hand, patient KT showed occasional organic aciduria and mental retardation with 40–100 mg/day of biotin administration. The reduction of Vmax seems to be an essential factor for pathophysiology and prognosis of HCS deficiency under treatments with large amounts of biotin.

Although the number of the samples is small, we observed a good correlation between genotype and phenotype in HCS deficiency. Thus, the determination of genotype in patients with HCS deficiency can be valuable for characterizing clinical phenotype.