THE JOURNAL OF BI~L~CICAL CHEMISTRY 0 1990 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 265, No. 34, Issue of December 5, pp. 20916~20922,199O Printed in U.S.A.

Determination of Functional Effects of Mutations in the Steroid 2 1-Hydroxylase Gene (CYPB 1) Using Recombinant Vaccinia Virus*

(Received for publication, February 21, 1990)

Maria-Teresa Tusie-LunaS, Paula Traktmang, and Perrin C. Whiten From the Division of Pediatric Endocrinolopv and the &Department of Microbiology, Cornell University Medical College, New York, New Yo;k 10021

“”

Steroid Zl-hydroxylase (P45Oc21) is absent or de- fective in more than 90% of patients with congenital adrenal hyperplasia. This disorder of cortisol biosyn- thesis occurs in a wide spectrum of clinical severity; specific mutations in the Zl-hydroxylase gene (CYPZ 1) have been found in association with particular clinical phenotypes. To determine the functional effects of mu- tations causing amino acid substitutions, normal P45Oc21 and three mutagenized P45Oc21 enzymes were expressed at high levels in cultured CO&l cells using recombinant vaccinia virus. A single amino acid substitution (Va12’l + Leu) present in patients with mild “nonclassical” 2 1 -hydroxylase deficiency resulted in an enzyme with ZO-50% of normal activity. A mu- tation (Ile’72 + Asn) identified in patients with the “simple virilizing” form (poor cortisol synthesis but adequate aldosterone synthesis) resulted in an enzyme with less than 2% of normal activity. Finally, a cluster mutation (Ile-Val-Glu-Met234-238 -, Asn-Glu-Glu-Lys) found in a patient with severe “salt wasting” Zl-hy- droxylase deficiency (inadequate aldosterone synthe- sis) results in an enzyme with no detectable activity. These data indicate that the severity of 2 1-hydroxylase deficiency correlates with the degree of enzymatic compromise.

Congenital adrenal hyperplasia due to steroid 21-hydrox- ylase deficiency is a common inherited disorder of cortisol biosynthesis which results in abnormal sexual differentiation and somatic growth (1, 2, 51). The disorder is inherited as a monogenic autosomal recessive trait closely linked to the HLA major histocompatibility complex on chromosome 6. The disease occurs in a wide spectrum of clinical variants, includ- ing a severely affected form with a concurrent defect in aldosterone biosynthesis (“salt wasting” type), a form with apparently normal aldosterone biosynthesis (“simple viriliz- ing” type), and a mild “nonclassical” form that may be asymp- tomatic or may be associated with symptoms of androgen excess developing during childhood or at puberty.

The 21-hydroxylase enzyme is a microsomal cytochrome P450 termed P45OXXI or P45Oc21. The structural gene en- coding P45Oc21 (CYP21 or CYPSlB) and a 98% identical pseudogene (CYPBlP or CYP2lA) are located in the HLA

* This work was supported in part by National Institutes of Health Grants DK37867 and HD00072. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “adoertisement” in accord- ance with 18 USC. Section 1734 solely to indicate this fact.

$ Supported by a Fellowship for Graduate Studies in Health Sci- ences from the Consejo National de Ciencia y Tecnologia (Mexico).

11 Irma T. Hirsch1 Trust Scholar. To whom correspondence should be addressed.

complex adjacent to and alternating with the C4B and C4A genes encoding the fourth component of serum complement (3-6). The close proximity between CYPPl and CYP2lP (30 kilobases) appears to generate frequent mutations in CYP21 by two mechanisms: unequal crossing-over during meiosis, resulting in a complete deletion of CYP21, and gene conver- sion events that result in the transfer of mutations from CYP2lP to CYP21 (7). These mutations include frameshift, nonsense, missense, and intron splicing mutations (8-14). Whereas deletions, frameshift, or nonsense mutations com- pletely prevent the synthesis of a functional enzyme and are associated with salt wasting disease (7, 9, 10, 12), different missense mutations causing functionally important amino acid substitutions may be associated with different phenotypic forms of 21-hydroxylase deficiency. For example, Val”’ ---, Leu has been identified in patients with the mild or nonclas- sical form (13), Ile17’ + Asn is described in patients with the simple virilizing form (ll), and a cluster of mutations Ile-Val- Glu-M&235-*38 + Asn-Glu-Glu-Lys is described in a patient with salt wasting disease (14). All of these mutations are presumed to have arisen in gene conversion events. Because synthesis of P45Oc21 is restricted to the adrenal gland, the direct effect of these mutations on the enzymatic activity of P45Oc21 in affected patients cannot be tested. A previous attempt to measure activity of mutant P45Oc21 in vitro used a plasmid expression system yielding low levels of enzymatic activity, and measurements of kinetic parameters were not performed (14). Therefore, in order to analyze the effects of missense mutations on enzymatic function, we have expressed wild-type and mutant P45Oc21 in eukaryotic cells using re- combinant vaccinia virus.

MATERIALS AND METHODS

Construction of a Full-length cDNA Encoding P450c21-Routine molecular manipulations were performed as described (15). The long- est available human cDNA encoding P45Oc21, pc21/3c (6), lacked the first 15 base pairs of the coding sequence. Two complementary oligonucleotides, a 49- and 47-mer, were synthesized corresponding to the coding and anticoding strand of the predicted full-length cDNA 5’ of the NarI site in codons 14 and 15. Several modifications were made in the sequence. The 5’-untranslated region was changed to 5’- GATCCACC(ATG)-3’ (the first four bases were left unpaired in the 49.mer to form a BamHI cohesive end), and codon 2 was changed from CTG, encoding leucine, to GTG, encoding valine. These changes created an optimal site for initiation of translation (16). The first two bases on the 5’ end of the 47-mer were also left unpaired to form a NurI cohesive end. Only the 47-mer was phosphorylated with poly- nucleotide kinase prior to annealing to the 49-mer.

Plasmid DNA was partially digested with NurI. The linear form was isolated by agaroie elect;oph&esis and binding to glass powder in the presence of sodium iodide (17). It was ligated with a 100-fold molar excess of the annealed oligonucleotidei and digested with BamHI. The fragment corresponding to the reconstructed full-length cDNA was isolated by agarose gel electrophoresis and ligated into the BamHI site of the pBluescriptKS+ vector (Stratagene). After trans-

20916

by guest on March 20, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Mutations in the Steroid 21 -Hydroxylase Gene 20917

formation, clones containing the full-length cDNA were identified by colony hybridization using both the P45Oc21 cDNA and 49-mer oligonucleotide probes. The sequence of one clone was verified at the 5’ end.

Introduction of the Full-length cDNA into Vaccinia Virus-Plasmid pTF7.5 was obtained from B. Moss (18). This shuttle vector contains a bacteriophage T7 promoter and terminator flanking a unique BamHI restriction site for the insertion of foreign genes, all of which are within a copy of the vaccinia thymidine kinase gene. Plasmid DNA was digested with BarnHI, dephosphorylated, and ligated to the isolated BamHI fragment carrying the reconstructed full-length P45Oc21. After transformation, a clone with the insert in the correct orientation termed p21-T7 was identified by colony hybridization and restriction mapping.

Cell lines were obtained from the American Type Culture Collec- tion and grown in Dulbecco’s modified Eagle’s medium (Sigma) supplemented with 10% newborn calf serum. CV-1 or BSC40 cells were infected with wild-type vaccinia virus (WR strain) at a multi- plicity of infection of 0.03% (19). Recombinant vaccinia virus was produced by cotransfection with linearized p21-T7 as described (20). Preparations were analyzed for the presence of recombinant virus by plating on TK-143 cell monolayers in the presence of 25-50 pg/ml of 5-bromodeoxyuridine (20, 21). Thymidine kinase negative recombi- nant viral clones were identified by dot-blot hybridization using P45Oc21 cDNA as a probe (22). One clone was termed vC21.

Site-directed Mutagenesis-A 514 HincII fragment from pDM1 (a gift of B. Kemper) (23) containing the replication origin from bacte- riophage fl was ligated to phosphorylated Hind111 linkers, digested with HindIII, and ligated to p21-T7 that had been digested with Hind111 and dephosphorylated. After transformation, recombinant plasmids containing the fl fragment were selected by colony hybrid- ization, and the orientation of the fl fragment in one clone, p21-T7- fl, was determined by restriction mapping.

Single-stranded copies of p21-T7-fl were synthesized in Esche- richia coli CJ236 (dut-, ung-, obtained from Bio-Rad) (24) in the presence of VSC-13 helper phage. Complementary strands were syn- thesized using primers containing each desired mutation (GAGGGGCAC’M’GCACATGGCT for Val’sl+ Leu, TGCAGCAT- CAACTGTTACTTC for Ile17’ + Asn, and AGGGATCACAACGAG- GAEAAGCAGCTGAAG for the Ile-Val-Glu-Met234~238 + &.n-Elu- Glu-Lys mutation) and modified T7 DNA polymerase (U. S. Bio- chemical). Double-stranded DNA was ligated and transformed into E. coli strain MV1190. Plasmids containing each desired mutation were selected by colony hybridization using each mutant oligonucle- otide as a probe (25). All mutations were verified by sequence analysis.

Each mutagenized version of p21-T7-fl was digested with Hind111 and the linearized plasmid separated from the fl fragment by electro- phoresis in agarose. Mutagenized plasmids were introduced into vac- cinia as described above.

Expression of P450c21-vTF7 (obtained from B. Moss) is a vacci- nia virus carrying the gene for bacteriophage T7 RNA polymerase under the control of a vaccinia promoter. COS-1 monkey kidney epithelial cells were infected with vTF7 at a multiplicity of infection of 500 with or without simultaneous infection with wild-type or mutagenized vC2I at a multiplicity of infection of 1. Cells were incubated for 18 h with Dulbecco’s modified Eagle’s medium supple- mented with 5% newborn calf serum, 10 fig/ml protoporphyrin IX, and 18 pg/ml transferrin.

Assays of Enzymatic Actiuity-Radiochemical purity of steroid precursors was confirmed by thin layer chromatography. Except where noted, enzymatic activity was assayed in whole cells or cell lysates 18 h after infection. To assay activity in whole cells, cells were incubated for 2 h with medium containing 17-[3H]hydroxyprogester- one or [3H]progesterone (Du Pont-New England Nuclear) and 2 pM unlabeled steroid (0.25 mCi/mM specific activity). The medium was removed and extracted with methylene chloride. Products and reac- tants were resolved by thin layer chromatography in chloro- form:acetone (70~30) (17-hydroxyprogesterone and ll-deoxycortisol) OT chloroform:ethyl acetate (80:20) (progesterone and ll-deoxycort- costerone). Radioactivity was measured using a thin layer chroma- tography analyzer (Berthold) or by liquid scintillation spectropho- tometry. To assay activity in cell lysates, the cells were washed twice with phosphate-buffered saline (GIBCO) and homogenized in hype- tonic buffer (10 mM Hepes,’ pH 6.2, 10 mM NaCl, 1.5 mM MgCl*) using a Dounce homogenizer. Nuclei were removed by centrifugation

i The abbreviations used are: Hepes, 4-(2.hydroxyethyI)-l-pipera- zineethanesulfonic acid; ACTH, corticotropin.

at 800 x g for 10 min. Assays were carried out using 250 rg of protein in 0.5 ml of 100 mM KHzPO+ pH 7.2, with 0.5 PCi of ‘H-substrate and 2 mM NADPH at 37 “C with or without 20% glycerol (26).

Statistical Analysis-Kinetic constants were derived from two to six determinations of enzymatic activity at each of five different substrate concentrations. Primary data were evaluated by nonlinear regression analysis using the Enzfitter program (Elsevier Science Publishers BV, Amsterdam). Parameters were optimized by iteration utilizing the enhanced algorithm of Marquart (27). Appropriate weighting for each set of data was determined by analysis of residual errors (AY/Y versus Y). Values were expressed as means + standard deviations.

Inhibitor models were evaluated using a weighted nonlinear least squares curve fitting program (28).

Radiolabeling and Zmmunoprecipitation-Cells were grown in 24. well plates to a density of 3 x lo4 cells/well and infected with vTF7 and, in some wells, with wild-type or mutagenized vC21 as described above. Eighteen h after infection, cells were incubated for 30 min in medium (1 ml/well) containing 100 &i of [““S]methionine/cysteine (1000 Ci/mmol) without additional unlabeled methionine and cys- teine. Cells were washed twice with phosphate-buffered saline and lysed by resuspension in 100 jd/well lysis buffer (0.05 M Tris-HCl, pH 7.5, 0.15 M NaCl, 1% Nonidet P-40, 5 mM EDTA, 0.25% sodium deoxycholate, 1 mM phenylmethylsulfonyl fluoride, 5 mM iodoaceta- mide, 100 units/ml aprotinin) for 30 min at 4 “C.

Aliquots of lysates (45 gg of protein) were subjected to electropho- resis in an 8.5% sodium dodecyl sulfate-polyacrylamide gel (29). For immunoprecipitation, lysates were precleared by three successive incubations with a 1:lOO dilution of preimmune rabbit serum and a 1:lO volume of protein A-Sepharose beads (Pharmacia LKB Biotech- nology Inc.). Precleared supernatants were incubated for 1 h at 4 “C with a 1:lOO dilution of rabbit anti-bovine P45Oc21 serum (30), and antigen-antibody complexes were adsorbed to protein A-Sepharose. Pellets were washed three times in RIPA buffer (50 mM Tris-HCl, pH 7.5,0.15 M NaCl, 1% Nonidet P-40, 0.1% sodium dodecyl sulfate, 1% sodium deoxycholate) (31), boiled in Laemmli sample buffer (29), and subjected to electrophoresis under the same conditions as unpre- cipitated lysates.

Gels were dried after electrophoresis and autoradiographed. Den- sitometry was performed using a Bio-Rad model 620 video densitom- eter.

Subcell&r Fractionation-Microsomal fractions were collected by differential centrifugation in 0.25 M sucrose (32) or using self-forming Percoll density gradients (33). Briefly, infected cells were washed twice with cold phosphate-buffered saline and harvested by scraping with a rubber policeman. Cells were homogenized in 5 volumes of homogenizing buffer (0.25 M sucrose, 1 mM EDTA, pH 6.8) in a Dounce homogenizer and centrifuged at 800 X g for 10 min. Nuclear free supernatants (5.5 ml in homogenizing buffer) were layered on a cushion of 1 ml of 60% sucrose and 4.4 ml of Percoll suspension (density 1.070 g/liter). Gradients were formed by centrifugation at 20,000 X g for 2 h at 4 “C in a Beckman Ti-80 rotor. Fractions of 250 ~1 were collected. Protein content and density were measured in each fraction. Each fraction was assayed for lactate dehydrogenase activity as described (34) and for 21-hydroxylase activity using 17-hydroxy- progesterone as a substrate. The latter assays were performed as described above on 50-~1 aliquots of each fraction.

Immunoblot analysis was performed in all fractions using antisera to bovine P45Oc21 or NADPH-cytochrome P450 reductase (a gift of K. C. Cheng) and lZ51-protein A (35).

RESULTS

Synthesis of Recombinant Human P450c21-We recon- structed a human full-length cDNA encoding functional P45Oc21. The introduction of two single-base changes around the translation initiation site (A at position -3 and G at position +4) were made to increase translational efficiency (16) and presumably did not significantly affect the enzymatic activity of the recombinant enzyme; Val’ also occurs in normal bovine (36) and porcine (37) P45Oc21. After mutagenesis of the cDNA and transfer into vaccinia virus, three mutant enzymes were expressed. These were termed P450c21-NC (nonclassical disease, ValZal ---, Leu), P45Oc21-SV (simple virilizing disease, Ile17’ ---, Asn), and P45Oc21-SW (salt wast- ing disease, Ile-Val-Glu-Met’34~*‘* -+ Asn-Glu-Glu-Lys).

by guest on March 20, 2020

http://ww

w.jbc.org/

Dow

nloaded from

20918 Mutations in the Steroid 21 -Hydroxylase Gene

We were able to synthesize large amounts of recombinant protein by infecting kidney epithelial cells with recombinant vaccinia virus containing wild-type or mutant P45Oc21 cDNA. Several parameters were optimized including multiplicities of infection for both vTF7 and vC21 and concentrations of transferrin and protoporphyrin IX in the medium. These compounds may have an effect on availability of heme iron required for the assembly of the holoenzyme (38,39), as shown by a 15% increase in enzymatic activity in their presence (data not shown).

Synthesis of the apoenzyme was maximal 24-48 h after infection as measured by Western blot analysis whereas 21- hydroxylase activity peaked 18 h postinfection (data not shown). Densitometric analysis showed that about 2% of total radiolabeled cellular protein was recombinant P45Oc21 (Fig. 1A). The identity of the radiolabeled recombinant protein was confirmed by immunoprecipitation (Fig. 1B) using a rabbit polyclonal antibody against bovine P45Oc21 (30). Sim- ilar amounts of the recombinant protein were obtained from cells infected with the wild-type or mutant constructs, indi- cating that the mutations did not have a major effect on synthesis of the apoenzyme.

Measurements of Enzymatic Activity in Whole Cells-Initial measurements of enzymatic activity were made in whole cells (Fig. 2). This provided a qualitative determination of the

A B

1 2 345 2 34 5

FIG. 1. Identification and quantification of wild-type and mutant P45Oc21. A, metabolic labeling. Cell lysates labeled with [JsS]methionine/cysteine were subjected to electrophoresis through an 8.5% polyacrylamide gel which was then dried and autoradi- ographed. Positions of molecular weight markers are indicated. Lanes are numbered: 1, infection with vTF7 alone; 2-5, infection with vTF7 and with vC21, respectively encoding wild-type P45Oc21, P45Oc21- NC, P45Oc21-SV, and P45Oc21-SW. Percentages of total incorporated radioactivity represented by P45Oc21, determined by scanning den- sitometry, are: wild type, 1.8; P450c21-NC, 3.6; P45Oc21-SV, 2.3; P45Oc21-SW, 1.8. B, immunoprecipitates of labeled cell lysates ana- lyzed in parallel with the unfractionated cell lysates. Lanes are numbered as in A (lysates from cells infected with vTF7 alone yield no band at 52 kDa). An arrow marks the position of P45Oc21.

% Conversion 100

so -

60

40 -

I L

: ‘-

20

0 .--. _L~ WT NC sv SW

FIG. 2. Enzymatic activity of wild-type and mutant human recombinant P45Oc2 1 in whole cells. Solid bars represent percent conversion of 17-hydroxyprogesterone to 11-deoxycortisol; hatched bars represent conversion of progesterone to 1 l-deoxycorticosterone (the percentage of substrate conversion was subtracted from the conversion exhibited by cells infected only with vTF7). WT, wild- type P45Oc21; NC, SV, SW are P450c21-NC, P45Oc21-SV, and P45Oc21-SW, respectively.

effects of different mutations and facilitated comparison of the present work with previous studies (14) of mutant P45Oc21. After a 2-h incubation, P450c21-NC had 75% of the wild-type activity for 17-hydroxyprogesterone but only about 30% of normal activity for progesterone. P45Oc21-SV had about 2% of normal activity for both substrates, and P45Oc21- SW had no detectable activity.

Subcellular Fractionation of Wild-type and Mutant P450c21-Subcellular location of P45Oc21 was determined by Western blot analysis of subcellular fractions and (except for P45Oc21-SW) by determinations of enzymatic activity (Fig. 3). Microsomal and cytosolic fractions were identified by density (1.06-1.11 and 1.038-1.040 g/ml, respectively); in addition, cytosolic fractions were identified by lactate dehy- drogenase activity, and microsomal fractions by the peak of NADPH-cytochrome P450 reductase immunoreactivity. Re- ductase immunoreactivity was also noted in cytosolic frac- tions; this may have represented newly synthesized enzyme that had not yet been inserted into the endoplasmic reticulum. Alternatively, small fragments of endoplasmic reticulum might have sedimented in cytosolic fractions.

Wild-type P45Oc21 immunoreactivity was distributed iden- tically to cytochrome P450 reductase, and the peak of enzy- matic activity was likewise found in the microsomal fractions. P450c21-NC activity was found in both microsomal and cy- tosolic fractions, suggesting that this mutant enzyme was not localized efficiently in microsomes. Presumably the cyto- chrome P450 reductase present in the cytosolic fractions was able to function as an electron donor. P45Oc21-SV localized poorly to microsomes, and a small amount of enzymatic activity was observed predominantly in the cytosolic fractions. Although P45Oc21-SW was not enzymatically active, its dis- tribution of immunoreactivity was not significantly different from the wild-type enzyme.

Kinetic Analysis of Wild-type and Mutant P450c21-Ap- parent K, and V,,,,, values for both 17-hydroxyprogesterone and progesterone were determined in cellular lysates (Fig. 4 and Table I). The K, of human recombinant wild-type P45Oc21 in cellular lysates was 1.2 pM for 17-hydroxyproges- terone and 2.8 pM for progesterone. These values are similar to those obtained for purified bovine (K,,, for l’l-hydroxypro- gesterone, 0.3 (26) to 7.9 pM (40)) or porcine (K, for 17- hydroxyprogesterone and progesterone, 5 pM (37)) enzymes or for recombinant bovine P45Oc21 expressed in COS cells (K, for 17-hydroxyprogesterone, 0.7 FM; K, for progesterone, 1.3 ELM) (38). Apparent V,,,,, of expressed human P45Oc21 was 53 pmol/min/mg of protein for 17-hydroxyprogesterone and 28 pmol/min/mg for progesterone. The V,,,,, for 17-hydroxy- progesterone reported for purified bovine P45Oc21 is 110 nmol/min/mg P45Oc21 (26). Considering that recombinant human P45Oc21 constitutes 1.5% of newly synthesized cellular protein (probably a smaller percentage of total protein), the specific activity obtained corresponds to a V,,,., of at least 3.5 nmol/min/mg of pure P45Oc21. The amount of functional enzyme may be limited by the availability of cytochrome P450 reductase, which decreases during vaccinia infection as judged by Western blot analysis (data not shown). Alternatively, a substantial amount of recombinant P45Oc21 may be apoen- zyme.

Kinetic analysis revealed differences between recombinant enzymes in the apparent K, and V,,,,, for each substrate (Table I). When activities were expressed as first-order rate constants V,,,.,/K,,,, P450c21-NC had about half the activity of wild-type P45Oc21 for 17-hydroxyprogesterone (44.2 vers’sus 21.2 pl/min/mg of total protein), whereas the activity of P45Oc21-SV was decreased 200-fold (0.25 pl/min/mg).

by guest on March 20, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Mutations in the Steroid 21-Hydroxylase Gene

FIG. 3. Subcellular fractionation of wild-type and mutant recombi- nant human P45Oc21. Pane/A, right, fractions of Percoll density gradients were analyzed for density, activity of lac- tate dehydrogenase (panel labeled LDH), and immunoreactivity on a West- ern blot (quantitated by scanning den- sitometry) using rabbit antiserum against NADPH-cytochrome P450 re- ductase (NADPH-reductase). Determi- nations on one of four gradients are shown. Panel A, left, fractions were as- sayed for 21-hydroxylase activity using 17-hydroxyprogesterone as a substrate. Wild-type P45Oc21 (P450c21-WT), P450c21-NC, and P45Oc21SV were ana- lyzed. Activities in fractions are ex- pressed as a percentage of total activity obtained by integrating under each curve. The actual activity of each enzyme differs. In the peak fraction of P45Oc21- WT, 20% of substrate was converted in a l-h incubation; in the peak fraction of P450c21-NC, 12% was converted in 2 h; and in the peak fraction of P45Oc21SV, 1.7% was converted in 5 h. Panel B, localization of P45Oc21 enzymes in the microsomal fractions detected by West- ern blotting (the remainder of the frac- tions are not shown). Fractions are num- bered as in the A. WT, wild-type P45Oc21; NC, P450c21-NC; SV, P45Oc21SV; SW, P45Oc21SW.

P450c21-WT

1 w Fraction x Fraction I

40 4 ? ‘I P450c21-NC

P45Oc21-SV ! NADPH-reductase

5 Fraction # Fraction #

B

Both wild-type P45Oc21 and P450c21-NC metabolized pro- gesterone more slowly than 17-hydroxyprogesterone (V,,,.,/ K,,, of 9.5 and 1.9 pl/min/mg, respectively); the difference was particularly pronounced forP450c21-NC, of which the first- order rate constant for progesterone was only one-tenth that for 17-hydroxyprogesterone (one-fifth the activity of the wild- type enzyme).

Kinetic measurements of P45Oc21-SV were not attempted using progesterone because of a relatively high apparent back- ground level of 21-hydroxylation of this substrate (4% con- version over 2 h) in lysates from COS-1 cells infected with

sv t

SW

vTF7 alone (the apparent background is not due to radi- ochemical impurity of the substrate, and it is not known if the background product in these cells is actually ll-deoxycor- ticosterone).

P45Oc21-SW had no detectable enzymatic activity in cel- lular lysates.

In addition to changes in intrinsic enzymatic activity, it seemed possible that the mutant enzymes might be relatively unstable on isolation from the cell. Therefore, the kinetic analyses of P450c21-NC and P45Oc21-SV (for 17-hydroxypro- gesterone) were repeated using cellular lysates prepared in the

by guest on March 20, 2020

http://ww

w.jbc.org/

Dow

nloaded from

20920 Mutationa in the Steroid 21 -Hydroxylase Gene

absence of glycerol, a known stabilizing agent for P45Oc21 (26, 42). These conditions resulted in a 3-fold decrease in the first-order rate constant for wild-type P45Oc21 but a 7-fold decrease for P450c21-NC. P45Oc21-SV had no detectable activity in the absence of glycerol.

Inhibition of Enzymatic Activity by Progesterone-Since P450c21-NC 21-hydroxylated progesterone only one-fifth as fast the wild-type enzyme, the ability of progesterone to inhibit 21-hydroxylation of 17-hydroxyprogesterone was tested by repeating kinetic measurements in the presence of 0, 0.5, or 2 pM progesterone. Both wild-type P45Oc21 and P450c21-NC were inhibited by 0.5 and 2 PM progesterone. The data (Fig. 5) were fitted to four models of inhibition:

300

P 5 200

E ,p 100 >

0

0 1 2 3 4 6 IS1

FIG. 4. Kinetic analysis of wild-type and mutant human P45Oc2 1. Michaelis-Menten plots of velocity ( V, in pmol/min/mg of total protein) against substrate concentration ([S], in PM). Closed circles, wild-type P45Oc21; open circks, P450c21-NC; squares, P45Oc21SV. A, conversion of 17-hydroxyprogesterone to ll-deoxy- cortisol in the presence of glycerol. Points represent means of six (wild-type P45Oc21 and P450c21-NC) or four (P45Oc21SV) deter- minations. B, conversion of 17-hydroxyprogesterone to ll-deoxycor- tisol in the absence of glycerol. Points are means of duplicate deter- minations. C, conversion of progesterone to ll-deoxycorticosterone in the presence of glycerol. Points are means of duplicate determi- nations. Note differences in scale in the three panels.

competitive (inhibitor and substrate bind to the enzyme in mutually exclusive manner), noncompetitive (inhibitor an< substrate bind independently of each other), uncompetitiv (inhibitor binds only to the enzyme-substrate complex), ant mixed competitive (inhibitor affects the affinity of the enzym for substrate) (28). For both enzymes, the data were equal1 consistent with competitive or mixed competitive inhibitior Additional kinetic data will be required to distinguish betwee: these possibilities. For the competitive inhibition model, & values were 0.4 and 0.8 pM for wild-type P45Oc21 an’ P450c21-NC, respectively.

DISCUSSION

Recombinant P45Oc21 enzymes carrying specific amin acid substitutions seen in patients with 21-hydroxylase defi ciency exhibit activities that correlate with the clinical sevei ity of the disease and with biochemical abnormalities such a 17-hydroxyprogesterone levels after ACTH stimulatio. (Table II).

Salt Wasting Versus Simple Virilizing Phenotypes-An im portant facet of the clinical variability of 21-hydroxylas deficiency concerns the ability to synthesize adequat amounts of the mineralocorticoid hormone aldosterone. Be cause aldosterone is normally secreted at a rate 100-100 times lower than that of cortisol, it is apparent that 21 hydroxylase activity would have to decrease to very low level before it became rate-limiting. The data presented here sug gest that as little as 0.6% of normal activity (measured b first-order rate constants, V,,,,,/K,) as seen in P45Oc21-S’ carrying the Ilei7’ + Asn mutation, allows adequate aldostei one synthesis to prevent significant salt wasting, thus resuli ing in the simple virilizing phenotype. In contrast, th P45Oc21-SW enzyme with three clustered nonconservativ amino acid substitutions lacks any detectable 21-hydroxylas activity, consistent with previously observed associations be tween deletions or nonsense mutations and salt wasting dif ease. It should be noted, however, that the distinction betwee the simple virilizing and salt wasting phenotypes is not at solute. One patient with the Ile17’ - Asn mutation has bee reported to have an elevated ratio of plasma renin to aldoE terone, consistent with mild salt wasting (ll), and HLA identical sibling pairs have been reported in which one siblin has salt wasting disease whereas the other can synthesin adequate amounts of aldosterone (43, 44). These finding suggest that additional epigenetic or nongenetic factors ca influence the presentation of the salt wasting phenotype.

Nonclassical Phenotype-As measured by V,,,/K,, th Val**i * Leu mutation results in an enzyme with about 50: of normal activity when 17-hydroxyprogesterone is the sul strate but only about 20% of normal activity for progesteron An individual homozygous for this mutation has nonclassic;

WT NC sv SW

TABLE I

Kinetic analysis of wild-type and mutant human P45Oc21 Measurements were made using 17-hydroxyprogesterone (17-OHP) as substrate in the absence (-) and presence

(+) of 20% glycerol, and progesterone in the presence of glycerol. K, values are in PM, and V,,, values are in pmol/min/mg of total protein. All values are means + standard deviations derived from regression analyses using replicate points as described in the legend to Fig. 4. First-order rate constants V,,,/K, are expressed in rl/min/ mg of total protein. NT, not tested. WT, wild type; NC, nonclassical; SV, simple virilizing; SW, salt wasting.

17-OHP - glycerol 17-OHP + glycerol Progesterone + glycerol

K”l V msx VmxlKm K”l V max VmlKm K.3 V ma/l Vm..lKn

1.2 + 0.3 17.5 f 1.2 14.6 1.2 * 0.3 53.0 + 4.5 44.2 2.8 -c 0.2 26.7 + 1.8 9.5 4.9 +- 0.4 15.2 -t 0.6 3.1 1.2 f 0.4 27.7 -e 1.0 21.2 5.9 k 1.0 11.3 2 1.7 1.9

0 0 0 2.4 + 0.4 0.6 + 0.1 0.25 NT NT NT NT NT NT NT NT

by guest on March 20, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Mutations in the Steroid 21 -Hydroxylase Gene 20921

21-hydroxylase deficiency with significant biochemical abnor- malities and variable symptoms of androgen excess. A heter- ozygous carrier of a salt wasting mutation might also be expected to have about 50% of normal 21-hydroxylase activ- ity, but such individuals are asymptomatic and have minimal biochemical abnormalities. This suggests that in uiuo 21- hydroxylase activity in patients with nonclassical 21-hydrox- ylase deficiency must actually be less than 50% of normal. This apparent paradox may be resolved by considering the kinetically abnormal activity of P450c21-NC for progesterone. The studies reported here demonstrate that 0.5-2 PM proges- terone inhibits 21-hydroxylation of 17-hydroxyprogesterone by either wild-type P45Oc21 or P450c21-NC. Although the intraadrenal concentration of progesterone has not been de- termined in patients with nonclassical 21-hydroxylase defi- ciency, the progesterone concentration in adrenal glands from normal individuals averages 2.5-4 PM (45). It is likely that progesterone will be metabolized poorly in adrenal glands expressing only P450c21-NC so that its concentration may be significantly elevated, depending on the activity of alternative metabolic pathways such as 17a-hydroxylase/l7,20-lyase. El- evated progesterone levels will strongly inhibit 21-hydroxyl-

< -0.1 l----L I 1

- -1 -0.5 0 0.6 1 1.5 2

l/S WUM)

, -o.lr ‘1 1 /

: -0.5 0 0.5 1 1.6 2

l/s (VIM FIG. 5. Inhibition by progesterone of the 21-hydroxylation

of 17-hydroxyprogesterone. Data are displayed as Lineweaver- Burk plots. Points are means of duplicate determinations. Triangles, measurements in the absence of progesterone; squares, measurements in the presence of 0.5 pM progesterone; circles, 2 pM progesterone. A, wild-type P45Oc21; B, P450c21-NC. First-order rate constants V,,,,,/ K,,, derived from these data at 0,0.5, and 2 PM progesterone are: wild- type P45Oc21, 38.7, 17.6, and 8.6; P450c21-NC, 18.6, 9.1, and 5.4, respectively.

ation of 17-hydroxyprogesterone. Thus, relatively small dif- ferences in intraadrenal progesterone concentration could account for much of the clinical variability that is a hallmark of nonclassical 21-hydroxylase deficiency.

The relatively low V,,, of P450c21-NC for progesterone is consistent with observations in Go; patients with nonclassi- cal 21-hydroxylase deficiency have nearly normal cortisol secretion after 3 days of stimulation with ACTH but decreased secretion of deoxycorticosterone (the 21-hydroxylated product of progesterone) similar to those seen in patients with the simple virilizing form of 21-hydroxylase deficiency (46).

Structural Effects of Mutations-It is not known how the mutations that we studied affect enzymatic activity. None involves regions of the enzyme thought to be involved in the active site or in substrate binding. Both P450c21-NC and P45Oc21-SV are not localized properly in microsomes, and both of these mutant enzymes are sensitive to isolation in the absence of glycerol, a known stabilizing agent for P45Oc21. These data suggest that the Ile17* + Asn and Valz81 + Leu mutations both affect the conformation of the enzyme in some manner. It has been speculated that the ValZ8’ -+ Leu mutation affects conformation by increasing the likelihood of forming an a-helix in a region that is not predicted to normally be in a helical conformation (13). The Ile17’ + Asn mutation involves a region that may normally interact with the mem- brane of the endoplasmic reticulum (47), and the isoleucine residue at this position is strongly conserved in many different P450 enzymes. Mutation of this hydrophobic residue to a polar residue might disrupt such an interaction, weakening the association of the enzyme with the endoplasmic reticulum (11).

The system used in this work (and in studies of other P450 enzymes (48)) allows expression of P450 enzymes at suffi- ciently high levels to study the functional effects of mutations over a wide range of severity. Other expression systems have recently been successfully employed in a similar manner (49). Additional mutagenesis of selected regions of P45Oc21 should permit further elucidation of the relationships between struc- ture and function of this physiologically important enzyme.

Addendum-After this manuscript was submitted, it was reported that the Ile17’ + Asn mutation decreased activity of P45Oc21 expressed in cultured cells (50).

Acknowledgments-We thank Drs. Bernard Moss and Byron Kem- per for gifts of vaccinia virus and plasmids, Dr. Carl Monder for helpful discussions and valuable assistance with the kinetic analysis, Dr. Phyllis Speiser for helpful discussions, and Drs. Monder and Maria New for reviewing the manuscript.

TABLE II

Comparison of functional effects of mutations in P45Oc21 Comparisons of typical values for normal individuals, heterozygous carriers for the salt-wasting or simple

virilizing forms of Pl-hydroxylase deficiency (Het.), or patients with nonclassical (NC), simple virilizing (SV), or salt wasting (SW) forms of 21-hydroxylase deficiency. Serum levels of 17-hydroxyprogesterone (17-OHP) measured 60 min after an intravenous bolus of corticotropin are in ng/lOO ml. Symptoms range from nil to severe (+++). Estimates of enzymatic activity in each individual as a percent of normal activity are derived from the present work.

Genotype Mutation 17.OHP Symptoms Activity

Normal Het. NC

% 200 0 100

1,000 0 50 VaYR’ + Leu 5,000 W+ 20 (progesterone)

50 (17-OHP) Ile”* -+ Asn 25,000 ++ 1 Codons 234-238 or deletion or Gln31R --) end 25,000 +++ 0

by guest on March 20, 2020

http://ww

w.jbc.org/

Dow

nloaded from

20922 Mutations in the Steroid 21-Hydroxylase Gene

1

2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16. 17.

18.

19.

20.

21.

22.

23.

REFERENCES

White, P. C., New, M. I. & DuPont, B. (1987) N. Eng2. J. Med. 316, 1519-1524

New, M. I., White, P. C., Pang, S., DuPont, B. & Speiser, P. W. (1989) in The Metabolic Basis of Inherited Disease (&river, C. R., Beaudet, A. L., Sly, W. S. & Valle, D., eds) pp. 1881-1918, McGraw-Hill Book Cbmpany, New York

White. P. C.. Grossberger. D.. Onufer. B. J.. Chaulin. D. D.. New. M. I:, Dupcmt, B. & &riminger, J.‘L. (1685) kroc: Natl.‘Acad: 5%. U. S. A. 82, 1089-1093

Carroll, M. C., Campbell, R. D. & Porter, R. R. (1985) Proc. N&l. Acad. Sci. U. S. A. 62, 521-525

Higashi, Y., Yoshioka, H., Yamane, M., Gotoh, 0. & Fujii- Kuriyama, Y. (1986) PFOC. Natl. Acad. Sci. U. S. A. 83, 2841- 2845

White, P. C., New, M. I. & DuPont, B. (1986) Proc. Natl. Acad. Sci. U. S. A. 83,5111-5115

White. P. C.. Vitek. A.. Dunont. B. & New. M. I. (1988) Proc. Nati Acad.‘Sci. UI S. A. 8k, 4436-4440 ’

Rodriguez, N. R., Dunham, I., Yung Yu, C., Carroll, M. C., Porter, R. R. & Campell, R. D. (1987) EMBO J. 6,1653-1661

Hinashi. Y.. Tanae. A.. Inoue. H. & Fuiii-Kurivama. Y. (1988) im. i. Hum. Genkt. 42, 17-i5 ” ”

Harada, F., Kimura, A., Iwanaga, T., Shimozawa, K., Yata, J. & Sasazuki, T. (1987) Proc. Natl. Acad. Sci. U. S. A. 84, 8091- 8094

Amor, M., Parker, K. L., Globerman, H., New, M. I. & White, P. C. (1988) PFOC. N&l. Acad. Sci. U. S. A. 85, 1600-1604

Globerman. H.. Amor. M.. Parker. K. L.. New. M. I. & White. P. C. (1988)‘J. &in. Z&es;. 82, 139-144

Speiser, P. W., New, M. I. &White, P. C. (1988) N. Engl. J. Med. 319,19-23

Higashi, Y., Tanae, A., Inoue, H., Hiromasa, T. & Fuji-Kuriyama, Y. (1988) PFOC. Natl. Acad. Sci. U. S. A. 85, 7486-7490

Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989) Molecular Cloning: A Laboratory Manual, 2nd Ed. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY

Kosak, M. (1986) Cell 44, 283-292 Vogelstein, B. & Gillespie, D. (1979) Proc. Natl. Acad. Sci. U. S.

A. 76,615-619 Fuerst, T. R., Earl, P. L. & Moss, B. (1987) Mol. Cell. Biol. 7,

2538-2544 Traktman, P., Kelvin, M. & Pacheco, S. (1989) J. Virol. 63,841-

846 Fuerst, T. R., Niles, E. G., Studier, F. W. & Moss, B. (1986) Proc.

N&l. Acad. Sci. U. S. A. 83, 8122-8126 Mackett, M., Smith, G. L. & Moss, B. (1984) J. Virol. 49, 857-

864 Mackett, M., Smith, G. L. & Moss, B. (1985) in DNA Cloning

(Glover, D. M., ed.) Vol. 2, pp. 191-211, IRL, Oxford Mead, A. D., Szczensa-Skourpa, E. & Kemper, B. (1985) Nucleic

Acids Res. 13, 1103-1118

24. Kunkel, T. A. (1985) Proc. Natl. Acad. Sci. U. S. A. 82,488-492 25. Wood, W. I., Gitschier, J., Lasky, L. A. & Lawn, R. M. (1985)

PFOC. Natl. Acad. Sci. U. S. A. 82, 1585-1588 26. Kominami, S., Ochi, H., Kobayashi, Y. & Takemori, S. (1980) J.

Biol. Chem. 255,3386-3394 27. Bevington, P. (1969) Data Reduction and Error Analysis for the

Physical Sciences, McGraw Hill, New York 28. Lutz, R. A., Bull, C. & Rodbard, D. (1986) Enzyme (Base11 36,

197-206 29. Laemmli, U. K. (1970) Nature 227,680-685 30. White, P. C., New, M. I. & DuPont, B. (1984) Proc. N&l. Acad.

Sci. U. S. A. 81,1986-1990 31. Chin, Y.-H., Rasmussen, R. A., Woodruff, J. J. & Easton, T. G.

(1986) J. Zmmunol. 136, 2556-2561 32. Zuber, M. X., Simpson, E. R., Hall, P. F. & Waterman, M. R.

(1985) J. Biol. Chem. 260. 1842-1848 33. Merion; M. & Poretz, R. D: (1981) J. Supramolec. Struct. Cell.

Biochem. 17,337-346 34. Gutmann, I. & Wakefeld, A. (1983) in Methods of Enzymatic

Analysis (Berg, H. V., ed) 2nd Ed., Vol. 2, pp. 1464-1468, Academic Press, New York

35. Burnett, W. N. (1981) Anal. Biochem. 112,195-204 36. Yoshioka, H., Morohashi, K., Sogawa, K., Yamane, M., Komi-

nami, S., Takemori, S., Okada, Y., Omura, T. & Fujii-Kuri- yama, Y. (1986) J. Biol. Chem. 261,4106-4109

37. Yuan, P. M., Nakajin, S., Haniu, M., Shinoda, M., Hall, P. F. & Shively, J. E. (1983) Biochemistry 22, 143-149

38. Dwarki, V. J., Francis, V. N. K., Bhat, G. J. & Padmanaban, G. (1987) J. Biol. Chem. 262,16958-16962

39. Barnes, D. & Sato, G. (1980) Cell 22, 649-655 40. Bryan, G. T., Lewis, A. M. & Harkins, J. B. (1974) Steroids 23,

185-201 41. Deleted in proof 42. Chasalow, F. I. & Lieberman, S. (1979) J. Biol. Chem. 25, 3777-

3781 43. Rosenbloom, A. L. & Smith, D. W. (1986) Pediatrics 38, 215-

219 44. Stoner, E., Dimartino-Nardi, J., Kuhnle, U., Levine, L. S., Ob-

erfield, S. E. & New, M. I. (1986) Clin. Endocrinol. 24, 9-20 45. Dickerman, Z., Grant, D. R., Faiman, C. &Winter, J. S. D. (1984)

J. Clin. Endocrinol. & Metab. 59. 1031-1036 46. Kater, C. E., Biglieri, E. G. & Wasjchenberg, B. (1985) J. Clin.

Endocrinol. & Metab. 60, 1057-1062 47. Monier, S., Van Luc, P., Kreibich, G., Sabatini, D. D. & Adesnik,

M. (1988) J. Cell Biol. 107.457-470 48. Aoyama, T:, Korzekwa, K., fiagata, K., Adesnik, M., Reiss, A.,

Lanenson. D. P.. Gillette. J.. Gelboin H. V.. Waxman. D. J. & Gdnzalez,‘F. J. (i989) J. &ok Chem. 264, i1327-21333

49. Yanase, T., Kagimoto, M., Suzuki, S., Hashiba, K., Simpson, E. R. & Waterman, M. R. (1989) J. Biol. Chem. 264,18076-18082

50. Chiou, S. H., Hu, M. C. & Chung, B. C. (1990) J. Biol. Chem. 265,3549-3552

51. White, P. C., New, M. I. & DuPont, B. (1987) N. Engl. J. Med. 316,1580-1586

by guest on March 20, 2020

http://ww

w.jbc.org/

Dow

nloaded from

M T Tusie-Luna, P Traktman and P C White(CYP21) using recombinant vaccinia virus.

Determination of functional effects of mutations in the steroid 21-hydroxylase gene

1990, 265:20916-20922.J. Biol. Chem. 

  http://www.jbc.org/content/265/34/20916Access the most updated version of this article at

 Alerts:

  When a correction for this article is posted• 

When this article is cited• 

to choose from all of JBC's e-mail alertsClick here

  http://www.jbc.org/content/265/34/20916.full.html#ref-list-1

This article cites 0 references, 0 of which can be accessed free at

by guest on March 20, 2020

http://ww

w.jbc.org/

Dow

nloaded from