Identification of novel proteins that interact with trans-activation domain of transcription factors

我们构建了表达VP16-P53融合蛋白的表达载体，并用哺乳动物双杂交方法(M2H)探索了相互作用伙伴(Gal4-融合蛋白)。根据荧光素酶报告的活性，我们检测到两个重要的候选伙伴：一个是MDM2，一个是已知的p53结合伙伴，另一个是一个17 kDa的蛋白，功能未知。我们尝试了使用标记蛋白的共免疫沉淀来确认M2H的结果，尽管我们没有确认后一种关联。由于我们开发了一种新的体外快速下拉试验方法，因此我们应用该方法进行了验证。然而，我们不能确定后一种相互作用。接下来，我们系统地研究了小鼠转录因子之间的蛋白质-蛋白质相互作用。为此，我们使用了大约1700个小鼠转录因子/转录调控因子的cDNA。自激活实验表明，14%的转录因子具有Gal4-融合蛋白的反式激活活性，显著高于其他蛋白2-3%的反式激活活性。我们使用M2H发现了大约4000个蛋白质-蛋白质相互作用。结果显示，每720次试验中发现1次交互作用，这也显著高于正常情况(每千次中有1次)。一般情况下，双混合数据中存在较多的假阳性交互作用。因此，我们使用上述体外下拉实验验证了我们的结果；对于已知的相互作用，验证的成功率非常高(90%)，对于未知的相互作用，验证的成功率也很高(60%)。这样，我们就可以尽可能地提取出真阳性的蛋白质-蛋白质相互作用。

We made an expression construct that expresses VP16-fused P53 protein followed by exploration of the interaction partners (Gal4-fusion proteins) using the mammalian two-hybrid methods (M2H). We detected two significant partner candidates judging from luciferase reporter activity ; one was mdm2, a known P53 binding partner, and the other was a 17 kDa protein without known functions. We tried co-immuno-precipitation using the tagged proteins to confirm the M2H result, although we did not confirm the latter association. Because we developed a novel method for rapid in vitro pull-down assay, we applied the method for the confirmation. However, we could not confirm the latter interaction.Next we systematically explored protein-protein interactions among mouse transcription factors. For this purpose, we used approximately 1,700 mouse cDNAs for transcription factors/transcription regulation factors. The self-activation experiment showed that 14% of transcription factors have trans-activation activity as Gal4-fusion proteins, which was significantly higher than 2-3% of trans-activation activity for other proteins. We found approximately 4,000 protein-protein interactions using the M2H. The result showed that one interaction was identified per 720 trials, which was also significantly higher value than normal case (one per thousands). Generally, there are many false-positive interactions in the two-hybrid data. Therefore, we validated our result using in vitro pull-down assay described above ; the validation was successful with very high rate (90%) for the publicly known interactions, and was also partially successful (60%) for the unknown interactions. Thus we could extract the true-positive protein-protein interactions as much as possible.