Study of the genetic variability in a Parkinson's Disease gene: EIF4G1

Highlights ► Mutations in the EIF4G1 gene have been found in patients with Parkinson's Disease (PD) [2]. ► We investigate the mutations in a familial PD cohort and in sub-Saharian Africans and Europeans. ► We failed to identify any mutation in the PD cohort. We identify the p.A502V (previously associated to PD) in Caucasians. ► Mutations in the EIF4G1 gene are a very rare cause of PD in European individuals. ► The protein can tolerate some extent of genetic variability.


a b s t r a c t
Chartier-Harlin and colleagues [2] recently reported mutations in the eukaryotic translation initiation factor 4-gamma (EIF4G1) gene in families with parkinsonism. Large-scale screening found two mutations (p.R1205H and p.A502V) only in affected individuals, although their relative frequency was very low. The aim of this study was to investigate EIF4G1 parkinsonism-related variants in two separate cohorts and study coding variability across the gene. We first screened a series of familial Parkinson's Disease (PD) patients in an attempt to confirm previous results by showing segregation. Then, to determine the extent of coding variation in the gene, we first screened a cohort of sub-Saharan African individuals from the Centre d'Etude du Polymorphisme Humain -Human Genome Diversity Cell Line Panel (HGDP) [1] and then analyzed data from 5350 individuals National Heart, Lung, and Blood Institute (NHLBI) exome sequencing project. We failed to identify any PD-related mutations in the familial samples. Conversely we found the p.A502V variant in the NHLBI population. We observed a high number of coding polymorphism in the exons where the two PD variants have been previously reported. We conclude that either EIF4G1 variants are an extremely rare cause of familial PD in Caucasian cohorts, or that A502V is in fact a rare benign variant not involved in PD aetiology. Our data also suggests that the protein can tolerate some extent of variability particularly at this point of the gene.
Recently, a new gene (EIF4G1) has been identified in one family with autosomal dominant late-onset Parkinson's Disease [2]. Linkage was ascribed to a region at chromosome 3q26-28 containing approximately 159 genes. Sequence analysis found only one novel coding variant (p.R1205H in the EIF4G1 gene) which segregated with disease, which was absent in 4050 controls and which was evolutionary conserved in mammals. Screening a cohort of about 4800 PD cases (familial and sporadic) identified nine additional patients of the p.R1205H mutation. Further molecular analysis of the EIF4G1 gene in a large case-control cohort (4500 cases and 3800 controls) identified another novel missense mutation (p.A502V) in three PD individuals, which was not found in controls. These data indicate that these variants are extremely rare in the PD population (0.2% for p.R1205H and 0.06% for p.A502V). Assignment of pathogenicity can be difficult when variants are very rare. With this background, we screened 150 familial PD cases from our UK familial Parkinson's Disease series, in which we have previously identified LRRK2, VPS35 and SNCA mutations [5,7] in order to determine whether we could provide further that this gene is indeed a PDrelated locus. We also assessed these coding positions in a set of African samples (Table 1) from the Human Diversity seriesa standard panel of African samples [1], as African samples have the greatest diversity and offer a rapid route to the identification of benign polymorphisms [4]. Briefly, exon 8 and exon 22 of the EIF4G1 gene (NM 182917.3) were PCR amplified and sequenced in the two cohorts for a total of 114 African samples and 150 familial PD cases.
To obtain a more exhaustive description of the pattern of variability in that gene we also extracted genotype data from the NHLBI exome sequencing project (Exome Variant Server, NHLBI Exome Sequencing Project (ESP), Seattle, WA (http://evs.gs.washington.edu/EVS/) [January 2011], which includes exome data for 3500 American individuals of European descent and 1850 African American. Frequencies were computed using VCFtools. We used the software ANNOVAR [8] to annotate the function of the variants.
We failed to identify any mutation previously reported to be associated with PD our familial cohort, but we identified one coding change (P486S) in two PD individuals. The P486S variant is reported in dbSNP (rs112545306). Interestingly it has been observed in African-Americans (http://snp.gs.washington.edu/EVS), with a frequency of 0.15%.  We identified six non-synonymous changes in exon 8 and in exon 22 in the African individuals. Of these, one is a novel change (P382L), the others are variants recently reported in dbSNP and found mainly in African populations (http://snp.gs.washington.edu/EVS). To predict the impact on protein function of these non-synonymous variants, we performed an in silico analysis using the software PolyPhen and SIFT [6] and all were predicted to be benign ( Table 2).
Analysis of the NHLBI samples allowed us to detect the A502V variant in two European-American individuals (frequency of 0.02%). We identified 95 nonsynonymous SNP over 32 exons   These data, combined with the presence of the A502V in the NHLBI population with 0.02% frequency and our failure to identify any PD mutation carrier in our familial cohort, are consistent with the interpretation that either EIF4G1 variants are an extremely rare cause of familial PD in Caucasian cohorts, or that A502V is in fact a rare benign variant not involved in PD aetiology.

Role of the funding source
This study was supported by Parkinson's Disease Society (AT) and by the Medical Research Council and Wellcome Trust Disease Centre (grant WT089698/Z/09/Z). DNA extraction work was undertaken at University College London Hospitals, University College London, who received a proportion of funding from the Department of Health's National Institute for Health Research Biomedical Research Centres funding.