Bonet Roman, Ramirez-Espain Ximena, Macias Maria J
The Institute for Research in Biomedicine, Barcelona (Protein NMR Group), Barcelona Science Park, Baldiri i Reixac, 10-12 08028 Barcelona, Spain.
Proteins. 2008 Dec;73(4):1001-9. doi: 10.1002/prot.22127.
FF domains are present in three protein families: the splicing factors formin binding protein 11 (FBP11), Prp40, and URN1, the transcription factor CA150, and the p190RhoGTPase-related proteins. This simplicity in distribution, however, is contrasted by the difficulty in defining their biological role. At best, the group of ligand FF domains can bind to form a motley crew with binding reports pointing also to negative/aromatic sequences, the tetratricopeptide repeat, the transcription factor TFII-I and even to RNA. To expand our knowledge on the FF domain, we selected the FF domain present in the URN1 yeast splicing factor as the subject for structural studies. The URN1 protein is one of the two known proteins containing only one FF domain, making it the most simplified representative of FF domain-containing splicing factors. The solution structure reveals that the domain adopts the classical FF fold, with a distinctive negatively charged patch on its surface. All available FF structures have a well-conserved fold but variable electrostatic patches on their surfaces. These patches are unconserved, even for domains with similar pK(a)s. To investigate potential binding sites in FF domains, we performed structural comparisons to other proteins with similar folds. In addition to the structures detected by SCOP, we included SURP domains, which also adopt the alpha1-alpha2-3(10)-alpha3 architecture. We observed that the main difference between all these structures resides in the orientation of the second helix. Remarkably, in DEK, SURP, and Prp40FF1 structures (the exception is the FBP11FF1 domain), the second helix participates in ligand recognition. Furthermore, SURP and Prp40FF1 binding sites also include the 3(10) helix, which forms a partially exposed hydrophobic cavity. This cavity is also present in at least CA150FF1 and FF2 structures. Thus, as with WW domains, the FF fold seems to have developed binding-site variations to accommodate an abundant and variable set of ligands.
FF结构域存在于三个蛋白质家族中：剪接因子formin结合蛋白11（FBP11）、Prp40和URN1，转录因子CA150，以及p190RhoGTP酶相关蛋白。然而，这种分布上的简单性与定义它们的生物学作用的困难形成了对比。充其量，配体FF结构域组可以结合形成一个混杂的群体，结合报告还指向负/芳香序列、四肽重复序列、转录因子TFII-I甚至RNA。为了扩展我们对FF结构域的认识，我们选择了URN1酵母剪接因子中存在的FF结构域作为结构研究的对象。URN1蛋白是已知的仅含有一个FF结构域的两种蛋白质之一，使其成为含FF结构域的剪接因子中最简化的代表。溶液结构表明，该结构域采用经典的FF折叠，其表面有一个独特的带负电荷区域。所有可用的FF结构都有一个保守的折叠，但表面的静电区域可变。即使对于具有相似pK(a)值的结构域，这些区域也不保守。为了研究FF结构域中的潜在结合位点，我们对其他具有相似折叠的蛋白质进行了结构比较。除了SCOP检测到的结构外，我们还纳入了SURP结构域，其也采用α1-α2-3(10)-α3结构。我们观察到所有这些结构之间的主要差异在于第二个螺旋的方向。值得注意的是，在DEK、SURP和Prp40FF1结构中（FBP11FF1结构域除外），第二个螺旋参与配体识别。此外，SURP和Prp40FF1结合位点还包括3(10)螺旋，其形成一个部分暴露的疏水腔。这个腔也至少存在于CA150FF1和FF2结构中。因此，与WW结构域一样，FF折叠似乎已经发展出结合位点的变化以容纳大量可变的配体。
