Main

HCS (EC 6.3.4.10) is an enzyme responsible for attaching biotin to mammalian mitochondrial carboxylases including pyruvate carboxylase, PCC, and methylcrotonyl-CoA carboxylase and to cytosolic ACC(1, 2). Inherited deficiency in HCS has been identified and shown to be responsible for early onset MCD. This is in contrast with biotinidase deficiency, which is the late onset form of MCD(2). The activities of these biotin-dependent carboxylases are important in gluconeogenesis, protein catabolism, and fatty acid synthesis(2). Decreases in their activities lead to metabolic ketoacidosis, hyperammonemia, and excretion of abnormal metabolites in MCD patients(2). Most patients with HCS deficiency manifest symptoms that include tachypnea, seizures, difficulties in feeding, and dermatitis in the early infantile period. In HCS patients mortality can occur without treatment because of severe metabolic ketoacidosis: however, clinical and biochemical symptoms are dramatically improved with pharmacologic doses of biotin(2). Specifically, although most patients have shown a good clinical response to daily administration of 10-20 mg of biotin, some patients only partially respond to biotin therapy and continue to excrete abnormal metabolites in their urine, even while receiving high doses of biotin(3, 4).

Studies designed to clarify the mechanism of biotin responsiveness have been carried out. The primary biochemical defect of HCS deficiency has been shown in a cohort of HCS patients using apo-PCC partially purified from biotin-starved rats(5). The kinetic properties of HCS in the fibroblasts of seven patients demonstrated that the Km for biotin ranged from 3 to 70 times the normal value, and theVmax decreased to various degrees(6). Another study showed that HCS Km values from a patient were 500 times that of controls by measuring holo-PCC formation using mitochondria-derived apo-PCC from a patient(7). Based on these and other studies, it has been proposed that an increasedKm for biotin was responsible for biotin responsiveness and that the age of onset was negatively correlated with the Km for biotin.

We and others have isolated the human cDNA encoding HCS(8, 9). The cDNA encodes a 726-amino acid polypeptide that displays homology with BirA, which is a biotin ligase inEscherichia coli(8, 10, 11). In one affected individual we identified a T to C substitution at nucleotide position 997 (L237P) in one allele and in another allele we identified a G nucleotide deletion at 1067 (delG 1067) that resulted in premature termination at amino acid 280(4, 8). However, neither the HCS activity of the mutant protein nor the kinetic properties of mutant HCS has been examined yet. We recently established a new HCS assay method based on measuring the incorporation of [3H]biotin into recombinant apo-CCP ofE. coli, a subunit corresponding to the catalytic domain of ACC in higher organisms(12). The assay method enabled us to perform kinetic analysis on HCS from patient's fibroblasts. In this study, we identified a HCS mutation in a patient and determined the kinetic properties of HCS proteins encoded by three different mutant genes expressed in a patient's fibroblasts to examine whether the mutant HCS proteins show abnormalities in biotin affinity.

METHODS

Patients. Patient KH was a Japanese girl whose parents were cousins. The patient was born as small for date at 39 wk and she developed normally. At 1 y, 8 mo, she was admitted to a hospital with symptoms of frequent vomiting, respiratory distress, and intermittent losses of consciousness(13). Blood gas analysis showed pH 7.076, Po2 111.5 mm Hg. Pco2 13.0 mm Hg, and base excess -24.0. The acidosis was improved by treatment with an infusion containing glucose, electrolytes, and sodium bicarbonate. There were frequent episodes of vomiting and metabolic acidosis accompanied with upper tract infection; however, her condition was improved by transfusion of glucose and sodium bicarbonate. When she was 6 y, 4 mo old, she was diagnosed as having multiple carboxylase deficiency by the analysis of urine organic acids and was treated with 10 mg/d of oral biotin. After receiving biotin, she never had an acidosis attack and the abnormal urine metabolites completely disappeared. She developed without mental retardation.

Clinical manifestation of patients MT [patient 4a(4)] and UW [patient 1(4)] are reported elsewhere(4, 14). HCS activity in lymphoblasts from patients MT and UW was less than 6% of normal control values by the method of monitoring the rate of holoenzyme synthesis using apo-PCC from HCS-deficient cells as the substrate(15). The genotypes of patients MT and UW have been analyzed, and patient MT was found to be a compound heterozygote for the L237P and delG 1067 and patient UW was a homozygote for the L237P mutation(4).

Measurement of HCS activity. To measure HCS activity, we used a method with apo-CCP, E. coli ACC, as a substrate(12). The apo-CCP was produced by a method described previously(12) except that the expressed protein was purified by a His-Tag purification system following the manufacture's recommendations (Novagen, Madison, WI). To measure HCS activity, incorporation of [3H]biotin into acid-insoluble material per unit time was monitored. In a standard assay, holo-CCP formation was performed in 20 μL of a mixture that contained HCS preparation (5 μL, 10-20 μg of protein), 100 μM apo-CCP, 125 mM Tris-HCl (pH 8.0), 8 mM MgCl2, 50 mM KCl, 1.4 mM EDTA, 3 mM ATP, 2.5 mM mercaptoethanol, and 750 nM [3H]biotin for lymphoblasts, or 1500 nM for transfected cells (specific activity = 14 Ci/mmol). The mixture was incubated at 30°C for 30 min. The reaction was stopped by adding 250 μL of 10% (wt/vol) trichloroacetic acid. The mixture was kept on ice for 15 min, then centrifuged at 18 000 × g for 15 min. The resultant pellet was washed two times with 250 μL of 10% trichloroacetic acid. Finally, the pellet was dissolved in 200 μL of 0.1 N NaOH and transferred to a scintillation vial. After adding 5 mL of scintillation cocktail (ACS-II, Amersham, Buckinghamshire, UK), radioactivity was measured in a Beckman LS-6500 scintillation counter. The HCS activity was calculated by subtracting nonspecific biotin binding to CCP at the same biotin concentration (zero time reaction). The protein concentration was determined by the method of Bradford(16).

For determination of the Km value for biotin, a serial dilution of [3H]biotin was made, and HCS assays were carried out. TheKm values of HCS for biotin and Vmax values were determined using double reciprocal plots.

Sequencing of cDNA. mRNA was isolated and cDNA was synthesized as described elsewhere(8). The entire coding region of HCS cDNA was amplified by nested PCR with four pairs of oligonucleotide primers in two overlapping fragments. The DNA was subcloned into theXho I-EcoRI site of pBlue-script II KS + (Stratagene, La Jolla, CA) and sequenced on an automated laser fluorescent (A.L.F.) sequencing apparatus (Pharmacia Biotech, Uppsala, Sweden).

Site-directed mutagenesis and transfection. The HCS cDNA was subcloned into pCAGGS, a mammalian expression vector that has a cytomegalovirus enhancer and a chicken β-actin promoter(17). The L237P was inserted by digestion withBbs I as described previously(4). The delG 1067 and a mutation detected in the patient KH (V550M) were generated using anin vitro mutagenesis kit (Clontech, Palo Alto, CA). cDNAs were excised from pBluescript KS + by digestion with XhoI and subcloned to pCAGGS. All mutagenized cDNAs were sequenced using the automated sequencing apparatus.

The wild-type or mutant HCS cDNA in the pCAGGS vector was transfected into SV40-transformed fibroblasts from patient MT using LipofectAMINE reagent according to the manufacturer's recommendations (Life Technologies, Inc., Gaithersburg, MD). Mock transfections of the SV40-transformed fibroblasts with parent pCAGGS vector were used as controls. Six micrograms of plasmid DNA were transfected with 30 μL of LipofectAMINE reagent per 90-mm dish. Two micrograms of pCAGGS-β-galactosidase, a vector containing the E. coli β-galactosidase cDNA, was cotransfected as an internal standard. The transfected cells were harvested by scraping 48 h after the transfection, washed twice with PBS, and stored at -80°C until use.

Measurement of PCC and β-galactosidase activities. To determine whether expressed HCS was capable of attaching biotin to apo-PCC in transfected cells, PCC activity in the cell lysates was measured by a radioactive assay entailing the fixation of [14C]bicarbonate(18). Lysate buffer [50 mM Tris-HCl (pH 8.0), 3 mM EDTA, and 2.5 mM reduced glutathione] was used instead of cell extracts as a blank assay. β-Galactosidase activity was measured using a β-galactosidase enzyme assay kit (Promega, Madison, WI). Biotin concentration in the medium of the transfected cells was determined as described previously(19).

RESULTS

HCS activities in lymphoblasts of patients. The onset of the disease, HCS activity of lymphoblasts of normal individuals and patients, and genotype of patients are shown in Table 1. The HCS activity from lymphoblasts derived from patient KH was 16% of that from normal controls. whereas activities of patient UW and MT were 4% and 1.6%, respectively.

Table 1 Onset of the disease, HCS activity, and genotype of control and patients' lymphoblasts
Full size table

Determination of mutations. Sequence analysis of the HCS cDNA extracted from lymphoblasts of patient KH revealed a G to A transition at position 1935 in all clones analyzed. The G to A transition converts a valine(GTG) codon to a methionine (ATG) codon (V550M) (Fig. 1). No other mutations were detected in the DNA sequencing analysis of this HCS cDNA. The G to A transition creates a new NlaIII site. PCR amplification of patient KH's and the patient's parents genomic DNA followed by NlaIII digestion indicated that the parents were heterozygous for the V550M mutation and the patient was homozygote for the mutation (data not shown). The V550M was the same mutation recently identified in three MCD patients(20).

Figure 1
figure 1

Nucleotide sequencing of cDNA from patient KH. A G to A transition at position 1935 was detected in five different cDNA clones(V550M), which converts valine (GTG) to methionine (ATG).

Full size image

Transient expression and kinetic analysis of expressed protein. To test whether the L237P, delG 1067, and V550M were the mutations responsible for decreased HCS activity, we transiently expressed the various mutant HCS genes in SV40-transformed fibroblasts from patient MT. Table 2 shows the HCS, PCC, and β-galactosidase activities in the transfected cells. HCS activity in cells transfected with the wild-type cDNA was about 100 times higher than that of normal fibroblasts. The mean activity in cells transfected with the V550M or L237P cDNA was 16 and 2.3 pmol/min/mg of protein, respectively. The mean activity in cells transfected with the delG 1067 cDNA was 33 fmol/min/mg of protein, which was similar to the value obtained in the mock transfection. There was no significant difference inβ-galactosidase activity observed between dishes. PCC activities were next measured to determine the ability of the mutant HCS proteins for attaching biotin to apo-carboxylase. Because apo-carboxylases have been shown to accumulate in an HCS-deficient patient(15, 21), PCC activity would be expected to increase if the mutant HCS protein could attach biotin to accumulated apo-PCC. To examine this possibility, cells were transfected with the V550M and L237P mutants, and the increased PCC activity was compared with that of mock transfection. Levels of PCC activity were then correlated with those of HCS activity. Western blot analysis of transfected fibroblasts revealed that the size and amount of the mutant enzyme expressed in the cells transfected with the V550M and the L237P were similar to those of the wild-type enzyme. However, the protein was barely detectable in cells transfected with the delG 1067 (data not shown).

Table 2 Enzyme activities in fibroblasts transfected with normal and mutant constructs
Full size table

Endogenous HCS activity in the transfected fibroblasts was 9 fmol/min/mg of protein when assay was carried out in 1500 nM of biotin concentration. In contrast, HCS activity in fibroblasts transfected with the wild-type cDNA was 15000 times higher than endogenous activity. The genotype of the cells used for transfection was a heterozygote for the L237P and delG 1067 mutation. Transfection of the L237P mutant into normal fibroblasts showed that the missence mutation did not affect HCS activity (data not shown), which suggested that the endogenous mutant protein does not interfere with the enzymatic activity of the expressed wild-type or mutant cDNA. To characterize the expressed mutant HCS, kinetic properties were examined using crude cell lysates from fibroblasts transfected with the V550M and the L237P mutants(Fig. 2). The Km for biotin of the V550M mutant was 6.5 times the normal control and the Vmax value was 17% of the value obtained from cells transfected with the wild-type cDNA. In contrast, the Km for biotin of the L237P mutant was 1.2 times the normal control (173 nM) and the Vmax was 4.3% of the value obtained from cells transfected with the wild-type cDNA.

Figure 2
figure 2

Kinetic study of wild-type and mutant HCS. The pCAGGS vector containing wild-type (A). L237P or V550M (B) cDNAs was transfected into SV40-transformed fibroblasts from patient MT. Incorporation of biotin into apo-CCP was measured at the indicated biotin concentrations. Inserts are double reciprocal plot representations of the same experiments. (C) The Km and Vmax values were determined from the double reciprocal plots shown in (A) and (B). The data represented here are from a single determination.

Full size image

DISCUSSION

In this study, we transiently expressed wild-type and mutant HCS cDNAs in transformed fibroblasts to examine and compare their kinetic properties. In fibroblasts transfected with the V550M mutant, an elevated Km value for biotin was observed supporting the assumption that elevatedKm values for biotin are related to biotin responsiveness in patients(5–7). HCS shows homology toE. coli BirA biotin ligase(8) and yeast biotin apoprotein ligase(20, 22). The BirA DNA binding and biotin-binding domains have been determined by x-ray crystallography(23). Homology to BirA suggested that the V550M mutation is located within the putative biotin-binding site(8, 20). Dupuis et al.(20) have identified three additional mutations within the biotin-binding region of the HCS gene in a genetic screen of nine biotin-responsive MCD patients. This clustering of mutations within the biotin-binding domain is highly suggestive that mutations in the putative biotin-binding site may affect affinity for biotin. In contrast, theKm value of the L237P mutant was similar to that found for the wild-type cDNA. In this regard, amino acid sequence comparison of L237P and the BirA gene indicates that the mutation is outside the putative biotin and ATP binding sites(8). Although the function of the region where the L237P mutation occurs in the HCS gene has not been clarified, the mutation may decrease the catalytic activity without changing the affinity for biotin. For instance, this region of the molecule may be important for catalytic activity because it mediates the association of HCS with carboxylases.

There are at least two possible mechanisms to explain biotin responsiveness in patients with HCS deficiency: 1) the mutant HCS has a decreased affinity for biotin (i.e. Km mutant) and 2) biotin may restore the stability to the mutant HCS protein. In this regard, we found that the Km value for biotin of the L237P mutant was essentially the same as the Km for the wild-type cDNA. In addition, the Vmax value for the L237P mutant was found to be 4.3% that of the wild-type cDNA. To determine whether high concentrations of biotin could restore enzymatic activity of the L237P mutant we measured the PCC activity of lymphoblasts cultured in media containing various concentration of biotin from patients UW (L237P/L237P) and KM(V550M/V550M). There was no significant difference in PCC activity over a 24-h period when patient UW's lymphoblasts were cultured in media containing 0.1, 1, 10, and 50 μM biotin. In contrast, lymphoblasts from patient KH increased from 14% of control at time zero to 19% of control after 24-h incubation in medium containing 1 μM biotin (data not shown). Our results show that the Km value of the L237P mutant was not elevated, and high doses of biotin were not able to stabilize the L237P mutant protein.

The frequency of the L237P mutation in Japanese patients appears to be high, and all of these patients were found to respond to biotin treatment(4). HCS activity was not observed in our expression studies with the delG 1067 cDNA, consistent with the mutant being a null allele. Therefore in L237P/delG 1067 heterozygotes and in the L237P/L237P homozygote, the L237P allele is likely responsible for biotin responsiveness. The question arises of why these patients responded to biotin treatment even though no significant increase in HCS activity was observed in lymphoblasts from patients cultured in high concentration of biotin. The concentration of biotin is between 0.8 and 3.0 nM in normal human plasma(24), 0.2-0.7 nM in erythrocytes(25), and 0.8-0.2 nM in whole blood(25). After receiving 20 mg of biotin per day, the plasma biotin concentration increases to 4830 nM(6). The plasma and intracellular biotin concentration may be lower than theKm value that HCS has for biotin that was measured in normal cultured fibroblasts(6, 12) or lymphoblasts(6). In this regard, it has been reported that PCC activity in lymphocytes from normal adults increases 2-3-fold with daily administration of 20 mg of biotin(26). These data suggest that HCS in normal cells does not show full activity under conditions of biotin intake obtained from a normal diet. It is possible that in patient UW (L237P/L237P), if mutant HCS does not show full activity without biotin administration, an increase in HCS activity would be obtained with biotin treatment. Further analysis of mutant HCSs will be necessary to confirm the hypothesis that the residual activity of mutant HCS is sufficient for biotin responsiveness.

The HCS activity of lymphoblasts from patient UW (L237P/L237P) was twice that from patient MT (L237P/delG 1067). The activity from cells transfected with the delG 1067 cDNA was the same as that for the mock transfection. Therefore, the delG 1067 mutant, which is a deletion resulting in a premature termination at amino acid 280 and truncation of the putative biotin-binding domain, is likely a null mutation. Moreover, in Western blotting analysis of the transfected delG 1067 cDNA, only low levels of the protein were detectable, indicating the protein is not stable. In five Japanese patients the delG 1067 mutant accounts for 3/10 of the mutant alleles. However, all of the patients with this allele are compound heterozygotes; two patients are heterozygous for of delG 1067 and L237P and one patient is a heterozygous for del G1067 and an unknown mutation(4) (our unpublished data). These data raise the possibility that homozygosity for the null mutation may be lethal in utero.