Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
Editor—Neurofibromatosis type 1 (NF1) is one of the most common autosomal dominant disorders with a population frequency of 1 in 3500.1 The disease is clinically characterised by multiple neurofibromas, café au lait spots and Lisch nodules of the iris. The NF1 gene, a tumour suppressor gene, resides on the proximal long arm of chromosome 17 (17q11.2). It spans approximately 350 kb of genomic DNA and, comprising 60 exons, encodes the protein neurofibromin.2 This protein, consisting of 2818 amino acids, contains a central domain that has homology with GTPase activating proteins (GAPs).3
A distinct feature of the NF1 gene is the very high spontaneous mutation rate (1 × 10-4 per gamete per generation), which is about 100-fold higher than the usual mutation rate for a single locus.1 Up to 50% of all NF1 cases are thought to result from de novo mutations. TheNF1 gene provides a large target for mutations because of its relatively large size, but this may only account for a factor of 10 in terms of increase in mutation rate.4 The presence of numerousNF1 pseudogenes has been proposed as an explanation for the high mutation rate in NF1.5 In the human genome, at least 12 different NF1related sequences have been identified on chromosomes 2, 12, 14, 15, 18, 21, and 22.5-13 Most of theNF1 pseudogenes have been mapped in pericentromeric regions. The chromosome 2NF1 pseudogene has been localised to region 2q21, which is known to contain the remnant of an ancestral centromere.14 Owing to the absence of selective pressure, mutations may accumulate in the NF1pseudogenes. Consequently, the pseudogenes could act as reservoirs of mutations, which might be crossed into the functionalNF1 gene by interchromosomal gene conversion.5 Gene conversion, the non-reciprocal transfer of genetic information between two related sequences, has been recognised as a mutational mechanism for several human genes.15-17 In all these cases, the conversions occurred between gene and pseudogene on the same chromosome. For NF1, interchromosomal gene conversion is required as none of theNF1 pseudogenes is located on chromosome 17. Interchromosomal gene conversion has been reported to occur between the von Willebrand factor gene on chromosome 12 and the von Willebrand pseudogene on chromosome 22.18
Gene conversion requires close contact between the functional gene and the corresponding pseudogene. The pericentromeric location of the functional NF1 gene and its pseudogenes may enable this close contact since centromeres tend to associate with each other in a non-random fashion.19 20 This is underlined by our finding that the NF1 pseudogenes on chromosomes 2, 14, and 22 have arisen by repeated transposition events between (peri)centromeric locations on these chromosomes (Luijtenet al, submitted).13 However, the high mutation rate in NF1 cannot be explained exclusively by interchromosomal gene conversion. Only a small part of the functionalNF1 gene is represented in theNF1 pseudogenes (see below), whileNF1 gene mutations are scattered over the entire gene. In this study, we investigated whether interchromosomal gene conversion contributes to the mutation rate in NF1.
First, we inventoried all available NF1 pseudogene sequences (table 1). These included not only the publishedNF1 pseudogene sequences, but also so far unidentified NF1 pseudogene sequences present in the first draft sequence of the human genome. The latter were detected by performing a BLAST search using the complete cDNA sequence of the NF1 gene (http://www.nf.org/nf1gene/nf1gene.cDNAtext.html). In a previous study, we elucidated the complete nucleotide sequence of theNF1 pseudogene on chromosome 22.13 Analysis of this sequence showed that sequences homologous to exons 10b, 12a-19a, and 27b are present in the chromosome 22 NF1 pseudogene. The same exons are represented in the NF1 pseudogenes on chromosomes 2 and 14. Chromosome 14 contains severalNF1 pseudogenes. Sequence analyses of all exons present in four chromosome 14 NF1pseudogene variants and in the one on chromosome 2 have been performed.13 The BLAST search yielded two clones each containing an additional pseudogene variant on chromosome 14. In one of them, sequences corresponding to exons 10b, 12a-19a, and 27b are present, while in the other only exons 12b-19a and 27b are represented. The chromosome 2 NF1 pseudogene was also found among the results from the BLAST search. In theNF1 pseudogene present in this particular clone, exons 12b-19a, and 27b are represented.
The complete sequence of the chromosome 21NF1 pseudogene has been determined (Weisset al, unpublished data).7 In this pseudogene, exons 7-9 and 11 are represented. For chromosome 15, three different NF1 pseudogenes have been reported.6 11 12 A fourth locus may be present on chromosome 15, but this could not be substantiated.11 The representation of the chromosome 15 NF1pseudogenes starts in intron 12b and ends in intron 27b. These pseudogenes are the only NF1 pseudogenes known so far with a nearly complete representation of the GAP related domain, which is encoded by exons 20-27a. Three chromosome 15 clones were found as a result of the BLAST search. Only one of them contains sequences corresponding to exons 24-27b, whereas in both the other two clones exons 13-23-1 and 24-27b are represented. The latter two clones contain the same NF1 pseudogene variant, but differ from the NF1 pseudogene present in the former. The BLAST search also yielded two unmapped clones, in which exons 13-23-1 and 24-27b are represented. TheNF1 pseudogene present in one of these unmapped clones is, with the exception of two nucleotides, identical to the NF1 pseudogene variant present in the two clones that have been mapped to chromosome 15. Considering this and the fact that in both the unmapped clones exons 20-23-1 and 24-27b are represented, we presume these NF1 related sequences to originate from chromosome 15. A number of exons of the published chromosome 15 NF1 pseudogene variants have been sequenced. None of the three pseudogene variants that were identified in the BLAST search is identical to the sequences available from the published chromosome 15NF1 pseudogene variants. The chromosome 12NF1 pseudogene maximally contains exons 12b-23-1,12 of which only exon 16 has been partially sequenced.8 One of the NF1pseudogenes that resulted from the BLAST search contains sequences corresponding to exons 16 and 17. The clone containing this pseudogene was derived from chromosome 6. However, alignment of exon 16 of this pseudogene to other NF1 pseudogenes showed a difference of only one nucleotide with the exon 16 sequence of the chromosome 12 NF1 pseudogene. Since no indications for a chromosome 6 NF1pseudogene exist so far, we assume this NF1related sequence to originate from chromosome 12 rather than from chromosome 6. The NF1 pseudogene on chromosome 18 consists at the most of exons 7-13.12Sequences of exons 8 and 9 have been reported.8 The BLAST search yielded an NF1 pseudogene, mapped to chromosome 18p11.2, in which exons 7-9 and 11 are represented. An overview of the exons represented in the variousNF1 pseudogenes is given in fig 1 and all accessible NF1 pseudogene sequences are listed in table 1. When compared to the functionalNF1 gene, substitutions, deletions, and insertions are present in all NF1 related sequences.
We continued by collecting all point mutations and/or minor lesions in the NF1 gene. In total, we found 225 different disease causing mutations in exons 7-9, 10b, 11-23-1, and 24-27b of the functional gene, the four regions known to be represented in the pseudogenes. These mutations included 30 novel disease causing mutations that were taken from our own data (Fahsoldet al, unpublished data; Nürnberget al, unpublished data). Only a very limited number (13 of 225, 5.8%) of these disease causing mutations appears to have a pseudogene equivalent (table 2). In every pseudogene involved, we detected within a short distance (ranging from 2 to 25 nucleotides) from the disease causing mutation at least one extra mutation compared with the functionalNF1 gene (table 2). If the 13 mutations had been generated by interchromosomal gene conversion, it is likely that these extra mutations would also have been transposed into the functional gene of the NF1 patients. However, none of them has been identified in the NF1 gene of these NF1 patients. Of the 13 mutations found, six reside in CpG sites resulting in C to T transitions. About 32% of all single base pair substitutions causing human genetic disease occur in CpG dinucleotides.21 As CpGs of theNF1 coding region are subject to methylation,22 these six C to T transitions are probably the result of spontaneous deamination of 5-methylcytosine rather than of interchromosomal gene conversion. This susceptibility for C to T transitions is substantiated by the fact that the six mutations found in both the functional gene and pseudogenes include four recurrent mutations of the NF1 gene (indicated in table 2). Taken together, these results imply that the contribution of interchromosomal gene conversion to the high mutation rate in NF1 is, at best, limited.