Natural membrane vesicles (MVs) derived from various types of cells play an essential role in transporting biological materials between cells. Here, we show that exogenous compounds are packaged in the MVs by engineering the parental cells via liposomes, and the MVs mediate autonomous intercellular migration of the compounds through multiple cancer cell layers. Hydrophobic compounds delivered selectively to the plasma membrane of cancer cells using synthetic membrane fusogenic liposomes were efficiently incorporated into the membrane of MVs secreted from the cells and then transferred to neighboring cells via the MVs. This liposome-mediated MV engineering strategy allowed hydrophobic photosensitizers to significantly penetrate both spheroids and in vivo tumors, thereby enhancing the therapeutic efficacy. These results suggest that innate biological transport systems can be in situ engineered via synthetic liposomes to guide the penetration of chemotherapeutics across challenging tissue barriers in solid tumors.

Using in situ transmission electron microscopy, we demonstrated that gold nanoparticles are unified via "oriented attachment" assisted either by nanoparticle rotation or grain boundary migration at the attachment interface. We also observed that the combined nanoparticle changes shape with stable facet planes via surface diffusion, along with recrystallization.

Engineering of extracellular vesicles (EVs) without affecting biological functions remains a challenge, limiting the broad applications of EVs in biomedicine. Here, we report a method to equip EVs with various functional agents, including fluorophores, drugs, lipids, and bio-orthogonal chemicals, in an efficient and controlled manner by engineering parental cells with membrane fusogenic liposomes, while keeping the EVs intact. As a demonstration of how this method can be applied, we prepared EVs containing azide-lipids, and conjugated them with targeting peptides using copper-free click chemistry to enhance targeting efficacy to cancer cells. We believe that this liposome-based cellular engineering method will find utility in studying the biological roles of EVs and delivering therapeutic agents through their innate pathway.

For practical device applications, monolayer transition metal dichalcogenide (TMD) films must meet key industry needs for batch processing, including the high‐throughput, large‐scale production of high‐quality, spatially uniform materials, and reliable integration into devices. Here, high‐throughput growth, completed in 12 min, of 6‐inch wafer‐scale monolayer MoS2 and WS2 is reported, which is directly compatible with scalable batch processing and device integration. Specifically, a pulsed metal–organic chemical vapor deposition process is developed, where periodic interruption of the precursor supply drives vertical Ostwald ripening, which prevents secondary nucleation despite high precursor concentrations. The as‐grown TMD films show excellent spatial homogeneity and well‐stitched grain boundaries, enabling facile transfer to various target substrates without degradation. Using these films, batch fabrication of high‐performance field‐effect transistor (FET) arrays in wafer‐scale is demonstrated, and the FETs show remarkable uniformity. The high‐throughput production and wafer‐scale automatable transfer will facilitate the integration of TMDs into Si‐complementary metal‐oxide‐semiconductor platforms.

Background: VRK1 phosphorylates mitotic histone H3 at Thr-3 and Ser-10, but its negative regulator was not elucidated during interphase. Results: The macrodomain of macroH2A1 interacts with VRK1, and this suppresses enzymatic activity of VRK1 during interphase. Conclusion: Specific binding between VRK1 and macroH2A1 is required to regulate the cell cycle-dependent histone H3 phosphorylation. Significance: Understanding epigenetic regulation of histone H3 during the cell cycle is important in cancer development.

No abstract

In this paper, we present a semi-supervised method for automatic speech act recognition in email and forums. The major challenge of this task is due to lack of labeled data in these two genres. Our method leverages labeled data in the Switchboard-DAMSL and the Meeting Recorder Dialog Act database and applies simple domain adaptation techniques over a large amount of unlabeled email and forum data to address this problem. Our method uses automatically extracted features such as phrases and dependency trees, called subtree features, for semi-supervised learning. Empirical results demonstrate that our model is effective in email and forum speech act recognition.

In natural language understanding (NLU), a user utterance can be labeled differently depending on the domain or application (e.g., weather vs. calendar). Standard domain adaptation techniques are not directly applicable to take advantage of the existing annotations because they assume that the label set is invariant. We propose a solution based on label embeddings induced from canonical correlation analysis (CCA) that reduces the problem to a standard domain adaptation task and allows use of a number of transfer learning techniques. We also introduce a new transfer learning technique based on pretraining of hidden-unit CRFs (HUCRFs). We perform extensive experiments on slot tagging on eight personal digital assistant domains and demonstrate that the proposed methods are superior to strong baselines.