(a) Genomic structure of the PRODH2 (Top and Bottom) and PRODH2-P locus (Middle); colored boxes indicate predicted coding exons; brackets indicate a large deletion that removes predicted exons 2, 3, 4, 5, and part of 6 from PRODH2-P. Introns are not to scale. The size of the PRODH2 exons (in nucleotides), as well as the number of exonic differences between PRODH2-P and PRODH2, is indicated. Black vertical lines indicate the approximate distribution of these differences. PRODH2 exons 7, 9, 10, 11, 13, and 14 were assayed for presence of a subset of these differences (n = 12; see Table 3) in the AS and COS patient samples. Red vertical lines (Bottom) indicate pseudogene-like substitutions detected in both samples. (b Top) Cross-species comparison of the Homo sapiens PRODH2 exon 11 predicted protein sequence by using the program clustal-w (as implemented in MacVector, Accelrys, San Diego). Boxed and shaded areas highlight identities or similarities. Bottom row indicates the predicted effect of pseudogene-like base changes on the exon 11 encoded protein sequence. * and † indicate the codons where the synonymous SNP PRODH2*1482 and the nonsynonymous SNP PRODH2*1496 reside, respectively. (Bottom) Exon 11 haplotypes determined by sequencing the exon from affected carriers and their parents. Seven distinct haplotypes were identified carrying single or clusters of multiple adjacent pseudogene-like substitutions. Filled circles indicate codons carrying pseudogene-like, synonymous (n = 2), and nonsynonymous (n = 6) base-pair substitutions. Solid lines indicate regions of identity to the PRODH2 reference sequence, shared by all haplotypes. COS patients from all ethnic backgrounds (21 European Americans, 12 African Americans, 8 of other ethnic background) were included in this analysis, whereas only European and African Americans were included in the analysis described in Table 1. (c) Southern blot hybridization on DNA isolated from hamster–human somatic cell hybrids retaining only one of the two chromosome 22 copies from a heterozygous carrier of haplotype D (m, maternal copy; p, paternal copy). The Arg-453→Cys substitution, present in the PRODH2-P and the paternal copy of PRODH2 in this patient, is detected by a BanII RFLP (BanII site absent). Genomic DNA was digested with BanII restriction endonuclease, and a 1-kb intronic fragment, indicated by a solid line, was used as a probe. (M, 1-kb ladder DNA marker; B, BanII). (d) Distribution of plasma proline level of three adult female carriers of exon 11 haplotypes A, B, and E and five sex- and age-matched noncarriers controls. Blood was drawn from the subjects in the morning, after overnight fasting. Proline levels were determined by the same qualified diagnostic laboratory (Mayo Clinic, Rochester, MN). Horizontal lines denote mean values and vertical lines, standard errors.

(a) Genomic structure of the PRODH2/DGCR6 locus. The approximate distribution of the identified polymorphisms in both genes is indicated (Top). Polymorphisms demonstrating significant overtransmission in the AS sample are indicated by arrowheads. Arrows indicate direction of transcription. The black horizontal bar signifies the PRODH2*1766/1945/2026 SNP combination that defines the risk haplotype presented in d. The approximate size of the locus is also indicated. CEN, centromere. (b and c) Linkage disequilibrium results in the AS and COS (b), as well as stratified AS families (c). In a total of 133 families, transmission to the affected probands of alleles and haplotypes defined by the PRODH2*1482, *1766, *1945, *2026, and DGCR6*959 SNPs deviates significantly from expectation under the null hypothesis and provides substantial support for a contribution of the PRODH2/DGCR6 locus variation to schizophrenia susceptibility. By contrast, no such association was detected in a sample of 80 North American families not afflicted with schizophrenia (not shown), arguing against a generalized transmission distortion at this locus. Color coding reflects the three stages of our analysis (see text). Markers are listed in order from centromere (Top) to telomere (Bottom). SNPs located in expressed sequences are referenced according to their position in corresponding cDNAs (accession nos. are available in the Appendix). Intronic SNPs are referenced according to their position in the University of California, Santa Cruz public database (October 7, 2000, freeze; http://genome.ucsc.edu). Polymorphism PRODH2*758 is not included in the analysis, because it was in complete LD with PRODH2*757 (see Fig. 3, which is published as supporting information on the PNAS web site). ND, the polymorphism was not typed in this group. Significant associations are shaded. Arrowhead indicates the relative position of microsatellite marker D22S1638. (d) Estimated risk (with 95% CI) for PRODH2*1766/1945/2026 2–2-1 haplotype relative to the other combinations. For each study, the circle represents the estimated risk, and the line indicates the extent of the 95% CI around this estimate. Data shown for odds ratios are based on haplotype counts (for the case-control study) and the transmission (HHRR) ratios in trios. AS(ChS), subset of AS patients with early childhood deviant behaviors (12); AS ≤ 18, subset of AS patients with DSM-IV disease onset before or at age 18; SAA, South African Afrikaner case/control sample.