Revealing the grammar of small RNA secretion using interpretable machine learning.

Small non-coding RNAs can be secreted through a variety of mechanisms, including exosomal sorting, in small extracellular vesicles, and within lipoprotein complexes. However, the mechanisms that govern their sorting and secretion are not well understood. Here, we present ExoGRU, a machine learning model that predicts small RNA secretion probabilities from primary RNA sequences. We experimentally validated the performance of this model through ExoGRU-guided mutagenesis and synthetic RNA sequence analysis. Additionally, we used ExoGRU to reveal cis and trans factors that underlie small RNA secretion, including known and novel RNA-binding proteins (RBPs), e.g., YBX1, HNRNPA2B1, and RBM24. We also developed a novel technique called exoCLIP, which reveals the RNA interactome of RBPs within the cell-free space. Together, our results demonstrate the power of machine learning in revealing novel biological mechanisms. In addition to providing deeper insight into small RNA secretion, this knowledge can be leveraged in therapeutic and synthetic biology applications.

A stress-induced tyrosine-tRNA depletion response mediates codon-based translational repression and growth suppression.

Eukaryotic transfer RNAs can become selectively fragmented upon various stresses, generating tRNA-derived small RNA fragments. Such fragmentation has been reported to impact a small fraction of the tRNA pool and thus presumed to not directly impact translation. We report that oxidative stress can rapidly generate tyrosine-tRNAGUA fragments in human cells-causing significant depletion of the precursor tRNA. Tyrosine-tRNAGUA depletion impaired translation of growth and metabolic genes enriched in cognate tyrosine codons. Depletion of tyrosine tRNAGUA or its translationally regulated targets USP3 and SCD repressed proliferation-revealing a dedicated tRNA-regulated growth-suppressive pathway for oxidative stress response. Tyrosine fragments are generated in a DIS3L2 exoribonuclease-dependent manner and inhibit hnRNPA1-mediated transcript destabilization. Moreover, tyrosine fragmentation is conserved in C. elegans. Thus, tRNA fragmentation can coordinately generate trans-acting small RNAs and functionally deplete a tRNA. Our findings reveal the existence of an underlying adaptive codon-based regulatory response inherent to the genetic code.

Efficacy of Amphotericin B in Corneal Preservation Media After Extended Frozen Storage.

PURPOSE: To investigate the antimycotic activity of amphotericin B deoxycholate that has been previously frozen for 28 days before supplementation of Optisol-GS. METHODS: Triplicate Optisol-GS samples were inoculated with 10 colony-forming units (CFU) of Candida albicans. Each set of triplicate cultures was supplemented with 2.5 µg/mL of amphotericin B that was either freshly resuspended and never frozen, frozen overnight at -20°C and thawed, or frozen at -20°C for 4 weeks and thawed. The cultures were stored at 4°C, with aliquots taken at 0, 6, 24, and 72 hours for quantification. The efficacy of each preparation of amphotericin B in reducing C. albicans growth was assessed at these time points. RESULTS: Six hours after antifungal supplementation, there was a 1.33 log10 CFU reduction with freshly resuspended amphotericin B, compared with a 1.31 log10 reduction with amphotericin B that was frozen overnight (P = 0.20) and a 1.18 log10 reduction with amphotericin B that was frozen for 4 weeks (P = 0.05). After 72 hours, there was a 2.72 log10 CFU reduction with freshly resuspended amphotericin B, a 2.64 log10 CFU reduction with amphotericin B that was frozen overnight (P = 0.45), and a 2.18 log10 CFU reduction with amphotericin B that was frozen for 4 weeks (P = 0.05). CONCLUSIONS: Previously frozen amphotericin B remains highly effective against C. albicans. Optisol-GS supplemented with 2.5 µg/mL amphotericin B that was frozen for 4 weeks at -20°C resulted in >90% CFU reduction by 6 hours and >99% reduction by 72 hours.

Identification of molecular determinants of primary and metastatic tumour re-initiation in breast cancer.

Through in vivo selection of multiple ER-negative human breast cancer populations for enhanced tumour-forming capacity, we have derived subpopulations that generate tumours more efficiently than their parental populations at low cell numbers. Tumorigenic-enriched subpopulations exhibited increased expression of LAMA4, FOXQ1 and NAP1L3­genes that are also expressed at greater levels by independently derived metastatic subpopulations. These genes promote metastatic efficiency. FOXQ1 promotes LAMA4 expression, and LAMA4 enhances clonal expansion following substratum detachment in vitro, tumour re-initiation in multiple organs, and disseminated metastatic cell proliferation and colonization. The promotion of cancer cell proliferation and tumour re-initiation by LAMA4 requires ß1-integrin. Increased LAMA4 expression marks the transition of human pre-malignant breast lesions to malignant carcinomas, and tumoral LAMA4 overexpression predicts reduced relapse-free survival in ER-negative patients. Our findings reveal common features that govern primary and metastatic tumour re-initiation and identify a key molecular determinant of these processes.

Architecture of the soluble receptor Aer2 indicates an in-line mechanism for PAS and HAMP domain signaling.

Bacterial receptors typically contain modular architectures with distinct functional domains that combine to send signals in response to stimuli. Although the properties of individual components have been investigated in many contexts, there is little information about how diverse sets of modules work together in full-length receptors. Here, we investigate the architecture of Aer2, a soluble gas-sensing receptor that has emerged as a model for PAS (Per-Arnt-Sim) and poly-HAMP (histidine kinase-adenylyl cyclase-methyl-accepting chemotaxis protein-phosphatase) domain signaling. The crystal structure of the heme-binding PAS domain in the ferric, ligand-free form, in comparison to the previously determined cyanide-bound state, identifies conformational changes induced by ligand binding that are likely essential for the signaling mechanism. Heme-pocket alternations share some similarities with the heme-based PAS sensors FixL and EcDOS but propagate to the Iß strand in a manner predicted to alter PAS-PAS associations and the downstream HAMP junction within full-length Aer2. Small-angle X-ray scattering of PAS and poly-HAMP domain fragments of increasing complexity allow unambiguous domain assignments and reveal a linear quaternary structure. The Aer2 PAS dimeric crystal structure fits well within ab initio small-angle X-ray scattering molecular envelopes, and pulsed dipolar ESR measurements of inter-PAS distances confirm the crystallographic PAS arrangement within Aer2. Spectroscopic and pull-down assays fail to detect direct interactions between the PAS and HAMP domains. Overall, the Aer2 signaling mechanism differs from the Escherichia coli Aer paradigm, where side-on PAS-HAMP contacts are key. We propose an in-line model for Aer2 signaling, where ligand binding induces alterations in PAS domain structure and subunit association that is relayed through the poly-HAMP junction to downstream domains.

Convergent multi-miRNA targeting of ApoE drives LRP1/LRP8-dependent melanoma metastasis and angiogenesis.

Through in vivo selection of human cancer cell populations, we uncover a convergent and cooperative miRNA network that drives melanoma metastasis. We identify miR-1908, miR-199a-5p, and miR-199a-3p as endogenous promoters of metastatic invasion, angiogenesis, and colonization in melanoma. These miRNAs convergently target apolipoprotein E (ApoE) and the heat shock factor DNAJA4. Cancer-secreted ApoE suppresses invasion and metastatic endothelial recruitment (MER) by engaging melanoma cell LRP1 and endothelial cell LRP8 receptors, respectively, while DNAJA4 promotes ApoE expression. Expression levels of these miRNAs and ApoE correlate with human metastatic progression outcomes. Treatment of cells with locked nucleic acids (LNAs) targeting these miRNAs inhibits metastasis to multiple organs, and therapeutic delivery of these LNAs strongly suppresses melanoma metastasis. We thus identify miRNAs with dual cell-intrinsic/cell-extrinsic roles in cancer, reveal convergent cooperativity in a metastatic miRNA network, identify ApoE as an anti-angiogenic and metastasis-suppressive factor, and uncover multiple prognostic miRNAs with synergistic combinatorial therapeutic potential in melanoma.