Abstract
Abstract 3358
Background:
Tissue factor pathway inhibitor (TFPI) is an essential anticoagulant protein located on endothelium and within platelets. Two TFPI isoforms, termed TFPIα and TFPIβ, are generated through alternative splicing toward the 3' end of the TFPI gene. Alternative splicing also occurs within the 5' untranslated region (UTR) of TFPI mRNA, resulting in the removal of exon 2 (X2). Previous studies have demonstrated that TFPIα and TFPIβ mRNA are produced, on average, in a ratio of 10:1 (TFPIα:TFPIβ) in human tissues. However, this does not appear to correlate with protein expression, suggesting that TFPI protein isoform production is regulated during translation. Biochemical studies were undertaken to determine how alternative splicing within the 5' or 3' UTRs of TFPI mRNA may function in the translational regulation of TFPI protein isoform production.
Methods:
Northern analysis of human lung RNA using probes directed toward Exon 1 (X1) or Exon 2 (both 5' UTR), Exon 6 (common to both TFPIα and TFPIβ), Exon 8 (TFPIβ specific) or Exons 9 and 10 (TFPIα specific), and RT-PCR analysis of multiple human tissue cDNAs using an exon 1 specific forward primer and TFPIα or TFPIβ specific reverse primers, were used to investigate isoform specific alternative splicing of exon 2. Polysome profiling of HUVEC and EA.hy926 cellular RNA was performed to investigate translational control of TFPI isoform expression. Luciferase (Luc) reporter assays were established to quantify the effect of exon 2 on TFPI isoform protein expression. Constructs were produced where both splice variants of the 5' UTR, and the TFPIα or TFPIβ specific 3' UTRs, were inserted at their respective ends of a Gaussia Luc expression cassette. Luc activity was assessed in stably transfected CHO or EA.hy926 cells and normalized to GFP production or Luc mRNA.
Results:
Northern analysis revealed the presence of two TFPIα specific bands at 4.1 and 1.4 kb, produced via alternative polyadenylation, and one TFPIβ specific band at 1.1 kb. Exon 1 and 2 probes bind both TFPIα species, demonstrating that 5' UTR alternative splicing occurs in TFPIα mRNA. RT-PCR analysis of human placenta, lung, and heart cDNA confirmed this, and demonstrated that exon 2 is also alternatively spliced in TFPIβ message. Polysome analysis identified TFPIβ-specific translational repression by exon 2. Luciferase reporter assays quantified this repression at 93% and 86% in CHO and EA.hy926 cells, respectively (see Table 1 for normalized data). In contrast, exon 2 had minimal effect on luciferase expression when the TFPIα 3' UTR was present. RT-PCR analysis of 15 human tissues revealed that testis, thymus, brain, and spleen had the highest relative amount of exon 2 containing TFPIβ message, although all tissues had more exon 2-absent TFPIβ message compared to that containing exon 2.
Conclusion:
Exon 2 is a strong translational repressor of TFPIβ production, and a much weaker repressor of TFPIα. To our knowledge, this is the first time that alternative splicing within a 5' UTR has been demonstrated to affect a specific protein isoform produced via alternative splicing at the 3' end of the same gene. While the mechanism of repression has yet to be elucidated, the repression of TFPIβ, and not TFPIα, by exon 2 suggests that regions within both exon 2 and the 3' UTR are required for this effect and is consistent with the hypothesis of mRNA circularization via interactions between the 5' and 3' UTRs. We hypothesize that the physiological relevance of the demonstrated repression will include tissue-specific regulation of TFPIβ expression and cellular regulation of TFPIβ expression in inflammatory disease.
Disclosures:
Mast: Novo Nordisk A/S: Research Funding.
Complex alternative splicing of human Endonuclease V mRNA, but evidence for only a single protein isoform