Abstract

One of the causes of chronic pancreatitis is the duplication and triplication of a ∼605 kb segment containing the trypsinogen locus. Employing array-comparative genomic hybridization, we fully characterized the triplication copy number mutation (CNM) and found it to be part of a complex rearrangement that also contains a triplicated ∼137 kb segment and 21 bp sequence tract. This triplication allele therefore constitutes a gain of two tandemly arranged composite duplication blocks, each comprising a copy of the ∼605 kb segment, a copy of the inverted ∼137 kb segment and a copy of the inverted 21 bp sequence tract. As such, it represents the first characterization of a human complex triplication CNM at the DNA sequence level. All triplications and duplications identified were found to arise from a common founder chromosome. A two-step process is proposed for the generation of this highly unusual triplication CNM. Thus, the first composite duplication block is envisaged to have been generated by break-induced serial replication slippage during mitosis. This duplication would have provided the sequence homology required to promote non-allelic homologous recombination (NAHR) during meiosis which would then, in a second step, have generated the complex triplication allele. Our data provide support for the view that many human germline copy number variants arise through replication-based mechanisms during the premeiotic mitotic divisions of germ cells. The low copy repeats thereby generated could then serve to promote NAHR during meiosis, giving rise to amplified DNA sequences which would themselves predispose to further recombinational events during both mitosis and meiosis.

INTRODUCTION

Over the last few years, we have witnessed dramatic advances in our understanding of copy number variation [CNV; a ≥1 kb DNA segment that differs in terms of its copy number with respect to a reference genome sequence ( 1 )] both in the context of inter-individual variation and with respect to human genetic disease ( 2–4 ). CNV is thought to arise by either non-allelic homologous recombination (NAHR) or non-homologous DNA end joining (NHEJ) ( 2 , 5 ). However, the extent of coverage of the human genome by CNVs is still largely unknown, partly due to the lack of standardization of the analytical techniques employed, partly to the lack of detailed characterization (by DNA sequencing) of the boundaries of identified CNVs in the vast majority of reported cases.

Recent analyses of the junction sequences of complex genomic rearrangements causing human inherited disease have led to the identification of a replication-based mechanism, known as ‘Serial Replication Slippage’ (SRS) ( 6–8 ), alternatively termed ‘Fork Stalling and Template Switching’ (FoSTeS) ( 9 , 10 ). The key features of these very similar models are that (i) the newly synthesized primer strand can, when it encounters a stalled replication fork, undergo multiple rounds of replication slippage or template switching and (ii) all steps of strand invasion or misalignment are microhomology-dependent. In short, both models assume SRS and stress the key importance of genomic architectural elements (such as palindromic DNA, stem-loop structures and repeats) in facilitating the initial stalling of the replication fork ( 11 ). SRS has also been proposed to occur illegitimately between non-homologous chromosomes ( 7 ).

More recently, a new model, termed ‘break-induced SRS’ (BISRS; 12 ) has been proposed; this model successfully integrated the key features of SRS ( 6 , 7 ) with those of the earlier ‘microhomology-dependent break-induced replication’ (BIR) model ( 13 ) to account for the generation of a double complex copy number mutation [CNM; a CNV that is found exclusively in association with a disease state but which is not evident in healthy controls ( 14 )] involving the F8 and FUNDC2 genes and causing severe haemophilia A. Although the BISRS model may be applied to complex genomic rearrangements, both prospectively and retrospectively ( 12 ), the aforementioned microhomology-dependent BIR model ( 13 ) may have explanatory value in the context of large simple deletions and duplications that are flanked by short direct repeats ( 6 ). Recently, taking the above observations into consideration, as well as the properties of DNA repair mechanisms in other organisms ( 15–17 ), microhomology-mediated BIR has been postulated to be of general importance in generating human CNVs ( 18 ).

In this study, we report the characterization of the trypsinogen locus triplication previously reported in association with hereditary pancreatitis ( 19 ). This triplication is much more complex than previously thought, comprising as it does two further triplicated segments which are additionally inverted. Its thorough characterization represents the first of any complex human CNM at the DNA sequence level. We propose a two-step mutational process, involving both BISRS and NAHR, which satisfactorily accounts for the generation of this highly unusual triplication. Finally, based upon this model, we propose a new generalized mechanism for the generation of human CNVs in vivo .

RESULTS AND DISCUSSION

The ∼605 kb trypsinogen locus triplication

We have previously reported, by means of quantitative fluorescence multiplex PCR (QFM-PCR) and FISH, the triplication of a ∼605 kb segment containing the trypsinogen locus on human chromosome 7q34 in five families with hereditary pancreatitis ( 19 ). Of the five tandemly arranged, highly homologous (sequence similarity, ∼91%) human trypsinogen genes, only two [i.e. PRSS1 (also known as T4) and PRSS2 (also known as T8)] are clearly functional, encoding cationic trypsinogen and anionic trypsinogen, respectively. The other three trypsinogen ‘genes’ represent pseudogenes, whether expressed at the mRNA level (T6) or clearly non-functional owing to the presence of splicing and nonsense mutations (T5 and T7) (for a recent review, 20 ). Two lines of evidence in particular, led us to conclude that the trypsinogen locus triplication causes hereditary pancreatitis through a gene dosage effect due to an increased number of copies of PRSS1 and PRSS2 ( 19 ). First, gain of function PRSS1 missense mutations had been previously reported to cause chronic pancreatitis ( 21–23 ). Second, loss of function variants in both PRSS1 ( 24 ) and PRSS2 ( 25 ) had been reported to be protective against the disease.

This triplication CNM, together with a duplication CNM involving precisely the same ∼605 kb segment, were further identified in 14 (triplication × 10, duplication × 4) apparently unrelated patients with idiopathic chronic pancreatitis (ICP) ( 26 ). More recently, we have also detected the triplication CNM in two further patients in one newly recruited hereditary pancreatitis family and the duplication CNM in five newly recruited ICP patients (unpublished data).

Array-comparative genomic hybridization reveals the presence of a second triplicated segment

Although, in the first report of the trypsinogen locus triplication, we succeeded in mapping each end of the triplication CNM to within a <3 kb region by QFM-PCR (Table  1 ), we were unable to clone the aberrant junctions of the triplication by long-range PCR ( 19 ). This provided the first hint that the ∼605 kb triplication might be more complex than originally thought. High-resolution array-comparative genome hybridization [array-comparative genomic hybridization (CGH)] was employed here in an attempt to resolve this puzzle. This method is based on the competitive hybridization of differentially labelled test and reference DNAs to probes spotted on a glass slide. Copy number changes are indicated by altered fluorescent ratios between labelled test and reference DNAs.

Table 1.

End regions of the two triplicated segments defined by QFM-PCR

End  Region a Size (bp) 
605 kb segment  
 Centromeric 141580825–141583339 2514 
 Telomeric 142188404–142191344 2940 
137 kb segment  
 Centromeric 142278162–142279548 1386 
 Telomeric 142415515–142417160 1645 
End  Region a Size (bp) 
605 kb segment  
 Centromeric 141580825–141583339 2514 
 Telomeric 142188404–142191344 2940 
137 kb segment  
 Centromeric 142278162–142279548 1386 
 Telomeric 142415515–142417160 1645 

a In accordance with the human chromosome 7 assembly, March 2006 (hg18).

Analysis of one of our previously reported triplication carriers (III.2 in pedigree C; 19 ) by 244 K Agilent oligonucleotide-based array-CGH served to identify an additional large triplication distal to that of the ∼605 kb triplication (Fig.  1 A). Using walking QFM-PCR, each end of the newly identified triplicated segment was then narrowed down to a ∼1.4–1.6 kb region (Table  1 ). In brief, this additional triplication was found to encompass a ∼137 kb segment located ∼90 kb telomeric to the ∼605 kb segment (Table  1 ). The array-CGH analysis did not reveal any other CNMs in the genome of this patient.

Figure 1.

Characterization of the trypsinogen locus triplication. ( A ) The two large triplications as revealed by array-CGH. Red data points indicate copy number gains whilst black data points indicate no copy-number change. The left hand blue block corresponds to the previously described ∼605 kb triplication ( 19 ) whereas the right hand blue block denotes the location of the newly identified ∼137 kb triplication. The two segments are separated by ∼90 kb on chromosome 7q34. ( B ) Relative positions of the four outward-facing primers (P1, P2, P3 and P4; see Table  2 for primer sequences) used for cloning the aberrant junctions by long-range PCR, in the context of the wild-type chromosome at 7q34 with the ∼605, ∼90 and ∼137 kb segments. ( C ) Electrophoretic separation of the P2/P4- and P1/P3-mediated long-range PCR products. M, DNA marker; WT, normal control; DUP, duplication carrier; TRI, triplication carrier. ( D ) Schema of the complex triplication (top panel) and duplication (lower panel) alleles (not drawn to scale). Primers (i.e. P5, P6, P7 and P8; see Table  3 for primer sequences) used for quantifying the number of aberrant junctions (i.e. J1, J2 and J3) are indicated. ( E ) Alignment of the sequenced junctions of the rearrangement and reference (wild-type) genomic sequences. Ref (+), plus strand of the reference sequence. Ref (−), minus strand of the reference sequence. A 6 bp microhomology (CAAGTA) is present at J1. The inverted repeats, putatively involved in the formation of J2 and J3, are indicated by coloured arrows. The extent of the 21 bp inversion is indicated by crossed red lines. The sequence has also been colour coded. ( F ) QFM-PCR analysis of the aberrant junctions in a triplication carrier [i.e. III.2 in pedigree C ( 19 ); red], a duplication carrier [i.e. patient 2193 (26); blue] and a control (green). The relative peak heights of the J1 and J2/J3 amplicons indicate the relative copy numbers of the respective junctions in the duplication (one copy) and triplication (two copies) carriers as well as in the healthy control (zero copies). ( G ) Rapid detection of the duplication and triplication mutations by conventional PCR. Top panel illustrates the relative positions of the two primer pairs (i.e. P9 /P10 and P9/P11; see Table  4 for primer sequences) whereas the lower panel shows the electrophoretic separation of the P9/P10- and P9/P11-mediated PCR products. M, DNA marker; WT, normal control; DUP, duplication carrier; TRI, triplication carrier.

Figure 1.

Characterization of the trypsinogen locus triplication. ( A ) The two large triplications as revealed by array-CGH. Red data points indicate copy number gains whilst black data points indicate no copy-number change. The left hand blue block corresponds to the previously described ∼605 kb triplication ( 19 ) whereas the right hand blue block denotes the location of the newly identified ∼137 kb triplication. The two segments are separated by ∼90 kb on chromosome 7q34. ( B ) Relative positions of the four outward-facing primers (P1, P2, P3 and P4; see Table  2 for primer sequences) used for cloning the aberrant junctions by long-range PCR, in the context of the wild-type chromosome at 7q34 with the ∼605, ∼90 and ∼137 kb segments. ( C ) Electrophoretic separation of the P2/P4- and P1/P3-mediated long-range PCR products. M, DNA marker; WT, normal control; DUP, duplication carrier; TRI, triplication carrier. ( D ) Schema of the complex triplication (top panel) and duplication (lower panel) alleles (not drawn to scale). Primers (i.e. P5, P6, P7 and P8; see Table  3 for primer sequences) used for quantifying the number of aberrant junctions (i.e. J1, J2 and J3) are indicated. ( E ) Alignment of the sequenced junctions of the rearrangement and reference (wild-type) genomic sequences. Ref (+), plus strand of the reference sequence. Ref (−), minus strand of the reference sequence. A 6 bp microhomology (CAAGTA) is present at J1. The inverted repeats, putatively involved in the formation of J2 and J3, are indicated by coloured arrows. The extent of the 21 bp inversion is indicated by crossed red lines. The sequence has also been colour coded. ( F ) QFM-PCR analysis of the aberrant junctions in a triplication carrier [i.e. III.2 in pedigree C ( 19 ); red], a duplication carrier [i.e. patient 2193 (26); blue] and a control (green). The relative peak heights of the J1 and J2/J3 amplicons indicate the relative copy numbers of the respective junctions in the duplication (one copy) and triplication (two copies) carriers as well as in the healthy control (zero copies). ( G ) Rapid detection of the duplication and triplication mutations by conventional PCR. Top panel illustrates the relative positions of the two primer pairs (i.e. P9 /P10 and P9/P11; see Table  4 for primer sequences) whereas the lower panel shows the electrophoretic separation of the P9/P10- and P9/P11-mediated PCR products. M, DNA marker; WT, normal control; DUP, duplication carrier; TRI, triplication carrier.

Characterization of the triplication junctions by long-range PCR and direct sequencing

Armed with the new information about the additionally triplicated sequence (Table  1 ), we again attempted to clone the triplication junctions by long-range PCR using the four primers listed in Table  2 and located as indicated in Figure  1 B. Ten possible primer combinations (see Materials and Methods) were tested in order to investigate the spatial organization of the two triplicated segments. That only two of the primer pairs, namely P2/P4 and P3/P1 (Fig.  1 B), yielded specific PCR products from the triplication carrier (Fig.  1 C) indicated that the ∼137 kb triplicated segment might lie in inverted orientation with respect to the ∼605 kb segment. Sequencing of the P2/P4-derived PCR product then allowed identification of the first junction (J1; Fig.  1 D), which displayed 6 base-pairs (bp) of microhomology between the telomeric end of the ∼605 kb segment and the centromeric end of the inverted ∼137 kb segment (Fig.  1 E). However, sequencing of the P3/P1-derived PCR product revealed that a small insertion had been introduced between the telomeric end of the inverted ∼137 kb segment and the centromeric end of the ∼605 kb segment. This micro-insertion was an inverted duplication of 21 nucleotides located immediately centromeric to the ∼605 kb segment (Fig.  1 D) and containing inverted repeats of 1 and 4-bp, respectively, at its ends (Fig.  1 E). Thus, the P3/P1-derived PCR product contained two telomeric junctions, J2 and J3, immediately flanking the 21 bp insertion (Fig.  1 D).

Table 2.

Primer sequences used for the long-range PCR performed to capture the triplication junctions

Primer a Sequence (5′ → 3′)  Region b Strand 
P1 CCCTCAACTGGATTGAAATCACC 141583430–141583452 − 
P2 TGGTTATGGCATTGCTCCCT 142188280–142188299 
P3 AGAGACCTGCGTGGGATAATCA 142279626–142279647 − 
P4 TTAGCAGCATCCACGGAAGA 142415515–142415534 
Primer a Sequence (5′ → 3′)  Region b Strand 
P1 CCCTCAACTGGATTGAAATCACC 141583430–141583452 − 
P2 TGGTTATGGCATTGCTCCCT 142188280–142188299 
P3 AGAGACCTGCGTGGGATAATCA 142279626–142279647 − 
P4 TTAGCAGCATCCACGGAAGA 142415515–142415534 

a See also Figure  1 B.

b In accordance with the human chromosome 7 assembly, March 2006 (hg18).

Employing the same long-range PCR procedure, analysis of patient 2193 (one of our previously reported PRSS1 gene duplication carriers; 26 ) also yielded PCR products but again only with the P2/P4 and P3/P1 primer pairs (Fig.  1 C). Direct sequencing of these PCR products revealed that the junctions of the duplication were identical to those identified in the triplication.

Structure of the complex triplication

The precise characterization of the aberrant junctions made it possible to draw up the overall structure of this complex triplication. As illustrated in Figure  1 D, the mutant allele was surmised to constitute a gain of two tandemly arranged, ∼840 kb composite blocks, each comprising one copy of the inverted ∼137 kb segment, one copy of the inverted 21 bp short sequence tract and one copy of the ∼605 kb segment. To provide further support for this predicted structure, we designed four QFM-PCR primers to allow precise quantification of the number of the aberrant junctions (Table  3 and Fig.  1 D) present in the aforementioned duplication (patient 2193; 26 ) and triplication (III.2 in pedigree C; 19 ) carriers. As expected, two copies of the short sequence tracts spanning J1 and J2/J3 were found in the triplication carrier whereas only one copy of each was detected in the duplication carrier. In contrast, these aberrant junctions were found to be entirely absent from 100 healthy controls (Fig.  1 F).

Table 3.

Primer sequences pertaining to the QFM-PCR used to determine the number of aberrant junctions in the duplication and triplication carriers

Primer a Sequence (5′ → 3′)  Region b Strand 
P5 TGGAGTGCAATGGTGCGAT 142190511–142190529 
P6 CCCATCTCTCCAATTTCCCACT 142416948–142416969 
P7 AGGGTTGGGTTTGGTGTG 142278875–142278892 − 
P8 TGCCCTCTGTGTGCATGAT 141580893–141580911 − 
Primer a Sequence (5′ → 3′)  Region b Strand 
P5 TGGAGTGCAATGGTGCGAT 142190511–142190529 
P6 CCCATCTCTCCAATTTCCCACT 142416948–142416969 
P7 AGGGTTGGGTTTGGTGTG 142278875–142278892 − 
P8 TGCCCTCTGTGTGCATGAT 141580893–141580911 − 

a See also Figure  1 D.

b In accordance with the human chromosome 7 assembly, March 2006 (hg18).

The QFM-PCR analysis was then extended to the remaining duplication ( n = 8) and triplication ( n = 28) carriers. One J1 and one J2/J3 sequence were invariably found in all the duplication carriers whereas two J1s and two J2/J3s were invariably present in all triplication carriers. Further, all breakpoints were sequenced from these 36 individuals and found to be identical. Finally, genotyping with a microsatellite marker (i.e. rs3222967), located toward the centromeric end of the ∼605 kb segment (refer to Fig.  1 b in 19 ), demonstrated that all duplication mutations invariably carried one 163 allele whereas all triplication mutations carried two 163 alleles; the 163 allele has a frequency of only ∼1.2% in the ethnically matched control population ( 19 , 26 ). These observations provided firm evidence that all the duplications and triplications tested had originated from the same common founder chromosome ( 19 , 26 ).

The newly identified ∼137 kb triplicated segment contains three known genes (i.e. TRPV6 , TRPV5 and KEL ) and one hypothetical gene (i.e. C7orf34 ). Both TRPV6 and TRPV5 encode calcium-permeable channels and are expressed at a high-level in the pancreas. We have been intrigued by our previous observation ( 19 , 26 ) that the triplication CNM is found exclusively in cases of either hereditary pancreatitis or ICP but never in cases of familial chronic pancreatitis [an immediate phenotype between the first two types of chronic pancreatitis ( 20 )]. Thus, it would be of potential interest to investigate the status of both TRPV6 and TRPV5 in the two groups of patients carrying a triplication CNM, with a view to identifying variants with modifier potential.

Finally, it should be noted that the precise characterization of the junctions of the duplications and triplications allows for easy screening using conventional PCR for the presence of these CNMs. The primer pair P9/P10 generated a band of 433 bp in both the duplication and triplication carriers as well as in healthy controls, whereas the primer pair P9/P11 generated a patient-specific band of 307 bp (Fig.  1 G; for primer sequences, see Table  4 ). This rapid method is not however capable of distinguishing a duplication from a triplication.

Table 4.

Primer sequences pertaining to the conventional PCR method used for the rapid detection of the presence of the duplication and triplication mutations

Primer a Sequence (5′ → 3′)  Region b Strand 
P9 GCATCTCTCACAGAGGCTGTT 142278852–142278872 − 
P10 ACTTTCCTTGTCTGGACTCACC 142278440–142278461 
P11 GTGCTTGCATTTGCACCTGT 141580924–141580943 − 
Primer a Sequence (5′ → 3′)  Region b Strand 
P9 GCATCTCTCACAGAGGCTGTT 142278852–142278872 − 
P10 ACTTTCCTTGTCTGGACTCACC 142278440–142278461 
P11 GTGCTTGCATTTGCACCTGT 141580924–141580943 − 

a See also Figure  1 G.

b In accordance with the human chromosome 7 assembly, March 2006 (hg18).

Exploring the origin of the complex triplication

The trypsinogen locus triplication analyzed here is one of the most complex CNMs known to cause human inherited disease and the first of its kind to be described at the DNA sequence level. Its analysis has provided new insights into the intricacy of the process underlying the generation of human CNVs. Our findings strongly support the view that the triplication allele was generated by two independent duplication events that gave rise to the first and second composite duplication blocks as indicated in Figure  1 D.

Origin of the first composite duplication block

The mechanism underlying the generation of the first duplication block is unlikely to have been homologous recombination since only very limited microhomologies were found at the junction sequences. The presence of microhomologies at aberrant chromosomal junctions is not inconsistent with NHEJ, a DNA-repair mechanism that is active throughout both the mitotic and meiotic cell cycles ( 27 ). However, it is difficult to explain how this composite block could have been generated by NHEJ alone. Moreover, both characteristic hallmarks of NHEJ, viz. untemplated sequence additions and microhomology-independent end joining, were absent at the aberrant junctions. In contrast, this composite duplication block can be readily accounted for by DNA replication-based mechanisms.

Before focusing on the precise details of the replication-based mechanisms involved, we would firstly like to address two issues. First, in principle, when the primer strand encounters a stalled DNA replication fork, it can switch to illegitimate templates which are located either on the same chromatid, the sister chromatid, the homologous chromosome, or even on non-homologous chromosomes. In the context of the composite duplication block studied here, the non-homologous chromosomes could be unequivocally excluded as potential illegitimate templates since all the duplicated sequences were derived from chromosome 7. Moreover, all the duplication mutations were found to carry one (rare) 163 allele of the microsatellite marker rs3222967 located within the ∼607 kb segment at 7q34; this suggested that at least the two copies of the ∼607 kb segment in the composite duplication block had originated from the two homologous chromosomes, respectively.

Second, replication-based mechanisms are thought to be microhomology-dependent. Indeed, mutational events have often been reported to involve a microhomology of ≥2 bp. This notwithstanding, although one may well argue that the microhomology of only 1 bp putatively involved in J2 formation (Fig.  1 D and E) may simply be coincidental, we could not formally exclude its functional significance. Further, a 1 bp microhomology may well play a role in mediating template switching, a possibility best exemplified by the 15.5 kb deletion/16 bp insertion component of the double complex mutation involving F8 and FUNDC2 ( 12 ). The 16 bp insertion could be divided into three short tracts, each representing a direct or reverse repeat of sequence flanking the rejoining junction. This insertion was found to be perfectly explicable in terms of three steps of replication slippage, each of which was mediated by a microhomology of 1 bp ( 12 ).

SRS/FoSTeS: This model, described in the left panel of Figure  2 , is proposed to involve four steps of replication slippage occurring between the homologous chromosomes. First, the nascent primer strand (containing the rare 163 allele of rs3222967; horizontal green arrowhead) would have dissociated from the original template strand before annealing to the telomeric end of the ∼137 kb segment, located on the homologous chromosome (containing a non-163 allele of rs3222967) in inverse orientation, through a 6 bp inverse repeat (i.e. 5′-CAAGTA-3′ on the primer strand and 5′-TACTTG-3′ on the misaligned template strand; see Fig.  1 E). New DNA synthesis would then have produced a strand complementary to the ∼137 kb template strand. Second, the primer strand is proposed to have then dissociated from the ∼137 kb template strand and annealed to the telomeric end of the 21 bp sequence tract with 1 bp microhomology (i.e. 5′-G-3′ in the primer strand and 5′-C-3′ in the misaligned template strand; see Fig.  1 E). New DNA synthesis would then have yielded the complementary strand of the 21 bp template sequence tract. Third, the primer strand would have further dissociated from the 21 bp template and annealed to the centromeric end of the ∼605 kb segment through a 4 bp direct repeat (i.e. 5′-ATTG-3′; see Fig.  1 E); new DNA synthesis would then have resulted in the formation of a copy of the ∼605 kb template segment. Finally, the primer strand would have re-annealed to the initial dissociation site leading to the resumption of normal DNA replication.

Figure 2.

Mechanisms underlying the generation of the trypsinogen locus duplication and triplication. ( A ) Two possible replication-based mechanisms, SRS/FoSTeS and BISRS, underlying the generation of the first composite duplication. In both models, the direction of replication of the nascent primer strand (5′ > 3′) is indicated by horizontal arrows (I). The different steps of replication slippage are numbered sequentially (I). The newly synthesized duplicated sequence is indicated by coloured dotted lines. The use of colour is consistent with the scheme as used in Figure  1 . In IV, the orientation of the duplicated sequences, with respect to the wild-type, is indicated by colour coded horizontal lines. The rare 163 allele of the microsatellite marker, rs3222967, is indicated by a star. See text for details as to how the second mutant strand could have been generated. ( B ) Schema of how the second composite duplication block could have been generated by NAHR giving rise to the complex triplication allele. Both the triplication allele and its alterative product (i.e. the predicted and inherently unobservable normal allele) of NAHR are shown. The composite first duplication block was presumed to have been generated from one of the scenarios described in A. ( C ) The clinically observed complex duplications could have been derived from the triplication allele through either intrachromatid or interchromatid NAHR. The alternative products of these events are also shown. The dotted lines are used only for illustration purposes and hence do not represent actual sequence.

Figure 2.

Mechanisms underlying the generation of the trypsinogen locus duplication and triplication. ( A ) Two possible replication-based mechanisms, SRS/FoSTeS and BISRS, underlying the generation of the first composite duplication. In both models, the direction of replication of the nascent primer strand (5′ > 3′) is indicated by horizontal arrows (I). The different steps of replication slippage are numbered sequentially (I). The newly synthesized duplicated sequence is indicated by coloured dotted lines. The use of colour is consistent with the scheme as used in Figure  1 . In IV, the orientation of the duplicated sequences, with respect to the wild-type, is indicated by colour coded horizontal lines. The rare 163 allele of the microsatellite marker, rs3222967, is indicated by a star. See text for details as to how the second mutant strand could have been generated. ( B ) Schema of how the second composite duplication block could have been generated by NAHR giving rise to the complex triplication allele. Both the triplication allele and its alterative product (i.e. the predicted and inherently unobservable normal allele) of NAHR are shown. The composite first duplication block was presumed to have been generated from one of the scenarios described in A. ( C ) The clinically observed complex duplications could have been derived from the triplication allele through either intrachromatid or interchromatid NAHR. The alternative products of these events are also shown. The dotted lines are used only for illustration purposes and hence do not represent actual sequence.

One issue should, however, be discussed in more detail. In accordance with SRS/FoSTeS, the nascent mutant strand should have been displaced from the illegitimate template immediately or shortly after its synthesis. This would have resulted in a large single-stranded/loop structure in the newly synthesized mutant strand (left panel, II, Fig.  2 ). Thus, replication against the nascent mutant strand would have been required to yield the double-stranded mutant molecule. This might have been achieved in two distinct ways ( 6 ). Firstly, the synthesis of, and replication against, the nascent mutant strand may have occurred within the same cell cycle (III1, Fig.  2 ); a process that would have required cleavage of the original template strand (indicated by a cross; Fig.  2 ) followed by DNA gap filling and ligation. Alternatively, the single-stranded/loop structure could have escaped the host repair system; DNA replication against the nascent mutant strand would then have occurred in the next cell cycle (III2, Fig.  2 ).

BISRS

The other replication-based mechanism, BISRS, provides an alternative means to generate the first composite block (right panel, Fig.  2 ). The defining features of this model, in contrast to SRS/FoSTeS, are that (i) initial replication slippage occurs when the primer strand encounters a collapsed replication fork rather than a stalled replication fork; (ii) SRS ends with the engagement of a misaligned template instead of reannealing to the original template and (iii) the synthesis of the second strand immediately follows the synthesis of the first strand ( 18 , 28 ).

SRS versus BISRS

We favor BISRS over SRS/FoSTeS in the context of the composite duplication for the following reasons. First, DNA replication is a frequent source of double strand breaks, either via replication across a single-stranded nick or through rupture of a DNA strand at a stalled replication fork ( 29 ). Second, microhomology-mediated BIR has been increasingly invoked as a potential mechanism for generating large genomic rearrangements in both experimental systems and human inherited disease ( 6 , 12 , 13 , 17 , 18 , 30 , 31 ). Third, serial template switching has been experimentally demonstrated to occur during BIR, at least in yeast ( 15 ). Fourth, BISRS seems to offer a more parsimonious explanation than SRS/FoSTeS since it involves only three misalignment steps instead of four. Finally, during the process of BIR, the synthesis of both the first and second strands of the newly synthesized DNA is thought to occur within a single cell cycle ( 18 , 28 ).

Timing of the origin of the first and second composite duplications

Although NAHR occurs predominantly during meiosis, NHEJ and replication-based mechanisms are mainly thought to occur during mitosis ( 32 ). Given that germ cells undergo repeated cycles of mitosis before entering meiosis, we speculate that the first composite block may have arisen during the premeiotic mitotic cell divisions. The tandemly duplicated ∼605-kb segments, derived from the aforementioned process, could then have promoted NAHR between sister chromatids during meiosis leading to the formation of the complex triplication allele. This duplication event, which putatively led to the generation of the second composite duplication, is indicated in Figure  2 B.

Relationship between the duplication and the triplication alleles

It is quite possible that both the first and second composite duplications were generated in the germline of the same individual (Fig.  3 A and B). Consequently, all examples of the clinically observed triplication and duplication mutations would ultimately have been derived from this common founder individual whereas the duplication alleles could have been derived from a triplication allele through NAHR during subsequent meioses (Figs  2 C and 3 C). It remains possible that the first composite duplication block that arose during the premeiotic cell divisions did not promote NAHR during a subsequent meiosis in the same individual but rather did so in a later direct descendant (Fig.  3 A-D-E). In this latter case, the observed duplication would not necessarily have been derived from a triplication allele(s) (Fig.  3 D-E-C or D-F). It should be noted that the triplication allele could only have been generated through NAHR between sister chromatids (Fig.  2 B) and not via an intrachromosomal mechanism. The duplication that was derived from a triplication allele could however have occurred either within one chromatid (intrachromatidal) or between sister chromatids (interchromatidal) (Fig.  2 C). In the latter context, we favor an intrachromatidal over an interchromatidal mechanism since, in four known NAHR hotspots, the former mechanism was shown, by means of a sperm-based assay, to occur more frequently than the latter ( 33 ). This may explain the observation that no quadruplication (lower panel, Fig.  2 C) has yet been identified. Alternatively, such a quadruplicated allele could be lethal in utero and hence may never come to clinical attention.

Figure 3.

Potential scenarios for how the clinically observed duplication and triplication rearrangements could have been generated. A - B - C indicates the possibility that the initial duplication and subsequent triplication were generated in the same individual; and the clinically observed duplications were then generated from the triplicated allele(s). A - D indicates the other possibility, namely the initially generated duplication was not converted to a triplication in the same individual, which may have led to two further alternative scenarios (i.e. D - E - C versus D - F ). See text for details.

Figure 3.

Potential scenarios for how the clinically observed duplication and triplication rearrangements could have been generated. A - B - C indicates the possibility that the initial duplication and subsequent triplication were generated in the same individual; and the clinically observed duplications were then generated from the triplicated allele(s). A - D indicates the other possibility, namely the initially generated duplication was not converted to a triplication in the same individual, which may have led to two further alternative scenarios (i.e. D - E - C versus D - F ). See text for details.

The possibility that the clinically observed duplications could have been derived from a triplication allele would have been supported by the identification of a duplication occurring de novo . However, parental samples were only obtainable in the case of three of the nine duplication carriers under study and, in these individuals, the duplication CNMs were all found to have been inherited (data not shown).

Conclusions and perspective

Employing high-resolution array CGH data, we were able to fully characterize the ∼605 kb trypsinogen gene triplication previously reported to cause hereditary pancreatitis. It turned out to be part of quite a complex rearrangement involving a second large (∼137 kb) triplication as well as a small 21 bp triplication. This complexity had initially rendered the locus refractory to analysis in the first report of the rearrangement ( 19 ). The newly identified triplicated sequences were both found to lie in inverted orientation as compared with the wild-type sequence. Furthermore, the ∼605 and ∼137 kb segments were separated by ∼90 kb in the genomic reference sequence. Characterization of this complex triplication also made it possible to speculate upon its origin. We have proposed a two-step pathway for the generation of this highly unusual complex triplication. Thus, we propose that BISRS could have firstly created a low copy repeat during mitosis that then provided the homology required for NAHR to occur during meiosis leading to the generation of the complex triplication allele. This triplication allele could then have undergone further NAHR during meiosis, leading to the formation of the clinically significant duplications (Fig.  3 ).

Based upon this and other studies, a general concept is now emerging that could account for the origin of CNVs in the human germline: NHEJ and various replication-based mechanisms (including SRS, BISRS and simple replication slippage or microhomology-mediated BIR that involves only a single step of template switching) constitute the two main driving forces responsible for generating duplicons or low copy repeats (and also deletions) during the mitotic divisions of diploid oogonia and spermatogonia. The low copy repeats then promote NAHR during meiosis whilst the consequent amplified sequences predispose to further recombination events during both mitosis and meiosis. This composite mechanism could potentially also explain the origins of other CNMs such as the PLP1 triplication and quadruplication causing severe Pelizaeus–Merzbacher disease ( 34 ), the SNCA triplication resulting in Parkinson's disease ( 35 ) and the MECP2 triplication giving rise to male mental retardation ( 36 ).

MATERIALS AND METHODS

Patients

Twenty nine trypsinogen ( PRSS1 ) gene triplication carriers [17 clinically affected individuals from five previously described hereditary pancreatitis families ( 19 ), two patients from one newly recruited hereditary pancreatitis family, and 10 subjects previously described with ICP ( 26 )] and nine trypsinogen duplication carriers (four ICP patients previously described in ( 26 ) and five newly recruited ICP patients) were investigated in this study. All patients are French Caucasians. A total of 100 controls (healthy individuals from the same population) were employed to analyze the aberrant junctions identified in the triplication and duplication mutations.

Genomic reference sequence

The human chromosome 7 assembly, March 2006 (hg18) ( http://genome.ucsc.edu/cgi-bin/hgGateway ) was used as the genomic reference sequence throughout.

Array-CGH

Array CGH was performed using the Agilent Human Genome CGH Microarray 244 K, with an overall median probe spacing of 8.9 kb (Agilent Technologies, Santa Clara, CA, USA). One microgram genomic DNA from a known carrier of a trypsinogen ( PRSS1 ) gene triplication or reference (pool of male DNAs) was double-digested with Alu I and Rsa I (New England Biolabs Corp., Ipswich, MA, USA) for 2 h at 37°C. Digested patient and reference DNAs were labelled by random priming with Cy5-dUTP and Cy3-dUTP, respectively, using the Agilent Genomic DNA Labelling Kit PLUS. Labelled products were purified with Microcon YM-30 filters (Millipore, Billerica, MA, USA). The yield and specific activity of each labelling reaction were measured with the NanoDrop UV–VIS ND-1000 spectrophotometer (NanoDrop Technologies, Wilmington, DE, USA). The combined Cy5- and Cy3-labelled DNAs were incubated with 50 µg human Cot-1 DNA (Invitrogen, Austin, TX, USA) for 40 h at 65°C in a rotating oven. Washing steps were performed according to the Agilent protocol using the Stabilization and Drying Solution. Arrays were scanned using the Agilent Scanner G2505B and analyzed using Agilent Feature Extraction Software v9.5.1 and Agilent CGH Analytics software v3.5.14.

Walking QFM-PCR

QFM-PCR was carried out as previously described ( 37 , 38 ). A series of walking primers were designed (data not shown) to narrow down the location of the centromeric and telomeric ends of the second triplicated segment as revealed by array-CGH (Fig.  1 A).

Long-range PCR

Long range PCR was employed to clone the aberrant junctions of the PRSS1 gene triplication. To this end, we designed four outward-facing primers that were located near the four end-regions of the two triplicated segments (Table  2 and Fig.  1 B) and tested all 10 possible primer combinations (i.e. P1/P1, P1/P2, P1/P3, P1/P4, P2/P2, P2/P3, P2/P4, P3/P3, P3/P4 and P4/P4) for amplification success. Long-range PCR was performed in a 50-µl reaction mixture containing 350 µ m each dNTP, 1.75 m m MgCl 2 , 300 n m each primer, 0.75 µl Taq DNA polymerase and 100 ng genomic DNA, using the Expand Long Template PCR system (Roche Diagnostics, Meylan, France). The PCR program comprised an initial denaturation at 94°C for 2 min; 10 cycles of denaturation at 94°C for 10 s, annealing at 60°C for 30 s and extension at 68°C for 4 min; 25 cycles of denaturation at 94°C for 15 s and annealing at 60°C for 4 min + 20 s (i.e. with an increase of 20 s/cycle); and a final extension at 68°C for 7 min.

DNA sequencing

Two long-range PCR products, generated from the primer combinations of P2/P4 and P3/P1, respectively, were purified by ExoSAP-IT (GE Healthcare, Orsay, France) and then sequenced using the ABI PRISM BigDye Terminator v3.1 Cycle Sequencing kit (Applied Biosystems, Foster City, CA, USA). A series of walking sequencing primers (data not shown) were designed in order to locate the aberrant junctions.

Quantifying the number of aberrant junctions

Four QFM-PCR primers were designed to permit quantitation of the number of aberrant junctions (Table  3 and Fig.  1 D) in all 9 duplication and 29 triplication carriers.

Rapid PCR-based detection of the duplication and triplication mutations

A conventional PCR, that co-amplifies a tract of wild-type sequence as well as a patient-specific tract spanning J2/J3, was developed for the rapid detection of the duplication and triplication CNMs. See Figure  1 G (top panel) for primer positions and Table  4 for primer sequences. PCR was performed with 2 m m MgCl 2 , 200 µ m each dNTP, 0.2 µ m each primer, 0.02 U AmpliTaq polymerase and 75 ng DNA template. The PCR program comprised an initial denaturation at 94°C for 5 min, 35 cycles of denaturation at 94°C for 30 s, annealing at 60°C for 30 s, extension at 72°C for 45 s and a final extension at 65°C for 10 min. PCR products were detected on a 2% agarose gel.

Genotyping of the microsatellite marker, rs3222967

A microsatellite marker, rs3222967, located within 7q34, was previously identified toward the centromeric end of the ∼607 kb segment (see Fig.  1 b in 19 ). This marker was genotyped in all the newly identified duplication and triplications carriers as previously described ( 19 ).

Conflict of Interest statement . None declared.

FUNDING

This work was supported by the INSERM (Institut National de la Santé et de la Recherche Médicale), the Programme Hospitalier de Recherche Clinique (grant PHRC R 08-04) the PICRI (Partenariats Institutions Citoyens pour la Recherche et l'Innovation) project, the GIS Institut des Maladies Rares (project no. A07092NS), and the APCH (Association des Pancréatites Chroniques Héréditaires), France.

REFERENCES

1
Scherer
S.W.
Lee
C.
Birney
E.
Altshuler
D.M.
Eichler
E.E.
Carter
N.P.
Hurles
M.E.
Feuk
L.
Challenges and standards in integrating surveys of structural variation
Nat. Genet.
 , 
2007
, vol. 
39
 (pg. 
S7
-
S15
)
2
Conrad
D.F.
Hurles
M.E.
The population genetics of structural variation
Nat. Genet.
 , 
2007
, vol. 
39
 (pg. 
S30
-
S36
)
3
Cooper
G.M.
Nickerson
D.A.
Eichler
E.E.
Mutational and selective effects on copy-number variants in the human genome
Nat. Genet.
 , 
2007
, vol. 
39
 (pg. 
S22
-
S29
)
4
Beckmann
J.S.
Sharp
A.J.
Antonarakis
S.E.
CNVs and genetic medicine (excitement and consequences of a rediscovery)
Cytogenet. Genome Res.
 , 
2008
, vol. 
123
 (pg. 
7
-
16
)
5
Shaw
C.J.
Lupski
J.R.
Implications of human genome architecture for rearrangement-based disorders: the genomic basis of disease
Hum. Mol. Genet.
 , 
2004
, vol. 
13
 
Spec no. 1
(pg. 
R57
-
R64
)
6
Chen
J.M.
Chuzhanova
N.
Stenson
P.D.
Férec
C.
Cooper
D.N.
Intrachromosomal serial replication slippage in trans gives rise to diverse genomic rearrangements involving inversions
Hum. Mutat.
 , 
2005
, vol. 
26
 (pg. 
362
-
373
)
7
Chen
J.M.
Chuzhanova
N.
Stenson
P.D.
Férec
C.
Cooper
D.N.
Complex gene rearrangements caused by serial replication slippage
Hum. Mutat.
 , 
2005
, vol. 
26
 (pg. 
125
-
134
)
8
Férec
C.
Casals
T.
Chuzhanova
N.
Macek
M.
Jr.
Bienvenu
T.
Holubova
A.
King
C.
McDevitt
T.
Castellani
C.
Farrell
P.M.
, et al.  . 
Gross genomic rearrangements involving deletions in the CFTR gene: characterization of six new events from a large cohort of hitherto unidentified cystic fibrosis chromosomes and meta-analysis of the underlying mechanisms
Eur. J. Hum. Genet.
 , 
2006
, vol. 
14
 (pg. 
567
-
576
)
9
Lee
J.A.
Carvalho
C.M.
Lupski
J.R.
A DNA replication mechanism for generating nonrecurrent rearrangements associated with genomic disorders
Cell
 , 
2007
, vol. 
131
 (pg. 
1235
-
1247
)
10
Carvalho
C.M.
Zhang
F.
Liu
P.
Patel
A.
Sahoo
T.
Bacino
C.A.
Shaw
C.
Peacock
S.
Pursley
A.
Tavyev
Y.J.
, et al.  . 
Complex rearrangements in patients with duplications of MECP2 can occur by fork stalling and template switching
Hum. Mol. Genet.
 , 
2009
, vol. 
18
 (pg. 
2188
-
2203
)
11
Gu
W.
Zhang
F.
Lupski
J.R.
Mechanisms for human genomic rearrangements
Pathogenetics
 , 
2008
, vol. 
1
 pg. 
4
 
12
Sheen
C.R.
Jewell
U.R.
Morris
C.M.
Brennan
S.O.
Férec
C.
George
P.M.
Smith
M.P.
Chen
J.M.
Double complex mutations involving F8 and FUNDC2 caused by distinct break-induced replication
Hum. Mutat.
 , 
2007
, vol. 
28
 (pg. 
1198
-
1206
)
13
Koszul
R.
Caburet
S.
Dujon
B.
Fischer
G.
Eucaryotic genome evolution through the spontaneous duplication of large chromosomal segments
Embo J.
 , 
2004
, vol. 
23
 (pg. 
234
-
243
)
14
Chen
J.M.
Masson
E.
Le Maréchal
C.
Férec
C.
Copy number variations in chronic pancreatitis
Cytogenet. Genome Res.
 , 
2008
, vol. 
123
 (pg. 
102
-
107
)
15
Smith
C.E.
Llorente
B.
Symington
L.S.
Template switching during break-induced replication
Nature
 , 
2007
, vol. 
447
 (pg. 
102
-
105
)
16
Slack
A.
Thornton
P.C.
Magner
D.B.
Rosenberg
S.M.
Hastings
P.J.
On the mechanism of gene amplification induced under stress in Escherichia coli.
PLoS Genet.
 , 
2006
, vol. 
2
 pg. 
e48
 
17
Payen
C.
Koszul
R.
Dujon
B.
Fischer
G.
Segmental duplications arise from Pol32-dependent repair of broken forks through two alternative replication-based mechanisms
PLoS Genet.
 , 
2008
, vol. 
4
 pg. 
e1000175
 
18
Hastings
P.J.
Ira
G.
Lupski
J.R.
A microhomology-mediated break-induced replication model for the origin of human copy number variation
PLoS Genet.
 , 
2009
, vol. 
5
 pg. 
e1000327
 
19
Le Maréchal
C.
Masson
E.
Chen
J.M.
Morel
F.
Ruszniewski
P.
Levy
P.
Férec
C.
Hereditary pancreatitis caused by triplication of the trypsinogen locus
Nat. Genet.
 , 
2006
, vol. 
38
 (pg. 
1372
-
1374
)
20
Chen
J.M.
Férec
C.
Chronic pancreatitis: genetics and pathogenesis
Annu. Rev. Genomics Hum. Genet.
 , 
2009
 
10.1146/annurev-genom-082908–150009
21
Whitcomb
D.C.
Gorry
M.C.
Preston
R.A.
Furey
W.
Sossenheimer
M.J.
Ulrich
C.D.
Martin
S.P.
Gates
L.K.
Jr.
Amann
S.T.
Toskes
P.P.
, et al.  . 
Hereditary pancreatitis is caused by a mutation in the cationic trypsinogen gene
Nat. Genet.
 , 
1996
, vol. 
14
 (pg. 
141
-
145
)
22
Gorry
M.C.
Gabbaizedeh
D.
Furey
W.
Gates
L.K.
Jr.
Preston
R.A.
Aston
C.E.
Zhang
Y.
Ulrich
C.
Ehrlich
G.D.
Whitcomb
D.C.
Mutations in the cationic trypsinogen gene are associated with recurrent acute and chronic pancreatitis
Gastroenterology
 , 
1997
, vol. 
113
 (pg. 
1063
-
1068
)
23
Chen
J.M.
Kukor
Z.
Le Maréchal
C.
Tóth
M.
Tsakiris
L.
Raguénès
O.
Férec
C.
Sahin-Tóth
M.
Evolution of trypsinogen activation peptides
Mol. Biol. Evol.
 , 
2003
, vol. 
20
 (pg. 
1767
-
1777
)
24
Chen
J.M.
Le Maréchal
C.
Lucas
D.
Raguénès
O.
Férec
C.
‘Loss of function’ mutations in the cationic trypsinogen gene ( PRSS1 ) may act as a protective factor against pancreatitis
Mol. Genet. Metab.
 , 
2003
, vol. 
79
 (pg. 
67
-
70
)
25
Witt
H.
Sahin-Tóth
M.
Landt
O.
Chen
J.M.
Kahne
T.
Drenth
J.P.
Kukor
Z.
Szepessy
E.
Halangk
W.
Dahm
S.
, et al.  . 
A degradation-sensitive anionic trypsinogen ( PRSS2 ) variant protects against chronic pancreatitis
Nat. Genet.
 , 
2006
, vol. 
38
 (pg. 
668
-
673
)
26
Masson
E.
Le Maréchal
C.
Chandak
G.R.
Lamoril
J.
Bezieau
S.
Mahurkar
S.
Bhaskar
S.
Reddy
D.N.
Chen
J.M.
Férec
C.
Trypsinogen copy number mutations in patients with idiopathic chronic pancreatitis
Clin. Gastroenterol. Hepatol.
 , 
2008
, vol. 
6
 (pg. 
82
-
88
)
27
Lieber
M.R.
The mechanism of human nonhomologous DNA end joining
J. Biol. Chem.
 , 
2008
, vol. 
283
 (pg. 
1
-
5
)
28
McEachern
M.J.
Haber
J.E.
Break-induced replication and recombinational telomere elongation in yeast
Annu. Rev. Biochem.
 , 
2006
, vol. 
75
 (pg. 
111
-
135
)
29
Haber
J.E.
DNA recombination: the replication connection
Trends Biochem. Sci.
 , 
1999
, vol. 
24
 (pg. 
271
-
275
)
30
Rooms
L.
Reyniers
E.
Kooy
R.F.
Diverse chromosome breakage mechanisms underlie subtelomeric rearrangements, a common cause of mental retardation
Hum. Mutat.
 , 
2007
, vol. 
28
 (pg. 
177
-
182
)
31
Bauters
M.
Van Esch
H.
Friez
M.J.
Boespflug-Tanguy
O.
Zenker
M.
Vianna-Morgante
A.M.
Rosenberg
C.
Ignatius
J.
Raynaud
M.
Hollanders
K.
, et al.  . 
Nonrecurrent MECP2 duplications mediated by genomic architecture-driven DNA breaks and break-induced replication repair
Genome Res.
 , 
2008
, vol. 
18
 (pg. 
847
-
858
)
32
Arlt
M.F.
Mulle
J.G.
Schaibley
V.M.
Ragland
R.L.
Durkin
S.G.
Warren
S.T.
Glover
T.W.
Replication stress induces genome-wide copy number changes in human cells that resemble polymorphic and pathogenic variants
Am. J. Hum. Genet.
 , 
2009
, vol. 
84
 (pg. 
339
-
350
)
33
Turner
D.J.
Miretti
M.
Rajan
D.
Fiegler
H.
Carter
N.P.
Blayney
M.L.
Beck
S.
Hurles
M.E.
Germline rates of de novo meiotic deletions and duplications causing several genomic disorders
Nat. Genet.
 , 
2008
, vol. 
40
 (pg. 
90
-
95
)
34
Wolf
N.I.
Sistermans
E.A.
Cundall
M.
Hobson
G.M.
Davis-Williams
A.P.
Palmer
R.
Stubbs
P.
Davies
S.
Endziniene
M.
Wu
Y.
, et al.  . 
Three or more copies of the proteolipid protein gene PLP1 cause severe Pelizaeus–Merzbacher disease
Brain
 , 
2005
, vol. 
128
 (pg. 
743
-
751
)
35
Singleton
A.B.
Farrer
M.
Johnson
J.
Singleton
A.
Hague
S.
Kachergus
J.
Hulihan
M.
Peuralinna
T.
Dutra
A.
Nussbaum
R.
, et al.  . 
alpha-Synuclein locus triplication causes Parkinson's disease
Science
 , 
2003
, vol. 
302
 pg. 
841
 
36
del Gaudio
D.
Fang
P.
Scaglia
F.
Ward
P.A.
Craigen
W.J.
Glaze
D.G.
Neul
J.L.
Patel
A.
Lee
J.A.
Irons
M.
, et al.  . 
Increased MECP2 gene copy number as the result of genomic duplication in neurodevelopmentally delayed males
Genet. Med.
 , 
2006
, vol. 
8
 (pg. 
784
-
792
)
37
Audrézet
M.P.
Chen
J.M.
Raguénès
O.
Chuzhanova
N.
Giteau
K.
Le Maréchal
C.
Quéré
I.
Cooper
D.N.
Férec
C.
Genomic rearrangements in the CFTR gene: extensive allelic heterogeneity and diverse mutational mechanisms
Hum. Mutat.
 , 
2004
, vol. 
23
 (pg. 
343
-
357
)
38
Masson
E.
Le Maréchal
C.
Delcenserie
R.
Chen
J.M.
Férec
C.
Hereditary pancreatitis caused by a double gain-of-function trypsinogen mutation
Hum. Genet.
 , 
2008
, vol. 
123
 (pg. 
521
-
529
)