Fundamental function and regulation of CTCF-BORIS and CT

从10-01-02到9-01-03，我们对CTCF作为一种具有许多重要功能的显着的多面蛋白的重要性的认识急剧增加。到2003年9月，通过NLM PubMed搜索以CTCF为关键字的标题，发现的原始出版物总数达到90篇。其中，三分之一的论文是V. Lobanenkov与MPS LIP的一名或几名现任或前任成员共同撰写的，以及/或与本报告中列出的合作者共同撰写的。同一时期发表的关于CTCF的三篇综述（见参考文献）证实CTCF是一种真正的多价多功能蛋白，它利用11个锌指（ZF）的不同亚群形成具有不同结构的复合物，具有不同的~50 bp的CTCF靶位点（CTS），介导不同的基因表达调节功能。这些功能包括上下文依赖的启动子抑制或激活，激素反应性基因沉默子的产生，以及所有已知的脊椎动物增强子阻断活性（EhBA）的形成，通常也被称为染色质绝缘子或边界。我们和其他人证明，甲基化敏感CTCF结合驱动的EhBA和沉默子子集在哺乳动物体细胞中读取控制x染色体失活（Xi）的等位基因特异性调控标记和印迹基因簇依赖于亲本的单等位基因表达的关键作用。LIP/MPS仍然是唯一的实验室，证明CTCF在体内能够区分H19基因的印迹控制区（ICR）的父本和母本等位基因，该区域通过甲基化敏感的EhBA调节等位基因Igf2的表达。此外，我们是第一个也是唯一一个解释由H19 CTS介导的cpg敏感CTCF- icr相互作用的序列特异性机制的实验室，通过表明它只能发生在这样的CTS上，这显示了具有me- dc残基的二核苷酸的配置与CTCF识别和结合任何给定CTS所必需的zf -接触核苷酸之间的完美匹配。结合ZF用于CTCF以实现多序列特异性的组合原则，后者的发现使我们能够预测，在整个哺乳动物基因组中，产生甲基化敏感（M-S） CTCF驱动的EnhBA的相同一般机制将被普遍采用。除了下面描述的我们与Ohlsson实验室的合作之外，也许证明这个想法既正确又有效的最好例子来自Paul Sadowski今年的HMG论文，该论文显示，在人类chr 11p15.5的一个很长（对于这种类型的研究）的16个基因印迹区域中，每一个不同甲基化的增强bas和沉默子中都存在CTCF位点。此外，作者还提出，该区域的各种缺失，与先天性过度生长障碍（BWS）和各种恶性肿瘤相关，可能是由这些众多CTS中的至少一些介导的CTCF功能丧失所解释的。有趣的是，之前V. Lobanenkov和他的前同事也提出了类似的将CTS-EnhBA的缺失与病理联系起来的想法，他们证明了DM1位点CTCF M-S结合缺失与严重的先天性肌强直性营养不良之间的功能关联（Nat. Genetics, 2001）。为了解决M-S CTS在表观遗传调控中的作用是否可能延伸到基因印迹调控之外，我们与Bill Paul（在IL-4位点上）和Barbara Birnstein（在IgH位点上）合作，绘制了两个构成型CTS依赖的EhBA（即定义这两个位点独立调控的假定边界），并确定了一些由变化数量驱动的M-S EhBA，并且通常组织成集群。差异甲基化调控区域的CTS。为了评估这些发现是否确实反映了CTCF靶点普遍的表观遗传学作用的具体情况，我们与Ohlsson实验室合作，通过生成DNA微阵列克隆来鉴定400多个新的CTS，这些克隆来源于用CTCF抗体免疫纯化的共享染色质片段的DNA。这些CTS，包括单拷贝和重复的，存在于涉及多种细胞功能的位点，如代谢、神经发生、生长和信号传导。使用基于毒素的enhba捕获载体，我们还发现这些靶点中的大多数介导染色质绝缘体功能。由于大多数位点以M-S类CTS为代表，我们相信基于ctcf的网络是表观遗传状态的主要决定因素。在哺乳动物发展瞄准机制所需的基因组表观遗传标记在某些地区的重建和阅读——毫无例外地证明港m CTCF网站所必需的一个子集parent-of-origin-dependent基因表达——哺乳动物CTCF征用一个新函数在gene-imprinting意味着前所未有的自复制和散度创造了对男性和女性germ-cell-specific CTCF-counterparts 11 ZF相同的域:睾丸特异性BORIS基因（去年由我们克隆并获得专利并发表）和卵母细胞特异性Natasha基因（今年8月部分克隆）。对这两个基因的研究在另一篇项目报告中进行了综述。只有让这三个哺乳动物基因在相同的DNA序列上一起工作，这些序列被所有三个因子共享的相同11 ZF结构域识别，才有可能实现母系（natasha导向）和父系（boris导向）甲基化标记在CTS上的种系传播（重置），并通过体细胞CTCF读取这些标记。这意味着我们在果蝇中发现的CTCF同源物可能编码了通用脊椎动物绝缘因子CTCF的唯一真正祖先。今年，由于Hanlim Moon博士（她完成了与V. Lobanenkov的培训并在她的祖国接受了高级职位）的杰出工作，以及与在苍蝇中进行绝缘体试验的Jumin Jhou博士建立合作，我们获得了支持这一观点的直接证据。证据是基于3个主要论点和实验结果的结合：[1]所有DNA序列(包括scs， scs？如FAB-8、珠蛋白HS4等)，能够结合果蝇CTCF和哺乳动物CTCF因子，在果蝇和哺乳动物系统中均表现出同样有效的EnhBA。[2]相比之下，这些DNA序列可以很好地与哺乳动物CTCF结合，但不能与drCTCF结合（在我们的研究中，用重组drCTCF筛选H19 ICR的2 kb结果为阴性），仅在哺乳动物细胞中表现出绝缘体功能，而在果蝇中则没有。[3]除drCTCF外，所有已知的3个参与果蝇绝缘子功能的其他基因，包括果蝇Su(Hw)， BEAF-32和Zw5，都没有脊椎动物的同源物，我们预计FOUNDER绝缘子基因存在同源物。因此，我们得出结论，drCTCF确实是脊椎动物普遍绝缘体基因的真正祖先。我们制作了一个drCTS文库（记住要创建一个由drCTS分离的基因簇的全基因组图谱），但注意到除了基因间位点（这显然是整个果蝇基因组中主要的drCTCF靶点）之外，在Notch和scs以及其他元件中也发现了带有EhBA的drCTS，这些元件除了EnhBA之外也具有已知的启动子活性。这一观察结果表明，drCTCF的EnhBA与作为诱饵启动子的能力密切相关。今年，我们与Elena Klenova博士完成了一个联合项目，以确认CTCF直接与Pol II全酶相互作用，启动基因间CTS的转录。

From 10-01-02 to 9-01-03, our understanding of the significance of CTCF as a remarkable multifaceted protein with many important functions, has increased dramatically. By September 2003, the total number of original publications, identified by searching NLM PubMed for titles containing CTCF as the key-word, reached 90. Among these, every third paper was (co)authored by V. Lobanenkov with one or several member(s) of the MPS LIP current or former stuff, and/or with collaborators listed in this report. Three reviews on CTCF published during the same time (see refs) established CTCF as a true multivalent multifunctional protein which utilizes different subsets of 11 Zn-fingers (ZF) to form structurally-distinct complexes with varying ~50 bp CTCF-target sites (CTS) that mediate distinct functions in regulation of gene expression. These functions include context-dependent promoter repression or activation, creation of hormone-responsive gene silencers, and formation of all known in vertebrates enhancer-blocking activities (EhBA), often also referred as chromatin insulators or boundaries. We, and others, demonstrated that a subset of EhBA and silencers driven by methylation-sensitive CTCF binding plays critical role in reading in mammalian somatic cells of allele-specific regulatory marks that control X-chromosome inactivation (Xi) and parent-of-origin-dependent monoallelic gene expression of imprinted gene clusters. The LIP/MPS remains the only lab, which demonstrated that CTCF in vivo is capable of discriminating paternal vs. maternal alleles of the imprinting control region (ICR) of the H19 gene that regulates allelic Igf2 expression by means of a methylation-sensitive EhBA. Moreover, we were the first and the only lab that explained the very nature of the sequence-specific mechanism for CpG-sensitive CTCF-ICR interactions mediated by the H19 CTS by showing that it can occur only for such CTS, which display a perfect match between dispositions of dinucleotides with the me-dC-residues and the ZF-contacting nucleotides necessary for CTCF to recognize, and bind to, any given CTS. Taken together with the combinatorial principals of ZF usage for CTCF to achieve multiple sequence-specificity, the latter finding allowed us to predict that the same general mechanism of creating methylation-sensitive (M-S) CTCF-driven EnhBA is to be employed universally throughout mammalian genomes. Besides our joint work with the Ohlsson lab described below, perhaps the best example to prove that the idea was both correct and productive, came from this year HMG paper by Paul Sadowski, which showed the presence of CTCF sites in EACH - from over a DOZEN in total - of differentially-methylated EnhBAs and silencers identified in a very long (for such type of studies) imprinted region with 16 genes in of human chr 11p15.5. Moreover, the authors also suggested that various deletions in this region, associated with the congenital overgrowth disorder called BWS and with various malignancies, may be explained by the loss of CTCF functions mediated by at least some of these numerous CTS. Interestingly, similar idea of linking the loss of an CTS-EnhBA with pathology was previously made by V. Lobanenkov and his former co-workers who demonstration (Nat. Genetics, 2001) functional association between the loss of CTCF M-S binding in the DM1 locus and severe congenital myotonic dystrophy. To address whether the role for M-S CTS in epigenetic regulation could possibly be extension beyond regulation of gene-imprinting, we mapped - in collaborations with Bill Paul (on the IL-4 locus) and with Barbara Birnstein (on the IgH) - both constitutive CTS-dependent EhBA (i.e. putative boundaries that define independent regulation of these 2 loci) and identified a number of M-S EhBA driven by the number of varying, and often organized into clusters, CTS in the differentially-methylated regulatory regions. To assess whether these findings may indeed reflect specific cases of the universal EPIGENETIC role for CTCF targets, we collaborated with the Ohlsson lab to identify more than 400 of new CTS by generating DNA microarrays of clones derived from DNA of shared chromatin fractions immunopurified with CTCF antibodies. These CTS, both single copy and repetitive, are found in loci involved in multiple cellular functions, such as metabolism, neurogenesis, growth and signaling. Using a toxin-based EnhBA-trapping vector, we also showed that the majority of these targets mediate chromatin-insulator functions. As the majority of the sites represented by the M-S class of CTS, we feel confident that a CTCF-based network emerges as a major determinant of epigenetic states. To develop in mammals a targeting mechanism required for re-setting and reading of epigenetic marks at certain genome regions - which without exception turned out to harbor a subset of M-S CTCF sites necessary for the parent-of-origin-dependent gene-expression - mammalian CTCF has COOPTED a new function in gene-imprinting by the means of an unprecedented self-duplication and divergence that created the pair of male and female germ-cell-specific CTCF-counterparts with the same 11 ZF domain: testis-specific BORIS gene (cloned, patented and published by us last year) and oocyte-specific Natasha gene (which we partially cloned in August this year), respectively. The work on these two genes is reviewed in the other project report. Only by having these three mammalian genes working together on the SAME DNA SEQUENCES recognized by the same 11 ZF domain shared by all three factors, it has become possible to achieve both germline transmission (resetting) of maternal (Natasha-directed) and of paternal (BORIS-directed) methylation marks at CTS, and reading of these marks by somatic CTCF. This implies that CTCF homologue that we identified in Drosophila, may encode THE ONLY TRUE ANCESTOR of the universal vertebrate insulator factor, CTCF. This year, due to the outstanding work of Dr. Hanlim Moon (who finished her training with V. Lobanenkov and accepted a senior position in her home country) and establishing of collaboration with Dr. Jumin Jhou who performed insulator assays in flies, we obtained direct evidence for this view. The evidence is based on the combination of the 3 major arguments and experimental results: [1] All DNA sequences ( including scs, scs?, FAB-8, globin HS4, etc ), which are capable of binding to BOTH Drosophila CTCF and to mammalian CTCF factors, manifested equally-efficient EnhBA in BOTH Drosophila and in mammalian systems. [2] In contrast, such DNA sequences, which perfectly well bind to mammalian CTCF but do NOT BIND to drCTCF (exemplified in our studies by the negative outcome of screening the 2 kb of the H19 ICR with the recombinant drcTCF), manifested insulator function only in mammalian cells BUT NOT in Drosophila. [3] Except drCTCF, all 3 known OTHER genes involved in the insulator function in flies, including Drosophila Su(Hw), BEAF-32 and Zw5, have NO vertebrate HOMOLOGUES, which one would expect to exist for a FOUNDER INSULATOR gene. Thus, we concluded that drCTCF is indeed the true ANCESTOR of the universal vertebrate insulator gene. We made a library of drCTS (with keeping in mind creating a genome-wide map of gene clusters separated by drCTS), but noticed that in addition to the intergenic sites (which are clearly the main drCTCF targets throughout the fly genome), drCTS with EhBA were also found in the Notch and scs' and other elements, which in addition to EnhBA also have known PROMOTER ACTIVITY. This observation suggested that the EnhBA of drCTCF is closely associated with the ability to function as a decoy promoter. This year, we finalized a joint project with Dr. Elena Klenova to confirm that CTCF directly interacts with of Pol II holoenzyme to initiate transcription from intergenic CTS.