High-resolution melting curve (HRM) analysis to establish CYP21A2 mutations converted from the CYP21A1P in congenital adrenal hyperplasia

(1)

UNCORRECTED PR

OOF

1

2

Q2

High-resolution melting curve (HRM) analysis to establish CYP21A2 mutations

3

converted from the CYP21A1P in congenital adrenal hyperplasia

Q3

4

Yi-Ching Lin

a,b

, Yu-Chih Lin

c

, Ta-Chi Liu

a,b,d

, Jan-Gowth Chang

a,b,e,f,

⁎

, Hsien-Hsiung Lee

g,

⁎⁎

5 a

Institute of Clinical Medicine, Kaohsiung Medical University, Kaohsiung 807, Taiwan

6 b

Department of Laboratory Medicine, Kaohsiung Medical University Hospital, Kaohsiung Medical University, Kaohsiung 807, Taiwan

7 c

Division of General Internal Medicine, Department of Internal Medicine, Kaohsiung Medical University Hospital, Kaohsiung Medical University, Kaohsiung 807, Taiwan

8 d

Division of Hematology and Oncology, Department of Interal Medicine, Kaohsioung Medical University Hospital, Kaohsioung Medical University, Kaohsioung, Taiwan

9 e

Center for Excellence in Environmental Medicine, Kaohsiung Medical University, Kaohsiung 807, Taiwan

10 f_{Cancer Center, Kaohsiung Medical University Hospital, Kaohsiung 807, Taiwan}

11 g_{School of Chinese Medicine, College of Chinese Medicine, China Medical University, 91 Hsueh-Shih Road, Taichung 404, Taiwan}

12 13

a b s t r a c t

a r t i c l e i n f o

14 Article history: 15 Received 2 June 2011 16 Accepted 24 June 2011 17 Available online xxxx 18 19 20 21 Keywords: 22 CYP21A2 23 CAH 24 HRM 25 PCR-based ampliﬁcation 26 Mutational detection 27 Heteroduplex 28 Homoduplex 29 Background: Congenital adrenal hyperplasia (CAH) is an autosomal recessive disease of an inborn error of

30 steroid metabolism in humans. More than 90% of CAH cases are caused by mutations of the steroid

21-31 hydroxylase (CYP21A2) gene, and approximately 75% of the defective CYP21A2 genes are generated through

32 an intergenic recombination with the neighboring CYP21A1P pseudogene.

33 Methods: A high-resolution melting (HRM) curve analysis was designed to characterize 11 mutation sites of

34 the CYP21A2 gene that commonly appeared in 21-hydroxylase deﬁciency. Among these 11 mutations, 9 were

35 found in CAH patients, and 2 were mutations created from normal individuals.

36 Results: From the HRM analysis using 6 fragments of amplicons, we have successfully identiﬁed these 11

37 common disease-causing mutations of the CYP21A2 gene, among which 3 showed 3 distinguishable melting

38 plots; the heteroduplexes showed an upcurved plot, a horizontal plot of homoduplexes of wild-type (WT),

39 and a downcurved plot of homoduplexes of compound mutations.

40 Conclusions: The HRM analysis is a 1-step of non-gel resolution technique which saves time and is a low-cost

41 method to undertake such a program for screening CAH patients with the 21-hydroxylase deﬁciency caused

42 by intergenic conversions from the neighboring CYP21A1P pseudogene.

49 Congenital adrenal hyperplasia (CAH) is an autosomal recessive 50 disease of an inborn error of steroid metabolism in humans. It may 51 produce excessive or deﬁcient sex steroids and can alter development of 52 primary and secondary sex characteristics. There are 6 enzymes, 53 cholesterol side-chain cleavage enzyme (CYP11A), CYP17 (17, 20-54 lyase), steroid 21-hydroxylase (CYP21A2), steroid 11-beta-hydroxylase 55 (CYP11B1), steroid 18-hydroxylase (CYP11B2), and 17 β-56 hydroxylsteroid dehydrogenase, that are required for the synthesis of 57 steroid hormones. However, more than 90%–95% of all CAH cases are 58 caused by a CYP21A2 deﬁciency[1]. There are 3 forms of CAH: the 59 classical salt-wasting, classical simple virilizing, and non-classical forms 60 [2,3]. The incidence of the classical form of CAH disease is reported to be

61 1:10,000–1:18,000, depending on race [1,4] while the non-classical

62 form is milder and commonly occurs in the general population at a rate

63 of 1:1700[3,5].

64 The gene coding for P450c21 is designated CYP21A2. A duplicate

65 copy designated CYP21A1P exists which shares 98% nucleotide

66 sequence homology with CYP21A2 in exons and 96% in non-coding

67 sequences [6,7]. These two genes are separated by 30 kb in

68 chromosome 6p21.3 adjacent to and alternating with the C4A and

69 C4B genes encoding the fourth component of the serum complement

70 and show great similarity. This seems the most likely reason for

71 misalignment and gene conversions to occur during meiosis [8].

72 Under this circumstance, genetic defects of the CYP21A2 gene in CAH

73 may commonly lead to 1 of 2 categories of (a) small-scale conversions

74 of the CYP21A1P sequence (commonly 1 of 11 mutations)[9]and (b)

75 chimeras of the chimeric CYP21A1P/CYP21A2 and TNXA/TNXB genes

76

[10–12]. The CYP21A1P is a nonfunctional gene which was thought to 77 carry 15 mutations[6,7]. However, a study of ethnic Chinese (i.e.,

78 Taiwanese)[13]indicated that not every healthy individual (n = 100)

79 bears these mutations, which had an approximately 90% in the

80 population frequency[13], and 4 loci of the I2 splice (including nt

81 707–714del), I172N, cluster E6, and F306ΛL307insT were processed

82 by “complete” selective pressure in evolution [13]. The CYP21A2

Clinica Chimica Acta xxx (2011) xxx–xxx

⁎ Correspondence to: J.-G. Chang, Department of Laboratory Medicine, Kaohsiung Medical University Hospital, 100 Shih-Chuan 1st Rd., Kaohsiung 870, Taiwan. Tel.: + 886 7 3115104; fax: + 886 7 3213931.

⁎⁎ Correspondence to: H.-H. Lee, School of Chinese Medicine, College of Chinese Medicine, China Medical University, 91 Hsueh-Shih Road, Taichung 404, Taiwan. Tel./fax: + 886 3 9389073.

doi:10.1016/j.cca.2011.06.033

Contents lists available atScienceDirect

Clinica Chimica Acta

(2)

UNCORRECTED PR

OOF

83 deﬁciency in our population (i. e., Taiwanese), approximately 81% of

84 which are defective CYP21A2 genes[14], is generated through an 85 intergenic recombination[9]with the neighboring CYP21A1P pseu-86 dogene. Among them, the 3 most common mutations of the CYP21A2 87 gene in ethnic Chinese (i.e., Taiwanese) of 69% frequencies are the I2 88 splice (nt 655, IV2-12A/C_{NG) (34%, n=400 chromosomes), I172N} 89 (23.5%), and R356W (11.8%)[14], which show high similar incidences 90 worldwide in different races [1,15,16]. The frequency of other 91 mutations such as Q318X, F306-L307insT, and cluster E6 are about 92 12%[14]. A mutation of V281L, the most common nonclassical disease 93 appearing in high frequency in patients in France, Austria, Italy, Spain, 94 Turkey, Argentine, and Portugal[15–17], was not found in Taiwanese 95 [18], Japanese[19], or Tunisian patients[20].

96 Polymerase chain reaction (PCR) amplification is an indispensable 97 tool for detecting a gene of interest in current molecular biology. The 98 molecular diagnosis of the CYP21A2 de_{ficiency through direct analysis} 99 of the CYP21A2 gene was proven to be feasible and accurate. To isolate 100 the CYP21A2 gene free from the CYP21A1P pseudogene, several 101 methods including 1-step[21,22]and 2-step methods [23_–25] for 102 amplification of the CYP21A2 gene were developed. These PCR 103 products with either 1 or 2 fragments as a template are subject to 104 known or unknown mutational detection using more-practical 105 methods, such as PCR/ligase detection[24], single-stranded confor-106 mation polymorphism (SSCP)[26], amplification-created restriction 107 site (ACRS)[27], real-time PCR [28], denaturing high-performance 108 liquid chromatography (DHPLC)[18], multiplex minisequencing[29], 109 laser desorption/ionization time-of-flight (MALDI-TOF) [30], and 110 multiple ligation-dependent amplification (MLPA) assay to detect 111 the CYP21A2 gene[31].

112 The aim of the present study was to use a high-resolution melting 113 curve (HRM) analysis to directly identify 11 nucleotide sequences 114 commonly appearing in the CYP21A1P gene, including p.P30L, the I2 115 splice, nt 707–714del, p.I172N, cluster E6, p.V281L, F306ΛL307inseT, 116 p.Q318X, and p.R356W and to establish such a rapid and precise 117 screening tool for CAH patients which account for 70%–80% of CAH cases. 118 2. Materials and methods

119 2.1. DNA samples

120 Genomic DNA was collected from 200 CAH patients in hospitals 121 across Taiwan from 1994 to 2006 [14]. All families requested an 122 extensive molecular diagnosis and provided informed consent. Among 123 these CAH patients, 9 mutations were from the unrelated patients which 124 accounted for about 81% of CAH cases[14]including the I2 splice where 125 G is substituted for A/C (designated B1), deletion of 8 base pairs (bps) in 126 exon 3 (nt707–714del, designated B2), isoleucine (ATC) at codon 172 127 substituted by asparagine (AAC) (p.I172N, designated C), cluster E6 128 (designated D), p.F306ΛL307insT (designated H2), glutamine (CAG) at 129 codon 318 substituted by a stop codon (TAG) (p.Q318X, designated J1), 130 and arginine (CGG) at codon 356 substituted by tryptophan (TGG) 131 (p.R356W, designated J2) (Fig. 1). The CYP21A2 mutations in these 132 patients were formerly determined by the ACRS method as previously 133 described[27]. In order to produce the heteroduplex DNA fragment for 134 the HRM analysis, patients with the haplotype of compound heterozy-135 gous mutations in the CYP21A2 allele were selected. Because of no 136 patient with the p.P30L (CCGNCTG) (designated A) or p.V281L 137 (GTGNTTG) (designated H1) mutations (Fig. 1) were found in our 138 population[18], we created these 2 mutations from a normal individual 139 as described previously[18].

140 2.2. A primary 3.5-kb differential PCR product of the CYP21A2 gene for 141 identifying 9 mutations converted from the CYP21A1P gene

142 To isolate the CYP21A2 free from the CYP21A1P genes, a 3.5-kb PCR 143 product covering 10 exons of the CYP21A2 gene was ampliﬁed with a

144 differential paired primer, BF1/21BR (Fig. 1), as described previously

145

[21]. To identify the CYP21A2 mutations converted from the CYP21A1P 146 gene, the 3.5-kb primary PCR products obtained from these CAH

147 samples were then used as templates to detect the 9 mutation sites.

148 2.3. A primary 3.0-kb PCR product containing a mixture of the CYP21A2

149 and CYP21A1P genes for creating P30L and V281L heterozygous

150 mutations in a normal individual

151 Because of no patient with the p.P30L or p.V281L mutations was

152 found in our population[18], a 3.0-kb PCR product was ampliﬁed with

153 a universal paired primer, CYP-270f/Ex10R[18](Fig. 1), to create the

154 p.P30L and p.V281L heterozygous mutations in 1 normal individual as

155 previously described [18]. The 3.0-kb PCR product contained a

156 mixture of the CYP21A2 and CYP21A1P genes which present the

157 haplotype of compound heterozygous mutations with 11 defective

158 alleles as does the CYP21A1P gene[6]. The 3.0-kb PCR product was

159 then used as a template to identify mutations of p.P30L (designated A)

160 and p.V281L (designated H1) (Fig. 1).

161 2.4. Secondary PCR ampliﬁcation of both the 3.5-kb and 3.0-kb PCR

162 products for the HRM analysis

163 The 3.5-kb PCR products ampliﬁed with the paired primer,

164 BF1/21BR, from these selected CAH samples and the 3.0-kb PCR product

165 ampliﬁed with the universal paired primer, CYP-270f/Ex10R, creating

166 p.P30L and p.V281L mutations from a normal individual were used as

167 templates for secondary PCR ampliﬁcation by HRM primers. There were

168 6 paired primers for the HRM analysis to detect 11 mutational loci. The

169 sequence and location of these HRM primers are listed inTable 1.

170 2.5. HRM analysis

171 The HRM analysis included a PCR reaction, DNA melting process, and

172 gene scanning for data analysis. These 3 programs can be performed on a

173 single instrument. The LightCycler® 480 Real-time PCR system (Roche

174 Diagnostics, Penzberg, Germany) with 96- or 384-well closed-tube

175 platforms is operated by the LightCycler® 480 Gene Scanning Software

176 (Vers. 1.5) which is an integrated, high-throughput real-time PCR

177 instrument, and these 3 programs can be completed within 1 h.

178 For the PCR program, the reaction mixture for 6 secondary HRM

179 primer PCR ampli_{ﬁcations contained a diluted primary PCR product}

180 (3.5- or 3.0-kb PCR product), 10μl of LightCycler®_{480 High Resolution}

181 Melting Master (commercially supplied, which contains FastStart Taq

182 DNA polymerase, 2× reaction buffer, dNTP, and High Resolution

183 Melting Dye) (Roche Diagnostics), 0.25μM of each primer, and

184 2.5 mM of MgCl2 in a ﬁnal volume of 20 μl. The High Resolution

185 Melting Dye only strongly binds to double-stranded (ds)DNA and has

186 nothing to bind single-stranded (ss)DNA. The PCR conditions

187 consisted of 2 steps: a denaturation–activation step at 95 °C for

188 10 min, and followed by a 45-cycle program (denaturation at 95 °C for

189 15 s, annealing at 60 °C for 15 s, and elongation at 72 °C for 15 s with

190 reading of theﬂuorescence; by a single acquisition mode).

191 The melting program in this study includes 3 steps:

denaturaliza-192 tion at 95 °C for 1 min, re-naturation at 40 °C for 1 min and then

193 melting with a continuousﬂuorescent reading from 60 to 90 °C at 25

194 acquisitions per °C. The software system can“watch” the processes of

195 dsDNA withﬂuorescence to a dissociated nothing-bound ssDNA and

196 then processes the raw melting curve data to form a different plot. The

197 plots obtained in the real-time stage with homozygous and

198 heterozygous samples, respectively, are signiﬁcantly different. The

199 shapes of difference plot curves of each DNA sample must be

200 reproducible in terms of both shape and peak height.

201 Gene scanning of the data analysis by the Gene Scanning Software

202 was comprised of 3 steps: normalization of the melting curves,

203 equilibrating to 100% as the initialﬂuorescence and to 0% as the

(3)

UNCORRECTED PR

OOF

204 ﬂuorescence remnant after DNA dissociation, and shifting of the

205 temperature axis of the normalized melting curves to a point where 206 the entire dsDNA was completely denatured. Then the difference plot 207 analyzes differences in melting curve shapes by subtracting the curves 208 from wild-type (WT) and mutated DNA (sequence variation), 209 therefore differences in the plots help cluster the samples into groups. 210 2.6. Confirmatory sequencing for secondary HRM PCR fragments 211 Before the HRM analysis, the secondary PCR products amplified 212 with the HRM paired primer (Table 1) (without using High Resolution 213 Melting Dye) for 11 mutation sites from unrelated patients were 214 confirmed by DNA sequencing (Supplemental Figs. 1, 2). The 215 sequence reaction was performed in afinal volume of 10 μl including 216 1μl of the purified PCR product, 0.8 μl of 2.5 μM of 1 of the PCR 217 primers, 2μl of the ABI PRISM terminator cycle sequencing kit v3.1 218 (Applied Biosystems, USA), and 2μl of 5× sequence buffer. The 219 sequencing program was a 25-cycle PCR program (denaturation 96 °C 220 for 10 s, annealing 50 °C for 5 s, and elongation 60 °C for 4 min), and 221 sequence detection was performed in the ABI Prism 3130 Genetic 222 Analyzer (Applied Biosystems).

223 3. Results

224 3.1. Use of the 3.5- and 3.0-kb PCR products for secondary HRM PCR 225 ampliﬁcation of 9 mutation sites in 9 unrelated patients and 2 created 226 mutation sites of P30L and V281L from a normal individual

227 To detect the 9 mutation sites of B1, B2, C, D (cluster E6), H2, J1 and J2 228 (Fig. 1) from 8 unrelated CAH patients with compound heterozygous

229 mutations (Supplemental Figs. 1, 2), a 3.5-kb primary PCR product

230 (Supplemental Fig. 3A, lane 1) (data from only 1 patient) was generated

231 by the paired primers, BF1/21BR. The 3.5-kb PCR product used as the

232 template was subjected to a secondary PCR ampliﬁcation (Supplemental

233 Fig. 3B) (data from only 1 patient) using the HRM paired primers

234

(Table 1) to produce 5 fragments of 226 bp (for loci B1 and B2

235 identi_{ﬁcation), 118 bp (for locus C identiﬁcation), 193 bp (for cluster E6}

236 identiﬁcation), 212 bp (for locus H2 identiﬁcation), and 283 bp (for loci

237 J1 and J2 identiﬁcation). On the other hand, the 3.0-kb PCR fragments

238 (Supplemental Fig. 3A, lane 2) ampli_{ﬁed with the paired primers,}

CYP-239 270f/Ex10R, were derived from 1 normal individual to detect 2 created

240 mutation sites of P30L and V281L which included 2 fragments of 182 bp

241 (for locus A identiﬁcation) and 212 bp (for locus H1 identiﬁcation)

242 (Supplemental Fig. 3B) generated by the secondary ampliﬁcation using

243 the HRM paired primers (Table 1). The HRM analysis was performed on 6

244 different secondary PCR fragments to cover these 11 mutation sites using

245 a 96-well plate of the LightCycler 480 system. In addition, 6 different

246 secondary PCR products of the WT prepared from a normal individual

247 were treated the same as those of CAH patients (data not shown).

248 3.2. HRM analysis of 11 different mutations in 6 different PCR fragments

249 Because a heterozygous DNA sample with a heteroduplex has 2

250 different rates of separation temperatures and while homoduplex has

251 1, the shapes of the melting curves obtained from these 2 samples,

252 respectively, are signiﬁcantly differed. The LightCycler® 480 Real-time

253 PCR system has the ability to monitor this process in high resolution

254 process to accurately document these changes. On the HRM analysis of

255 the 182-bp amplicon (Fig. 2A) with the created heterozygous mutation

256 of p.P30L (CCG/CTG) from the normal individual (Sc) (Supplemental

Table 1 t1:1

Primers for secondary PCR ampliﬁcation and the HRM analysis of the CYP21A2 gene. t1:2

t1:3 Designation Primer (5′–N3′) Location (nt)a

Amplicon (bp) Detection locusb

t1:4 1A2 CTGCTGGCTGGCGCCCGCCT 31–50 182 p.P30L (A)

t1:5 C100 GAAGAAG GTCAGGCCCTC 602–619 226 I2 splicec

(B1) and In3R CTTACCTCACAGAACTCCTG808–827 707–714del (B2)

t1:6 E4r AGGCACCTTGATCTTGTCTCC 808–827

t1:7 In3 TCTCCACAGCGCATGAGAGC 920–939 118 p.I172N (C)

t1:8 E4r GAGGCACCTTGATCTTGTCTCC 1016–1037

t1:9 Ex6 TCATGCTTCCTGCCGCAGTTC 1304–1324 193 Cluster E6d

(D)

t1:10 C8 TGCAAAAGAACCCGCCTCATAG 1475–1496

t1:11 C9 TGCAGGAGAGCCTCGTGGCAGG 1573–1594 212 pV281L (H1) and pF306ΛL307insT (H2)

t1:12 S7r GACGCACCTCAGGGTGGTGA 1764–1785

t1:13 In7-1 In7-1 CACTCAGGCTCACTGGGTTGC 1890–1910 283 pQ318X (J1) and R356W (J2)

t1:14 C12-1 ACCCTCGGGAGTCACCTGCTG 2152–2172

a_{Based on Higashi et al.}_[6]_.

t1:15

b

Designation of A to J2 is corresponding toFig. 1. t1:16

c

I2 splice, IVS2−12A/CNG or nt 655. t1:17

d

Cluster E6 represents I236N, V236E, and M239K. t1:18

Fig. 1. Diagram of 11 CYP21A2 mutations converted from the neighboring CYP21A1 pseudogene and primer sequences, and locations of the ampliﬁcation of the CYP21A2 and CYP21A1P genes. The paired primers, BF1/21BR, were used to amplify a 3.5-kb PCR product of the CYP21A2 gene. The universal paired primers, CYP-270f/Ex10R, were used to amplify a 3.0-kb PCR product of the mixture of the CYP21A2 and CYP21A1P genes. The structure of the CYP21A2 gene is indicated by a white box. Designations of A to J2 indicate the 11 mutation sites converted from the CYP21A1P pseudogene[18].

(4)

UNCORRECTED PR

OOF

257 Fig. 1A), it showed that the difference plot of the created heterozygous 258 mutation of pP30L (CCG/CTG) (sample Sc) differentiated the one of 259 WT subjects (CCG/CCG) (WT) (n = 3). Obviously, unambiguous 260 differences were present in the shapes of the melting curves for the

261 heteroduplexes and homoduplexes. On analysis of the 226-bp

262 amplicon with mutations of the I2 splice (IVS2-12A/CNG) (B1) and

263 707–714del (B2) (Fig. 2B) from 2 unrelated CAH patients

(Supple-264 mental Figs. 1B1, 1B2), sample #81 was heterozygous for the I2 splice

Fig. 2. Normalized and temperature-shifted difference plots of the HRM analysis for detecting 11 mutation sites of the CYP21A2 gene from different CAH patients. Sequences A to J are designated inFig. 1. Plot A represents a created heterozygous mutation of p.P30L in a normal individual (Sc). Plot B represents sample #81 with a heterozygous mutation of the I2 splice (B1), and sample #81 with a heterozygous mutation of 707–714del (B2). One sample with homozygous 707–714del mutations was included. Plot C represents samples #250 and #419 with a heterozygous mutation of p.I172N and sample #443 with a homozygous p.I172N mutation. Plot D represents sample #249 (D1) with a heterozygous mutation of p.I236N combined with p.V237E and sample # 393 (D2) with a heterozygous mutation of p.I236N, and pV237E combined with p.M239K. In addition, 1 sample with a homozygous mutation of p.I236N, and pV237E combined with p.M239K was included. Plot H represents a created heterozygous mutation of p.V281L in a normal individual (H1) (Sc) and sample #C13 with a heterozygous mutation of p.F306ΛL307insT (H2). Plot J represents sample #708 with a heterozygous mutation of p.Q318X, and sample #579 with a heterozygous mutation of p.R356W. One sample #89-1 with a heterozygous mutation of p.R316X was included. WT, wild-type subject; Sc, sample created; #, patient ID number.

(5)

UNCORRECTED PR

OOF

265 mutation which could easily distinguish it from WT subjects (n = 12)

266 and heterozygous for the 707–714del mutation of sample #109. A 267 homozygous 707_{–714del was identified as a downcurved plot which} 268 differed from the horizontal plot of the WT and sample #109 with an 269 upcurved plot. On analysis of the 118-bp amplicon with mutations of 270 p.I172N (Table 1), the HRM analysis (Fig. 2C) showed that sample 271 #250 (SupplementalFig. 1C) (and sample #419) had a heterozygous 272 mutation of p.I172N distinguished by a downcurved melting plot of 273 the homozygous p.I172N mutation of sample #443 (sequencing data 274 not shown) and a horizontal plot of WT subjects (n = 14). When 275 analyzing cluster E6 (I236, V237, and M239) (Fig. 1) of the 193-bp 276 amplicon (Fig. 2D), there were 2 mutational types shown in Taiwanese 277 CAH patients[14,32]. Sample #249 with heterozygous mutations of 278 p.I236N and p.V237E (Supplemental Fig. 1D1) and sample #393 with 279 heterozygous mutations of p.I236N, p.V237E, and p.M239K (Supple-280 mental Fig. 1D2) showed different melting curves and were identi_fied 281 as different groups from WT subjects (n = 3) by the HRM analysis. 282 Obviously, these 2 different samples (samples #249 and #393) with 1 283 nucleotide difference at M239 could be distinguished. From the 212-284 bp amplicon for the p.V281L and p.F306ΛL307insT (Table 1) HRM 285 analysis (Fig. 2E), the created heterozygous mutation of p.V281L of 286 sample Sc (Supplemental Fig. 2H1) and heterozygous mutation of 287 p.F306ΛL307insT of sample #C13 (Supplemental Fig. 2H2) could easily 288 be distinguished from WT subjects (n = 21), and different groups 289 could be identified from each other. When analyzing p.Q318X and 290 p.R356W in the exon 8 region (Fig. 1) of the 283-bp amplicon

291 (Table 1), sample #708 with heterozygous mutations (Supplemental

292 Fig. 2J1) of p.Q318X and sample #579 with heterozygous mutations of 293 p.Q318X (SupplementalFig. 2J2) presented upcurved plots which 294 differed from the horizontal plot of WT subjects (n = 12) as different 295 groups from each other using the HRM analysis (Fig. 2F).

296 Obviously, the HRM analysis of the CYP21A2 gene with 11 different 297 mutations converted from the CYP21A1P pseudogene showed 3 298 distinguishable melting plots which included the heteroduplexes 299 that showed an upcurved plot, a horizontal plot of homoduplexes of 300 WT, and a downcurved plot of homoduplexes of compound 301 mutations. In addition, polymorphic sites which inﬂuenced the 302 heteroduplex form in the collected amplicon (Table 1) for identifying 303 the CYP21A2 gene are listed in Table 3

304 4. Discussion

305 CAH is a term that describes several inheritable disturbances in 306 steroid hormone metabolism. Gene conversion, i.e., changing part of 1 307 gene to the sequence of a nearby homologous gene (often its 308 pseudogene), is often the cause of genetic defects and the issue of 309 small-scale conversions generating the defective CYP21A2 gene is the 310 most frequent of the 21-hydroxylase deﬁciencies in CAH. The wide 311 range of CAH phenotypes is associated with multiple mutations 312 known to affect 21-hydroxylase enzymatic activity. Clinically, muta-313 tions of the I2splice, 707–714del, the cluster E6 (I236N and V237E) 314 [33], F306ΛL307insT, Q318X, and R356W produce a picture of the 315 classic salt-wasting form in most patients and I172N produces the 316 classic simple virilizing form in patients[34].

317 To date, PCR amplification provides the majority of samples for 318 throughput mutational analyses. Methods for detecting a single 319 nucleotide substitution for positional determination include ASO, 320 PCR/ligase, ACRS, and MLPA while the SSCP and DHPLC analyses are 321 used for non-positional detection; all of these except in the MLPA 322 method require an agarose or PAGE preparation, and the result relies 323 on a gel-staining or labeling process. Although direct DNA sequencing 324 is considered the gold standard method for mutation analysis, it 325 entails significant costs and labor and does not show the absolute 326 sensitivity or specificity for detecting tuberous sclerosis (TSC) patients 327 with somatic mosaicism in low-level mutant alleles[35,36]. The HRM 328 analysis is a non-positional technique and a non-gel-based system in a

329 closed-tube to detect mutations including polymorphisms and

330 epigenetic differences in dsDNA samples existing in heteroduplexes

331 and homoduplexes. Additional applications such as quantitative

332 analysis of copy number variants, purity of PCR products, and clone

333 identity determinations make HRM a versatile multipurpose

analyt-334 ical tool[37]. Compared to DNA sequencing, the HRM analysis offers

335 cost-effectiveness for larger-scale gene screening such as DMD with

336 79 exons which cost€140 per patient, compared to a total of ~€800

337 using a direct sequencing analysis[37].

338 The HRM analysis was successfully applied to analyze more than

339 50 genes documented in the literature[38]. However, it has never

340 been applied to detect mutations of the CYP21A2 gene. The

341 dependence of the scanning accuracy on the PCR product length

342 was studied, and more errors were reported to occur as the length

343 increases above 400 bp[39]. For high sensitivity, fragments of 150–

344 250 bp are generally used. However, there was a successful case of

345 scanning BRCA1 mutations up to a 600-bp amplicon[40]. Because

346 large fragments may have more than 1 melting domain, this increases

347 the chance that not all variants are detected. For this, the HRM

348 analysis for CYP21A2 mutations used a 217-bp PCR fragments on

349 average (Table 1). In addition, SNP existing in the target gene might

350 interfere with genotyping as described elsewhere [41]. We have

351 pointed out that the most polymorphic region between the CYP21A2

352 and CYP21A1P genes is located in intron 2 (IVS2) which shows an

353 11.2% (31/278) rate of sequence polymorphism [18]. From DNA

354 sequencing (Supplemental Figs. 1, 2) and the TaqI analysis of the

3.5-355 kb PCR product (data not shown), sample #81 with the I2 splice and

356 sample #109 with 707–714del mutations did not have a TaqI site

357 [TCGA] at nt −198 [6]. This indicates that these 2 mutations

358 independently resulted from an intergenic conversion. As described

359 in another study[9], mutation of the I2 splice (IVS2−12A/CNG) in

360 combination with 707-714del (without the P30L mutation) was

361 caused by multiple gene deletions (~30-kb deletion). Therefore, these

362 polymorphic sites of nts 620, 624, 629–630, S108 (TCCNTCG), and

363 S113 (TCCNTCT) (Table 2) in IVS2 were not presented in the 226-bp

364 amplicon (Table 1) ampliﬁed with the paired primers, C100/In3R, and

365 did not inﬂuence the HRM analysis (Table 2). In addition, the HRM

366 proﬁle (Fig. 2D) of the cluster E6 mutation in 2 (D1, I236N and V237E,

367 sample #249) and 3 (D2, I236N, V237E, and M239K, sample #393)

368 mutated sites showed two different melting plots. This indicated that

369 the sequence with the heterozygous variant might show either an

370 upcurved (sample #393) or a downcurved (sample #249) plots in this

371 case. We are not sure that whether the polymorphic site of D234

372 (GATNGAC), which is always bounded (SupplementalFig. 1, D1, D2),

373 can be attributed to the production of 2 different melting types. The

374 polymorphic sites of nts 1420 (ANG) and 1421(CNT) not being

375 included (data of DNA sequencing not shown) indicates that the

376 occurrence of the intergenic conversion did not extend to these 2

377 polymorphic sites in these 2 mutation types.

378 In addition, the inﬂuence of different template concentrations in

379 the HRM analysis should be considered in our study. In order to

380 separate the CYP21A2 gene from the CYP21A1P pseudogene, a primary

381 3.5-kb primary PCR product of the CYP21A2 gene should be ampliﬁed

382 ﬁrst, and then the primary PCR product can be used as a template for

383 the secondary PCR ampliﬁcation by the HRM analysis. A nested PCR

384 was carried out to identify mutations of the CYP21A2 gene, and the

385 concentration of the primary PCR product was difﬁcult to calculate. It

386 was reported that a deviating curve can occasionally occur due to

387 input of a higher amount of DNA (2.5×) that might give rise to a

false-388 positive result[42].

389 In conclusion, a rapid, sensitive, and reliable strategy for mutation

390 scanning of the CYP21A2 gene using an HRM analysis was

documen-391 ted. As indicated above, we established a standard proﬁle for the most

392 common 11 mutation sites of the CYP21A2 gene. This protocol can be

393 used as a tool for screening most patients with CAH caused by defects

394 of the CYP21A2 gene converted from the CYP21A1P pseudogene.

(6)

UNCORRECTED PR

OOF

395 Supplementary materials related to this article can be found online

396 atdoi:10.1016/j.cca.2011.06.033. 397 Acknowledgements

398 This study was supported by a grant from Kaohsiung Medical 399 University Hospital (KMUH98-8G68).

400 References

401 [1] White PC, Speiser PW. Congenital adrenal hyperplasia due to 21-hydroxylase

402 deﬁciency. Endocr Rev 2000;21:245–91.

403 [2] New MI, Wilson RC. Steroid disorders in children: congenital adrenal hyperplasia

404 and apparent mineral corticoid excess. Pro Natl Acad Sci U S A 1999;96:12790–7.

405 [3] New MI. Extensive clinical experience nonclassical 21-hydroxylase deﬁciency. J

406 Clin Endocrinol Metab 2006;91:4205–14.

407 [4] Therrel BL. Newborn screening for congenital adrenal hyperplasia. Endocrinol

408 Metab Clin North Am 2001;30:15–30.

409 [5] Fitness J, Dixit N, Webster D, et al. Genotyping of CYP21, linked chromospme 6p

410 marker, and a sex-speciﬁc gene in neonatal screening for congenital adrenal

411 hyperplasia. J Clin Endocrinol Metab 1999;84:960–6.

412 [6] Higashi Y, Yoshioka H, Yamane M, Gotoh O, Fujii-Kuriyama Y. Complete nucleotide

413 sequence of two steroid 21-hydroxylase genes tandemly arranged in human

414 chromosome: a pseudogene and genuine gene. Proc Natl Acad Sci U S A 1986;83:

415 2841–5.

416 [7] White PC, New MI, Doupont B. Structure of human steroid 21-hydroxylase genes.

417 Proc Natl Acad Sci U S A 1986;83:5111–55.

418 [8] YusiÉ-Luna MT, White PC. Gene conversion and unequal crossovers between

419 CYP21 (steroid 21-hydroxylase gene) and CYP21P involve different mechanisms.

420 Proc Natl Acad Sci U S A 1995;92:10796–800.

421 [9] Chang SF, Lee HH. Analysis of the CYP21A2 gene with intergenic recombination

422 and multiple gene deletions in the RCCX module. Genet Test Mol Biomarkers

423 2011;15:35–42.

424 [10] Koppens PFJ, Hoogenboezem T, Degenhart HJ. Carriership of a defective tenascin-X

425 gene in steroid 21-hydroxylase deﬁciency patients: TNXB–TNXA hybrids in

426 apparent large-scale gene conversions. Hum Mol Genet 2002;11:2581–90.

427 [11] Lee HH. Chimeric CYP21P/CYP21 and TNXA/TNXB genes in the RCCX module. Mol

428 Genet Metab 2005;84:4–8.

429 [12] Vrzalová Z, Hrubá Z, Hrabincová ES, et al. Chimeric CYP21A1P/CYP21A2 genes

430 identiﬁed in Czech patients with congenital adrenal hyperplasia. Eur J Med Genet

431 2011;54:112–7.

432 [13] Tsai LP, Cheng CF, Chuang SH, Lee HH. Analysis of the CYP21A1P pseudogene:

433 indication of mutational diversity and CYP21A2-like and duplicated CYP21A2

434 genes. Anal Biochem 2011;413:133–41.

435 [14] Lee HH, Lee YJ, Wang YM, et al. Low frequency of the CYP21A2 deletion in ethnic

436 Chinese (Taiwanese) patients with 21-hydroxylase deﬁciency. Mol Genet Metab

437 2008;93:450–7.

438 [15] Dain LB, Buzzalino ND, Oneto A, et al. Classical and nonclassical 21-hydroxylase

439 deﬁciency: a molecular study of Argentine patients. Clin Endocrinol 2002;56:

440 239–45.

441

[16] Stikkelbroeck NMML, Hoefsloot LH, de Wijs IJ, Otten BJ, Hermus ARMM,

442

Sistermans EA. CYP21 Gene mutation analysis in 198 patients with

21-443

hydroxylase deﬁciency in the Netherlands: six novel mutations and a speciﬁc

444

cluster of four mutations. J Clin Endocrinol Metab 2003;88:3852–9.

445

[17] Friaes A, Rego AT, Aragues JM, et al. CYP21A2 mutations in Portuguese patients

446

with congenital adrenal hyperplasia: identiﬁcation of two novel mutations and

447

characterization of four different partial gene conversions. Mol Genet Metab

448

2006;88:58–65.

449

[18] Tsai LP, Cheng CF, Hsieh JP, Teng MS, Lee HH. Application of the DHPLC method for

450

mutational detection of the CYP21A2 gene in congenital adrenal hyperplasia. Clin

451

Chim Acta 2009;410(1–2):48–53.

452

[19] Koyama S, Toyoura T, Saisho S, Shmozawa K, Yata J. Genetic analysis of Japanese

453

patients with steroid 21-hydroxylase deﬁciency: identiﬁcation of a patient with a

454

new mutation of a homozygous deletion of adenine at codon 246 and patients

455

without demonstrable mutations within the structural gene for CYP21. J Clin

456

Endocrinol Metab 2002;87:2668–73.

457

[20] Kharrat M, Tardy V, M'rad R, et al. Molecular genetic analysis of Tunisian patients

458

with a classic form of 21-hydroxylase deﬁciency: identiﬁcation of four novel

459

mutations and high prevalence of Q318X mutation. J Clin Endocrinol Metab

460

2004;89:368–74.

461

[21] Lee HH. CYP21 mutations and congenital adrenal hyperplasia. Clin Genet 2001;59:

462

293–301.

463

[22] Keen-Kim D, Redman JB, Alanes RU, Eachus MM, Wilson RC, New MI, et al.

464

Validation and clinical application of a locus-speciﬁc polymerase chain

reaction-465

and minisequencing based assay for congenital adrenal hyperplasia

(21-466

hydroxylase deﬁciency). J Mol Diagn 2005;7:236–46.

467

[23] Owerbach D, Crawford YM, Draznin MB. Direct analysis of CYP21B genes in

21-468

hydroxylase deﬁciency using polymerase chain reaction ampliﬁcation. Mol

469

Endocrinol 1990;4:125–31.

470

[24] Day DJ, Speiser PW, White PC, Barany F. Detection of steroid 21-hydroxylase

471

alleles using gene-speciﬁc PCR and a multiplexed ligation detection reaction.

472

Genomics 1995;29:152–62.

473

[25] Loidi L, Quinteiro C, Parajes S, et al. High variability in CYP21A2 mutated alleles in

474

Spanish 21-hydroxylase deﬁciency patients, six novel mutations and a founder

475

effect. Clin Endocrinol 2006;64:330–6.

476

[26] Tajima T, Fujieda K, Nakayama K, Fujii-Kuriyama Y. Molecular analysis of patients

477

and carrier genes with congenital steroid 21-hydroxylase deﬁciency by using

478

polymerase chain reaction and single strand conformational polymorphism. J Clin

479

Invest 1993;92:2182–90.

480

[27] Lee HH, Chao HT, Ng HT, Choo KB. Direct molecular diagnosis of CYP21 mutations

481

in congenital adrenal hyperplasia. J Med Genet 1996;33:371–5.

482

[28] Olney RC, Mougey EB, Wang J, Shulman DI, Sylvester JE. Using real-time,

483

quantitative PCR for rapid genotyping of the steroid 21-hydroxylase gene in a

484

north Florida population. J Clin Endocrinol Metab 2002;87:735–41.

485

[29] Krone N, Braun A, Weinert S, Peter M, Roscher AA, Partsch CJ, et al. Multiplex

486

minisequencing of the 21-hydroxylase gene as a rapid strategy to conﬁrm

487

congenital adrenal hyperplasia. Clin Chem 2002;48:818–25.

488

[30] Zeng X, Witchel SF, Dobrowolski SF, Moulder PV, Jarvik JW, Telmer CA. Detection

489

and assignment of CYP21 mutations using peptide mass signature genotyping.

490

Mol Genet Metab 2004;82:38–47.

491

[31] Concolino P, Mello E, Toscano V, Ameglio F, Zuppi C, Capoluongo E. Multiple

492

ligation-dependent ampliﬁcation (MLPA) assay for the detection of CYP21A2 gene

493

deletions/duplications in congenital adrenal hyperplasia:ﬁrst technical report.

494

Clinca Chimica Acta 2009;402:164–70.

495

[32] Lee HH, Chao MC, Lee YJ. Only two amino acid substitutions of I236N and V237E in

496

exon 6 are converted to the CYP21 gene in a Chinese patient with congenital

497

adrenal hyperplasia. Clin Endocrinol 2006;64:227–9.

498

[33] Robins T, Barbaro M, Lajic S, Wedell A. Not all amino acid substitutions of the

499

common cluster E6 mutation in CYP21 cause congenital adrenal hyperplasia. J Clin

500

Endocrinol Metab 2005;90:2148–53.

501

[34] New MI. An update of congenital adrenal hyperplasia. Ann N Y Acad Sci

502

2004;1038:14–43.

503

[35] Jones AC, Sampson JR, Cheadle JP. Low level mosaicism detectable by DHPLC but

504

not by direct sequencing. Hum Mutat 2001;17:233–4.

505

[36] Dobrowolski S, Gray J, Miller T, Sears M. Identifying sequence variants in the

506

human mitochondrial genome using high-resolution melt (HRM) proﬁling. Hum

507

Mutat 2009;30:891–8.

508

[37] Vossen RHAM, Aten E, Roos A, den Dunnen JT. High-resolution melting analysis

509

(HRMA)—more than just sequence variant screening. Hum Mutat 2009;30:860–6.

510

[38] Leiden Genome Technology Center (LGTC) Published assays using high-resolution

511

melting analysis (HRMA)..http://www.LGTC.nl/HRMA.

512

[39] Reed GH, Wittwer CT. Sensitivity and speciﬁcity of single-nucleotide

polymor-513

phism scanning by high-resolution melting analysis. Clin Chem 2004;50:1748–54.

514

[40] Takano EA, Mitchell G, Fox1 SB, Dobrovic A. Rapid detection of carriers with BRCA1

515

and BRCA2 mutations using high resolution melting analysis. BMC Cancer 2008;8:

516

59.

517

[41] Shih HC, Er TK, Chang TJ, Chang YS, Liu TC, Chang JG. Rapid identiﬁcation of HBB

518

gene mutations by high-resolution melting analysis. Clin Biochem 2009;42:

519

1667–76.

520

[42] van der Stoep N, van Paridon CD, Janssens T, Krenkova P, Stambergova A, Macek M,

521

et al. Diagnostic guidelines for high-resolution melting curve (HRM) analysis: an

522

interlaboratory validation of BRCA1 mutation scanning using the 96-well

523

LightScanner. Hum Mutat 2009;30:899–909.

524

Table 2 t2:1

Polymorphic sites inﬂuencing the heteroduplex form in a speciﬁc fragment of the CYP21A2 gene using the HRM analysis.

t2:2 t2:3 Mutational locus Fragment (paired primer ampliﬁcation) Polymorphic site (exon/nucleotide)a Interchange t2:4 CYP21A2 CYP21A1P t2:5 P30L 1A2/1AR L39 (TTG) (CTG) ? t2:6 P45 (CCA) (CCC) ? t2:7 I2 spliceb C100/In3R nt 620 (A) (G) No t2:8 nt 624 (G) (T) No t2:9 nt 629/630 (C/A) (G/G) No

t2:10 707–714 del C100/In3R S108 (TCC) (TCG) Yes

t2:11 S113 (TCC) (TCT) Yes

t2:12 I172N In3-1/E4r – – –

t2:13 Cluster E6c

Ex6/C8 D234 (GAT) (GAC) Yes

t2:14 nt 1420/21 (A/C) (G/T) No t2:15 V281L, C9/S7r – – – t2:16 F306ΛL307insT C9/S7r – – – t2:17 Q318X In7-1/C12-1 – – t2:18 R356W In7-1/C12-1 – – – a

Based on Higashi et al.[6]. t2:19

b

I2 splice, IVS2−12A/CNG, or nt 655. t2:20

c

Cluster E6 represents I236N, V237E, and M239K. t2:21