V490M, a common mutation in 3-phosphoglycerate dehydrogenase deficiency, causes enzyme deficiency by decreasing the yield of mature enzyme.

A deficiency of 3-phosphoglycerate dehydrogenase (PHGDH) is a disorder of serine biosynthesis identified in children with congenital microcephaly, seizures, and severe psychomotor retardation. We report here the identification of the 1468G-->A (V490M) mutation of this gene in two siblings of an Ashkenazi Jewish family, providing further evidence that the V490M mutation is a common, panethnic cause of this deficiency. Using a novel, DNA-based diagnostic test, the mutation was not detected in 400 non-Jewish controls; one heterozygote was found among 400 persons of Ashkenazi Jewish ethnicity. Extensive biochemical studies were undertaken to characterize the effect of this mutation on enzyme activity, turnover, and stability. The V490M PHGDH yielded less than 35% of the activity observed for the wild-type enzyme when overexpressed by transient transfection or when comparing the endogenous activity in fibroblast cells from the patients with controls. Immunoblotting studies showed a comparable reduction in the level of immunoreactive PHGDH in cells expressing the mutant enzyme. Pulse-chase experiments with metabolically labeled PHGDH indicated that this resulted from an increased rate of degradation of the mutant enzyme following its synthesis. Thermolability analyses of mutant and wild-type enzyme activity revealed no significant differences. While others have proposed that the V490M mutation decreases the V(max) of the enzyme, we conclude that this mutation impairs the folding and/or assembly of PHGDH but has minimal effects on the activity or stability of that portion of the V490M mutant that reaches a mature conformation.

A deficiency of 3-phosphoglycerate dehydrogenase (PHGDH) is a disorder of serine biosynthesis identified in children with congenital microcephaly, seizures, and severe psychomotor retardation. We report here the identification of the 1468G3 A (V490M) mutation of this gene in two siblings of an Ashkenazi Jewish family, providing further evidence that the V490M mutation is a common, panethnic cause of this deficiency. Using a novel, DNA-based diagnostic test, the mutation was not detected in 400 non-Jewish controls; one heterozygote was found among 400 persons of Ashkenazi Jewish ethnicity. Extensive biochemical studies were undertaken to characterize the effect of this mutation on enzyme activity, turnover, and stability. The V490M PHGDH yielded less than 35% of the activity observed for the wild-type enzyme when overexpressed by transient transfection or when comparing the endogenous activity in fibroblast cells from the patients with controls. Immunoblotting studies showed a comparable reduction in the level of immunoreactive PHGDH in cells expressing the mutant enzyme. Pulse-chase experiments with metabolically labeled PHGDH indicated that this resulted from an increased rate of degradation of the mutant enzyme following its synthesis. Thermolability analyses of mutant and wild-type enzyme activity revealed no significant differences. While others have proposed that the V490M mutation decreases the V max of the enzyme, we conclude that this mutation impairs the folding and/or assembly of PHGDH but has minimal effects on the activity or stability of that portion of the V490M mutant that reaches a mature conformation.
A deficiency of the enzyme 3-phosphoglycerate dehydrogenase (PHGDH 1 ; EC 1.1.1.95) was identified as an inborn error of metabolism associated with congenital microcephaly (MIM 601815) by Jaeken et al. in 1996 (1). Siblings from a consanguineous family of Turkish origin were found to have abnormally low concentrations of serine and, to a lesser extent, glycine, in their cerebrospinal fluid. Fibroblasts from the probands displayed decreased activity of PHGDH (22 and 13% of control), the first enzyme in the de novo biosynthesis of serine from carbohydrates. Similar findings have since been reported on a second (unrelated) Turkish family (2), a Moroccan family (3), and another European family of undefined ethnicity (4). In addition to microcephaly, major clinical findings in patients with this disorder include severe psychomotor retardation, postnatal growth retardation, and intractable seizures. Magnetic resonance imaging of the brain shows cerebral atrophy and abnormal myelination (1)(2)(3)5). Importantly, replacement therapy with serine can result in some neurological and developmental benefits, such as improved seizure control (reviewed in Ref. 6) and increased central nervous system myelination (5).
A reduced capacity to synthesize L-serine has potentially serious consequences for cellular metabolism. Serine is incorporated directly into proteins and is a precursor for the de novo biosynthesis of glycine, cysteine, and the nonstandard amino acid, selenocysteine. It is also essential for the biosynthesis of phosphatidylserine and sphingolipids. Serine can be converted to pyruvate in the liver and kidney, with subsequent utilization for gluconeogenesis, lipogenesis, or energy production. L-serine is also the precursor of D-serine (7,8), an enantiomer prominent in the brain, where it functions as a neuromodulator of the strychnine-insensitive N-methyl-D-aspartate receptor (9 -12). Finally, serine serves as the major intracellular source of onecarbon units in the cell, through the action of L-serine hydroxymethyltransferase. The products of this reaction, 5,10methylenetetrahydrofolate and glycine, are critical metabolites involved in folate and bile acid metabolism and in the de novo biosynthesis of creatine, porphyrins, thymidylate, and purine nucleotides. Glycine also functions as a neurotransmitter in the central nervous system.
Serine is classified as a nutritionally nonessential amino acid because it can be synthesized de novo, from both glucose and glycine. PHGDH is the first enzyme in a widely distributed, cytoplasmic pathway that synthesizes serine from the glycolytic intermediate D-3-phosphoglycerate (reviewed in Refs. 13 and 14). PHGDH catalyzes the oxidation of this substrate to 3-phosphohydroxypyruvate, with NAD ϩ required as a cofactor. Transamination of 3-phosphohydroxypyruvate with glutamate * This work was supported by operating grants from the Canadian Institutes of Health Research (CIHR; Grant MT-14065) and the Manitoba Health Research Council (MHRC), studentships from the Children's Hospital Foundation of Manitoba and the MHRC (to J. M.), and a New Investigator Award from the CIHR (to S. P.). The PHGDH antibody was generated in and subsidized by the Canadian Genetic Diseases Network Immunoprobes Facility (managed by J. A. W.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
In the final step, dephosphorylation of 3-phosphoserine by phosphoserine phosphatase produces L-serine. Only the final reaction in this pathway is irreversible. In some cells, serine can also be produced from glycine (13). This process requires the enzymes of the glycine cleavage complex, found in mitochondria, and serine hydroxymethyltransferase, an enzyme with one isoform localized in the cytoplasm and another in mitochondria. The mitochondrial isozyme may operate reversibly, producing either serine or glycine, depending upon metabolite levels and the presence of the glycine cleavage system. In contrast, the cytoplasmic isozyme appears to operate in one direction only, producing serine (15)(16)(17). While the relative contributions of these alternate synthesis pathways to serine homeostasis remain to be determined, the neurodevelopmental phenotype of patients with a deficiency of PHGDH highlights the importance of that pathway in the central nervous system. The sequence of the cDNA encoding human PHGDH (4,18), its localization to the 1p12 (19) or 1q12 (4) pericentromeric region of chromosome one, and the identification of two mutations causing PHGDH deficiency (4) have recently been reported. Klomp et al. (4) determined that a mutation in the Moroccan family resulted in a substitution of the amino acid valine at position 425 with methionine (V425M). The remainder of the families that have been described have a V490M substitution (4). The distribution and autosomal recessive inheritance pattern of these mutations support the conclusion that they are disease-causing. However, following synthesis in vitro, these mutations caused only modest reductions in PHGDH activity. These reductions were attributed to a reduced V max of the mutant enzymes, whereas the K m did not differ from the wild-type in this system. To reconcile the dramatic phenotype observed in the patients with the modest effects of these mutations on enzyme activity, the authors speculated that in addition to the lower catalytic turnover, the mutations may result in decreased expression of the enzyme in vivo. Measurements of PHGDH mRNA levels in fibroblasts showed no differences between patients and controls (4); other approaches to investigate reduced enzyme activity, such as measurements of protein levels, turnover, and stability were not reported.
In this paper, we report the independent identification of the V490M mutation in two siblings born to nonconsanguineous parents of Ashkenazi Jewish ethnicity from New England. Detailed characterization of the enzyme demonstrated an ϳ70% reduction of enzyme activity and a similar reduction of immunoreactive protein in cells expressing V490M PHGDH. In contrast to the work of Klomp et al. (4), we find that these reductions resulted from an increased degradation of the mutant protein, probably due to less efficient folding and/or assembly following synthesis. Our results provide a molecular explanation for the low levels of PHGDH and serine observed in our patients and confirm a role for this enzyme deficiency in the genesis of a neurodevelopmental phenotype of congenital microcephaly, severe psychomotor retardation, and seizures.

EXPERIMENTAL PROCEDURES
Clinical Samples-The probands, male and female siblings with congenital microcephaly, severe global developmental delays, seizures, and spastic quadriparesis, were born to healthy, unrelated parents of Ashkenazi Jewish ethnicity. They were found to have low levels of serine in cerebrospinal fluid, as will be described in detail elsewhere. 2 Heparinized blood and skin fibroblasts were obtained with informed consent.
Leukocyte pellets were prepared from heparinized blood as described (20). Blood specimens were obtained through informed consent from 400 non-Jewish and 400 Ashkenazi Jewish individuals. Samples were included in the latter group if all four grandparents were of Ashkenazi Jewish ethnicity. Cell Culture-Cells were cultured at 37°C and 5% CO 2 using reagents from Invitrogen. HeLa and BHK-21 cells (American Type Culture Collection, Manassas, VA) and skin fibroblasts were maintained in DMEM (high glucose) containing 8% fetal bovine serum, 100 units/ml penicillin, and 100 g/ml streptomycin. Control fibroblasts (GM 08399A) were obtained from the NIGMS (National Institutes of Health) Human Genetic Mutant Cell Repository (Camden, NJ).
Isolation of cDNAs Encoding PHGDH-Overlapping expressed sequence tags encoding the entire PHGDH cDNA were identified by sequence similarity to the rat liver cDNA (21) using the Basic Local Alignment Search Tool (available on the World Wide Web at www.ncbi.nlm.nih.gov/BLAST/) (22). Primers based upon this sequence were used to amplify the human cDNA by reverse transcriptase-PCR. 100 ng of total RNA prepared from HeLa cells and proband fibroblasts (Trizol; Invitrogen) was reverse transcribed using 200 units of Superscript II RNase H Ϫ reverse transcriptase (Invitrogen) and 2.5 pmol of a primer complementary to the sequence 54 -74 nucleotides downstream of the predicted stop codon (5Ј-TCT CTC CCT ATT GAT CAC AGT GG-3Ј). The 5Ј PCR primer (5Ј-TTA GGT ACT TCT ACT CAC AGC GGC-3Ј) begins 47 nucleotides upstream of the predicted start codon. The 3Ј PCR primer (5Ј-CAA GTG GAT CCA GGT TAG AAG TGG AAC TGG AAG GC-3Ј) is complementary to the carboxyl-terminal region of the protein and contains a BamHI restriction site (in italics) added after the stop codon (underlined) for cloning purposes. Amplification conditions were as follows: 94°C for 5 min, followed by 36 cycles of 94°C for 45 s, 60°C for 45 s, and 72°C for 3 min 30 s, with a final 10 min at 72°C, using Pfu DNA polymerase (Stratagene). An ϳ1.6-kb product was purified (Qiaex II kit; Qiagen) from an agarose gel and digested with EagI and BamHI. This was then cloned into the pcDNA 3.1(Ϫ) expression vector (Invitrogen), which had been digested with NotI and BamHI. The resulting clones were assessed by dideoxy sequencing using T 7 DNA polymerase (Amersham Biosciences, Inc.).
Isolation of Genomic DNA-Tissue culture cell pellets and white blood cell pellets were incubated overnight at 50°C in 50 mM Tris-HCl, pH 7.5, 10 mM EDTA, 1% SDS, and 200 g/ml proteinase K (Roche Molecular Biochemicals). DNA was then purified by sequential extractions with phenol, phenol/chloroform/isoamyl alcohol (25:24:1), and chloroform/isoamyl alcohol (24:1), followed by precipitation with ethanol (modification of Ref. 23) Detection of the 1468G3 A Mutation-A PCR-based assay was developed to survey for the 1468G3 A mutation in genomic DNA. Preliminary analysis indicated the presence of an ϳ800-bp intron close to this mutation site. This region was cloned, and the exon/intron boundaries (between 1447 and 1448) were determined. 3 For detection of the 1468G3 A mutation, the 5Ј PCR primer (5Ј-CTC CGC TCA TTG CAC CTT TGA-3Ј) is situated in the intron, starting 115 bp upstream of 1448. The 3Ј primer (5Ј-ATG CTG CTT CCA CGC TTC CA-3Ј) is complementary to nucleotides 1553-1572. A 238-bp product was amplified from ϳ100 ng of genomic DNA using Taq DNA polymerase (Invitrogen) and the following amplification conditions: 94°C for 5 min, followed by 30 cycles of 94°C for 30 s, 65°C for 30 s, and 72°C for 30 s, followed by a final 10-min extension at 72°C. Aliquots were digested with Hsp92II (Promega), separated on 10% acrylamide gels, and stained with ethidium bromide. Fragments of 45 and 193 bp were observed for control DNA and 45, 57, and 136 bp if the 1468G3 A mutation was present.
Transient Expression and Enzyme Assays-The activity of the V490M PHGDH was compared with the wild-type enzyme following transient expression in BHK-21 cells. In some experiments, the PHGDH-encoding vectors were cotransfected with pSV-␤-galactosidase (Promega) in order to monitor the transfection efficiency. DNA-lipid complexes were prepared with 9 g of pcDNA-PHGDH (wild-type or V490M), or 7 g of pcDNA-PHGDH plus 2 g of pSV-␤-galactosidase and 20 l of LipofectAMINE 2000 in DMEM, according to the manufacturer's instructions (Invitrogen). 24 h following transfection, the cells were lysed by sonication in 25 mM Hepes, pH 7.1, 400 mM KCl, 1 mM DTT, and 0.2% Triton X-100 containing a mammalian cell proteinase inhibitor mixture (Sigma). The soluble fraction was obtained by centrifugation at 20,000 ϫ g for 10 min, and the protein concentration was determined using the Coomassie Plus reagent (Pierce). ␤-Galactosidase activity was measured using ortho-nitrophenyl-D-galactopyranoside as substrate (24). PHGDH was assayed in the direction of NADH oxidation by monitoring the decrease in absorbance at 340 nm. Assays were performed at 30°C in 25 mM Hepes, pH 7.1, 400 mM KCl, 1 mM DTT, 100 M NADH, and 50 M phosphohydroxypyruvate in a final volume of 1 ml (1,21). Absorbance measurements were taken every 5 s for 5-10 min using an Ultrospec 3000 spectrophotometer (Amersham Biosciences), and the initial slopes of the decrease in absorbance were determined. One unit of enzyme activity is defined as the amount that oxidizes 1 mol of NADH/min under these assay conditions. Corrections were made for transfection efficiency by dividing the specific activity of PHGDH in a dish of cells by a factor determined by dividing the activity of ␤-galactosidase in that same dish by the average ␤-galactosidase activity in all dishes on that day.
Expression and Purification of PHGDH-An NcoI restriction site was introduced into the pcDNA-PHGDH construct by site-specific mutagenesis (QuikChange™ kit; Stratagene) to change the A at position Ϫ1 of the PHGDH cDNA to a C. The forward primer used for mutagenesis was 5Ј-CCGAGGCCAACTCCAGCCATGGCTTTTGCAAATCTGCGG-3Ј, with the nucleotide change highlighted in italics and the resulting NcoI site underlined. The reverse primer for mutagenesis was complementary to the forward primer. A 212-bp XbaI/BstEII fragment that encompassed this region was removed from the mutagenesis product and used to replace this region in another aliquot of pcDNA-PHGDH that had not undergone mutagenesis. It was sequenced to ensure that only the desired mutation had been introduced. The entire PHGDH coding region was then removed from this vector by digestion with NcoI and BamHI and inserted into the same sites of the pTrc 99 A prokaryotic expression vector (Amersham Biosciences). The resulting plasmid was introduced into competent BL-21 CodonPlus™ bacteria (Stratagene) and grown in Terrific broth at room temperature until reaching an OD 660 of 0.5-0.6. Expression of PHGDH was induced with 1 mM isopropyl-1-thio-␤-D-galactopyranoside for a further 16 h at room temperature. Bacterial cells were collected by centrifugation and stored as a pellet at Ϫ80°C.
PHGDH was purified from the bacterial cells using differential centrifugation and column chromatography. Cells were lysed by sonication in 10 volumes of buffer 1 (25 mM Tris-HCl, 2 mM EDTA, 20 mM NaCl, and 1 mM DTT, pH 7.5, containing a bacterial cell proteinase inhibitor mixture (Sigma) and 1 mg/ml lysozyme). Following the removal of particulate matter by centrifugation, the lysate was mixed with 0.3 volumes of 5% streptomycin sulfate and clarified by centrifugation. A 30 -50% ammonium sulfate pellet prepared from the supernatant was solubilized in buffer 2 (25 mM Hepes, 150 mM NaCl, and 1 mM DTT, pH 7.5) and desalted using a column of Cellulofine GH-25 (Amicon) equilibrated in the same buffer. The excluded fraction from this column was then applied to a column of 5Ј AMP-Sepharose 4B (25). Following extensive washing with buffer 2, proteins were eluted with 1 mM ␤-NADH in buffer 2. Fractions containing protein were analyzed by SDS-PAGE and staining with Coomassie Blue R-250. Those containing PHGDH were concentrated and applied to a column of Sephacryl S-300HR (Amersham Biosciences) equilibrated in buffer 3 (25 mM NaH 2 PO 4 , 25 mM Na 2 HPO 4 , 100 mM NaCl, 1 mM EDTA, and 1 mM DTT, pH 7.2). Fractions containing purified PHGDH were pooled and stored at Ϫ80°C following the addition of glycerol to a final concentration of 50%. PHGDH prepared in this manner was greater than 95% pure (not shown).
Preparation of Monoclonal Antibodies to PHGDH-Monoclonal antibodies specific for the purified PHGDH were produced from BALB/c mice (26) by the Canadian Genetic Diseases Network Immunoprobes Facility (Winnipeg, Canada). Hybridomas that produced antibodies to PHGDH were identified by enzyme-linked immunosorbent assay, and positive clones were tested for their ability to detect PHGDH on immunoblots and to immunoprecipitate labeled PHGDH from HeLa cells. Clone 13D5 performed well in all of these assays.
Immunoblotting-Proteins were separated by SDS-PAGE in 10% minigels and transferred to 0.2-m nitrocellulose membranes for 2 h at 35 V and 4°C in 250 mM Tris-HCl, 192 mM glycine, and 15% methanol (modified from Ref. 27). Membranes were blocked for 1 h in phosphatebuffered saline containing 0.1% Tween 20 and 5% skim milk powder. The 13D5 antibody was diluted in this same buffer and incubated with the membranes overnight at 4°C. Following incubation with a horseradish peroxidase-linked goat anti-mouse secondary antibody, bound complexes were detected by enhanced chemiluminescence (SuperSignal West Dura; Pierce).
Radiolabeling and Immunoprecipitation-BHK cells were transfected with 9 g of pcDNA-PHGDH (wild-type or mutant) as described above. 24 h following transfection, cells were preincubated in Met/Cysfree DMEM (Invitrogen) for 30 min at 37°C. Cells were pulse-labeled for 20 min with 100 Ci/ml L-[ 35 S]Met/Cys (Ͼ1000 Ci/mmol; PerkinElmer Life Sciences) in the same medium. For chase samples, the labeling medium was replaced with complete DMEM containing 8% fetal bovine serum, 1 mM Met, and 1 mM Cys. Cell monolayers were washed twice with ice-cold phosphate-buffered saline and lysed with 1 ml of lysis buffer (25 mM Hepes, 150 mM NaCl, 0.2% Triton X-100, 0.2% digitonin, 1 mM DTT, pH 7.5, containing a mixture of proteinase inhibitors). After scraping from the plates, insoluble material was removed from the lysates by centrifugation at 17,000 ϫ g for 10 min. Supernatants were incubated overnight at 4°C with the 13D5 antibody preadsorbed onto protein G-agarose (Sigma). Immune complexes were washed six times with radioimmune precipitation buffer (25 mM Hepes, 150 mM NaCl, 1% Triton X-100, 1% deoxycholate, and 0.1% SDS, pH 7.5) and eluted and denatured with Laemmli sample buffer (28). Samples were separated by SDS-PAGE using 10% gels, fixed, dried, and exposed to film or a phosphorimaging screen (Bio-Rad). Images were analyzed using a Personal Molecular Imager FX and quantified using Quantity One software (Bio-Rad).

RESULTS
Characterization of the PHGDH cDNA-Based upon their similarity to the sequence of rat liver PHGDH (21), a series of overlapping expressed sequence tags was identified that predicted the entire cDNA encoding human PHGDH. Primers were designed to amplify the coding region and 47 nucleotides of the 5Ј-untranslated region of this cDNA from human HeLa cells using reverse transcriptase-PCR. A product of the expected size (ϳ1.6 kb) was obtained and cloned into the pcDNA 3.1 expression vector. The sequence obtained for the coding region of the HeLa cell PHGDH was identical to that predicted from the overlapping expressed sequence tags (not shown). It was 88% identical to the rat sequence through the coding region and predicted a 533-amino acid protein that was 94% identical to the rat enzyme. With the exception of a single base substitution (G 3 T) at position 75, our sequence was also identical to the sequence of PHGDH that was subsequently deposited into GenBank TM (accession number AF006043; Ref. 18). This substitution changes amino acid 25 to Asp rather than Glu as reported by Cho et al. (18). The same substitution is also present in our family members and in all of the expressed sequence tags we examined as well as the sequence recently reported by Klomp et al. (4).
Identification of the 1468G3 A Mutation-cDNAs encoding PHGDH were also cloned and sequenced from skin fibroblasts derived from the probands, using the strategy described in the preceding paragraph. The only difference noted in these samples was a 1468G3 A transition, resulting in a substitution of the Val at amino acid 490 with Met (V490M). This change also introduced an Hsp92II restriction enzyme site into the DNA, providing a rapid method to screen for this mutation. A 238-bp fragment that included one Hsp92II site in control samples and two sites if the 1468G3 A mutation was present, was amplified from genomic DNA (Fig. 1). Restriction digests of control DNA yielded bands of 193 and 45 bp (Fig. 1B, lane 2). In samples containing the 1468G3 A mutation, the 193-bp product was cleaved in two, producing 136-and 57-bp fragments. Lanes 4 and 5 show that the probands were homozygous for the 1468G3 A mutation, with only the 136-, 57-, and 45-bp fragments remaining following digestion. Samples from the parents (lanes 3 and 6) were heterozygous, since all four fragments were present following digestion. These results are consistent with an autosomal recessive mode of inheritance for this mutation.
Our results provide an independent identification of the V490M mutation and support its assignment as a diseasecausing mutation (4). Since eight of the first nine patients reported with this enzyme deficiency carry this mutation and come from differing ethnic backgrounds, it suggests that the 1468G3 A mutation is a common cause of this deficiency. To estimate the frequency of this allele in the general population, we evaluated 400 randomly selected, non-Jewish individuals. None were found with the 1468G3 A transition. Because the parents in our family were both Ashkenazi Jewish and had the same mutation, it was important to determine whether this mutation was present at a higher frequency in this population. Four hundred individuals with four Jewish grandparents were screened; one individual was found to be heterozygous for the 1468G3 A mutation. Based upon this finding, we estimated that the 95% confidence limits for the frequency of heterozygosity in this population would be 1/70 to 1/16,000 (True Epistat Software; Epistat Services, Richardson, TX) Effect of the V490M Mutation on Enzyme Activity-Transient transfection studies were employed to characterize the effect of the V490M substitution on the enzymatic activity of the mutant PHGDH in vivo. For these experiments, wild-type and 1468G3 A mutant cDNAs were transfected into BHK cells and assayed 24 h later. Fig. 2A shows that transfection of the cDNA encoding the wild-type PHGDH dramatically increased the enzyme activity observed in cells, as compared with the untransfected control. Transfection of the 1468G3 A mutant also increased the amount of enzyme activity in the cells, but much less so than the wild-type construct. The amount of enzyme activity recovered from replicate plates was very reproducible within a single experiment but did vary somewhat from day to day (not shown). Nevertheless, in six different experiments, the yield of PHGDH activity from the mutant construct was always less than 35% of that observed with the wild-type construct. A similar ratio was also found when the constructs were transfected into cells using calcium phosphate complexes or when the constructs were transfected into Chinese hamster ovary cells (not shown). These results indicate that the V490M mutation reduces but does not eliminate the amount of active enzyme produced in the transfected cells. This observation is consistent with the residual enzyme activity noted in patients' fibroblasts (Ref. 1; see Fig. 3) and following in vitro translation (4), confirming that this mutation is responsible for the enzyme deficiency and that the overexpression system reproduces the effects of the mutation.
Effect of the V490M Mutation on the Steady-state Concentration of the Enzyme-Theoretically, the V490M mutation could reduce the yield of PHGDH activity through a direct effect on reaction kinetics or through an indirect effect, by reducing the steady-state concentration of enzyme in the cells. Recent results from in vitro translation studies suggest that the mutation does not alter the K m of the mutant enzyme but rather decreases its V max (4). However, since accurate V max estimates are dependent upon knowing the exact amount of native enzyme assayed, confirmation of these results awaits further kinetic measurements using purified wild-type and mutant PHGDH. To evaluate the effect of the V490M mutation on the steady-state concentration of PHGDH in our transfected cells, equal amounts of protein from the cell lysates described in Fig.  2A were separated by electrophoresis and analyzed by immunoblotting with a monoclonal antibody specific for PHGDH (characterized further in Fig. 3B). Fig. 2B shows that the cells transfected with V490M PHGDH (lanes 6 -10) had significantly lower levels of immunoreactive PHGDH than the cells transfected with the wild-type enzyme (lanes 1-5). Similar results were observed in five other transfection experiments, indicating that the V490M mutation decreases the activity of PHGDH by reducing its steady-state concentration in the cell.
To confirm that the results observed following transient transfection reflected the fates of wild-type and V490M PHGDH when expressed at physiological levels, similar experiments were performed on skin fibroblasts obtained from a control individual and from the probands' family. Once again, lower levels of PHGDH activity were obtained in samples homozygous for the V490M mutation when compared with heterozygotes or the control fibroblasts (Fig. 3A, compare lanes 4  and 5 with lanes 1-3). Fig. 3B demonstrates that there was a correspondingly lower level of immunoreactive protein in the fibroblasts homozygous for the V490M mutation, when compared with the control fibroblasts or those from the parents (Fig. 3B, compare lanes 6 and 7 with lanes 3-5). The specificity

FIG. 2. Transient expression of wild-type and V490M PHGDH in BHK cells. BHK cells were transfected with plasmids encoding wild-type or V490M PHGDH and harvested following a 24-h incubation.
A, the graph shows the specific activities of PHGDH detected in lysates from cells transfected with the wild-type or V490M constructs. Values are the mean Ϯ S.D. of measurements made from five independently transfected plates of cells. Three plates of untransfected control cells were also assayed. Results were normalized for transfection efficiency as described under "Experimental Procedures." Typically, the efficiency of transfection was very similar between individual plates, and the normalization had a minimal effect on the results. B, 2 g of total protein from each transfected cell lysate in A was separated by SDS-PAGE, transferred to nitrocellulose, and visualized by immunoblotting with 0.1 g/ml 13D5. Lanes 1-5, plates transfected with wild-type PHGDH; lanes 6 -10, plates transfected with V490M PHGDH. of this antibody for PHGDH was confirmed by its reaction with purified PHGDH (lane 1) and a single, identically sized protein of ϳ55 kDa in the HeLa and control fibroblast samples (lanes 2  and 3). This is the first report of the PHGDH protein levels in patients' cells and provides additional support for the hypothesis that the V490M mutation decreases the steady-state concentration of PHGDH in cells.
By visual inspection, it appears that the levels of immunoreactive V490M PHGDH in Figs. 2 and 3 were reduced to a similar extent as the reductions in activity observed for the mutant enzyme. However, it is difficult to accurately quantify the chemiluminescence blots to verify this speculation. An alternative method of comparing these data is presented in Fig.  4. Five replicate plates were transfected with wild-type or V490M mutant PHGDH and assayed as described in Fig. 2. The immunoblot shown in Fig. 4 was prepared by loading equivalent units of enzyme activity per lane on the gel rather than equivalent total protein. The signals obtained from each of the lanes were very similar to one another; specifically, the signal from the wild-type samples (lanes 1-5) was indistinguishable from the mutant samples (lanes 6 -10). This shows that when equal units of enzyme activity were loaded on the gel, equal levels of immunoreactive PHGDH were observed. Taken together, these experiments indicate that the V490M mutation reduces the steady-state concentration of PHGDH in cells but has a much smaller effect on the initial rate kinetics of the mutant.
Effect of the V490M Mutation on Enzyme Stability and Turnover-Since the cellular level of a protein is governed by its rates of synthesis and degradation, we wished to determine whether the V490M mutation affected this balance. Previous work reported that the 1468G3 A substitution does not alter the rate of synthesis or stability of the mRNA encoding PHGDH (4), suggesting that the rate of protein synthesis is also not affected. In Fig. 5 we demonstrate, through pulsechase and immunoprecipitation analyses, that the V490M substitution increased the degradation of this mutant. Replicate plates of BHK cells transfected with constructs encoding wildtype or V490M PHGDH were labeled by a 20-min pulse of [ 35 S]Met/Cys and then chased for 0 -6 h in complete medium. PHGDH was immunoprecipitated from these cells, and quantified by phosphorimaging analysis following SDS-PAGE. Inspection of the autoradiograph indicated that a significant band representing PHGDH was present in all lanes transfected with the constructs (Fig. 5A, lanes 1-5 and 7-11) and was absent from the control (lane 6). It was also apparent that the labeling of the mutant protein (lanes 7-11) decreased much greater during the chase period than did the wild-type protein (lanes 1-5). There was a significant decrease in the labeling of both proteins during the first hour of the chase, after which the rate of decrease was much slower (Fig. 5B). Approximately 25% of the mutant protein remained following 6 h of chase, compared with ϳ70% of the wild-type protein.
Further experiments probed whether the V490M mutation decreased the stability of PHGDH, which could account for the increased rate of degradation. As an initial attempt to investigate this problem, the thermal inactivation profiles of wild-type and V490M PHGDH were compared. Cell lysates from the transfection experiments described in Fig. 4 were incubated at increasing temperatures for 5 min (Fig. 6A) or for increasing times at 45°C (Fig. 6B), and then the initial rates of enzyme activity remaining in the extracts were determined. The graphs show that the V490M PHGDH was inactivated to a similar extent as the wild-type PHGDH under all of the test conditions, with the inactivation curves following pseudo-first-order kinetics as a function of time. These results indicate that the active forms of V490M and wild-type PHGDH display a similar stability to thermal inactivation. These data further suggest that the increased degradation of the mutant enzyme observed in Fig. 5 did not arise due to stability differences between the "functional" forms of the wild-type versus the mutant enzyme.

DISCUSSION
The results presented in this report represent an independent confirmation of the 1468G3 A (V490M) mutation in PHGDH and affirms its role in the pathogenesis of enzyme deficiency (4). Our identification of the same mutation in the Ashkenazi Jewish population that was first described in the Turkish and European families suggests that this is a common cause of this deficiency. Thus, there is a strong rationale to screen for this mutation in any patients suspected to have a serine biosynthesis deficiency. The diagnostic test described in our report using genomic DNA offers a rapid and accurate means to this end. PHGDH. PHGDH-specific activities were measured in lysates prepared from BHK cells transfected with wild-type (five plates of cells, lanes [1][2][3][4][5] or V490M (five plates of cells, lanes 6 -10) PHGDH. 3000 microunits of PHGDH activity from each lysate were separated by SDS-PAGE, transferred to nitrocellulose, and immunoblotted, using 0.1 g of 13D5 as the primary antibody. The amount of total protein loaded in each lane ranged from 2.6 to 2.9 g for lanes 1-5 and 7.6 -8.3 g for lanes 6 -10. Possible mechanisms for how the V490M mutation reduces PHGDH activity can be proposed from our knowledge of this enzyme from other organisms. PHGDH from Escherichia coli is a 410-amino acid protein (29) whose three-dimensional structure has been solved (30). It is a homotetramer, with each monomer being made up of three distinct domains: a nucleotide-binding domain (residues 108 -294), a substrate-binding domain (residues 7-107 and 295-336), and a regulatory domain that binds serine (residues 337-410). The main contact points between the subunits are at the level of the nucleotide-binding domains and the regulatory domains. Sequence alignments comparing the enzyme from E. coli to other organisms indicate that the highest homology is found among the various nucleotide-and substrate-binding domains, whereas the regulatory domains are less conserved (4,21,31). 4 Indeed, the regulatory domains from most other organisms sequenced to date are ϳ100 -140 amino acids longer than that of the enzyme from E. coli (the human enzyme is 533 amino acids in length). The V490M mutation appears unlikely to influence the nucleotideor substrate-binding sites of the human enzyme directly, due to its location 43 amino acids from the carboxyl terminus. However, in E. coli, serine acts as an allosteric inhibitor of PHGDH by binding to the carboxyl-terminal regulatory domain and decreasing the V max of the active site, some 30 Å distant (30,32). A flexible Gly-Gly hinge region between the regulatory and the substrate-binding domains has been proposed to undergo a conformational change that influences the enzyme V max ; there is experimental evidence to support this hypothesis (33)(34)(35). Although the rat and human enzymes are not regulated by serine (21), 4 it is conceivable that the V490M mutation could cause a conformational change to decrease V max , as was observed following in vitro translation (4). However, our results, obtained following in vivo expression of the enzymes, do not support this model. We found that the principal effect of the V490M mutation was to decrease its steady-state concentration in cells. With the possible caveat that we examined soluble cell lysates and not purified protein, our results also indicate that when the levels of immunoreactive wild-type and mutant PHGDH were normalized, the initial rate kinetics of the mutant enzyme were very similar to the wild-type enzyme. In addition, the active forms of both enzymes displayed similar stability to thermal denaturation, suggesting that the V490M PHGDH was not less stable than the wild-type enzyme once it achieved a mature conformation. Thus, by all criteria that we measured, the mature, functional forms of the two enzymes were very similar. The differences in the conclusions reached in this report and by other investigators (4) are probably due to differences in the expression systems utilized. In vitro translations are performed at temperatures below 37°C, which can result in more efficient protein folding, and they are designed 4 S. Pind, unpublished results. to prevent rapid proteolysis, which could lead to accumulation of misfolded, inactive proteins.
Rather than affecting the stability or the activity of PHGDH, several lines of evidence suggest that the V490M mutation impairs the folding and/or assembly of the enzyme. Misfolding mutations are common and often result in degradation of the affected protein, reducing its steady-state concentration within the cell (for recent reviews, see Refs. 36 -38). The regulatory domain of E. coli PHGDH forms an extensive interface (ϳ1000 Å) for formation of the tetramers (30). Achouri et al. (21) have shown that removal of the carboxyl-terminal 209 amino acids of the rat enzyme lowers but does not abolish enzyme activity but does block the ability of enzyme dimers to form tetramers. The increased rate of degradation of the V490M mutant enzyme in the first hour following synthesis suggests that a lower percentage of the newly synthesized mutant enzyme achieved a "proteinase-resistant" or mature conformation, thus resulting in its degradation. This lower efficiency of folding results in the lower steady-state concentration that we noted in both the transfected cells and the fibroblast cells. We also observed that the mutant PHGDH could not be recovered from the soluble fraction when expressed in bacteria under any conditions that were tried. In contrast, more than 75% of the wild-type enzyme could be recovered in the soluble fraction using this expression system (results not shown). Taken together, our results suggest that the V490M mutation lowers the efficiency of folding of the mutant enzyme into its mature, enzymatically active conformation but does not reduce the activity of any mutant enzyme that does fold correctly. Precedents for this type of mutation include the cystic fibrosis transmembrane conductance regulator, where the ⌬F508 deletion prevents the folding of this mutant, but following purification and reconstitution in vitro the mutant Cl Ϫ -channel functions very similarly to the wild type (39). In addition, the R147W mutation of short-chain acyl-CoA dehydrogenase has been reported to inhibit folding of the enzyme but not affect the stability or the activity of that portion of the enzyme that does fold properly (37).
Clinical Implications-It is particularly striking that the neurodevelopmental phenotype of the PHGDH deficiency is so severe, given that the two known amino acid substitutions in that enzyme are relatively conservative and the mutant proteins characterized to date retain significant residual activity. The central nervous system appears to be particularly sensitive to the low concentration of serine that results from a deficiency in its production through the PHGDH pathway. The severity of the neurological and developmental impairments in children with this deficiency indicates that the production of serine through alternate pathways, such as proteolysis or through the glycine-cleavage complex and the reverse reaction of serine hydroxymethyltransferase, is not adequate to meet the requirements of the brain. Whether this is due to a lower capacity of the alternative pathways, low expression in this organ, or an inadequate supply of the glycine precursor is not known. In addition, although serine could be provided from other tissues in the body or from the diet, studies in rats have shown that it is transported across the blood-brain barrier relatively inefficiently (40). Serine transport by the neutral amino acid carrier is limited at normal plasma concentrations of amino acids, since the transporter prefers the larger, more hydrophobic amino acids (40 -42).
The importance of, if not requirement for, serine in specific cell types in the central nervous system has been documented using several model systems. For example, neurons may be dependent upon an exogenous supply of serine. Serine promotes the morphological differentiation of chicken dorsal root ganglion neurons in vitro (43). Hippocampal neurons cultured in the absence of exogenous serine or glycine showed a greatly diminished capacity to synthesize phosphatidylserine and sphingolipids (44). Further studies have shown that serine is released from rat astroglia-rich cultures and is a trophic factor for the survival and growth of cultured neurons (45)(46)(47). These in vitro studies are supported by recent immunolocalization results, showing that PHGDH is not expressed in Purkinje neurons in the rat cerebellar cortex but is highly expressed in the Bergman glia, a native astroglia in this region (47). More extensive results have been reported for mouse brain, where PHGDH is expressed highest in the radial glia/astrocyte lineage and in the olfactory ensheathing glia (48).
Conclusions-We have characterized a common, panethnic mutation in PHGDH that reduces its activity in cells. In contrast to the work of others, we noted that the reduction of enzyme activity associated with the V490M mutation was due to an increased turnover of the mutant PHGDH, most likely as a consequence of impaired protein folding and/or assembly. These results provide a genetic and biochemical explanation for the origin of the disease in our probands and verify that the low levels of serine observed are the result of a deficiency of this enzyme. It remains to be determined how low serine concentrations impair function in the central nervous system. Possible mechanisms include global limitations on protein and/or lipid synthesis, insufficient levels of the neurotransmitters glycine and D-serine, and the absence of serine as a trophic factor for neuronal growth, migration, and survival. In addition, the roles of radial glial cells and excitatory amino acids in neuronal development (49,50) may well be impaired in a state of serine insufficiency. While resolution of these possibilities requires more study, it remains important to identify patients with deficiencies in serine biosynthesis, since they may benefit from replacement therapy.