The Genom of Homo sapiens.pdf

184 GEORGES AND ANDERSSONbe common between syntenic markers several tens ofcentimorgans apart, and gametic association was evenfound between nonsyntenic markers, especially in syntheticpopulations derived by the recent admixture of distinctbreeds (N. Harmegnies and M. Georges, unpubl.).To provide additional support for the direct involvementof IGF2 in the SSC2 QTL, we therefore used an LDbasedIBD mapping method to refine the map position ofthe QTL. The approach chosen was very similar to theone previously described for the fine-mapping of theBTA14 QTL in dairy cattle. To increase the QTL detectance,the QTL genotype of a series of boars sampled inmultiple populations was determined with high accuracyby marker-assisted segregation analysis using large cohortsof offspring. Heterozygous Qq sires were predictedto carry an IBD Q QTL allele having appeared by mutationor migration on a founder chromosome g generations agoand having undergone a selective sweep due to its favorableeffect on muscle mass. As a consequence, a sharedmarker haplotype spanning the QTL was predicted for allQ-bearing chromosomes. We developed a panel of 51SNPs, and 5 microsatellite markers uniformly spanningthe KVLQT1–H19 imprinted domain and genotyped heterozygousQq boars for all of these including sufficientoffspring to establish marker phase. As predicted, all Q-bearing chromosomes were indeed sharing a haplotypespanning 250 kb and containing INS and IGF2 as onlyknown paternally expressed genes. Such a shared haplotypewas not found among q chromosomes, which wereshowing a much higher degree of genetic diversity, as expected.This provided strong additional support for a directinvolvement of IGF2 (Nezer et al. 2003). Contrary to theDGAT1 story, however, sequencing the Q and q alleles forthe coding portions of IGF2 did not reveal any structuraldifference. This suggested that the QTL effect was mostlikely due to a regulatory rather than a structural mutation.Extensive Resequencing of a 28.5-kbChromosome Segment Encompassing theIGF2 Gene Identifies the QTNTo identify the hypothetic regulatory mutation, we sequenced28.5 kb (encompassing the last exon of the THgene, the complete INS gene, and the complete IGF2 geneup to the SWC9 microsatellite marker in its 3´UTR correspondingto the distal end of the marker haplotype sharedby all Q alleles) for three Q chromosomes and seven qchromosomes, each carrying a distinct marker haplotypeacross the KVLQT1–H19 domain (Van Laere et al. 2003).We identified a staggering 258 SNPs, corresponding toone polymorphic base pair every 111 nucleotides. A phylogenetictree describing the relationship between the tenanalyzed chromosomes was constructed by neighborjoininganalysis. It showed that all three Q chromosomeswere virtually identical to each other, whereas the q chromosomesessentially fell into two distinct clades. One ofthese included the EWB, suggesting that the chromosomesin this cluster were of European descent. Chromosomesin this cluster exhibited an average nucleotide diversityof 0.0007 among themselves and of 0.0039 withthe chromosomes from the Q clade. The other q cladecomprised two more distantly related chromosomes exhibitinga nucleotide diversity of 0.0025 among themselves,of 0.0032 with the Q clade, and 0.0034 with theother q clade.For 33 of the 258 identified SNP positions, the three Qchromosomes shared a nucleotide not encountered on anyof the analyzed q chromosomes. These would most likelycorrespond to mutations having occurred on the branchseparating the Q from the other chromosomes and wouldtherefore be prime candidates for the QTN. Two of these33 SNPs reside in evolutionary footprints, regions exhibitinga high degree of conservation between human,mouse, and pig, which were detected in a preliminary sequenceexploration of the porcine IGF2 and H19 genes(Amarger et al. 2002).We then screened a large number of sire families usingone of the 33 candidate SNPs (located between TH andINS) as being diagnostic of the Q versus q status. Weidentified two sire families that were not showing any evidencefor the segregation of a QTL despite the fact thatthe corresponding boars were clearly heterozygous forthe SNP: one Hampshire boar and one F 1 M x LW boar.We completely sequenced the 28.5-kb segment for thefour chromosomes of the corresponding boars. TheHampshire boar proved to carry one chromosome that belongedto the Q clade and one chromosome that wasclearly a recombinant chromosome, being from the qclade between TH and the middle of the IGF2 intron 1and from the Q clade for the remainder of the interval.This allowed us to exclude the entire segment for whichthis boar was heterozygous, eliminating 9 of the 33 candidateSNPs. The F 1 M x LW boar proved to carry onechromosome—of Meishan origin—belonging apparentlyto the Q clade over its entire 28.5-kb length, the other—of large white origin—belonging to the q clade over theentire 28.5 kb. Intriguingly, however, the Meishan chromosomediffered from all other known Q chromosomesat one position, corresponding to one of the two candidateSNPs located in an evolutionary footprint: a CpG islandin the third intron of IGF2. The Meishan chromosomewas sharing an A residue with all q chromosomes at thatposition whereas all other Q chromosomes had a Gresidue. The q LW chromosome, on the contrary, sharedthe same residue with all other q chromosomes at all the33 QTN candidate positions. The easiest interpretation ofthese results is that the chromosome of the M x LW F 1boar clustering within the Q clade was in fact functionallyof q type differing with the genuine Q chromosomes atthe actual “holy grail”: the QTN. The functional q statusof this Meishan chromosome was in agreement with aprevious publication detecting the imprinted SSC2 QTLin a M x LW intercross, and in which the Meishan chromosomewas associated with a decrease in muscle massand concomitant increase in fat deposition (de Koning etal. 2000). Extending our screen, we identified one boarthat was heterozygous at the QTN only: The QTL provedto segregate in its offspring, providing very strong supportthat the G residue at the QTN was necessary for achromosome to be functionally Q.