Structural Studies Of The C-myc Gene Regulation

人c-myc原癌基因的转录受启动子上游和下游多个顺式元件的调控。C-myc启动子上游的一个顺式元件是CT元件。这是一个富含CT的序列，位于P1启动子上游100个碱基。结合CT元件编码链的一种蛋白质是hnRNPk。HnRNPk与CT元件结合上调c-myc转录。HnRNPk中有三个同源重复，称为KH结构域。HnRNPk蛋白的86个残基C-末端组成hnRNPk的第三个KH基序(Kh3)。除了常规的蛋白质核磁共振方法外，还使用了液晶技术的核磁共振波谱来确定该化合物的三维结构。液晶环境在体系中产生了轻微的有序，重新引入了偶极耦合，提供了有用的结构信息。使用偶极耦合信息的添加计算的最终结构族导致0.17埃的族的均方根偏差(RMSD)。相比之下，在没有偶极耦合的情况下进行的计算的RMSD为0.32埃。这为结构的精度提供了定量的评估，RMSD值越小，意味着精度越高。这种比较是在理想的条件下进行的：所有核磁共振实验的信噪比都很好，一些核磁共振实验重复两次以进行误差估计和一致性检查，并且所有距离估计都是保守的，以考虑到任何可能的系统误差。因此，对于典型的非理想条件下的核磁共振结构，偶极耦合信息的加入对RMSD值的增加应该要大得多。在确定Kh3结构的同时，我们还利用偶极耦合信息开发了一种直接的结构精化协议。该方案将提供给任何希望利用偶极耦合信息的实验室。还对HH3结构域进行了动态研究。在经历快速(PS)波动的Kh3结构域中发现了一个柔性环(L52-R56)。与之前对FMR1类似KH结构域的研究相反，保守的环(G30-G33)没有显示出任何灵活性。为了通过CT元件更好地了解c-myc的调控，已经启动了一个平行项目来确定细胞核酸结合蛋白(CNBP)的结构。该蛋白与hnRNPk结合位点相反的链结合。CNBP上调CT元件活性。构建了全长(176个残基)的CNBP。它由七个锌指组成。目前正在探索一种仍然保持核苷酸结合亲和力的CNBP的最低结构。我们完成了突变型(Gly26-Arg)Kh3、野生型Kh3和Kh3+ss-DNA复合体的核磁共振骨架动力学研究。我们已经证明突变的Kh3没有ss-DNA结合活性。我们还对KH3+DNA络合物进行了滴定研究，以绘制DNA结合部位。在这一点上，我们可以访问所有的动态参数和DNA结合位点的地图。目前，我们正在比较所有这些参数，以基于结构和动态信息来表征Kh3ss-DNA相互作用。我们还开展了以KH3为模型体系的侧链动力学研究。我们开展了一项实验，可以探测谷氨酰胺和天冬氨酸侧链上的NH_2部分。比较自由和结合形式下NH2基团的动力学将提供关于侧链与ss-DNA靶标相互作用的信息。这一方法将被扩展到观察蛋白质中的CH、CH2和CH3部分。去年，我们一直在尝试进行实验，以识别含有Kh3的蛋白质中的结构域之间的运动，以及完成我们关于hnRNPk的Kh3中的侧链运动的项目。为了研究域间运动，我们使用了一个已被广泛研究的模型系统来测试我们的分析模型。所选择的模型系统是一种钙结合蛋白，即钙调蛋白。我们已经能够证实，为了量化缓慢的域间运动，需要一系列在不同场强下获得的核磁共振主干弛豫数据。此外，我们的简单模型代表了问题的一阶方法，似乎足以提供这一运动的幅度和时间尺度。通过对钙调蛋白中的结构域运动进行立体模拟，证实了估计的运动幅度。我们现在正在应用这个简单的运动模型来研究Kh3运动。我们改进了可以用来描述蛋白质中缓慢的网际结构域运动的运动模型。我们最新的模型使用了锥体内摆动的Pade近似，这导致了用三个指数项来描述内部运动。到目前为止，我们的测试表明，当运动较大(拐角为50度或更大)时，该模型是更合适的模型。在小幅度慢动作的极限情况下，较早且数学上较简单的模型就足够了。

The transcription of the human c-myc proto-oncogene is regulated by multiple cis-elements upstream as well as downstream of the promoter sites. One cis-element upstream from the c-myc promoters is the CT element. This is a CT rich sequence which is located 100 bases upstream from the P1 promoter. One protein that binds the coding strand of the CT element is hnRNPk. Binding of hnRNPk to the CT element upregulates c-myc transcription. There are three homology repeats in hnRNPk called the KH domains. The 86 residue C-terminal segment of the hnRNPk protein comprises the third KH motif (KH3) in hnRNPk. The three-dimensional structure of this has been determined by NMR spectroscopy using the liquid crystal technique in addition to the conventional protein NMR method. The liquid crystal environment produces a slight order in the system which reintroduces dipolar coupling providing useful structural information. The final family of structures calculated using the addition of dipolar coupling information results in root mean square deviation (RMSD) for the family of 0.17 Angstrom. In contrast, calculations carried out without the dipolar couplings has an RMSD of 0.32 Angstrom. These provide a quantitative evaluation of the precision of the structures with the smaller RMSD value implying higher precision. This comparison was made under ideal conditions: good signal to noise for all of the NMR experiments, some NMR experiments repeated twice for error estimates as well as consistency check, and all distance estimates done conservatively to take into account any possible systematic error. Thus, for a typical NMR structure under non-ideal conditions, the increase of the RMSD value with the addition of dipolar coupling information should be much greater. In parallel to the KH3 structure determination we also developed a straight forward protocol of structure refinement using the dipolar coupling information. This protocol is being provided to any laboratory which wishes to take advantage of dipolar coupling information. A dynamic study of KH3 domain has also been carried out. One flexible loop (L52-R56) was identified in the KH3 domain which undergoes rapid (ps) fluctuation. In contrast to a previous study of a similar KH domain of FMR1 the conserved loop (G30-G33) does not show any flexibility. In order to get a better understanding of c-myc regulation through the CT element, a parallel project has been initiated to determine the structure of the cellular nucleic acid binding protein (CNBP). This protein binds the strand opposite to the hnRNPk binding site. CNBP upregulates the CT element activity. A construct of the full length (176 residues) CNBP has been made. It consists of seven zinc fingers. A minimum construct of CNBP which still retains the nucleotide binding affinity is being probed. We have completed the NMR backbone dynamic studies of mutant (Gly26-Arg) KH3, wild type KH3, and KH3+ss-DNA complex. We have shown that there is no ss-DNA binding activity for the mutant KH3. We have also done titration studies on the KH3+DNA complex to map the DNA binding site. At this point, we have access to all dynamic parameters and maps of the DNA binding site. We are currently comparing all of these parameters to characterize the KH3 ss-DNA interaction based on structure as well as dynamic information. We also have initiated a study on side chain dynamics using KH3 as a model system. We have developed an experiment where we can probe the NH2 moiety on the side chain of Gln and Asn. Comparison of the dynamics of this NH2 group in the free and bound form would provide information on side chain interaction with the ss-DNA target. This methodology will be extended to look at CH, CH2, and CH3 moieties in the protein. This last year we have been trying to carry out experiments to identify inter-domain motions in KH3-containing protein as well as to finish our project on the side chain motion in KH3 of hnRNPk. In order to study the inter-domain motions we have used a model system that has been studied extensively to test our analysis model. The model system chosen was a calcium binding protein, calmodulin. We have been able to confirm that in order to quantify slow inter-domain motions a series of NMR backbone relaxation data obtained at different field strengths would be required. Furthermore, our simple model that represents a first order approach to the problem, seems to be adequate to provide the amplitude as well as the time scale of this motion. The estimated amplitude of motion has been confirmed by sterically modeling the domain motion in calmodulin. We are now applying this simple motional model to study KH3 motion. We have refined the motional model that can be used to describe slow internam domain motions in protein. Our latest model use a Pade approximation of wobbled in a cone which results in a description of the internal motion by three exponential terms. So far our tests have shown this model to be a more suitable model when the motion is large (angle of inflection of 50 degrees or larger). In the limit of small amplitude slow motion the earlier and mathematically simpler model is sufficient.