Identification of Large Rearrangements of the PCDH15 Gene by Combined MLPA and a CGH: Large Duplications Are Responsible for Usher Syndrome

Abstract

Purpose.: PCDH15, encoding protocadherin 15, is mutated in Usher syndrome type 1F (USH1F) patients. Not only point mutations, but also large deletions have been detected within this gene. However, the detection and characterization of gross deletions in the USH1F locus have been difficult. The purpose of the present work was to identify large genomic rearrangements of PCDH15 in a cohort of patients and to accurately identify the location of the junction breakpoints of the detected rearrangements.

Methods.: A PCDH15 MLPA (multiplex ligation-dependent probe amplification) commercial kit was used, combined with a customized oligonucleotide array–based CGH analysis (aCGH), containing almost 20,000 probes tiling the nonrepetitive sequence of the PCDH15 gene.

Results.: Two large intragenic rearrangements were identified—one deletion of 55 kb and one direct duplication of 82 kb—in 3 (13%) families from a cohort of 23 USH cases. The patients had been screened for mutations in the five known USH1 genes and were found to carry one or none of the pathogenic mutations in PCDH15. The exact breakpoints of both rearrangements were identified.

Conclusions.: This is the first time that large duplications have been associated with Usher syndrome. USH patients have not been extensively tested for large genomic rearrangements such as duplications and deletions. This type of mutation easily escapes detection by traditional PCR-based methods. Thus, a combination of PCR-based mutation screening, together with deletion and duplication analysis, is mandatory for the accurate screening of the PCDH15 gene in Usher patients.

Usher syndrome (USH) is a genetically and clinically heterogeneous autosomal recessive disorder with an estimated prevalence of 3 to 6.2 per 100,000. ¹ It accounts for more than 50% of cases of hereditary forms of deafness with blindness. ²  

Three distinct types of USH, I, II, and III (USH1, -2, and -3), are recognized, all of them characterized by hearing loss, retinitis pigmentosa (RP) and, in some cases, vestibular dysfunction. Although this classification is generally deemed adequate, atypical clinical types have also been described. 

Usher syndrome type I (USH1) is the most severe form of Usher syndrome ³ and is characterized by congenital profound deafness, vestibular areflexia, and prepuberal onset of RP. To date, seven loci (USH1B–USH1H) have been mapped, and five genes have been identified: MYO7A (USH1B), USH1C (USH1C), CDH23 (USH1D), PCDH15 (USH1F), and USH1G (USH1G). ^4,5  

PCDH15 is involved in 11% to 19% of USH1 cases. ^{6 –8} The PCDH15 gene, on chromosome 10q21-22, was initially described to span more than 980 kb of genomic DNA and code for several transcripts. The longest transcript (isoform A) consisted of 33 exons and was predicted to encode a 1955-amino-acid transmembrane protein, containing 11 cadherin repeats, one transmembrane domain, and a cytoplasmic domain (CD1) with two proline-rich regions. ^6,9 However, Ahmed et al. ¹⁰ characterized multiple alternative Pcdh15 transcripts in the mouse inner ear containing additional exons that were also detected in human retina. Building on these observations, Ahmed et al. ¹¹ described an updated gene structure with six additional exons: exon 11a (encoding new cadherin repeats), exons 25a and 25b (which code for a new 5′ UTR and a signal peptide), exon 34, and exons 35 and 36 (which code for the cytoplasmic domains CD2 and CD3). 

The first functional studies showed that, together with other USH1 proteins, protocadherin-15 ensures hair bundle morphogenesis in the inner ear. ¹² More recently, Kazmierczak et al. ¹³ demonstrated that protocadherin-15 and cadherin-23 interact to form the tip links of stereocilia bundle in the inner hair cells. The exact function of protocadherin-15 in the retina has not been determined. Seiler et al. ¹⁴ showed that Pcdh15b, one of the two protocadherin orthologues in zebrafish, is required for alignment and interdigitation of photoreceptor outer segments with the pigment epithelium. 

Mutations of PCDH15 can cause either Usher syndrome type I (USH1F) or nonsyndromic hearing loss (DFNB23). ⁶ Approximately 30 different point mutations in the PCDH15 gene have been identified by numerous mutation screenings (UMD-PCDH15 Locus Specific Databases ¹⁵ ). However, recently, Le Guédard et al. ¹⁶ characterized three different large deletions within the PCDH15 gene, showing that gross deletions, which can escape routine screening methods, form a significant proportion of PCDH15 mutations (30% in this aforementioned author's cohort). These results indicate that comprehensive screening of the PCDH15 gene should include a search for gross genomic rearrangements. 

In recent years, the newly introduced MLPA (multiplex ligation-dependent probe amplification) and oligonucleotide array-based CGH (comparative genomic hybridization; aCGH) techniques have facilitated the detection of total and partial gene deletions and duplications that may escape conventional PCR screening methods. 

In the present study, we used the MLPA technique to screen our cohort of USH patients for large deletions and duplications affecting the PCDH15 gene. Subsequently, we applied a customized oligonucleotide aCGH assay, specific for the USH1F locus, to confirm the presence of rearrangements and accurately determine the location of their breakpoints. 

Methods

Patients

Twenty-three unrelated patients were included in the study. Twenty-two were of Spanish origin and were recruited from the Federación de Asociaciones de Afectados de Retinosis Pigmentaria del Estado Español (FAARPEE) and from the Ophthalmology and ENT Services of several Spanish hospitals. One patient was referred from Italy. Seventeen patients were classified as USHI on the basis of ophthalmic studies, including visual acuity, visual field, and fundus ophthalmoscopy, electroretinography, pure-tone and speech audiometry, and vestibular evaluation. Detailed clinical data could not be obtained for six patients and they remained nonclassified. 

The patients were selected for this study because previous mutation screenings had discarded mutations in other USH1 genes as the underlying cause of the disease. ^{17 –20}  

For each patient, samples from parents as well as from siblings were obtained, when possible. The research adhered to the tenets of the Declaration of Helsinki. The study was approved by the Hospital La Fe Ethics Committee, and consent for genetic testing was obtained from adult probands or the parents of the minors. 

Control DNA Groups

Two control DNA pools (one male and one female) were used as a reference for both the MLPA and the oligonucleotide aCGH. Each pool comprised 10 different DNAs randomly chosen from a healthy control group. This control group did not have symptoms or a history of Usher syndrome or related disorders. 

MLPA Analysis

The MLPA analysis used 50 ng of DNA, diluted in 5 μL of TE buffer (P292-A1 SALSA MLPA kit; MRC Holland, Amsterdam, The Netherlands). The P292-A1 probemix contained probes for exons 1-32 of PCDH15 (two probes each for exons 1 and 2) and this probemix also included eight reference probes detecting sequences on autosomal chromosomes. 

The MLPA reaction was performed according to the manufacturer's recommendations (http://www.mlpa.com). One microliter of each reaction product was separated on POP-7 polymer by capillary electrophoresis with a DNA Analyzer (model 3730; Applied Biosystems, Inc. [ABI], Foster City, CA). Freely available software provided by MRC Holland was used to analyze the MLPA data (Coffalyser; MRC Holland; http://www.mlpa.com). 

Oligonucleotide aCGH

Genomic DNA was quantified and its purity confirmed by spectrometry (ND1000; NanoDrop Technologies, Wilmington, DE). Nonamplification labeling of DNA was performed by the direct method (Oligonucleotide Array-Based CGH for Genomic DNA Analysis, ver. 4.0 cat. no. G4410-90010; Agilent Technologies). In one restriction digestion step, 1400 ng of experimental and reference genomic DNA samples were fragmented. Digestion was confirmed and evaluated by DNA assay (7500 Bioanalyzer; Agilent Technologies). Cyanine 3-dUTP and cyanine 5-dUTP were used for fluorescent labeling of, respectively, test- and reference-digested gDNAs with a genomic DNA-labeling kit (PLUS; cat. no. 5188-5309; Agilent) according to the manufacturer's instructions. Labeled DNA was hybridized with a CGH microarray (Human Custom CGH Microarray 44K, cat. no. G4426A-023890; Agilent). This array contains 20,357 unique probes from the HD-CGH oligo Agilent database covering the whole human genome (NCBI build 36.3; National Center for Biotechnology Information, Bethesda, MD), 2,768 normalization and control probes, and 19,978 probes tiling the nonrepetitive sequence of the PCDH15 gene. Arrays were scanned in a microarray scanner (model G2565BA; Agilent), according to the manufacturer's protocol, and the data were extracted (Feature Extraction Software 9.5.3.1; Agilent, according to the Agilent protocol CGH-v4_95_Feb07 and the QC Metric Set CGH_QCM_Feb07). 

Identification of Breakpoints

PCR Amplification of Deletion Breakpoints.

One pair of primers was designed to specifically amplify patients' genomic DNA in the region where deletion breakpoints were indicated by the aCGH analysis: direct primer, 5′-gtatagttccattatcactatg-3′ (located at intron 2); reverse primer: 5′-gtatttggctgggttaacca-3′ (located at intron 3). 

Two PCR rounds (Expand Long-Template PCR System; Roche Applied Science, Mannheim, Germany) were applied to amplify the fragment, expected to have a length of ±5 kb. 

PCR Amplification of Direct Duplication Breakpoints.

A primer pair was designed to amplify the expected junction of the duplication, based on aCGH results. Direct primer: 5′-tattagcagggtgaaggtcag-3′ (located at intron 6); reverse primer: 5′-gtatcctctcagttaagctcc-3′ (located at intron 3). A standard PCR protocol was used to amplify the genomic DNA fragment of patients, expected to have a length of 250 to 650 bp. 

DNA Sequencing.

An additional internal primer was designed to sequence the amplified DNA fragment containing the exact location of the deletion breakpoint: 5′-cacacggatgcagttatagg-3′ (located at intron 2). The reverse PCR primer was used to sequence the amplified DNA fragment containing the duplication junction,. The genomic PCDH15 reference sequence used for nomenclature was NG_009191.1. The PCDH15 gene starts at nucleotide position 5001 in this sequence. 

Haplotype Analysis

Haplotypes were constructed from three microsatellite markers in the USH1F locus: D10S1124, D10S2536, and D10S546. Fluorescent PCR products of microsatellites were examined on a DNA analyzer (3730 DNA Analyzer; ABI). 

Results

The present study led to the identification of intragenic PCDH15 rearrangements in 3 (13%) of 23 cases. One intragenic deletion of 55 kb, affecting part of intron 2, exon 3, and part of intron 3, was heterozygously found in one USH1 patient. Furthermore, a duplication of 82 kb affecting exons 4, 5, and 6 and their respective flanking introns was also detected. This duplication was found, a priori, in two unrelated USH1 patients. One patient was found to be heterozygous for this duplication, whereas the other was homozygous. 

MLPA and Segregation Analysis

MLPA studies indicated the presence of a heterozygous deletion affecting exon 3 in patient RP-1034; this patient had already been found to carry the mutation p.R245X in PCDH15 (Jaijo T, unpublished results, 2009). Segregation analysis showed that this deletion was inherited from the father and the nonsense mutation from the mother. A healthy sister carried only the deletion (Fig. 1). 

Figure 1.

 

Segregation analysis performed in families of patients RP-1034, RP-367, and RP-982.

MLPA results also indicated the presence of a copy gain affecting exons 4, 5, and 6, in two patients. One patient (RP-367) carried the duplication in the heterozygous state. The amino acid change p.R134Q, affecting protocadherin-15, had previously been detected in this case (Jaijo T, unpublished results, 2009). Segregation analysis indicated that the duplication was inherited from the father and the point mutation from the mother. The two healthy sisters carried only the duplication (Fig. 1). Patient RP-982 carried the duplication in a homozygous state. Segregation analysis showed that, in both parents and in a healthy brother, the duplication was heterozygous (Fig. 1). 

Oligonucleotide aCGH

The P292-A1 probemix used for the MLPA analysis detects only loss or gain of PCDH15 exonic sequences. Thus, an aCGH analysis was performed to characterize the deletion and duplications found. This array contained 19,978 probes tiling the nonrepetitive genomic sequence of the PCDH15 gene, indicating size, position, and breakpoints of deletions and duplications. 

The PCDH15-specific aCGH study was performed with genomic DNA from patient RP-1034, heterozygous for the deletion of exon 3; patient RP-367, heterozygous for the duplication of exons 4, 5 and 6; and the parents of patient RP-982, both heterozygous for the duplication of exons 4, 5, and 6. This study revealed that patient RP-1034 had a deletion spanning 55 to 60 kb from introns 2 to 3. The breakpoint position in intron 2 was found to be located in a repetitive region covering ±5 kb, from g.259881 to g.264801. In intron 3, the breakpoint was predicted to be in the region of nucleotide positions g.319760 and g.319848 (Fig. 2A). 

Figure 2.

 

Representation of the deletion and duplication detected in this study. (A) One primer pair was designed to amplify the DNA fragment containing the deletion breakpoints based on aCGH predictions. A 5-kb PCR product could be obtained only from chromosomes carrying the deletion. (B) To amplify the duplication junction a primer pair was designed based on aCGH results. Direct primer was designed to hybridize 100 bp upstream of nt g.456533 (intron 6), and reverse primer was designed to hybridize 100 bp downstream of nt g.374668 (intron 3). These primers hybridize twice in those chromosomes carrying the duplication, but only the DNA fragment containing the duplication junction could be amplified, obtaining a PCR product of 800 bp.

Similar results were obtained for patient RP-367 and the parents of patient RP-982. aCGH revealed that all three carried a duplication of approximately 82 kb, spanning a region from introns 3 to 6. The duplication was shown to start somewhere between nucleotide positions g.374668 and g.375063 and to finish around nucleotides g.456460 and g.456533 (Fig. 2B). 

Determination of the Breakpoints

To locate the exact position of the deletion and duplication breakpoints, different primer pairs were designed based on the aCGH results. 

Deletion Breakpoints.

We designed a pair of primers to specifically amplify the genomic DNA fragment of patient RP-1034, where deletion breakpoints were expected to be located. Direct primer was located at intron 2, approximately ±70 bp upstream of g.259881. Reverse primer was located ±70 bp downstream of g.319848, at intron 3 (Fig. 2A). PCR showed a DNA fragment of ±5 kb. A sequencing reaction with the internal primer 5′-cacacggatgcagttatagg-3′ revealed the exact breakpoint positions: g.264636 (intron 2) and g.319905 (intron 3; Fig. 3). These results indicate that the deletion detected in this patient is exactly 55,269 nt in length. 

Figure 3.

 

Sequences surrounding the deletion and duplication breakpoints. The shaded nucleotides are those located directly adjacent to the breakpoints. Those sequences overrepresented within ±15 bp of GRaBD translocation breakpoints are highlighted inside a box. ²¹

The loss of exon 3 would lead to an in-frame deletion of 23 amino acids at protein level (D31-G53del), but the domain structure of protocadherin-15 would not be disrupted. 

Duplication Breakpoints.

We assumed that patients RP-367 and RP-982 carried a direct in situ duplication. To verify this hypothesis, a specific primer pair was designed to amplify the expected junction of the duplication. The direct primer was located at intron 6, ±100 bp upstream of g.456533. The reverse primer was located ±100 bp downstream of g.374668, at intron 3 (Fig. 2B). A standard PCR reaction amplified a DNA fragment of approximately 800 bp in length. A sequencing reaction with the reverse primer revealed the exact location of the duplication junction: g.374569 (intron 3) and g.456513 (intron 6) (Fig. 3). Thus, we confirmed that it was a direct duplication of a sequence with an exact length of 81,944 bp. 

The duplication of exons 4 to 6 would lead to a frameshift in residue 99, generating a truncated protein of 106 amino acids. This protein would conserve only the first cadherin repeat domain. 

Haplotype Analysis

Three microsatellite polymorphic markers were used to perform this study in the three patients carrying the deletion and duplication. As a result, we found that all chromosomes bearing the 82-kb duplication shared the same haplotype. This haplotype differed from that linked to the deletion and the other two point mutations detected in PCDH15 in these families (Fig. 1). 

Discussion

In this study, we identified and characterized a large deletion and a large direct duplication within the PCDH15 gene, underlying Usher syndrome in three Spanish USH1 patients. The deletion spans a region of 55 kb and affects the coding exon 3. The duplication spans 82 kb; affects coding exons 4, 5, and 6; and generates a tandem repeat of these exons. These exons are present in most PCDH15 isoforms described. ¹¹ However, it is difficult to predict the repercussion of these rearrangements on RNA and protein level, as it would be tissue dependent. 

The deletion was found in patient RP-1034, and segregation analysis showed that it was in trans together with the previously detected p.R245X PCDH15 point mutation. Le Guédard et al. ¹⁶ found three different deletions within the PCDH15 gene, but none of them coincides with the deletion detected in our patient. Thus, it seems that these deletions are nonrecurrent events. 

The duplication was found in two unrelated patients. Patient RP-367 carried the duplication in trans with the amino acid change p.R134Q, and patient RP-982 had the duplication in homozygous state. That all chromosomes carrying the duplication shared the same haplotype (Fig. 1) indicates that it is not a recurrent event, but has a common origin. Furthermore, both patients carrying the duplication came from the same geographic region in central Spain (Guadalajara), which reinforced the hypothesis of a common origin. 

Large deletions and duplications are well known disease-causing mutations in human disorders. Among the USH genes, PCDH15 appears to be the one in which large deletions most frequently cause Usher syndrome ¹⁶ ; however, in this study we show for the first time that large duplications can also be responsible for Usher syndrome. 

To investigate the mechanisms involved in the deletion and duplication detected in our patients, the intervals surrounding the breakpoints were analyzed through RepeatMasker (http://www.repeatmasker.org/). Several recombinogenic elements, short interspersed elements (SINEs), long interspersed elements (LINEs), and long terminal repeats (LTRs) were found in all cases (Table 1). 

Table 1.

 

Localization of the Deletion and Duplication Breakpoints and Their Analysis with the RepeatMasker Program

5′ Breakpoint	Repeat Position	Repeat Type	3′ Breakpoint	Repeat Position	Repeat Type
Deletion
	g.263618_g.263936	L1P4_orf2 (LINE/L1)
	g.263937_g,263993	L1PA16_3end (LINE/L1)		g.318571_g.318740	HAL1 (LINE/L1)
	g.263994_g.264303	AluY(SINE/Alu)		g.319546_g.319657	LIME4a_3end (LINE/L1)
g.264636	g.264304_g.264792	L1PA16_3end (LINE/L1)	g.319905
	g.265037_g.265178	MIR (SINE/MIR)		g.320237_g.320283	ME3 (SINE/MIR)
	g.265555_g.265863	AluY(SINE/Alu)
Duplication
				g.455458_g.455558	Aluy (SINE/Alu)
	g.373715_g.374022	AluSx(SINE/ALU)		g.455560_g.455655	Aluy (SINE/Alu)
	g.374312_g.374373	MIRb(SINE/MIR)		g.456064_g.456158	MIRb (SINE/MIR)
g.374569			g.456513
	g.374723_g.375059	MER1B (DNA/hAT-Charlie)		g.457484_g.457645	L1PA5_3end (LINE/L1)
	g.375739_g.376158	MSTC (LTR/ERVL-MaLR)

In addition, significantly overrepresented motifs within ±15 bp of GRaBD translocations breakpoints ²¹ were also sought. Regarding deletion, only the GAG sequence (DNA polymerase α pause site core) was found (Fig. 3). When we studied the duplication, however, we found the motif GCC directly adjacent to the breakpoint (Fig. 3). This triplet is also a DNA polymerase α pause site core, significantly overrepresented and immediately adjacent to the GRaBD translocation breakpoints at position +1. Furthermore, GCC is also known to be associated with replication slippage in vitro and in vivo. ^22,23 Thus, replication slippage could be a plausible explanation for the mechanism involved in duplication. However, the relative sequence homology existing between deletion breakpoints (Fig. 4) could indicate unequal homologous recombination is the cause of deletion. 

Figure 4.

 

Representation of the homology existing between sequences surrounding the deletion breakpoints.

In the present study, the MLPA analysis identified intragenic PCDH15 rearrangements in 3 (13%) of 23 cases. However, it is still possible that some of the 20 remaining cases also carried deletions and duplications affecting the PCDH15 gene, as the commercial P292-A1 probemix contains only probes for exons 1 to 32 and not for exon 33 or for the newly described exons 11a, 25a, 25b, 34, 35, and 36. ¹¹ Therefore, the P292-A1 probemix must be updated to improve both the efficiency and detection rate of rearrangements in PCDH15 by MLPA. 

Le Guédard et al. ¹⁶ failed to detect the exact breakpoint junctions of the deletions heterozygously present in their patients by using semiquantitative PCR walking methods. In our case, however, the aCGH assay specific for the PCDH15 gene enabled us to accurately narrow down the deletion and duplication breakpoints, initially detected by MLPA. A subsequent PCR, based on array results, enabled us to identify the exact breakpoint junctions in our patients. 

USH genes have not been extensively tested for large genomic rearrangements such as insertions and deletions. This type of mutation easily escapes detection by traditional PCR-based methods. The present report and the previous one by Le Guédard et al. ¹⁶ have shown that large deletions and duplications can cause USH1F. Therefore, it would be interesting to apply a high-density aCGH assay to study all USH genes. 

Determination of the underlying genetic defects in a patient is a prerequisite for gene or mutation-specific therapy. Recent preliminary results in gene therapy for retinal degeneration have been promising. Therefore, the development of efficient and cost-effective strategies to detect mutation is mandatory. 

Footnotes

 Supported by Research Projects PI07/0558 and PI08/90311 (Institute of Health Carlos III, Spanish Ministry of Science) and Project GE-013/10 (Conselleria de Sanitat).

Footnotes

 Disclosure: E. Aller, None; T. Jaijo, None; G. García-García, None; M.J. Aparisi, None; D. Blesa, None; M. Díaz-Llopis, None; C. Ayuso, None; J.M. Millán, None

The authors thank the participating patients and their relatives and Anne-Françoise Roux for providing the positive DNA controls carrying the PCDH15 deletions.