Introduction

Congenital adrenal hyperplasia (CAH) is a common autosomal-recessive disorder caused mainly by defects in the steroid 21-hydroxylase (CYP21) gene. More than 90% of CAH cases are caused by mutation of the CYP21 gene (White and Speiser 2000). To date, 72 different CYP21 mutations have been reported (Human Gene Mutation Database 2003), among which 57 are spontaneous mutations. The remaining 15 mutations (Lee 2001) are believed to be products of intergenic recombinations of DNA sequences between the CYP21 gene and the highly homologous CYP21P pseudogene. These mutations can be identified by allele-specific oligonucleotide (ASO) (Speiser et al. 1994) and reverse dot-blot hybridizations (Yang et al. 2001), direct DNA sequencing (Tajima et al. 1993), PCR/ligase (Day et al. 1995), and the amplification-created restriction site (ACRS) method (Lee et al. 1996). For some unknown defective loci somewhere within the CYP21 gene, SSCP (Tajima et al. 1993; Lee et al. 1998) and denatured gradient gel electrophoresis (DGGE) (Ohlsson et al. 1999) may be applied to possibly find novel mutations in the CYP21 gene. However, all these applications for mutational detection should be carried out using the functional CYP21 gene. Therefore, the primary PCR product of the CYP21 gene has to be prepared by differential PCR amplification to eliminate any CYP21P contamination prior to mutational analysis. Most primary PCR products are derived by multiple allele-specific primer sets to generate the CYP21 gene in two PCR fragments using two-to-three-step PCR amplification (Day et al. 1996; Ordonez-Sanchez et al. 1998). For convenience and for amplification efficiency, a single, complete CYP21-specific amplification performed in a single reaction mixture is used (Lee et al. 1996). This PCR product contains the entire CYP21 gene, which can be cloned and characterized by an expression analysis (Lee and Chang 2001).

Identification of the 30-kb gross gene deletion (White et al. 1984) encompassing the C4A and CYP21 genes in the C4-CYP21 repeat module was controversial in the past (Miller 1988; White et al. 1988; Morel et al. 1989), even though the deletion had been demonstrated by pulse field electrophoresis (Collier et al. 1989). The ambiguity was mainly caused by the lack of detailed information on the extent of gene conversion or deletion in the C4-CYP21 repeat module or RCCX module (Shen et al. 1994) in chromosome 6p21.3. This module includes CYP21P, XA (Gitelman et al. 1992), the serine/threonine nuclear protein kinase RP (Yang et al. 1998), C4, CYP21, and tenascin B (TNXB) (Bristow et al. 1993) genes (Fig. 1A). The C4 protein is coded by two genes, C4A and C4B. Long gene (20.4 kb) or short gene (14.1 kb) of the C4 is due to the presence of an endogenous retrovirus sequence (6.7 kb), HERV-K (C4), in intron 9 (Yu 1991). TNXB, in the downstream CYP21 gene, is partially duplicated in the downstream CYP21P gene, where it is termed XA gene (TNXA). Both XA and TNXB are transcribed on the opposite strand. The RP gene contains two duplicated genes, RP1 and RP2. The RP2 gene is truncated and corresponding to RP1 adjacent to TNXA (Yang et al. 1999). These tandemly arranged genes, CYP21P-XA-RP2-C4B-CYP21-TNXB, are designated as the RCCX module. At present, such a deletion is considered to result from an unequal crossover in meiosis occurring in 20% of alleles in most populations (White and Speiser 2000). However, the population frequency is dependent on the population studied (White and Speiser 2000). Three reports (Levo and Partanen 1997; Koppens et al. 2000;L’Allemand et al. 2000) pointed out that such a gross 30-kb deletion consisted of a fused CYP21 gene, with its 5’ and 3’ ends corresponding to CYP21P and CYP21 respectively, and the product appearing as a 3.2-kb Taq I fragment in Southern blot analysis. In a recent study (Lee et al. 2003a), it demonstrated that such a 30-kb gene deletion in fact is a chimeric CYP21P/CYP21 formation caused by multiple gene deletions, including XA, RP2, and C4B, and between unequal parts of the CYP21P and CYP21 genes in the C4-CYP21 repeat module (Fig. 1A). Such a gene deletion or gene conversion is traditionally detected by Southern blotting with multiple isotope-labeled probes and RFLP analysis; Taq I generates the 3.7-kb (functional) and 3.2-kb (pseudogene) fragments, while Bgl II produces the 11-kb (functional) and 12-kb (pseudogene) fragments. These two approaches to fragment analysis have been used since 1984 (White et al. 1984; Donohoue et al. 1989; Koppens et al. 2000; L’Allemand et al. 2000). However, the method is laborious and indirect, and densitometric screening of fragments is error prone, although a nonisotopic Southern procedure was later described (Krone et al. 1998).

Fig. 1
figure 1

A Strategy for identification of the chimeric CYP21P/CYP21 in the RCCX module of the human histocompatibility (MHC) class III gene. The structure of the normal gene is shown in the white box; the black box represents a nonfunctional gene (CYP21P and XA) or a different functional gene (C4A). The deleted region of the RCCX module is indicated by a solid line, which represents a 26- or 32-kb deletion as described in the text published previously (Lee et al. 2003a). The chimeric CYP21P/CYP21 is indicated by a half-white, half-black box. A 120-bp deletion in exon 36 of the TNXA gene and 707–714delGAGACTAC in the CYP21P gene are marked with an asterisk (*). Vertical arrows show the location of the restriction endonuclease cleavages. Deletion of the XA, RP2, and C4B genes led to the formation of a chimeric CYP21P/CYP21 gene in a 9.3-kb fragment digested by AseI and NdeI endonucleases by Southern analysis, which was described previously (Lee et al. 2003a). B Restriction analysis of the amplification-created restriction site (ACRS) amplification product (Lee et al. 1996) and the map for detection of mutational loci in three CAH carriers with distinct chimeric CYP21P/CYP21. The three distinct chimeras identified were designated CH-1, CH-2, and CH-3. ACRS analysis was carried out on 2.5% Metaphor (FMC Bioproducts, USA). The ACRS primers used for the detection of P30L (a), IVS2 −12A/C>G (b1), I172 N (c), I236 N (d), V281L (g), Q318X (j1), and R356 W (j2) have been described previously (Lee et al. 1996; 2002). On the gel, lanes 2, 6, 10, 14, 18, 22, and 25 were a carrier with CH-1. Lanes 3, 7, 11, 15, 19, 23, and 26 were a carrier with CH-2. Lanes 4, 8, 12, 16, 20, 24, and 27 were a carrier with CH-3. Each amplification product was either untreated (”−” lanes) or treated (”+”) with an appropriate restriction enzyme (RE) (Lee et al. 1996). The marker used was a 100-bp molecular ladder (lane “mk”). The schematic representation is a result of three identified chimeras detected by the ACRS method. Exons for the CYP21P and CYP21 genes are shown as solid blocks and open boxes respectively. The vertical arrow indicates the position of the locus corresponding to the 2.5% metaphor analysis. A horizontal arrow indicates the position of the chi-like sequence and the minisatellite consensus. C A schematic diagram of the strategy for PCR amplification of a 6.2-kb fragment by allele-specific primers (CYP779f/Tena 36F2) encompassing the TNXB gene to the CYP21 gene (Lee et al. 2003a, 2003b). The structure of the CYP21 and C4B genes is shown by a white box, while the black box represents the C4A gene and exons of the TNXB gene. The chimeric CYP21P/CYP21 gene is indicated as described in A. Horizontal arrows represent the direction and location of the primers CYP779f and Tena36F2. Primer Tena36F2 is located in exon 36 of the TNXB gene in which a 120-bp fragment deleted in TNXA is marked by an asterisk (*)

Full size image

Causes of the formation of the chimeric CYP21P/CYP21 gene

Basically, the chimera features fused genes. In the case of P450c21, chimeric CYP21P/CYP21 is formed by recombination of the duplicated CYP21P and CYP21 genes, but the cause of formation is unclear. However, several possibilities have been proposed:

Proximity to the human major histocompatibility complex (MHC)

In order to increase diversity of transplantation antigens and to maximize versatility of immune responses in humans, there is a high recombination rate in the MHC of the HLA complex. Both the CYP21P and CYP21 genes are on chromosome 6p21.3, adjacent to and alternating with the C4A and C4B genes encoding the fourth components of the serum complement (White et al. 1984), located within the HLA complex. As a result, the frequency of recombination events between these two genes is relatively high in human populations.

Homologous recombination

Because the exon sequence of the CYP21P pseudogene shares 98% nucleotide sequence homology with functional CYP21 (Higashi et al. 1986; White et al. 1986), homologous recombination of misalignment may occur at meiosis (Tusie-Luna and White 1995). It has been noted that the region between intron 2 and the 3’ end of exon 3 in CYP21 is considered a hotspot for recombinations and microconversions (Tusie-Luna and White 1995).

Initiation by the existence of a chi-like sequence and tandem-repetitive minisatellite consensus

Sequences such as chi (GCTGGTGG; Smith et al. 1981) and the tandem-repetitive minisatellite consensus (GGGCAGGAXG; Jeffreys et al. 1985) are important for genetic recombination in eukaryotic cells. These sequences may be recognition sites for the recombinase complex responsible for gene recombination. It has been pointed out that the chi-like sequence GCTGGGG is present several times in the CYP21P and CYP21 gene (Table 1). Most of them appear in intron sequence. Two distinct areas of nucleotides (nt), 470–999 and 1375–1993 (Higashi et al. 1986), of CYP21 genes have been proposed to be related to crossover regions (Donohoue et al. 1989). On the other hand, there are three minisatellite consensus sequences located in the CYP21 gene at nt 1113–1121 (IVS4), 1950–1959 (IVS7), and the 3’ untranslated region from nt 2710–2719 (Urabe et al. 1990) (Table 1). One report suggested unequal crossover in exon 7 sequences (Chu et al. 1992).

Table 1 Distribution of the chi-like sequence and minisatellite consensus in the CYP21P and CYP21 genes
Full size table

Types of chimeric CYP21P/CYP21

There have been various studies (White et al. 1988; Sinnott et al. 1990; Levo and Partanen 1997; Koppens et al. 2000) indicating hybrid genes with the 30-kb deletion between CYP21P and CYP21 genes in Caucasians. Several studies (Lee et al. 2002; Lee et al. 2003a) have shown that there are three distinct chimeric CYP21P/CYP21s in the CYP21 gene in the ethnic Chinese population in Taiwan. These three molecules are designated CH-1, CH-2, and CH-3 (Fig. 1B) (Table 2). Results indicated that CH-1 has an identical sequence to the CYP21P gene from exons 1–3 without the mutation at I172 N (nt 999) (Fig. 1B) (Table 2), CH-2 is identical to the CYP21P gene from exons 1–5 of D183E (nt 1117), and CH-3 is identical in sequence to the CYP21P gene from exons 1–8 of Q318X (nt 1994) without the R356 W mutation (Fig. 1B) (Table 2) (Lee et al. 2002; Lee et al. 2003a). In addition to these, the three chimeric CYP21P/CYP21s possess identical 5’ ends of the CYP21P gene (Table 2) (Lee et al. 2002; Lee et al. 2003a). Obviously, structural differences of the three chimeras show that these molecules are composed of unequal sequences between the CYP21P and CYP21 genes leading to the production of a variety of structural arrangements.

Table 2 Sequence analysis of the three chimeric CYP21P/CYP21 genes from ethnic Chinese congenital adrenal hyperplasia (CAH) patients. nt nucleotide, aa amino acid, Del deletion
Full size table

Properties of the chimeric CYP21P/CYP21 gene

Allele dropout in PCR amplification for detecting the IVS2 −12A/C>G mutation

As previously reported (Day et al. 1996), the allele dropout that occurs in the detection of the IVS2 −12A/C>G mutation at nt 655 (Higashi et al 1986) is due to preferential amplification of DNA segments and is an artifact caused by polymerase (Schulze et al. 1998). This leads to an excessively high frequency of apparent homozygosity and genotyping discordance between parents and probands. A previous study (Lee et al. 2000) pointed out that allele dropout was associated with unequal amplification in the presence of the chimeric CYP21P/CYP21, which lacks a specific primer for amplification during analysis of the mutation IVS2 −12A/C>G. In addition, a recent study (Lee et al. 2003b) found that a haplotype of the CYP21 gene with the mutations of IVS2 −12A/C>G and 707–714delGAGACTAC (Higashi et al. 1986) is a case of gene deletion including the CYP21P, XA, RP2, and C4B genes in the C4-CYP21 repeat module. This finding further confirms that allele PCR dropout in previous PCR analyses was in fact caused by the presence of the 5’ end region of the CYP21P sequence, which is consistent with our previous suggestion (Lee et al. 2000). Since an antisense primer anchored at nt 707–714 of CYP21 has been used in most primary PCR amplifications (Tajima et al. 1993; Day et al. 1996; Koppens et al. 2000) to eliminate CYP21P contamination, it may have led to PCR dropout in detecting CYP21 mutations of IVS2–12A/C>G combined with707–714delGAGACTAC and the chimeric CYP21P/CYP21.

Production of a 3.2-kb fragment by Taq I digestion

Since three chimeric CYP21P/CYP21 genes contain an identical sequence in the 5’ end of CYP21P (Table 2) resulting in a frameshift of the sequence upstream of CYP21P, each of the chimeras has the base T at nt −209 replaced by C (Higashi et al. 1986) (Table 2), producing a Taq I restriction site in these defective CYP21 alleles and subsequently leading to a 3.2-kb Taq I fragment (Fig. 2A). Therefore, the assumption that the 3.2-kb fragment from conventional Southern analysis is a deletion product of the CYP21 gene is incorrect.

Fig. 2
figure 2

A Analysis of the 6.2-kb PCR product by Taq I digestion on a 0.65% agarose gel. Lane 1: the 6.2-kb PCR product amplified with paired primers Tena36F2/CYP779f from a normal individual. Lanes 2 and 3 are the 6.2-kb PCR product digested by Taq I from a normal individual and a congenital adrenal hyperplasia (CAH) patient with CH-1 chimera in one of the chromosomes respectively. “mk” was a Lambda DNA-BstE II-digested molecular marker (New England BioLabs, Beverly, MA, USA). B Southern blot analysis of Ase I and Nde I digestion of genomic DNA. The hybridization probe used consisted of the 2271-bp PCR product derived using the paired primers Tena36F2/Tena43R (Lee et al. 2003a). A normal individual (lane N) had two fragments of 11.3 and 21.6 kb, and one patient with a CH-1 chimera in one of the chromosomes (lane M) had an additional 9.3-kb fragment as well as having the 11.3- and 21.6-kb fragments

Full size image

Production of a 9.3-kb fragment resulting from deletion of the TNXA, RP2, and C4B genes

In the traditional Southern blot analysis, Taq I generates a 3.7-kb (functional) and a 3.2-kb (pseudogene) fragment, and Bgl II produces an 11-kb (functional) and a 12-kb (pseudogene) fragment, which cannot indicate detailed information on the extent of gene conversion or deletion in the RCCX module. Therefore, a study (Lee et al. 2003a) using two restriction endonucleases, Ase I and Nde I, with a single probe was used to identify the interchange region. Results revealed that deletion of the TNXA, RP2, and C4B genes (Fig. 1A) leads to formation of the chimeric CYP21P/CYP21 in a 9.3-kb fragment (Fig. 2B, lane M). This improves the detection protocol for gene deletions and conversions in the RCCX module (Fig. 1A). Given the potential for variations and a lack of information about the C4B gene (Yu 1991) in Chinese populations, the size of the RCCX modular deletion may either be 26 or 32 kb long (Fig. 1A) (Lee et al. 2003a).

Biological activity of the chimeric CYP21P/CYP21

Because the 5’ end of the three different chimeric CYP21P/CYP21 genes have identical sequences with that of CYP21P (Table 2), nucleotide substitutions at these sites from CYP21P- to CYP21-specific sequences at nt −103, −110, and −123 (Higashi et al. 1986) cause a five-fold decrease in transcriptional activity and in the ability of the promoter to bind with the Sp1 protein (Chang and Chung 1995). At nt 655 (Higashi et al. 1986) of the IVS2 -12A/C>G mutation, the aberrant splicing site results in very low but measurable enzyme levels (Higashi et al. 1988a). In addition, deletion of 707–714delGAGACTAC leads to a frameshift mutation, which forms a TGA stop codon downstream at nt 830 (Higashi et al. 1986) and produces a truncated protein in translation. Taken together, these three kinds of chimeric CYP21P/CYP21 have no steroid 21-hydroxylase activity. However, the hybrid gene reported by the study (L’Allemand et al. 2000) without these two mutations had low 21-hydroxylase activity with a defect in P30L.

Identification of the chimeric CYP21P/CYP21 gene

Identification of gene deletions and conversions is still being studied by Southern blot analysis with oligonucleotides (Higashi et al 1988b; Donohoue et al. 1989; Helmberg et al. 1992; Tusie-Luna and White 1995) or DNA probes (Krone et al. 1998; Koppens et al. 2000; L’Allemand et al. 2000) for hybridization. There are two established strategies of PCR amplification for identifying the CYP21 gene:

  1. 1.

    Mixed-primer amplification of the 3.5-kb PCR product: To collect the chimeric CYP21P/CYP21, three primers (Table 3) containing a 3’-specific primer for the CYP21 gene and two different 5’-primers for both CYP21 and CYP21P used to amplify the normal CYP21, and the chimera have been successfully used for CAH diagnosis (Lee et al. 2000; 2002). The 3.5-kb fragments (Lee et al. 2000; 2002) covering sequences of the CYP21 gene from −313 to 3170 (Higashi et al. 1986) were generated. The PCR products were analyzed using the ACRS primer for mutational analysis (Fig. 1B) (Lee et al. 1996).

    Table 3 Primers for the amplification of chimeric CYP21P/CYP21 gene
    Full size table
  2. 2.

    The 6.2-kb PCR product encompassing the TNXB gene to the 3’-end of the CYP21 gene: In order to understand the status of the CYP21 gene within the RCCX module after gene deletion or conversion, a strategy of using a 6.2-kb PCR product amplified with locus-specific primers from the tenascin (TNXB) gene to the 5’ end of CYP21P or CYP21 has been successfully achieved (Figs. 1C, 2A) (Lee et al. 2003a,2003b). The sense primer located in tenascin B (TNXB) containing a 120-bp sequence deleted from tenascin A (TNXA) and an antisense primer in the 5’ end of the CYP21P and CYP21 genes were used (Table 3) (Fig. 1C). The 6.2-kb PCR product was further identified either by the ACRS primer to detect CYP21 mutations (Fig. 1B) (Lee et al.1996) or by Taq I digestion directly analyzed on an agarose gel to evaluate the 3.2- and 3.7-kb fragments (Fig. 2A).

Conclusions

Questions regarding gross gene deletion of 30 kb encompassing the C4A and CYP21 genes have arisen because there is no information on the deleted region in the RCCX module. The chimeric CYP21P/CYP21 is a consequence of the deletion of the complete XA, RP2, and C4B, and partial CYP21P/CYP21 genes of 26 or 32 kb, depending on whether C4B is a long or short gene. Therefore, presentation of a 3.2-kb fragment produced by Taq I digestion is not only a single CYP21 deletion but there is also a chimera included. A previous strategy (Krone et al. 1998) using the Taq I restriction enzyme was applied to specifically disrupt the CYP21P gene for identification of the CYP21 mutation, which may possibly cause a loss of amplification of the chimera. Results of a recent study (Lee et al. 2003b) indicated that the haplotype of the CYP21 allele with the mutation IVS2–12A/C>G combined with 707–714delGAGACTAC is not a chimeric CYP21P/CYP21 gene but does show a 3.2-kb fragment by Taq I digestion. Therefore, we believe that the diversity of haplotypes of the CYP21 allele with the 3.2-kb fragment may frequently occur in CAH patients with steroid 21-hydroxylase deficiency.