Introduction

Several hundred distinct BRCA1 and BRCA2 mutations have been identified in hereditary breast/ovarian cancer (HBOC) families, including large genomic rearrangements (LGRs). The high content of Alu elements in the intronic sequences of both BRCA1 and BRCA2 genes contributes to the occurrence of rearrangements in these genes.1, 2 These genomic rearrangements include primarily deletions and duplications of one or more exons and represent 10–15% of all deleterious germline mutations in the BRCA1 gene and 1–7% in the BRCA2 gene.3, 4 The contribution of these LGRs to BRCA1/2 mutation-positive families varies in a population-dependent manner and numerous BRCA1 founder rearrangements have been identified in particular ethnic groups.5, 6, 7 On the other hand, the proportion of rearrangements in the BRCA2 gene is generally much smaller than that in the BRCA1 gene, which can be explained by the lower content of repetitive Alu sequences in the former compared with the latter. The only exception to this rule occurs in Portugal, where the founder rearrangement BRCA2 c.156_157insAlu (an Alu insertion in exon 38) accounts for more than one fourth of HBOC families with deleterious BRCA1/BRCA2 mutations and about half of those with BRCA2 mutations.9

The detection of LGRs requires specific techniques, like multiplex ligation probe amplification (MLPA) or quantitative PCR, which only give information of the exonic region involved in the rearrangement and not on the exact genomic breakpoints. Therefore, rearrangements with distinct breakpoints but involving the same exons may be described similarly, making it difficult to assess whether a given rearrangement is new or even specific of a given population.10 To determine whether a LGR is actually novel and to gain insight about the mutational mechanism responsible for its occurrence, molecular characterization with breakpoint identification is mandatory. About 81 distinct BRCA1 LGRs have been described, the majority of them being deletions, followed by rearrangements with both an insertion and a deletion, duplications, one insertion and one triplication.4 We here report the characterization of the genomic breakpoints of two BRCA1 LGRs detected in Portuguese HBOC families, namely, a deletion involving BRCA1 to the nearby gene NBR1 and an intragenic BRCA1 deletion.

Materials and methods

Patients and MLPA analyses

Two patients from two HBOC families, initially shown to be negative for BRCA1 and BRCA2 point mutations, were included in this study after being shown to carry BRCA1 exonic rearrangements by MLPA. This study was approved by the Institutional Review Board of the Portuguese Oncology Institute-Porto, and written informed consent was obtained from all patients after genetic counseling. DNA was isolated from peripheral blood samples using the salt–chloroform extraction method. Routine screening for BRCA1 and BRCA2 exonic rearrangements was performed by MLPA using the P002 kit and the P045 kit, respectively (MRC-Holland, Amsterdam, The Netherlands), according to the manufacturer’s instructions. One index patient presenting a deletion from exon 1 to exon 7, c.-232-?_441+?del (NM_007294.3) according to the Human Genome Variation Society nomenclature, had bilateral breast cancer at age 32 and 36, gastric cancer at age 38 and ovarian cancer at age 61, besides having a first-degree relative with breast cancer diagnosed at age 37. The second index patient presented a BRCA1 deletion including exon 11 to exon 15, described as c.671-?_4675+?del (NM_007294.3) according to the Human Genome Variation Society nomenclature and initially reported in Peixoto et al.,11 had an ovarian borderline tumor at age 38 and breast cancer at age 46, with a first-degree relative presenting breast cancer diagnosed at age 50. These two genomic rearrangements account for two of 48 HBOC families with deleterious BRCA1 mutations so far identified at the Department of Genetics of the Portuguese Oncology Institute, Porto, Portugal.

Breakpoint identification of the BRCA1 exons 1–7 deletion

To map the 5′ breakpoint region of this BRCA1 c.-232-?_441+?del mutation, we used the MLPA kit P239 (MRC-Holland) to evaluate the BRCA1 promoter region, the BRCA1 pseudogene (ΨBRCA1), and the NBR2 (exons 1, 3 and 5) and NBR1 (exons 3, 7 and 11) genes, followed by a semi-quantitative multiplex PCR method spanning intron 11 to exon 22 of the NBR1 gene (NM_005899.3). Four multiplex reactions were designed, each one containing four controls (exons 10 and 24 of the BRCA1 gene and exons 11 and 27 of the BRCA2 gene) and three different NBR1 regions (Table 1). In addition, all these NBR1 fragments were designed to include single nucleotide polymorphisms (SNPs) to question, at the same time, their heterozygosity status. The primers for multiplex amplification were designed using the online Primer-BLAST tool (http://www.ncbi.nlm.nih.gov/tools/primer-blast/). AutoDimer (http://www.cstl.nist.gov/strbase/NIJ/AutoDimer.htm) was used to test for potential hairpin structures and primer dimers. Subsequently, primers were designed spanning the putative breakpoints and PCR was carried out using standard conditions. PCR fragments containing the suspected weight were sequenced with BigDye Terminator cycle sequencing chemistry on an ABI PRISM 310 automatic sequencer (Applied Biosystems, Foster City, CA, USA), according to the manufacturer’s recommendations.

Table 1 NBR1 gene analysis using SNPs and a semi-quantitative multiplex PCR method to characterize the BRCA1 exons 1–7 deletion

Breakpoint identification of the BRCA1 exons 11–15 deletion

The strategy for breakpoint identification of the BRCA1 c.671-?_4675+?del mutation was based on the heterozygosity status of a set of SNP markers localized in introns 10 and 15 of BRCA1. The selected SNPs were rs7503154, rs8176214, rs77152443 and rs72130855. Subsequently, primers were designed flanking the putative breakpoints and PCR was carried out using standard conditions. PCR fragments were sequenced with BigDye Terminator cycle sequencing chemistry on an ABI PRISM 310 automatic sequencer (Applied Biosystems), according to the manufacturer’s recommendations.

Breakpoint sequence context analysis

Breakpoints were defined as a set of coordinates in the genome spanning the genomic sequence of the deletion. Bioinformatic analyses were carried out to assess the genomic context of the region. Both Censor Server (http://www.girinst.org/)12 and the RepeatMasker software (http://www.repeatmasker.org) were used to search for low-complexity DNA sequences and interspersed repeats in both breakpoint regions. All sequence alignments were performed using the complementary strands to match the Alu consensus sequence.13

Results

Characterization of the BRCA1 exons 1–7 deletion

Analysis by the MLPA kit P002 revealed a genomic deletion encompassing BRCA1 exons 1–7. Analysis with the MLPA kit P239 of the BRCA1 promoter region, ΨBRCA1 and the NBR2 (exons 1, 3 and 5) and NBR1 (exon 3, 7 and 11) genes revealed that all these regions were also deleted. To analyze the remaining NBR1 exons, we performed a semi-quantitative multiplex PCR method spanning intron 11 to exon 22 of the NBR1 gene. All these NBR1 fragments included SNPs to question, at the same time, their heterozygosity status (Table 1). The results showed that the NBR1 region encompassing rs112693602 (intron 11/ exon 12) to rs13119 (exon 22) was not deleted, narrowing the distance to the breakpoint. After long-range PCR with putative primers spanning intron 7 of the BRCA1 gene and intron 11 of the NBR1 gene, we obtained an 1300-bp fragment in the sample with the BRCA1 rearrangement. Sequence analysis of this PCR product revealed that the case harbored a 89 664-bp deletion, comprising exon 7 of the BRCA1 gene to exon 11 of the NBR1 gene, with the 5′ and 3′ breakpoints located 1724-bp upstream of BRCA1 exon 7 and 480-bp upstream of NBR1 exon 11, respectively, corresponding to the mutation BRCA1 c.441+1724_oNBR1:c.1073+480del (Figure 1). The 5′ and 3′ breakpoint flanking regions presented a complete homology sequence of 20 bp. Two Alu elements were found in the genomic sequences flanking the deletion breakpoints in BRCA1 intron 7 (AluSc) and NBR1 intron 11 (AluSg), and sequence alignment indicated that these two elements were highly homologous (Figure 2a).

Figure 1
figure 1

Scheme representing the molecular characterization of the BRCA1 c.441+1724_oNBR1:c.1073+480del and c.671-319_4677-578delinsAlu mutations. (a) Genomic region encompassing the NBR1 and BRCA1 genes, with the deleted regions and flanking Alu elements represented for both mutations. (b) BRCA1 c.441+1724_oNBR1:c.1073+480del mutation with the 5′ and 3′ breakpoints located in intron 7 and 11 of BRCA1 and NBR1, respectively. The deletion breakpoints are flanked by two AluSc/AluSg elements with a complete homology sequence of 20 bp within these Alu repeats. (c) Sequence electrophorogram of the BRCA1 c.441+1724_oNBR1:c.1073+480del breakpoint with the overlapping sequence highlighted. (d) BRCA1 c.671-319_4677-578delinsAlu mutation, with the breakpoints in introns 10 and 15 of the BRCA1 gene and presenting an Alu insertion. The 5′ breakpoint of the deletion was near an AluJb element, whereas the 3′ breakpoint occurred in an AluY element. (e) Sequence electrophorogram of the 5′ breakpoint region of the BRCA1 c.671-319_4677-578delinsAlu, showing a stretch of the poly-A inserted Alu element. A full color version of this figure is available at the Journal of Human Genetics journal online.

Figure 2
figure 2

Alignment of the Alu elements in the BRCA1 rearrangements. (a) Alignment of the Alu elements involved in the BRCA1 c.441+1724_oNBR1:c.1073+480del rearrangement, namely, an AluSc in BRCA1 intron 7 (INT7) and an AluSg in NBR1 intron 11 (INT11). (b) Alignment of the Alu elements involved in the BRCA1 c.671-319_4677-578delinsAlu rearrangement, an AluJb in BRCA1 intron 10 (INT10), an AluY in intron 15 (INT15) and the Alu element inserted (Alu insertion). Each dot stands for the same nucleotide and dashes indicate gaps introduced so as to maximize alignment.

Characterization of the BRCA1 exons 11–15 deletion

The breakpoint identification of the BRCA1 exons 11–15 deletion was based on the information obtained through the study of SNPs located in introns 10 (rs7503154) and 15 (rs8176214, rs77152443 and rs72130855) of the BRCA1 gene. Markers rs7503154 and rs72130855 were heterozygous, indicating that these regions were not deleted. Markers rs8176214 and rs77152443 were not heterozygous, possibly suggesting that they were located within the deleted region. After long-range PCR with the forward primer of marker rs7503154 and the reverse primer of marker rs72130855, we obtained a fragment of 1100-bp in the sample with the BRCA1 rearrangement. Sequence analysis of this PCR product revealed the breakpoint region in the mutated allele. The case harbored a 23 363-bp deletion, comprising exons 11–15 of the BRCA1 gene with the 5′ and 3′ breakpoints located 319-bp upstream of exon 11 and 578-bp upstream of exon 16 (Figure 1). In addition, at the deletion breakpoint, a sequence of 250-bp was inserted (Figure 1). Although the exact length could not be assigned because of the inherent difficulties in sequencing long poly-A tracts, this fragment contained 60-bp poly-A. The Censor Server and RepeatMasker software showed that this sequence was an Alu element belonging to the AluY family, which allowed us to describe this rearrangement as BRCA1 c.671-319_4677-578delinsAlu. The 5′ breakpoint of the deletion was 120-bp downstream from an Alu element (AluJb), whereas the 3′ breakpoint occurred in another Alu element (AluY). Sequence alignment of the three Alu elements supposedly involved in this rearrangement is represented in Figure 2b. The inserted Alu sequence presented a higher homology to the intron 15 breakpoint region of the BRCA1 gene.

Discussion

According to Sluiter and van Rensburg,4 81 and 17 different LGRs have been characterized for the BRCA1 and BRCA2 genes, respectively. One of the mechanisms causing major gene rearrangements in humans involves Alu elements, which are retrotransposable interspersed repetitive sequences that use the L1-encoded transposition machinery and constitute 10% of the human genome.14 Alu sequences comprise three major subfamilies, namely, the J (oldest) and S (intermediate) subfamilies that are responsible for the large majority of Alu copies currently present in the human genome, and the Y subfamily (youngest) that is responsible for all of the recently integrated Alu sequences in the human genome.15, 16 These repetitive sequences are responsible for creating genetic variation and also human diseases, not only because Alu sequences are known to mediate the occurrence of genetic recombination events as a direct result of their abundance and sequence homology, but also due to the occurrence of insertion mutations as a consequence of Alu retransposition events.17

We characterized two large deletions involving the BRCA1 gene, which account for about 4% (2/48) of all BRCA1 HBOC families identified in our population. The first rearrangement we here report harbored a deletion comprising exon 7 of the BRCA1 gene to exon 11 of the NBR1 gene (c.441+1724_oNBR1:c.1073+480del). Two highly homologous Alu elements were found in the genomic sequences flanking the deletion breakpoint, namely, an AluSc in BRCA1 intron 7 and an AluSg in NBR1 intron 11. The alignment of these intronic sequences suggested that the most likely mechanism for the occurrence of this rearrangement was nonallelic homologous recombination between these Alu repeats, substantiated by the presence of a 20-bp overlapping sequence at the breakpoint junction. We hypothesize that this sequence contains a recombinogenic hotspot as assumed for other sequences in Alu elements, as also postulated before.18, 19 Rüdiger et al.,19 suggested that four different rearrangements in the low density lipoprotein receptor gene presented a 26-bp common sequence that functions as a recombinational hotspot. This core sequence contains a pentanucleotide motif (CCAGC), which is also part of chi, an 8-bp sequence known to stimulate recBC-mediated recombination in Escherichia coli.19 We compared our 20-bp overlapping sequence with the sequences involved in the 81 BRCA1 rearrangements described by Sluiter and van Rensburg.4 This 20-bp sequence was observed (up to a maximum of three mismatches without affecting the pentanucleotide motif) at the breakpoint region in four rearrangements (NG_005905.2: g.114552_121641del, g.148011_151126del, g.153110_161767dup and g.166375_170153delins), supporting the possibility that these rearrangements were induced by a specific sequence with a recombinogenic nature (Figure 3).4, 19 However, this 20-bp sequence is highly conserved as it is present in the consensus sequence of the three major Alu subfamilies (J, S and Y), which alone can explain the high incidence of homologous recombination involving these sequences.15 A deletion of exons 1–7 of the BRCA1 gene was previously described by Engert et al.,20 but the genomic breakpoint was not determined. However, they confirmed by comparative genomic hybridization that the minimum size of the deletion was about 200 kb and the maximum size was about 285 kb encompassing the entire NBR1 region, so one can assume that the BRCA1 c.441+1724_oNBR1:c.1073+480del we here report is a novel genomic rearrangement. This BRCA1 LGR encompasses three more loci, namely, NBR1, ΨBRCA1 and NBR2. No recognizable biological function has been assigned to ΨBRCA1 and NBR2,21 whereas the NBR1 gene encodes a highly conserved multidomain scaffold protein involved in selective autophagy.22 The family here reported does not show any remarkable phenotypic trait apart from being predisposed to breast and ovarian cancers, indicating that the relevant genetic defect is the inactivation of the BRCA1 gene. A contiguous gene deletion syndrome was also excluded by Gad et al.,23 in one family presenting a BRCA1 deletion of exons 1–22 that also encompassed the NBR1, ΨBRCA1 and NBR2 genes. The absence of mutations in the NBR1 and NBR2 genes in breast and ovarian cancer patients also indicates that BRCA1 inactivation is the relevant consequence of this LGR.24, 25

Figure 3
figure 3

Flanking sequences of four different BRCA1 gene rearrangements presenting the 20-bp overlapping sequence (gray colored boxes) near or at the breakpoint region, which have been previously reported. The genomic breakpoints (underlined) were obtained from Sluiter and van Rensburg using the RefSeqGene record, NG_005905.2.4

The second rearrangement fully characterized at the nucleotide level was the BRCA1 mutation c.671-319_4677-578delinsAlu, previously described as EX11_15del by our group.11 The 3′ breakpoint occurred in an AluY element and the 5′ breakpoint was 120 bp downstream from a second Alu element (AluJb) in the BRCA1 gene. In addition, an 250-bp Alu fragment was inserted at the breakpoints of the deleted region. This Alu fragment presents a high homology to the intron 15 breakpoint region of the BRCA1 gene and, according to the Censor Server and RepeatMasker software, this sequence is an Alu element belonging to the AluY family. Alu-mediated unequal homologous recombination is the most common mechanism causing BRCA1 LGRs. The deletion/insertion event here described could be due to homologous recombination with exchange of genetic material.26 However, another possible explanation for the observed rearrangement could involve an insertion-mediated deletion mechanism caused by the retrotransposition of Alu elements at this genomic site leading to the deletion of adjacent genomic regions.26, 27, 28, 29 This hypothesis is corroborated by the fact that the inserted Alu element belongs to a still active AluY family. There are only a few reports of large deletions caused by the retrotransposition of an Alu element: for instance, a complex deletion–insertion rearrangement has been reported in a patient with lipoprotein lipase deficiency disorder caused by an Alu retrotransposition-mediated mechanism29 and the TP53 gene has been shown to be rearranged by a complex rearrangement resulting in deletion of exons 2–4 with an insertion of a truncated Alu fragment.26 Only eight different deletion–insertion rearrangements have been reported in BRCA1, but none attributable to an Alu-mediated retrotransposition mechanism,4 so this would be the first description of such a mechanism of mutation of the BRCA1 gene. The deletion of exons 11–15 of the BRCA1 gene was also described by de la Hoya et al.;30 however, the mutational mechanism and the genomic breakpoint (L78833:g.33450_56562del) are different from those we here report, emphasizing the importance of breakpoint characterization to determine whether rearrangements involving the same exons are related or have occurred de novo.

In conclusion, we have identified the breakpoints of two novel large deletions involving the BRCA1 gene, one encompassing exons 1–7 in addition to the two adjacent genes NBR2 and NBR1 (c.441+1724_oNBR1:c.1073+480del) and another comprising exons 11–15 (c.671-319_4677-578delinsAlu). Both rearrangements most likely were mediated by Alu elements, either by homologous recombination between Alu repeats or by the retrotransposition of an Alu element.