Experiment
In the current study a major quantitative trait locus (QTL) influencing femur size was dissected. The QTL was originally identified in an F2 cross between B6.Cg-Socs2hg (HG) and CAST/EiJ (CAST) strains and was referred to as femur length in high growth mice 2, QTL Feml2. The HG strain is a C57BL/6J (B6) mouse that is homozygous for a deletion encompassing the Socs2 gene [the high growth (hg) locus], a negative regulator of growth hormone signaling, that results in increased growth and body size. In contrast, CAST mice are genetically divergent wild-derived inbred mice that are small and lean. Feml2 was located on Chr 9 at 20 cM, explained 10.7% of the variance in femur length, and was independent of the hg locus (J:68441).
To identify the gene(s) responsible for Feml2, the current study generated the HG.CAST-(D9Mit249-D9Mit133)/Jfmuc (HG9) congenic strain that possessed CAST alleles from 9 to 84 Mbp on an HG background. Two HG9 F2 intercrosses were used to fine map Feml2. HG9 x HG (n=283) and HG9 x B6 (n=457) male and female F2 mice were generated by intercrossing F1 mice.
At 9 weeks of age male HG9 congenic (n=7) and HG control (n=14) mice were anesthetized and body weights and lengths were measured. Mice were killed and each femur length and width were measured. F2 mice were phenotyped. Mice were genotyped using microsatellite markers. (HG9 x B6) F2 mice were genotyped for the hg locus.
The R/qtl package was used to preform linkage analysis. Both F2 crossed were combined for the linkage analysis. The scanone function, using the Haley Knott regression algorithm, was used to perform interval mapping using a model that included sex, body weight, and cross type (HG9 x HG) or (HG9 x B6) terms as additive covariates. LOD significance for all models tested were empirically determined using 1000 permutations.
Femur length was first mapped in each cross separately. No differences in the LOD score profile of peak positions were observed (Figure 1, A and B). As a result, both crosses were combined to increase mapping resolution (Figure 1C). Feml2 mapped to 30.1 cM with a peak LOD score of 40.0 The effects of Feml2 on femur length were additive with each CAST allele decreasing femur length by 0.23 mm (Figure 1D). Feml2 explained 7.8% of the total variance in femur length. The 95% conficence interval for Feml2 extended from 28.0 to 31.7 cM, which in physical distance equated to the 5.7 Mbp region extending from 57.6 to 63.3 Mbp (GRCm38/mm10).
Feml2 contains a total of 69 RefSeq protein-coding genes (Table S1). Potentially high impact variants were found in 37 of the 69 Feml2 genes. The initiator codon variant is in the enhancer of mRNA decapping 3 (Edc3) gene; however, a second in-frame ATG is located 6 bp downstream. Two of the stop-gain and the stop-retained variants were found in the unclassified gene 1700036A12Rik. The other stop-gain variant was in another unclassified gene, Gm10657 (Table S1).
To identify Feml2 genes whose expression was under genetic regulation, allele-specific-expression was quantified in growth plate tissue in (CAST x B6) F1 mice. Six of the 69 genes from the Feml2 region were found to be expressed in an allele-specific manner, demonstrating higher transcript levels originating from either the B6 or CAST chromosomes. The genes were Gramd2, Larp6, Adpgk, Ubl7, Rec114 and Glce; all except Adgpk were preferentially expressed from the CAST allele as compared to the B6 allele.
Analysis Tools