The functional interaction on in vitro gene expression of APOA5 SNPs, defining haplotype APOA5⁎2, and their paradoxical association with plasma triglyceride but not plasma apoAV levels

Abstract

Plasma triglyceride (TG) and apoAV levels are reported to be positively correlated, yet SNPs defining haplotype APOA5⁎2 have consistently shown association with elevated plasma triglyceride (TG) but not plasma apoAV levels. We previously reported that individually − 1131T>C, − 3A>G and + 1891T>C did not influence luciferase activity or in vitro translation efficiency. To investigate the combined effect of these SNPs additional constructs were examined. Compared to the wildtype − 1131T/− 3A/+ 1891T (TAT), the triple rare allele construct − 1131 C/− 3 G/+ 1891 C ( CGC) conferred 46% lower luciferase activity ( p < 0.0001), showing these SNPs are acting co-operatively. Although only these two combinations occur in vivo, we experimentally altered the TAT construct one site at a time; − 3G (T GT) had the largest effect (94% lower luciferase), with lesser effects from CAT (− 77%) and TA C (− 70.3%) (all p < 0.0001). Deletion constructs excluding one site at a time showed that − 3 G/1891 C ( – GC) in combination, compared to − AT, was having the largest effect on luciferase activity (− 59%, p = 0.055). Using sequence homology and EMSA analysis no transcription factor binding at − 1131 or + 1891 was identified, though + 1891 lies within a putative mRNA stability motif. Taken together, these data identify − 3A>G in the Kozak sequence as functional, affecting translation initiation and driving the haplotype effects, while showing interaction with + 1891T>C and to a lesser extent − 1131T>C. A paradox arises since these results predict that APOA5⁎2 will lead to reduced apoAV with concomitant reduced LPL activation or lipoprotein-receptor interaction, resulting in higher plasma TG levels. We conclude that APOA5 expression, and not circulating plasma apoAV levels, is causatively associated with plasma TG levels.