Data_Sheet_3_Widespread Separation of the Polypyrimidine Tract From 3′ AG by G Tracts in Association With Alternative Exons in Metazoa and Plants.pdf

At the end of introns, the polypyrimidine tract (Py) is often close to the 3′ AG in a consensus (Y)20NCAGgt in humans. Interestingly, we have found that they could also be separated by purine-rich elements including G tracts in thousands of human genes. These regulatory elements between the Py and 3′ AG (REPA) mainly regulate alternative 3′ splice sites (3′ SS) and intron retention. Here we show their widespread distribution and special properties across kingdoms. The purine-rich 3′ SS are found in up to about 60% of the introns among more than 1,000 species/lineages by whole genome analysis, and up to 18% of these introns contain the REPA G-tracts (REPAG) in about 0.6 million of 3′ SS in total. In particular, they are significantly enriched over their 3′ SS and genome backgrounds in metazoa and plants, and highly associated with alternative splicing of genes in diverse functional clusters. Cryptic splice sites harboring such G- and the other purine-triplets tend to be enriched (2–9 folds over the disrupted canonical 3′ SS) and aberrantly used in cancer patients carrying mutations of the SF3B1 or U2AF35, factors critical for branch point (BP) or 3′ AG recognition, respectively. Moreover, the REPAGs are significantly associated with reduced occurrences of BP motifs between the −24 and −4 positions, in particular absent between the −7 and −5 positions in several model organisms examined. The more distant BPs are associated with increased occurrences of alternative splicing in humans and zebrafish. The REPAGs appear to have evolved in a species- or phylum-specific way. Thus, there is widespread separation of the Py and 3′ AG by REPAGs that have evolved differentially. This special 3′ SS arrangement likely contributes to the generation of diverse transcript or protein isoforms in biological functions or diseases through alternative or aberrant splicing.

在人类中，内含子末端的多嘧啶束（polypyrimidine tract，Py）通常与3′端的AG序列邻近，其保守序列为(Y)₂₀NCAGgt。有趣的是，我们发现在数千个人类基因中，二者之间也可被富含嘌呤的元件（包括G串）分隔开来。这些位于Py与3′AG之间的调控元件（regulatory elements between the Py and 3′ AG，REPA）主要调控可变3′剪接位点（alternative 3′ splice sites，3′SS）和内含子滞留。本研究揭示了这类元件在整个生物界的广泛分布与独特特性。通过全基因组分析，在超过1000个物种类群中，多达约60%的内含子含有富含嘌呤的3′剪接位点，其中总计约60万个3′剪接位点中，有高达18%的内含子携带REPA G串（REPAG）。尤为值得注意的是，在后生动物与植物中，这类元件在其3′剪接位点及全基因组背景下均显著富集，且与多种功能簇基因的可变剪接高度相关。携带此类G串及其他嘌呤三联体的隐蔽剪接位点，往往呈现出相较于被破坏的经典3′剪接位点高2-9倍的富集特征，并在携带SF3B1或U2AF35突变的癌症患者中被异常使用——这两种因子分别是识别分支点（branch point，BP）与3′AG的关键蛋白。此外，在多个被检测的模式生物中，REPAG与-24至-4位区域内分支点基序的出现频率降低显著相关，尤其在-7至-5位区域完全缺失。在人类与斑马鱼中，位置更远的分支点与可变剪接发生频率升高相关。REPAG似乎以物种或门特异性的方式演化。综上，Py与3′AG之间普遍存在由差异化演化而来的REPAG所分隔的现象。这种特殊的3′剪接位点排布，可能通过可变剪接或异常剪接参与生成多样的转录本或蛋白质异构体，进而影响生物学功能或疾病进程。