Minor lesion mutational spectrum of the entire NF1 gene does not explain its high mutability but points to a functional domain upstream of the GAP-related domain

Abstract

More than 500 unrelated patients with neurofibromatosis type 1 (NF1) were screened for mutations in the NF1 gene. For each patient, the whole coding sequence and all splice sites were studied for aberrations, either by the protein truncation test (PTT), temperature-gradient gel electrophoresis (TGGE) of genomic PCR products, or, most often, by direct genomic sequencing (DGS) of all individual exons. A total of 301 sequence variants, including 278 bona fide pathogenic mutations, were identified. As many as 216 or 183 of the genuine mutations, comprising 179 or 161 different ones, can be considered novel when compared to the recent findings of Upadhyaya and Cooper, or to the NNFF mutation database. Mutation-detection efficiencies of the various screening methods were similar: 47.1% for PTT, 53.7% for TGGE, and 54.9% for DGS. Some 224 mutations (80.2%) yielded directly or indirectly premature termination codons. These mutations showed even distribution over the whole gene from exon 1 to exon 47. Of all sequence variants determined in our study, <20% represent C-->T or G-->A transitions within a CpG dinucleotide, and only six different mutations also occur in NF1 pseudogenes, with five being typical C-->T transitions in a CpG. Thus, neither frequent deamination of 5-methylcytosines nor interchromosomal gene conversion may account for the high mutation rate of the NF1 gene. As opposed to the truncating mutations, the 28 (10.1%) missense or single-amino-acid-deletion mutations identified clustered in two distinct regions, the GAP-related domain (GRD) and an upstream gene segment comprising exons 11-17. The latter forms a so-called cysteine/serine-rich domain with three cysteine pairs suggestive of ATP binding, as well as three potential cAMP-dependent protein kinase (PKA) recognition sites obviously phosphorylated by PKA. Coincidence of mutated amino acids and those conserved between human and Drosophila strongly suggest significant functional relevance of this region, with major roles played by exons 12a and 15 and part of exon 16.

Figure 1

Histogram of number of mutations, exon by exon. Black columns represent the 278 pathogenic mutations of table 4. White columns represent the mutations reported to the NF1 Genetic Analysis Consortium as of November 1997 (Korf 1998).

Figure 2

Weighed distribution of mutations over the NF1 gene. For each exon, the number of pathogenic mutations was divided by the number of base pairs (bp). Ten bp were added to each exon, to allow for splice-site mutations. Values shown are ratios between the exon-specific mutation densities and the average mutation density for the whole gene (278/9,204 bp). A, linear presentation. B, logarithmic presentation.

Figure 3

Weighed distribution of missense mutations over the NF1 gene. For each exon, the number of genuine missense or single-amino-acid-deletion mutations was divided by the number of base pairs (bp). Values shown are ratios between the exon-specific mutation densities and the average mutation density for the whole gene (28/8457 bp). Location of the GRD is indicated by the box.

Figure 4

Missense mutations in the similarity plot of human versus Drosophila NF1 protein. The program PLOTSIMILARITY of the GCG-Wisconsin package, version 9 was used to calculate the amino-acid-sequence similarity profile of the human and Drosophila neurofibromins (accession numbers AAA59925 and AAB58976, respectively). A sliding window of 40 amino acids was chosen. Note that there are 2,963 instead of 2,839 amino acid positions in the alignment (program PILEUP; gap weight = 12 and gap-length weight = 4). For the first 20 and the last 20 positions, a value could not be calculated, because of the chosen window size. Mutations are located primarily in regions of higher similarity. Usually, helices and beta sheets have a higher local sequence similarity than coils.

Figure 5

Synopsis of all known NF1 missense mutations and deletions of 1 or 2 amino acids. Mutations on the left and right sides are data from this study and from the literature, respectively. Most published mutations have been reviewed in Upadhyaya and Cooper (1998). Some have been published very recently: R1204G (Krkljus et al. 1997); C39Y(1), Y489C, L847P, and 991delM(1) (Messiaen et al. 1998); and L2317P (Wu et al. 1999). Others have been submitted to the consortium but not published at this time. They are reported here with the consent of the consortium members who identified these mutations: 2366delNF(2) and R2616Q (L. Messiaen, personal communication); R1849G and L1932P (M. Upadhyaya, personal communication); 1658delIY and P2046R (D. Vidaud, personal communication). Mutations not representing genuine missense mutations including polymorphisms, splice mutations, or loss of the initiation codon are printed in italics. Number in parentheses after the mutation symbol indicates recurrent identifications. GRD region is shown in gray, whereas the exons 11–17, supposed to code for a new functional domain, are shown in black with a central light effect. To facilitate orientation, the numbers of some of the larger exons are given.