Accurately predicting changes in protein stability due to mutations is important for protein engineering and for understanding the functional consequences of missense mutations in proteins. We have developed DeepDDG, a neural network-based method, for use in the prediction of changes in the stability of proteins due to point mutations. The neural network was trained on more than 5700 manually curated experimental data points and was able to obtain a Pearson correlation coefficient of 0.48−0.56 for three independent test sets, which outperformed 11 other methods. Detailed analysis of the input features shows that the solvent accessible surface area of the mutated residue is the most important feature, which suggests that the buried hydrophobic area is the major determinant of protein stability. We expect this method to be useful for large-scale design and engineering of protein stability. The neural network is freely available to academic users at http:// protein.org.cn/ddg.html.

Cancer-associated mesenchymal stem cells (MSCs) play a pivotal role in modulating tumor progression. However, the interactions between liver cancer-associated MSCs (LC-MSCs) and hepatocellular carcinoma (HCC) remain unreported. Here, we identified the presence of MSCs in HCC tissues. We also showed that LC-MSCs significantly enhanced tumor growth in vivo and promoted tumor sphere formation in vitro. LC-MSCs also promoted HCC metastasis in an orthotopic liver transplantation model. Complementary DNA (cDNA) microarray analysis showed that S100A4 expression was significantly higher in LC-MSCs compared with liver normal MSCs (LN-MSCs) from adjacent cancer-free tissues. Importantly, the inhibition of S100A4 led to a reduction of proliferation and invasion of HCC cells, while exogenous S100A4 expression in HCC cells resulted in heavier tumors and more metastasis sites. Our results indicate that S100A4 secreted from LC-MSCs can promote HCC cell proliferation and invasion. We then found the expression of oncogenic micro-RNA (miR)-155 in HCC cells was significantly up-regulated by coculture with LCMSCs and by S100A4 ectopic overexpression. The invasion-promoting effects of S100A4 were significantly attenuated by a miR-155 inhibitor. These results suggest that S100A4 exerts its effects through the regulation of miR-155 expression in HCC cells. We demonstrate that S100A4 secreted from LC-MSCs promotes the expression of miR-155, which mediates the down-regulation of suppressor of cytokine signaling 1, leading to the subsequent activation of STAT3 signaling. This promotes the expression of matrix metalloproteinases 9, which results in increased tumor invasiveness. Conclusion: S100A4 secreted from LC-MSCs is involved in the modulation of HCC progression, and may be a potential therapeutic target. (HEPATOLOGY 2013;57:2274-2286 T he tumor microenvironment plays an important role in modulating cancer and cancer stem cell progression. 1,2 Recently, mesenchymal stem cells (MSCs), as a pivotal part of the tumor stroma, have attracted great attention for their ability to participate in tumor proliferation 3 and metastasis. 4 Although several lines of evidence demonstrate that MSCs can be activated by cancer cells and contribute to tumor progression, the Abbreviations:: cDNA, complementary DNA; ELISA, enzyme-linked immunosorbent assay; HCC, hepatocellular carcinoma; IHC, immunohistochemistry; LCMSCs, liver cancer-associated MSCs; LN-MSCs, liver normal MSCs; miRNA, microRNA; miR-155, microRNA-155; MMP9, matrix metalloproteinases 9; MSCs, mesenchymal stem cells; qRT-PCR, quantitative real time polymerase chain reaction; siRNA, small interfering RNA; SOCS1, suppressor of cytokine signaling 1; STAT, signal transducer and activator of transcription.From the

Computational protein design has a wide variety of applications. Despite its remarkable success, designing a protein for a given structure and function is still a challenging task. On the other hand, the number of solved protein structures is rapidly increasing while the number of unique protein folds has reached a steady number, suggesting more structural information is being accumulated on each fold. Deep learning neural network is a powerful method to learn such big data set and has shown superior performance in many machine learning fields. In this study, we applied the deep learning neural network approach to computational protein design for predicting the probability of 20 natural amino acids on each residue in a protein. A large set of protein structures was collected and a multi-layer neural network was constructed. A number of structural properties were extracted as input features and the best network achieved an accuracy of 38.3%. Using the network output as residue type restraints improves the average sequence identity in designing three natural proteins using Rosetta. Moreover, the predictions from our network show ~3% higher sequence identity than a previous method. Results from this study may benefit further development of computational protein design methods.

Metrics & MoreArticle Recommendations CONSPECTUS:The fine design and regulation of catalysts play critical roles in the development of catalysis. The microenvironment, which gives rise to unique spatial structures and electronic properties around catalytic sites, has been proven to dramatically regulate catalytic behavior in enzymes and homogeneous catalysis. However, understanding the microenvironment modulation (MEM) of catalytic sites remains challenging and very limited in heterogeneous catalysis mainly due to the lack of structural precision and/or tailorability of traditional solid catalysts. Among diverse materials, metal−organic frameworks (MOFs), a class of porous crystalline solids, have been intensively studied as heterogeneous catalysts in recent years. The atomically precise and well tunable structures of MOFs make them an ideal platform for rationally regulating the microenvironment surrounding catalytic sites. Accordingly, their well-defined structures hold great promise for elucidating how the microenvironment modulation affects the resulting catalytic performance. Nevertheless, the investigations of accurate control over the microenvironment of catalytic sites in MOFs for modulated catalysis are still very limited. Therefore, it is of great importance to summarize the related results and provide in-depth insights into microenvironment modulation in MOF-based catalysis, accelerating the future development of this emerging research topic.In this Account, we have presented a summary of our recent attempts to optimize the catalytic performance of MOF-based materials via microenvironment modulation. In view of the unique component and structural advantages of MOFs, we deliver the general fundamentals for rational control over the microenvironment in MOF-based catalysis. Initially, the great opportunities brought about by MOFs for accurate control over microenvironment engineering, including the origin of abundant active sites, flexible regulation strategies, and well-defined structure, are introduced in detail. In the next section, we focus on the specific strategies of microenvironment modulation in MOF-based catalysis, which dominate the molecular/electron-transfer process and regulate the intrinsic activity of catalytic sites. Meanwhile, the related chemical basis and underlying structure−property relationship behind the enhanced catalytic performance will be highlighted. Finally, the major challenges and future outlooks on the microenvironment modulation in MOF-based catalysis will be further discussed. It is expected that this Account would provide an understanding of the importance of microenvironment modulation around catalytic sites in MOF-based catalysts and afford significant inspiration toward enhanced performance by microenvironment engineering in heterogeneous catalysis.