Overexpression of the normal phosphoribosylpyrophosphate synthetase 1 isoform underlies catalytic superactivity of human phosphoribosylpyrophosphate synthetase.

To define the enzymatic and genetic basis of X-linked phosphoribosylpyrophosphate synthetase (PRS) catalytic superactivity, we measured concentrations of X-linked PRS1 and PRS2 isoforms in cultured fibroblasts and lymphoblasts by immunoblotting after separation by polyacrylamide-urea isoelectric focusing. PRS1 comprised >80% of measurable PRS isoforms in all fibroblast strains, but PRS1 concentrations in cells from six affected males exceeded those in normal cells by 2-6-fold. PRS absolute specific activities (activity per mg of PRS isoforms) were comparable in all fibroblast strains and in purified recombinant normal PRS1, confirming selectively increased levels of PRS1 isoform as the enzymatic basis of PRS catalytic superactivity. Cloning, sequencing, and expression of normal subject- and patient-derived PRS cDNAs predicted normal translated region sequences for both PRS isoforms and revealed no differences in catalytic properties of recombinant PRS1. Normal and patient PRPS1 transcribed but untranslated DNA sequences were also identical. Northern blot analysis showed selective increase in relative concentrations of PRS1 transcripts in patient fibroblasts. In PRS catalytic superactivity, overexpression of the normal PRS1 isoform thus appears to result from an altered pretranslational mechanism of PRPS1 expression. In lymphoblasts, however, expression of this alteration is attenuated, explaining the absence of phenotypic expression of PRS catalytic superactivity in these cells.

In a second and, in our experience, more frequently encountered class of PRS superactivity, regulation of enzyme activity by nucleotide inhibitors is normal, as are affinities for substrates and activators (Mg 2ϩ , P i ) (16 -19). Increased maximal reaction velocity (V max ) is the only identifiable kinetic alteration in the PRS of such individuals (catalytic superactivity) (20). Although indirect evidence for structural alterations in PRS has been reported in catalytic superactivity (20,21), specific structural defects in a PRS isoform have not been documented. In fact, studies supporting a role for translated region mutation in catalytic superactivity of PRS were carried out before recognition of the existence of multiple PRS isoforms encoded by separate PRPS genes (11)(12)(13)(14)(15). We have carried out studies aimed at assessment of the enzymatic and genetic basis of catalytic superactivity of PRS. The results of these studies indicate that, in contrast to PRS superactivity associated with defective allosteric regulatory properties, inherited catalytic superactivity is unassociated with alteration in the translated sequences of either PRS1 or PRS2 cDNA. Rather, catalytic overactivity of PRS appears to reflect increased intracellular concentrations of the normal PRS1 isoform. Moreover, accompanying increases in levels of PRS1 transcript with entirely normal sequence in cells from affected individuals suggest derangement of a pretranslational mechanism regulating the expression of PRPS1 in catalytic superactivity of PRS.

EXPERIMENTAL PROCEDURES
Cell Lines-Fibroblast strains were initiated from skin biopsies obtained from five normal individuals and affected male patients in six families with PRS catalytic superactivity (16,17,19). Fibroblasts were propagated in monolayer in Eagle's minimal essential medium containing 10% fetal bovine serum, 2 mM L-glutamine, and nonessential amino acids. B lymphoblast lines, derived from peripheral blood mononuclear cells exposed to Epstein-Barr virus lysates (21), were initiated and propagated from two normal individuals and two of the unrelated males in whom PRS catalytic superactivity was demonstrable in erythrocytes, lymphocytes, and fibroblasts (16,19,21). The growth medium was RPMI 1640 containing 10% fetal bovine serum and 2 mM L-glutamine. Conditions for growth of fibroblast and lymphoblast cultures were as described by Becker and co-workers (3,21). Fibroblast cultures were studied in late log phase of growth (corresponding to light monolayer confluence), and lymphoblast cultures were studied during log phase growth at cell densities ranging from 0.5 to 1.5 ϫ 10 6 cells/ml. Prior to extraction of total RNA for reverse transcription and PRS cDNA amplification or for Northern blot analysis, cells were washed twice in ice-cold serum-free medium. Procedures for harvesting cells and for preparing cell-free crude, dialyzed, and chromatographed extracts for PRS activity determinations and PRS isoform analyses were as described by Becker and co-workers (3,21). The extraction buffer was 8 mM sodium phosphate, 1 mM dithiothreitol, 1 mM EDTA (pH 7.5).
PRS Activity and Isoform Analyses-PRS activities were determined by a two-step procedure previously described in detail (20). Studies of purified recombinant human PRS1 and PRS2 have identified differences between these structurally similar isoforms (95% amino acid sequence identity) with respect to substrate and activator affinities, purine nucleotide inhibitor responsiveness, pH optima, specific catalytic activities, and stability of activity upon dilution (13). In the current studies, PRS assays were carried out (except where specifically stated) at pH 7.5 in the presence of saturating substrate concentrations (500 M MgATP; 350 M ribose 5-phosphate), 5.0 mM MgCl 2 , and 32.0 mM P i . Dilution of recombinant PRS1 or PRS2 was made in a stabilizing buffer at pH 7.5 containing 50 mM Tris-HCl, 0.3 mM ATP, 6.0 mM MgCl 2 , 1.0 mM sodium phosphate, and 1 mg/ml bovine serum albumin. Activities of PRS are expressed in units, where 1 unit is defined as 1 mol of PRPP formed per min at 37°C.
A difference in pIs of human PRS1 and PRS2 (8) was exploited in the development of an isoelectric focusing (IEF)-immunoblotting procedure for separation and quantitation of the isoforms in human cell extracts. Supernatant layers of fibroblast and lymphoblast extracts (5-150 g of protein), transformed bacterial cell extracts (1.5-4.0 g of protein), and purified human recombinant PRS1 and PRS2 (2-60 ng) were electrofocused at 23°C (600 -800 V for 1 h and then 1000 V for 15 h) on pretreated (500 V; 30 min) 6% polyacrylamide, 9.2 M urea IEF gels containing 0.6 ml of ampholytes (0.2 ml of pH 5-8; 0.4 ml of pH 3.5-10) per 10 ml of gel solution. After electrophoretic separation, proteins were electroblotted (MilliBlot-SDE transfer system, Millipore Corp., Milford, MA) onto polyvinylidene difluoride membranes, which were fixed, blocked, and incubated for 1 h each with rabbit anti-human PRS IgG (protein concentration, 6.13 g/ml, in a dilution buffer of 20 mM Trissaline, pH 7.6, with 0.1% Nonidet P-40 and 1% nonfat dry milk) and then horseradish peroxidase-linked donkey anti-rabbit Ig (Amersham Life Science Inc.), diluted 5000-fold in dilution buffer. The membranes were washed in dilution buffer, immersed in ECL Western blot detection reagent (Amersham Life Science Inc.), and exposed to x-ray film for 2-5 min before to development of the film. The intensity of bands identified in samples of cell extracts with mobilities corresponding to those identified for authentic purified recombinant PRS1 and PRS2 isoforms were quantified on a Molecular Dynamics (Sunnyvale, CA) computing densitometer. Cell extract band densities were related to those of the respective purified recombinant PRS isoform and to the amount of protein in the sample applied to the IEF gel, permitting determination of the concentration of each isoform in a corresponding sample. Protein concentrations in all studies were determined by the method of Lowry et al. (22). Combining PRS activity and isoform content measurements allowed comparison of the specific activities of PRS isoforms (absolute specific activities) in cell extracts with those measured for the purified recombinant PRS1 and PRS2 isoforms (13).
Preparation, Sequencing, and Expression of PRS cDNA-Total cellular RNA was isolated from 1-2 ϫ 10 7 fibroblasts and from 5-10 ϫ 10 7 lymphoblasts (23), and reverse transcription-polymerase chain reaction (PCR) amplification (24) of PRS cDNA was carried out as described (10), resulting in amplification of the full translated regions of both PRS1 and PRS2 cDNA. Sequencing of amplified PRS1 and PRS2 cDNA was performed directly from PCR mixtures, using a modification of the dideoxynucleotide chain termination method (25). The entire translated regions of both strands of PRS1 and PRS2 cDNA were sequenced with the specific amplification primers and a series of PRS consensus sequencing primers that hybridized to PRS1-and PRS2-coding sequences internal to the PCR amplification primers (10).
Preparation of PRS1 cDNA for expression in an Escherichia coli strain was with a nested primer PCR procedure recently reported for the introduction of the translated region of PRS1 cDNA into the expression vector pSPRBS (8). After transformation of E. coli strain DH5 and confirmation of the sequence of both strands of the human PRS1 cDNA-translated and adjacent regions by standard dideoxy sequencing (26), recombinant human PRS1 expression was induced in expression construct-transformed E. coli BL21 (DE3/pLysS) (27) by addition of isopropylthiogalactoside, under conditions previously described (10). Recombinant human PRS1, extracted from bacterial host cells (13), were analyzed by electrophoresis on 12% SDS-polyacrylamide gel electrophoresis slab gels and on polyacrylamide-urea IEF gels and were assayed for PRS activities all as described (8,10). Purification of several of the recombinant PRS1 to Ͼ98% homogeneity was carried out, also as described (13).
PRPS1 Genomic DNA Sequencing-Genomic DNA was isolated from two normal fibroblast strains and five strains derived from individuals with PRS catalytic superactivity (1-2 ϫ 10 7 cells each) (26). Sequential PCR amplifications, utilizing a nested primer approach when necessary, were used to prepare PRPS1 genomic DNA segments inclusive of the 5Ј-and 3Ј-transcribed but untranslated regions of the gene (14,15,28). 2 The amplification primers utilized are listed in Table I, along with their respective locations relative to the PRS1 cDNA translation initiation site. Sequencing of the 137-and 122-bp 5Ј-DNA segments extending from the PRPS1 transcription initiation sites (28) to the translation initiation codon was carried out on both DNA strands from all strains, utilizing appropriately oriented primer sequences (Table I) from the proximal PRPS1 promoter region and exon 1 of PRPS1 (28). The 997-bp 3Ј-untranslated segments of the transcribed region of two normal and two patient cell-derived PRPS1 genomic DNAs were sequenced by a series of appropriately oriented oligonucleotide primers (Table I) (14,28), except that primers PRPS1-3, PRPS1-8, PRPS1-25, and R-11 incorporate previously unreported genomic DNA sequences. b ϩ, sense strand; Ϫ, antisense strand.
PRS1 and PRS2 Transcript Levels-Steady state levels of PRS1 and PRS2 transcripts were estimated by Northern blot analysis after electrophoresis (4 h at 150 V) of samples of fibroblast and lymphoblast total RNA on 1% agarose-formaldehyde denaturing gels and transfer of RNA to nitrocellulose membranes (26). After prehybridization for 4 h at 42°C, membranes were hybridized at 42°C for 18 h with oligo-32 Plabeled human PRS1 cDNA (2.3 kilobase pairs) or PRS2 cDNA (2.7 kilobase pairs) (or with both probes together) and with a human glyceraldehyde-3-phosphate dehydrogenase cDNA probe (1.8 kilobase pairs). Specific radioactivities of labeled PRS1 and PRS2 cDNA were identical when used to probe a single filter or duplicate filters. Blots were then washed at suitable stringency (26), and radioactivities in the regions of the membrane corresponding to PRS and control transcripts were quantified on a PhosphorImager (Molecular Dynamics) (12 h) before exposure to x-ray film for 24 -72 h at Ϫ70°C. Values for PRS1 and PRS2 transcript levels in a cultured cell total RNA sample are expressed relative to the glyceraldehyde-3-phosphate dehydrogenase transcript level measured in that sample.

PRS Activities in Cell
Extracts-PRS activities in extracts of fibroblasts derived from six individuals with purine nucleotide and uric acid overproduction and previously described PRS catalytic superactivity (16,17,19) exceeded the mean value of PRS activity in cell extracts from five normal individuals by 2.0 -5.0-fold (Table II). Each patient-derived strain exhibited a constant relative increase in PRS activity over the entire range of P i tested (0.2-50 mM), and comparable differences among strains were demonstrable in crude as well as dialyzed and Sephadex G-25-chromatographed extracts. Patient and normal PRS activities were indistinguishable with regard to Michaelis constants (K m ) for MgATP, ribose 5-phosphate, and Mg 2ϩ ; apparent activation constants for P i ; and inhibitory constants (I 0.5 ) for ADP and GDP (tested in chromatographed extracts, as described previously (8,20)). Thus, normal and patient PRS differed only in maximal reaction velocities, confirming previously reported observations (17,19,20) defining the class of PRS catalytic superactivity in fibroblasts. PRS activities in lymphoblasts derived from patients TB and AD were modestly but consistently increased (1.4 -1.6-fold) relative to those in normal lymphoblasts (Table III). Increased V max was again the only identifiable difference in kinetic constants.
PRS Isoform Levels in Fibroblast and Lymphoblast Extracts-Separation and quantitation of PRS1 and PRS2 isoforms in extracts of cultured cells were achieved by immunoblotting of polyvinylidene difluoride membranes to which extract proteins were transferred after polyacrylamide-urea IEF gel electrophoresis. In preliminary studies, samples containing known amounts of highly purified recombinant normal human PRS1 and PRS2 (13) were separated and immunoblotted, and the bands corresponding to the respective isoforms were measured on a computing densitometer. These studies confirmed quantitative entry of the proteins into the gel as well as the range of the linear relationship between the amounts of each isoform applied and total densities of the immunospecific bands (Fig. 1). Both recombinant isoforms consistently appeared as doublet bands in this system, with additional bands (comprising up to 15% of total density) appearing when larger amounts of recombinant PRS1 were applied (Fig. 2). The sums of the densities of the bands were used in estimating quantities of the respective isoform. In initial measurements of cellular PRS isoform contents, equivalent quantities of protein from the supernatant layers of fibroblast and lymphoblast extracts were applied to the IEF gel along with a series of purified recombinant PRS1 and PRS2 standards. In subsequent fibroblast studies, appropriate dilutions of individual extracts were made (in extraction buffer containing 1 mg/ml bovine serum albumin) to permit the respective band densities to fall in the linear range of the standard curves. (Curves resulting from serial dilutions of extract samples and recombinant PRS isoform standards were nearly identical.) Bands identified in samples of all cell extracts corresponded in mobilities to those observed with purified recombinant normal PRS1 and PRS2.
A representative IEF-immunoblot analysis of fibroblast PRS isoforms is shown in Fig. 2A. Extracts of normal and patient fibroblasts contained substantially higher concentrations of PRS1 than PRS2 (Table II). Fibroblast PRS1 concentrations in extracts derived from patients with PRS catalytic superactivity consistently exceeded those from normal individuals, and PRS2 concentrations in normal and patient fibroblast extracts were comparable. For each normal and patient cell strain, fibroblast PRS activity corresponded closely with total PRS isoform concentration, so that calculated absolute specific activities of PRS isoforms (milliunits/mg PRS isoforms) were quite similar in patient and normal fibroblast extracts (Table II). These calculated specific activities also agreed closely with the reported specific activity of highly purified recombinant normal human PRS1 (25.1 units/mg protein (13)). These findings provide strong evidence that PRS superactivity in fibroblasts from these patients results from a selective increase in levels of the PRS1 isoform and do not support the idea (20) that PRS1 in the cells of these patients display increased activity per immunoreactive enzyme molecule.
For each lymphoblast line, PRS2 isoform concentration (Fig.  2B) comprised a greater proportion of total PRS isoform concentration than in fibroblasts (Table III), permitting more accurate quantitation of PRS2 levels in cell extracts when samples of equivalent protein content were tested. Normal and patient lymphoblast lines contained comparable PRS2 isoform  concentrations in each of four determinations carried out in extracts of cells harvested over a range of log phase culture growth from 0.5-1.5 ϫ 10 6 cells/ml. PRS1 isoform concentrations in these extracts differed, however, with consistent increases, averaging 1.7-and 2.0-fold, respectively, in extracts of cells derived from patients TB and AD (Table III). When PRS absolute specific activities were calculated from the results of PRS activity and isoform assays, values for normal, TB, and AD PRS were similar, approaching those of purified recombinant normal PRS1. As is the case for fibroblasts, the increments in PRS activities in TB and AD lymphoblasts, although of lesser magnitude, are explainable by selective increases in PRS1 isoform concentrations in these cells.
PRS1 and PRS2 cDNA Sequences and Expression of Recombinant PRS1 Isoforms-The 954-bp translated regions of PRS1 and PRS2 cDNA derived from fibroblast and lymphoblast total RNAs of five normal individuals and four patients with catalytic superactivity (TB, SS, AD, ZB) were sequenced directly from PCR pools. The sequences of both PRS1 and PRS2 cDNA derived from patients were identical with those of the respective normal PRS cDNA (14,15). This finding contrasts with the demonstration, by means of the identical reverse transcription-PCR amplification and sequencing strategies used here, of single base substitutions in the PRS1 cDNA derived from patients with allosteric regulatory defects in PRS (8,10).
Because the results of PRS isoform analyses suggested that selectively increased PRPS1 gene expression provides the basis of PRS catalytic superactivity, normal and patient PRS1 cDNA were cloned into pSPRBS, and both the plasmid and recombi-nant derivatives were used to transform E. coli BL21 (DE3/ pLysS). During induction with isopropyl-1-thio-␤-D-galactopyranoside, recombinant normal and patient PRS1 cDNA were expressed, resulting in high levels of PRS activity in bacterial cell extracts and the appearance of a 34.5-kDa band on SDSpolyacrylamide gel electrophoresis, identified as human PRS1 by immunoblot analysis (8). Recombinant human PRS1 comprised 4 -10% of total bacterial cell protein in the supernatant fraction of bacterial cell lysates. Based on enzyme activities and on the proportion of total bacterial lysate protein represented by recombinant PRS1 isoforms, provisional estimates of the absolute specific activities of recombinant normal and patient PRS1 were made, indicating comparable values. When recombinant normal and patient (TB and AD) PRS1 were purified to Ͼ98% homogeneity, the specific activities of each of the purified recombinant enzymes were nearly identical (recombinant normal PRS1, 23.4 -26.7 units/mg protein; TB, 27.2 units/ mg; AD, 24.1 units/mg). Together, these studies provide evidence against the view that catalytic superactivity of PRS results from mutation in the protein-encoding regions of PRPS1 and favor an alternative possibility, increased expression of the normal PRPS1 gene product.
PRS1 Transcript Sequence-Segments of the PRPS1 gene including the 5Ј-transcribed but not translated region and the 3Ј-untranslated portion of the last exon (exon 7) were amplified by PCR from normal and patient fibroblast genomic DNA and sequenced. Identical sequences corresponding to those previously reported (28) were confirmed for the gene segment inclusive of both previously demonstrated transcription initiation sites and extending to the initiation codon (Ϫ141 to Ϫ1) in two  3, 6, and 7) when increased amounts of PRS1 are applied. Normal PRS2 migrates as a doublet (bracket) with pIs near 6.6. normal and five patient genomic DNA preparations. Similarly, the 997-bp terminal portions of exon 7 and the adjacent 3Ј-DNA, extending from the translation termination codon (955-957) to 33 bp beyond the polyadenylation signal sequence (1928 -1933), were identical in the two normal and two patient (TB and AD) fibroblast genomic DNAs studied. In conjunction with the normal translated region sequence demonstrated in PRS1 cDNA from patients with PRS catalytic superactivity, these findings support the view that increased expression of the PRS1 isoform in this disorder is unlikely to result from primary alteration in the 2.1-kilobase portion of the 2.3-kilobase mature PRS1 transcript encoded by the PRPS1 gene.
PRS Transcript Levels-Northern blot analyses of PRS transcripts in total RNAs extracted from cultured cells of normal individuals and patients with PRS catalytic superactivity are shown in Fig. 3. The relative abundance of 2.7-kilobase PRS2 transcripts is substantially greater in lymphoblasts than in fibroblasts, but within each cell type normal and patient cells contain comparable steady state levels of this transcript relative to glyceraldehyde-3-phosphate dehydrogenase mRNA (Table IV). In contrast, relative levels of 2.3-kilobase PRS1 transcripts in patient fibroblasts exceeded those in normal fibroblasts by 1.7-6.6-fold (Table IV). In addition, a nearly coordinate relationship between the increase in relative PRS1 mRNA abundance and the relative increment in PRS1 isoform content and enzyme activity was demonstrable in each patient fibroblast strain. More modest but similarly selective increases in PRS1 transcript levels were also found in lymphoblasts from patients TB and AD (Table IV). As was the case in fibroblasts, relative increases in PRS1 transcript levels in these cell lines correlated well with relative increases in PRS1 isoform concentrations. DISCUSSION These studies establish the enzymatic basis of human Xlinked PRS catalytic superactivity as selective overexpression of the PRS1 isoform. In six fibroblast strains from unrelated individuals in whom PRS catalytic superactivity is associated with uric acid overproduction and accelerated fibroblast PRPP and purine nucleotide synthesis, the concentrations of PRS1 (but not PRS2) substantially exceeded those in normal fibroblast strains. Moreover, in all normal and patient fibroblast strains, PRS activities and total immunoreactive PRS isoform concentrations were closely related, so that calculated PRS absolute specific activities were comparable and were, in fact, very similar to the specific activities of highly purified recombinant normal PRS1 measured here and previously reported (13). Three inferences may be drawn from these findings: (a) the immunoblotting procedure appears to provide quantitative or nearly quantitative measurements of PRS isoforms in cell extracts; (b) with PRS1 constituting the great majority of total PRS isoform content in fibroblasts, the nearly uniform absolute specific activities calculated for PRS in normal and patient cells indicates that PRS catalytic superactivity must derive from increased concentrations of PRS1 with normal or near normal maximal reaction velocities rather than, as previously suggested (20), from PRS1 with increased catalytic activity per immunoreactive molecule; and (c) the comparable values of PRS absolute specific activities calculated for fibroblast strains and measured for recombinant normal PRS1 make it unlikely that the molecular basis for PRS superactivity involves a discrepancy between normal and patient PRS1 in binding to an effector of PRS activity, such as the inhibitory PRS-associated protein described in rat liver (29).
The enzymatic basis of both PRS catalytic superactivity and PRS superactivity associated with defective allosteric regulation of enzyme activity is overexpression of the PRS1 isoform, but the genetic defects underlying these classes of inherited enzyme overactivity are distinct. Defects in purine nucleotide inhibition and P i activation of PRS1 are the consequences of single amino acid substitutions reflecting point mutations in the translated sequence of PRPS1 (8,10). In contrast, the allosteric properties of PRSs in cells of patients with catalytic superactivity are normal, and the translated sequences of PRS1 and PRS2 cDNA derived from fibroblasts of affected  individuals are also normal. In addition, both crude and highly purified preparations of recombinant PRS1 expressed in E. coli transformed with PRS1 cDNA derived from these patients display normal catalytic and allosteric properties, in distinction to the aberrant allosteric properties identified in recombinant PRS1 expressed from the PRS1 cDNA of patients with point mutations in the PRS1 cDNA coding sequence (8). These data are consistent with the view that excessive expression of PRS1 activity in the fibroblasts of patients with PRS catalytic superactivity results from increased concentrations of the normal PRS1 isoform.
The level of regulation of PRPS1 expression altered in inherited PRS1 superactivity remains to be defined. We have shown evidence, however, that the sequence of the mature 2.3-kilobase PRS1 transcript is identical in normal fibroblasts and in fibroblasts from two of the affected individuals. Moreover, by Northern blot analysis, selective increases in the PRS1 transcript were demonstrable in the fibroblast total RNA of the five patient strains tested, and overall there was a close relationship between relative increases in levels of PRS1 transcript and in PRS1 activity and isoform contents. The results of these studies thus support the contention that a pretranslational defect (or defects) in regulation of PRPS1 expression underlies PRS catalytic superactivity. Whether such a defect acts through increases in rates of PRPS1 transcription or through altered PRS1 transcript processing or stability and whether, in fact, PRS1 catalytic superactivity reflects mutations in the PRPS1 locus or, alternatively, in a gene modifying PRPS1 expression are issues requiring additional studies aimed at defining the regulation of expression of this gene.
Prior studies of PRS transcript levels (30,31) have established organ-and cell type-specific differences in the expression of PRPS genes. The current measurements of X-linked PRS transcripts and isoforms extend these observations to a substantially greater relative contribution of PRPS2 to total PRPS gene expression in normal lymphoblasts than in normal fibroblasts. Not only do PRS2 transcript levels exceed those of PRS1 transcript in normal lymphoblasts but also PRS2 constitutes 40% of total PRS isoforms in these cells, compared with Ͻ20% in normal fibroblasts. Of greater interest in the current context, however, is the apparent attenuation of selective overexpression of PRS1 in the lymphoblasts of patients (TB and AD) whose PRS superactivity was substantially greater (4-and 3-fold, respectively) in fibroblasts. Although consistent but low level overexpression of PRS1 was detectable in the lymphoblasts of these patients (1.4-and 1.6-fold, respectively), as reported by Losman et al. (21), rates of PRPP and purine nucleotide synthesis were normal. It seems likely that the normal metabolic phenotype of the lymphoblasts from these patients reflects the low level of overexpression of PRS1 in these cells (compared with fibroblasts), as a consequence of which increased PRPP availability sufficient to activate the pathway of purine synthesis de novo to an abnormal rate is not achieved. PRS catalytic superactivity may thus be an example of a gene-regulatory defect in which cell-specific differences in phenotypic expression reflect variation among cell types in mechanisms modulating aberrant gene expression.