Identification of FOXP2 truncation as a novel cause of developmental speech and language deficits

Abstract

FOXP2, the first gene to have been implicated in a developmental communication disorder, offers a unique entry point into neuromolecular mechanisms influencing human speech and language acquisition. In multiple members of the well-studied KE family, a heterozygous missense mutation in FOXP2 causes problems in sequencing muscle movements required for articulating speech (developmental verbal dyspraxia), accompanied by wider deficits in linguistic and grammatical processing. Chromosomal rearrangements involving this locus have also been identified. Analyses of FOXP2 coding sequence in typical forms of specific language impairment (SLI), autism, and dyslexia have not uncovered any etiological variants. However, no previous study has performed mutation screening of children with a primary diagnosis of verbal dyspraxia, the most overt feature of the disorder in affected members of the KE family. Here, we report investigations of the entire coding region of FOXP2, including alternatively spliced exons, in 49 probands affected with verbal dyspraxia. We detected variants that alter FOXP2 protein sequence in three probands. One such variant is a heterozygous nonsense mutation that yields a dramatically truncated protein product and cosegregates with speech and language difficulties in the proband, his affected sibling, and their mother. Our discovery of the first nonsense mutation in FOXP2 now opens the door for detailed investigations of neurodevelopment in people carrying different etiological variants of the gene. This endeavor will be crucial for gaining insight into the role of FOXP2 in human cognition.

Figure 1

Schematic of the human FOXP2 locus, which spans >600 kb of genomic DNA, showing sequence variants identified in subjects with verbal dyspraxia. Black shading indicates translated exons; “atg” and “tga” denote positions of initiation and termination codons. Known domains encoded by exons include polyglutamine tracts (Q₄₀ and Q₁₀), a zinc-finger motif (ZnF), a leucine zipper (LeuZ), the forkhead domain (FOX), and an acidic C-terminus (Acidic). Exons 3b and 4a are alternatively spliced coding exons yielding amino acid insertions, whereas alternatively spliced exons 2a, 2b, and 3a are predicted to be noncoding. Exons s1–s3 and 1 represent alternative 5′ UTR regions that have not been found in the same human transcript; the position of the 5′ end of exon 1 is based on currently available EST data. For more information on splicing and isoforms, see Lai et al. (2001) and Bruce and Margolis (2002). All known exons were screened for mutations, with the exception of two noncoding exons: s1 (5′ CpG-rich UTR) and 2a (alternatively spliced and untranslated). Coding variants are shown above the locus, with resulting codon and amino acid changes indicated. For reference, the KE mutation is also included. Noncoding variants are shown below the locus, with information regarding relative position (with respect to the exon in question) and frequency in probands (number of heterozygous probands/total probands screened). An asterisk (*) denotes intronic variants that correspond to those previously detected by Newbury et al. (2002), and rs numbers are indicated for polymorphisms also present in dbSNP. Locus schematic is adapted from Fisher et al. (2003).

Figure 2

A, Amino acid sequence of main isoform of human FOXP2, showing all coding changes found thus far in individuals with developmental verbal dyspraxia (Lai et al. ; present study). Each change was absent from large numbers of control chromosomes, but only the R328X and R553H mutations are found in multiple affected members in families segregating developmental speech and language disorder. B, Nucleotide sequence of normal and expanded polyglutamine-encoding region identified in one of the probands in this study.