Functional and Structural Consequences of a Novel Point Mutation in the CYP21A2 Gene Causing Congenital Adrenal Hyperplasia: Potential Relevance of Helix C for P450 Oxidoreductase-21-Hydroxylase Interaction

Abstract

Background: Congenital adrenal hyperplasia is caused by insufficient adrenal steroid biosynthesis due to impaired steroidogenic enzymes. The majority of patients suffer from deficiency of 21-hydroxylase (CYP21) coded by the CYP21A2 gene.

Objective: Our objective was to study the functional and structural consequences of the novel CYP21A2 missense mutation c.364A > C (K121Q) detected in a female patient with nonclassical 21-hydroxylase deficiency. The patient was compound heterozygous for the novel K121Q mutation and the mild P453S mutation.

Results:In vitro expression analysis of the mutant K121Q enzyme in transiently transfected COS-7 cells revealed reduced CYP21 activity of 14.0 ± 5% for the conversion of 17-hydroxyprogesterone and 19.5 ± 4% for the conversion of progesterone. K121 is located on helix C in the CYP21 protein, which is part of the heme coordinating system. In addition, helix C is involved in the interaction with the electron-providing enzyme P450 oxidoreductase. Protein modeling revealed that the substitution of glutamine for the basic amino acid lysine introduces an electrostatic change on the surface of CYP21 and may additionally change heme coordination. We hypothesize that the electron flux between P450 oxidoreductase and CYP21 is impaired and, moreover, that substrate affinity is altered due to heme dislocation with K121Q.

Conclusion: Both the interaction of P450 oxidoreductase and CYP21 as well as heme coordination are likely to be disturbed due to the K121Q mutation. Our data exemplify how the combination of in vitro expression and structural protein analysis provide novel insights into molecular mechanisms of reduced CYP21 activity, eventually explaining the patient’s phenotype.

Congenital adrenal hyperplasia (CAH) is caused by the complete loss or severe decrease in activity of one of the five steroidogenic enzymes involved in glucocorticoid biosynthesis or defects in the electron donor P450 oxidoreductase (1). The commonest form is 21-hydroxylase deficiency (21OHD) caused by inactivating mutations in the CYP21A2 gene (2). The mild nonclassical 21OHD form is reported to occur in a frequency up to 1 in 500 (3). Patients are asymptomatic at birth and manifest in later childhood or adolescence with premature pubarche, precocious pseudopuberty or hirsutism, and decreased fertility. The most frequent CYP21A2 mutations derive from unequal crossing over or gene conversion between the active CYP21A2 gene and the inactive CYP21A1P pseudogene (4). Two thirds of 21OHD patients are compound heterozygous.

Many oxidative reactions are catalyzed by heme-containing cytochrome P450 enzymes (P450). P450 type II enzymes receive electrons from its electron donor P450 oxidoreductase (POR) (5). Reduced or absent activity of POR leads to a recently described form of CAH (POR deficiency, OMIM no. 201750) (6).

We describe a patient suffering from nonclassical 21OHD in whom a novel CYP21A2 point mutation was detected. The residual 21-hydroxylase activity was assessed in vitro, and the putative structural consequences were studied in silico using our three-dimensional molecular CYP21 model revealing a change in surface charge as a possible pathogenic mechanism.

Patients and Methods

Patient, clinical presentation, and hormonal analyses

Physical examination of the female patient originating from Northern Germany at age 7 yr revealed pubic hair stage P2, no additional signs of virilization, and no breast development. Her height was 135.6 cm (+1.5 sd score) with a significantly accelerated bone age of 10 yr (Greulich-Pyle). The weight was 40.3 kg, resulting in a body mass index of 21.9 kg/m² (+2.3 sd score). Plasma dehydroepiandrosterone-sulfate was increased to a pubertal concentration of 5552 nmol/liter (normal range for age, 156-2321 nmol/liter). Morning baseline steroid levels were 17-hydroxyprogesterone (17-OHP) 11 nmol/liter (normal range for age, 0–2.5 nmol/liter), 21-desoxycortisol 5.3 nmol/liter (normal range for age, 0.6–3.6 nmol/liter), and cortisol (F) 119 nmol/liter (normal range for age: 98–510 nmol/liter). Sixty min after stimulation with 250 μg ACTH 17-OHP was 45 nmol/liter (normal range for age: 0.7–11 nmol/liter, range for nonclassical 21OHD: 29–61 nmol/liter), 21-desoxycortisol 28 nmol/liter (normal range for age: 0.1–18 nmol/liter, range for nonclassical 21OHD: 7–35 nmol/liter) and F 643 nmol/liter (normal range for age: 441-1600 nmol/liter, range for nonclassical 21OHD: 298-1256 nmol/liter). Adrenal steroids were measured by multisteroid analysis including solvent extraction, chromatography and RIA (7). Based on the clinical and hormonal data the diagnosis of nonclassical CAH was established and treatment with hydrocortisone was started.

Mutation analysis and plasmid construction

Screening for the most common CYP21A2 point mutations was performed by multiplex minisequencing (8) followed by direct DNA sequencing (9). Informed consent was obtained from the parents in advance. Mutagenesis was performed from a pGEM3Z-CYP21 construct kindly provided by W. L. Miller (Department of Pediatrics, University of California, San Francisco) by standard procedures. Introduction of the K121Q and the P453S mutation was verified by sequencing the entire construct. Transfer into a pcDNA3 expression vector was performed as described previously (10).

In vitro expression and assay of enzyme activity

In vitro expression in COS-7 cells was performed as described recently (10). COS-7 cells were transiently transfected with 1 μg wild-type or mutant pcDNA3-CYP21A2 construct together with 400 ng pRK-TK (Promega, Mannheim, Germany) coding for Renilla luciferase. CYP21 activity in intact COS-7 cells was determined 48 h after transfection by the conversion of ³H-labeled substrate [17-OHP or progesterone (P)] measured directly from thin-layer chromatography plates (10). Cellular protein content was determined and luciferase activity was measured following a standard protocol (Promega). For the determination of apparent kinetic constants, intact COS-7 cells were incubated with 0.5, 1.0, 2.0, 4.0, 8.0, or 12.0 μmol/liter unlabeled steroid.

CYP21 activities were expressed as described (10). The apparent kinetic constants were calculated from the measurements of 21-hydroxylase activity in intact COS-7 cells at each of the different substrate concentrations with the GraphPad Prism software version 4.0. An antihuman-CYP21 rabbit polyclonal antiserum provided by Dr. W. L. Miller was used for Western blot analysis to ensure the expression and translation of the intact CYP21 wild-type and mutant protein.

Immunofluorescence

Immunofluorescence was performed with the antihuman-CYP21 rabbit polyclonal antiserum in combination with a mouse anti-KDEL antibody (BIOMOL, Hamburg, Germany) as marker for the smooth endoplasmic reticulum (ER) (10).

Molecular modeling

The three-dimensional CYP21 model recently built by our group using the x-ray structure of the mammalian cytochrome P4502C5 (PDB accession code 1DT6) as template (10) was used to study the localization of the amino acid residue K121. The structural representation was generated with the Deep View/Swiss-PDB Viewer program and POV-Ray (www.povray.org).

Results

Mutation analysis

Mutational screening by multiplex minisequencing revealed the mutation c.1361C>T (P453S) in the heterozygous state in the index patient and her father. DNA sequencing of the CYP21A2 gene detected the mutation c.364A>C (K121Q) in exon 3 in the heterozygous state in the index patient and her mother [nucleotide numbering according to Higashi et al. (11)]. Therefore, the patient was compound heterozygous for P453S (paternal allele) and K121Q (maternal allele). Two older brothers were heterozygous for P453S and K121Q, respectively. The younger sister showed wild-type sequences at both positions.

Enzyme activity

The K121Q mutant demonstrated a reduced CYP21 activity of 19.5 ± 4% (sd) of wild-type activity for the conversion of P to deoxycorticosterone and of 14.0 ± 5% of wild-type activity for the conversion for 17-OHP to 11-deoxycortisol detected at 2 μm of the respective steroid.

The P453S mutant revealed a reduced CYP21 activity of 36 ± 5% (sd) of wild-type activity for the conversion of P to deoxycorticosterone and of 44 ± 3% of wild-type activity for the conversion for 17-OHP to 11-deoxycortisol detected at 2 μm of the respective steroid. The relative enzymatic activity of P453S for both steroids was significantly higher than the K121Q activity (t test, P < 0.01; n = 9).

Determination of the apparent kinetic constants revealed a decreased affinity of K121Q for both substrates and a decreased maximal velocity for the synthesis of deoxycorticosterone and 11-deoxycortisol. Substrate affinity of P453S was not different from the wild type, but the maximal velocity for both substrates was decreased (Table 1 and Fig. 1).

Western blot analysis of the wild-type and mutant proteins demonstrated that both mutations did not affect the translation efficiency (data not shown). Expression and intracellular localization in the ER of mutant protein was not different from wild-type CYP21 protein (data not shown).

Discussion

21OHD is caused by inactivating mutations in the CYP21A2 gene (4). Only 3–5% of all mutations are unique novel mutations (12). The investigation of these novel mutations is important to estimate their severity and elucidate the consequences on phenotypic expression. Furthermore, studying these novel defects provides novel insights into the structure-function relationship of CYP21 and other P450 enzymes.

The patient presented with the typical symptoms of nonclassical CAH. The paternal P453S mutation is typically associated with nonclassical CAH and has a residual activity of 20% for the conversion of P and 50% for the conversion of 17-OHP (13). In our in vitro system, the residual activity of P453S was 36% for the conversion of P and 44% for the conversion of 17-OHP. The novel mutation K121Q has a residual enzymatic activity of 19.5% for the conversion of P and of 14.0% for the conversion of 17-OHP. Depending on the residual activity, CYP21 mutations are classified in four groups (0, A, B, and C). Mutation group C includes mutations with a residual activity of 20–50% resulting in nonclassical CAH, and mutation group B includes mutations with a residual activity of 1–5% (14). However, the classification is more or less arbitrary, because a continuum of residual activity with different mutations can be expected in a biological system. We therefore suppose that the K121Q mutation is correlated with an intermediate severity resulting in a phenotype between nonclassical and simple virilizing forms. This has been described for other mutations such as P30L or A434V (10, 15). In contrast to mutants associated with nonclassical 21OHD having an apparent Michaelis-Menten constant comparable to the wild-type enzyme, the K121Q mutation reveals an increased K_m as found with more severe mutations (16). The phenotype of the presented patient depends on the residual activity of the P453S mutation, because this has the higher residual activity.

The activity of P450s depends on the electron flux from its redox partners. Cytochrome P450 type II enzymes such as CYP21 receive electrons from POR. POR and type II enzymes are aligned by hydrophobic interactions with the ER membrane. The interaction of POR and type II enzymes is facilitated by positively charged basic amino acid residues in the binding site of the type II enzyme and a negatively charged acidic surface on the POR surface (17). A cluster of basic amino acid residues, including K121 in helix C, can be found on the surface of the redox partner binding site of CYP21 (Fig. 2C). Other superficial basic residues are K117, R132, R339, K414, and R435 (18). We propose that mutations of these basic amino acid residues forming the redox-partner binding site are likely to cause electrostatic disturbances impairing or preventing proper electron flux from POR to CYP21.

However, impaired electron flux should disturb maximal velocity of the enzymatic reaction without influencing substrate affinity. Because K121Q causes both a decreased reaction velocity and reduced substrate affinity, additional molecular mechanisms are possibly involved causing reduced enzyme activity. Helix C is involved in the coordination of the heme group through a conserved WXXXK motif (19). W116 and K120 coordinate the heme propionate ring D (Fig. 2, A and B). P450 enzymes alter their conformation in specific regions in response to the ligand (20). The four regions with the highest flexibility are helices B′, C, F, G, and I and the β₄ system (21). Flexibility is possible through surrounding GXG motifs. Crystal structures of CYP2B4 in a ligand-bound and unbound state revealed that the heme coordinating residues in helix C are severely displaced or even exchanged with ligand binding (21). In addition, the conformation of the redox partner binding site is significantly different after ligand-induced fit as a result of the different placement of helix C and the C/D loop. Because of these experimental data, we assume that similar changes may occur in CYP21 in the steroid-bound state, which are likely to be changed by the K121Q mutation explaining the change in substrate affinity.

A unique novel missense mutation of the CYP21A2 gene has been characterized, emphasizing the importance of the integrity of the redox partner binding site and helix C. Recent insights in the interaction of POR and type II cytochrome P450 enzymes and the induced structure of P450 enzymes with ligand binding may provide an explanation for the molecular cause of CYP21 inactivation caused by the K121Q mutation.

Acknowledgments

We are grateful to Dr. W. L. Miller for providing the CYP21A2 cDNA and the antihuman-CYP21 rabbit polyclonal antiserum. We appreciate the technical assistance of Tanja Dahm and Gisela Hohmann.

Disclosure Statement: The authors have nothing to disclose.

Abbreviations

 
• CAH,

  Congenital adrenal hyperplasia;

 
• P450,

  cytochrome P450 enzymes;

 
• ER,

  endoplasmic reticulum;

 
• 21OHD,

  21-hydroxylase deficiency;

 
• 17-OHP,

  17-hydroxyprogesterone;

 
• P,

  progesterone;

 
• POR,

  P450 oxidoreductase.

References