Introduction

Alport syndrome (AS) is the most common hereditary nephropathy, characterized by a progressive renal failure, sensorineural deafness and ocular abnormalities. The most common X-linked form of this disease (XLAS, OMIM: 301050) results from mutations in the COL4A5 gene, which encodes the α5 chain of type IV collagen.1 Alport syndrome-diffuse leiomyomatosis (AS-DL, OMIM: 308940) occurs as a rare variant of XLAS that shows overgrowth of visceral smooth muscles in the gastrointestinal, respiratory and female reproductive tracts, in addition to renal symptoms.1, 2, 3 COL4A5 is located on the long arm of the X chromosome (Xq22) and is head-to-head with the COL4A6 gene. The COL4A6 gene encodes the α6 chain of type IV collagen, which is mainly expressed in heart, human esophagus, aorta and bladder smooth muscle basement membrane.3, 4, 5

AS-DL patients exhibit contiguous gene deletions at the COL4A5–COL4A6 locus.5 Sixteen AS-DL patients reported so far have been found to have a deletion that encompassed the 5′ end of the COL4A5 and COL4A6 genes, and included the bidirectional promoter (the Human Genome Mutation Database). The COL4A6 deletion breakpoints have been consistently found within intron 2, whereas the COL4A5 breakpoints usually occur in intron 1.6 However, the breakpoints were characterized at a single-nucleotide level only in four cases.6, 7, 8, 9

Patients and methods

In our laboratory, we conduct a comprehensive molecular diagnostics of inherited kidney diseases, including AS. So far, 415 suspected AS patients underwent molecular genetic tests, five of them with clinical signs of leiomyomatosis. The pedigrees of the five cases are shown in Supplementary Figure 1.

Genomic DNA was isolated from peripheral blood leukocytes. Screening of contiguous gene deletions was performed with the multiplex ligation probe amplification (MLPA) using the SALSA P191/192 Alport MLPA assay (MRC-Holland, Amsterdam, the Netherlands).2, 10, 11 Long-range PCR amplification and direct sequencing of COL4A5 and COL4A6 was conducted to characterize the deletion breakpoints. The human reference sequence of NG_011977.1 and NG_012059.2 for COL4A5 and COL4A6 was used, respectively.

To detect a large heterozygous deletion in Case 2, we conducted a semi-quantitative PCR amplification using capillary electrophoresis.12, 13

Results

The MLPA analysis of five patients with AS-DL revealed hetero- or hemizygous deletions in each case (Supplementary Figure 2). This analysis was followed by long-range PCRs and direct sequencing to identify each breakpoint (Figures 1a–e, Supplementary Figure 3 and Supplementary Data 1). For a female patient (Case 2, Supplementary Figure 2B), we also conducted a semi-quantitative PCR assay (Supplementary Figure 4).

Figure 1
figure 1

Sequence chromatograms of contiguous gene deletion breakpoints in five AS-DL cases. (a) Case 1 (c.66+5840 of COL4A6 and c.81+8068 of COL4A5). (b) Case 2 (c.66+25107 of COL4A6 and c.81+18040 of COL4A5). (c) Case 3 (c.66+85676 of COL4A6 and c.276+3257 of COL4A5). (d) Case 4 (c.66+119476 of COL4A6 and c.3246+6706 of COL4A5). (e) Case 5 (c.66+84055 of COL4A6 and c.3246+66915 of COL4A5). COL4A6 and COL4A5 sequences are shown as black and open rectangles, respectively. Homologous sequence in each case is shown in red. A full color version of this figure is available at the Journal of Human Genetics journal online.

Full size image

Deletions are schematically shown in Supplementary Figure 3.6, 7, 8, 9 To examine the genomic context of deletion breakpoints, we analyzed their flanking sequences using RepeatMasker (Figure 2). Their alignments with Repbase entries15 revealed the presence of transposed elements (TEs) in 13/18 (72%) breakpoints in nine cases (Figure 2). Cases p2 and p3 showed COL4A5-side breakpoints at different positions within the same long interspersed element, family L1. Similarly, Cases c3 and p2 exhibited COL4A6-side breakpoints at different positions in the same L1 copy. Homologous sequences at the recombination breakpoints were apparent in eight of nine patients. The length of overlapping sequences varied from 2 to 42 bp (Supplementary Data 1). No breakpoints were located in the annotated duplicated segment (UCSC genome browser) of COL4A6 intron 2 (Figure 2).

Figure 2
figure 2

Schematics of novel and previously reported deletions Location of both sides of the breakpoints in the in COL4A5 and COL4A6. Deletions are shown as dark rectangles. Transposable elements in cenetromeric (COL4A6) and telomeric (COL4A5) breakpoints are shown to the right. The segmental duplication in intron 2 is marked as a vertical rectangle at the top of the figure. Exons are numbered at the top. c1−c5, our case 1−5, p1 (ref. 4), p2 (ref. 14), p3 (ref. 15), p4 (ref. 13). The number of homologous nucleotides (Figure 1 and Supplementary Data 1) is shown in parentheses.

Full size image

MLPA analysis of maternal samples of Cases 1, 3 and 4 failed to detect any deletion, suggesting that, unlike Cases 2 and 5, these are sporadic cases.

Discussion

This work has more than doubled the number of sequence-characterized COL4A5 and COL4A6 breakpoints in AS-DL. Although eight out of nine cases had homologous sequences at the deletion breakpoints, only two cases (c3 and p2) showed relatively long homologous sequences in the L1 family of long interspersed elements, consistent with a well-known L1-mediated recombination mechanism.16 Almost a half of the human genome is occupied by recognizable TEs, with L1 occupying ~17%.17 TEs have been shown to provide a source of new exons, genes and regulatory sequences, dramatically influencing evolutionary history, exon–intron structure, speciation and regulation of gene expression. TEs facilitate non-allelic homologous recombination events leading to inherited diseases,14 including AS-DL (Figure 2).8 As compared to c3 and p2, Cases c1, c4, c5, p1 and p3 revealed shorter homologous sequences indicative of the same mechanism. The number of breakpoints in TEs 13/18 (72%) appears to be higher than expected since TEs represent only 40% and 62% of the genomic sequences of COL4A6 and COL4A5, respectively suggesting that TEs might have some roles for causing recombination in this disease.

All deletion breakpoints characterized at the single-nucleotide level in COL4A6 took place in intron 2.6, 7, 8, 9 Recently, an AS-DL case was identified with a deletion extending into COL4A6 beyond this intron,2 but the precise breakpoint was not characterized. On the COL4A5 side, the deletion breakpoint usually maps to intron 1; however, some cases, including two of ours (c3 and c4), showed breakpoints in intron 36 and intron 4, respectively. Recently, Sa et al. reported an AS-DL case with a COL4A5-only gene deletion, which encompassed exons 1−51 but did not include the common promoter region or exon 1 of COL4A6.18 This report may still be compatible with the requirement for inactivation of both genes in AS-DL, as first proposed by Zhou et al.,5 because the authors did not analyze COL4A6 expression nor did they exclude inactivation of this gene by other mechanisms. Recent studies in yeasts revealed that deletion of many genes were associated with altered mRNA levels of the neighboring genes.19 These studies are particularly relevant for bidirectional promoters which generate products of two adjacent, often related genes in stoichiometric quantities, ensuring their co-expression in the same or similar biological pathway.

In conclusion, we have more than doubled the number of large contiguous deletions in AS-DL characterized at the single-nucleotide level. Our results show that most deletions were mediated by transposons via homologous recombination events and support the original proposal5 that inactivation of both genes is required for the development of leiomyomas in AS-DL.