Genetic and functional confirmation of the causality of the DGAT1 K232A quantitative trait nucleotide in affecting milk yield and composition

Abstract

We recently used a positional cloning approach to identify a nonconservative lysine to alanine substitution (K232A) in the bovine DGAT1 gene that was proposed to be the causative quantitative trait nucleotide underlying a quantitative trait locus (QTL) affecting milk fat composition, previously mapped to the centromeric end of bovine chromosome 14. We herein generate genetic and functional data that confirm the causality of the DGAT1 K232A mutation. We have constructed a high-density single-nucleotide polymorphism map of the 3.8-centimorgan BULGE30-BULGE9 interval containing the QTL and show that the association with milk fat percentage maximizes at the DGAT1 gene. We provide evidence that the K allele has undergone a selective sweep. By using a baculovirus expression system, we have expressed both DGAT1 alleles in Sf9 cells and show that the K allele, causing an increase in milk fat percentage in the live animal, is characterized by a higher Vmax in producing triglycerides than the A allele.

Fig. 3.

Posttranscriptional effect associated with the DGAT1 K232A mutation. (A) Schematic representation of part of the DGAT1 gene showing exons III to XVII in red, the 3′ UTR in green, the position of the K232A and Nt1501(C-T) mutations in yellow, the alternative splicing pattern, and the different primers (P1–P5) used for PCR and RT-PCR amplification in blue. Primer combinations corresponding to the G-I, G-II, G-III, RT-I, RT-II, and RT-III products are given. (B) Ethidium bromide staining of the RT-PCR products (RT-IV) obtained from mammary gland mRNA of K/K and A/A individuals by using primers P1 and P4 and size-separated by agarose gel eletrophoresis. The most prominent band corresponds to the major splicing variant; the minor band marked by the arrow corresponds to the alternatively spliced variant. (C) HEX/6-FAM fluorescence ratios obtained by OLA genotyping of PCR products G-I, G-II, and G-III obtained from genomic DNA of heterozygous K/A individuals, and RT-PCR products RT-I, RT-II, and RT-III obtained from mammary gland mRNA of the same.

Fig. 1.

(A) BAC contig (horizontal lines; see ref. 5), STS content (box), and linkage map of the BULGE30–BULGE9 interval. Polymorphic STS are given in red, and monomorphic STS are given in black. Dots on the linkage map correspond to blocks of nonrecombining markers; distance between adjacent marker blocks is given in centimorgan. (B and D) LD map in the Dutch and New Zealand populations, respectively. Linkage disequilibrium between all marker pairs was measured by using r² values (see Materials and Mehtods), which are shown on a black-to-white scale. The heterozygosity (“Het”) of each marker in the respective populations, measured as (1 – ∑ p² i i), where p_i corresponds to the frequency of allele i, is given on the right of the LD square. Markers with more than two alleles are underlined. These are either microsatellite markers or DGAT1 for which the four SNPs [K232A, Nt984 + 8(A–G), Nt848 + 26(C–T) and Nt1501(C–T)] were considered separately when computing r² but as haplotypes when computing Het. (C and E) Statistical significance (measured as a LOD score) of the effects of each individual marker on “milk fat percentage” computed by using a reml model, without (vertical red bars) or with (vertical blue bars) the K232A genotype as fixed effect in the mixed model (see Materials and Methods), computed, respectively, for the Dutch (C) and New Zealand (E) samples. EHH was computed at each marker position for the two and three major DGAT1 core haplotypes that are encountered, respectively, in the Dutch and New Zealand dairy cattle populations. ▪ corresponds to the sH^q haplotype, ▴ corresponds to the sH^Q-D haplotype, and Δ corresponds to the sH^Q-NZ haplotype. For ▪ and ▴, EHH values are flanked by error bars corresponding to ± 1.96, the standard error of the estimate computed as , where t is the number of usable chromosomes carrying the considered DGAT1 core haplotype. In the New Zealand population, EHH values for the sH^Q-NZ haplotype (Δ) were significantly different from neither those of the sH^Q-D haplotype nor those of the sH^q haplotype; error bars have thus been omitted for clarity.

Fig. 2.

Effect of the K232A mutation on the V_max of DGAT1. Amounts of synthesized triglycerides (TG) were estimated from the intensity of the TG spot on a TLC plate by phosphorimaging. Decreasing amounts of total microsomal protein (50, 25, and 12.5 μg) were incubated for increasing lengths of time (2, 4, and 8 min) with diacylglycerol and ¹⁴C-labeled oleoyl-CoA to assay DGAT1 activity under apparent V_max conditions. Total microsomal protein was prepared from uninfected Sf9 cells (SF9: •), Sf9 cells infected with wild-type pFastBac-1 baculovirus (WTFASTBAC1: ♦), Sf9 cells infected with pFastBac-1 baculovirus expressing the alternative spliced DGAT1 form (DGAT1–AS: ▪), Sf9 cells infected with pFast-Bac1 baculovirus expressing the DGAT1 K allele (DGAT1–K: ▴), and Sf9 cells infected with pFastBac-1 baculovirus expressing the DGAT1 A allele (DGAT1–A: ▪). Small open symbols correspond to individual measurements; large filled symbols correspond to the mean of the corresponding group.