A heterotrimeric SMARCB1-SMARCC2 subcomplex is required for the assembly and tumor suppression function of the BAF chromatin-remodeling complex.
A heterotrimeric SMARCB1-SMARCC2 subcomplex is required for the assembly and tumor suppression function of the BAF chromatin-remodeling complex.
复制标题
BAF 染色质重塑复合物的组装和肿瘤抑制功能需要异三聚体 SMARCB1–SMARCC2 亚复合物
DOI：
10.1038/s41421-020-00196-4
复制
发表时间：
2020
期刊：
影响因子：
33.5
通讯作者：
Long J
中科院分区：
文献类型：
--
作者：
Chen G;Zhou H;Liu B;Wang Y;Zhao J;Giancotti FG;Long J
尊敬的编辑，SWI/SNF 复合体利用其 ATP 依赖性染色质重塑活性来动员核小体，从而调节核小体模板处的 DNA 可及性。 BAF（哺乳动物 SWI/SNF）复合体由大约 15 个蛋白质亚基 1、2 组成。BAF 的一个重要亚基由两个相互排斥的催化 ATP 酶亚基（SMARCA2/Brm 和 SMARCA4/Brg1）之一、SMARCB1/BAF47/INI1/SNF5、SMARCC1/BAF155 和 SMARCC2/BAF170 3. 其中，SMARCC1 和 SMARCC2 彼此高度相似（补充图 S1a），被认为是组装和调节 BAF 4 的关键支架蛋白，BAF 4 在多种癌症中发生突变 2, 5。有趣的是，SMARCB1 在约 98% 的恶性横纹肌样肿瘤中被发现双等位基因失活，具有侵袭性 儿科肿瘤 6, 7。最近，确定了核小体结合-SWI/SNF、-RSC 或-BAF 复合物的冷冻电子显微镜 (cryo-EM) 结构 8-10。然而，这些文章中存在的子复合物 Snf5-Swi3、SMARCB1-SMARCC2 和 SFH1-RSC8 的组装和结构是有限的。与此同时，特定 BAF 亚基突变导致不同癌症类型发展的机制尚不清楚。首先，我们使用基于截断的方法逐步绘制 SMARCB1 和 SMARCC2 的最小结合区域（补充图 S1b-g）。我们发现SMARCB1的一个区域（aa 169-385，SMARCB1（169-385））和SMARCC2的SWIRM结构域（aa 423-518，SMARCC2（423-518））（图1a）形成一个稳定的亚复合体（补充图S1g）。我们成功结晶了包含 SMARCC2 片段 SMARCC2 (325-518) 和 SMARCB1 (169-385) 的复合物（图 1a 和补充图 S1h）。最后，我们以2.60 Å的分辨率确定了人类SMARCB1-SMARCC2子复合物的晶体结构（补充表S1），并发现该子复合物组装成异源三聚体（图1b）。一致地，等温滴定量热法（ITC）显示SMARCC2（325-518）遵循具有两个计算的Kd值的双位点结合模型与SMARCB1（169-385）结合（补充图S1i）。在最终模型中，SMARCB1（169-385）从aa 184解析到aa 356，包括Rpt1和Rpt2基序，SMARCC2（325-518）从423解析到514，包括SWIRM结构域（图1b和补充图S2a-b）。然而，SMARCC2（325-518）的N端残基325-422缺乏可观察到的密度，可能是因为它们在结晶过程中被降解（补充图S1j）。 Rpt1（由双链反平行β-折叠和两个α螺旋组成）和Rpt2（由三链反平行β-折叠和两个α-螺旋组成）可以很好地叠加，47个Cα原子的均方根偏差为0.553 Å（图1b和补充图S3a-b）。每个Rpt基序与SMARCC2的一个单独的SWIRM结构域结合，形成两个单独的子复合物（图1b）（定义为Rpt1/SWIRM-1和Rpt2/SWIRM-2）。这
Dear Editor, The SWI/SNF complex utilizes its ATP-dependent chromatin-remodeling activity to mobilize nucleosomes and thus regulates DNA accessibility at nucleosomal templates. The BAF (mammalian SWI/SNF) complex is composed of approximately 15 protein subunits 1, 2. An important subcomplex of BAF consists of one of two mutually exclusive catalytic ATPase subunits (SMARCA2/Brm and SMARCA4/Brg1), SMARCB1/BAF47/INI1/SNF5, SMARCC1/BAF155, and SMARCC2/BAF170 3. Among them, SMARCC1 and SMARCC2 are highly similar to each other (Supplementary Fig. S1a) and considered as the key scaffold proteins in assembling and regulating the BAF 4, which are mutated in several cancers 2, 5. Intriguingly, SMARCB1 was found to be biallelically inactivated in~ 98% of all malignant rhabdoid tumors, aggressive pediatric tumors 6, 7. Recently, the cryo-electron microscopy (cryo-EM) structures of nucleosome bound-SWI/SNF,-RSC, or-BAF complexes were determined 8–10. However, the assembly and structure of subcomplexes Snf5-Swi3, SMARCB1-SMARCC2, and SFH1-RSC8 present in these articles are limited. Meantime, the mechanisms through which mutations in specific BAF subunits underlie the development of different cancer types are not well known.Firstly, we used a truncation-based approach to gradually map the minimal binding region of SMARCB1 and SMARCC2 (Supplementary Fig. S1b-g). We found that a region of SMARCB1 (aa 169–385, SMARCB1 (169–385)), and the SWIRM domain of SMARCC2 (aa 423–518, SMARCC2 (423–518))(Fig. 1 a) form a stable subcomplex (Supplementary Fig. S1g). We successfully crystallized a complex comprising a SMARCC2 fragment SMARCC2 (325–518) and SMARCB1 (169–385)(Fig. 1 a and Supplementary Fig. S1h). Finally, we determined the crystal structure of the human SMARCB1–SMARCC2 subcomplex at a resolution of 2.60 Å (Supplementary Table S1) and found that this subcomplex assembles into a heterotrimer (Fig. 1 b). Consistently, isothermal titration calorimetry (ITC) showed that SMARCC2 (325–518) binds to SMARCB1 (169–385) following a two-site binding model with two calculated Kd values (Supplementary Fig. S1i). In the final model, SMARCB1 (169-385) was resolved from aa 184 to aa 356 including the Rpt1 and Rpt2 motifs and SMARCC2 (325–518) was well resolved from 423 to 514 including the SWIRM domain (Fig. 1 b and Supplementary Fig. S2a-b). However, the N-terminal residues 325–422 of SMARCC2 (325–518) lacked observable density, possibly because they were degraded during crystallization (Supplementary Fig. S1j). Rpt1 (consisting of a two-stranded antiparallel β-sheet followed by two αhelices) and Rpt2 (consisting of a three-stranded antiparallel β-sheet and two α-helices) can be superimposed well with a root-mean-square deviation of 0.553 Å for 47 Cα atoms (Fig. 1 b and Supplementary Fig. S3a-b). Each Rpt motif binds to a separate SWIRM domain of SMARCC2 to form two separate subcomplexes (Fig. 1 b)(defined as Rpt1/SWIRM-1 and Rpt2/SWIRM-2). The