Background: Our previous analysis demonstrated that two intronic SNPs (rs1861972 and rs1861973) in the homeobox transcription factor, ENGRAILED 2 (EN2), are significantly associated both individually and as a haplotype with Autism Spectrum Disorder (ASD). This was observed in three separate datasets indicating that EN2 is a likely ASD susceptibility gene (Gharani et al., 2004; Benayed et al., 2005). Objectives: The goal of our study was two-fold: to identify candidate risk alleles by extending our linkage disequilibrium (LD) map and testing these candidates for function by standard molecular genetic approaches. Methods: To extend our LD map additional EN2 polymorphisms were typed and publicly available Hapmap data analyzed. Candidate risk alleles were then tested for functional differences by luciferase assays and EMSAs. Results: The LD mapping determined that the associated SNPs are not in high LD (r2<.370) with any typed polymorphisms, identifying rs1861972 and rs1861973 as candidate risk alleles. The EN2 intron was then tested for transcriptional activity by transfecting appropriate luciferase constructs in three different cell types: primary mouse cerebellar neurons, a neuronal cell line (PC12) and a non-neuronal cell line (HEK293T). This analysis demonstrated that the EN2 intron acts as a transcriptional repressor in all three cell types (P<.00005; two tailed paired Student's T test). To test for functional differences between haplotypes, both versions of the intron were cloned into luciferase constructs with two different promoters and transfected into the same three cell types. In all cases the autism-associated alleles result in weaker repressor activity (P<.005). EMSAs were performed for both SNPs individually and uncovered differential binding of proteins to both associated alleles, providing a molecular mechanism for the transcriptional difference. . Conclusions: These studies present additional evidence that the associated EN2 haplotype is functional and a likely common genetic risk factor for ASD.