Search for: All records

Starvation causes the accumulation of lipid droplets in the liver, a somewhat counterintuitive phenomenon that is nevertheless conserved from flies to humans. Much like fatty liver resulting from overfeeding, hepatic lipid accumulation (steatosis) during undernourishment can lead to lipotoxicity and atrophy of the liver. Here, we found that although surface populations ofAstyanax mexicanusundergo this evolutionarily conserved response to starvation, the starvation-resistant cavefish larvae of the same species do not display an accumulation of lipid droplets upon starvation. Moreover, cavefish are resistant to liver atrophy during starvation, providing a unique system to explore strategies for liver protection. Using comparative transcriptomics between zebrafish, surface fish, and cavefish, we identified the fatty acid transporter slc27a2a/fatp2 to be correlated with the development of fatty liver. Pharmacological inhibition of slc27a2a in zebrafish rescues steatosis and atrophy of the liver upon starvation. Furthermore, down-regulation of FATP2 in Drosophila larvae inhibits the development of starvation-induced steatosis, suggesting the evolutionarily conserved importance of the gene in regulating fatty liver upon nutrition deprivation. Overall, our study identifies a conserved, druggable target to protect the liver from atrophy during starvation. 

Replicating human-like dexterity in robot hands represents one of the largest open problems in robotics. Reinforcement learning is a promising approach that has achieved impressive progress in the last few years; however, the class of problems it has typically addressed corresponds to a rather narrow definition of dexterity as compared to human capabilities. To address this gap, we investigate piano-playing, a skill that challenges even the human limits of dexterity, as a means to test high-dimensional control, and which requires high spatial and temporal precision, and complex finger coordination and planning. We introduce RoboPianist, a system that enables simulated anthropomorphic hands to learn an extensive repertoire of 150 piano pieces where traditional model-based optimization struggles. We additionally introduce an open-sourced environment, benchmark of tasks, interpretable evaluation metrics, and open challenges for future study. 

Large language models (LLMs) exhibit a wide range of promising capabilities -- from step-by-step planning to commonsense reasoning -- that may provide utility for robots, but remain prone to confidently hallucinated predictions. In this work, we present KnowNo, which is a framework for measuring and aligning the uncertainty of LLM-based planners such that they know when they don't know and ask for help when needed. KnowNo builds on the theory of conformal prediction to provide statistical guarantees on task completion while minimizing human help in complex multi-step planning settings. Experiments across a variety of simulated and real robot setups that involve tasks with different modes of ambiguity (e.g., from spatial to numeric uncertainties, from human preferences to Winograd schemas) show that KnowNo performs favorably over modern baselines (which may involve ensembles or extensive prompt tuning) in terms of improving efficiency and autonomy, while providing formal assurances. KnowNo can be used with LLMs out of the box without model-finetuning, and suggests a promising lightweight approach to modeling uncertainty that can complement and scale with the growing capabilities of foundation models. 

The tumor suppressor and master gene regulator protein p53 has been the subject of intense investigation for several decades due to its mutation in about half of all human cancers. However, mechanistic studies of p53 in cells are complicated by its many dynamic binding partners and heterogeneous post-translational modifications. The design of therapeutics that rescue p53 functions in cells requires a mechanistic understanding of its protein–protein interactions in specific protein complexes and identifying changes in p53 activity by diverse post-translational modifications. This review highlights the important roles that peptide and protein chemistry have played in biophysical and biochemical studies aimed at elucidating p53 regulation by several key binding partners. The design of various peptide inhibitors that rescue p53 function in cells and new opportunities in targeting p53-protein interactions are discussed. In addition, the review highlights the importance of a protein semisynthesis approach to comprehend the role of site-specific PTMs in p53 regulation. 

Abstract The proteasome, the primary protease for ubiquitin-dependent proteolysis in eukaryotes, is usually found as a mixture of 30S, 26S, and 20S complexes. These complexes have common catalytic sites, which makes it challenging to determine their distinctive roles in intracellular proteolysis. Here, we chemically synthesize a panel of homogenous ubiquitinated proteins, and use them to compare 20S and 26S proteasomes with respect to substrate selection and peptide-product generation. We show that 20S proteasomes can degrade the ubiquitin tag along with the conjugated substrate. Ubiquitin remnants on branched peptide products identified by LC-MS/MS, and flexibility in the 20S gate observed by cryo-EM, reflect the ability of the 20S proteasome to proteolyze an isopeptide-linked ubiquitin-conjugate. Peptidomics identifies proteasome-trapped ubiquitin-derived peptides and peptides of potential 20S substrates in Hi20S cells, hypoxic cells, and human failing-heart. Moreover, elevated levels of 20S proteasomes appear to contribute to cell survival under stress associated with damaged proteins.