Formation of synthetic RNA protein granules using engineered phage-coat-protein -RNA complexes.

Liquid-solid transition, also known as gelation, is a specific form of phase separation in which molecules cross-link to form a highly interconnected compartment with solid - like dynamical properties. Here, we utilize RNA hairpin coat-protein binding sites to form synthetic RNA based gel-like granules via liquid-solid phase transition. We show both in-vitro and in-vivo that hairpin containing synthetic long non-coding RNA (slncRNA) molecules granulate into bright localized puncta. We further demonstrate that upon introduction of the coat-proteins, less-condensed gel-like granules form with the RNA creating an outer shell with the proteins mostly present inside the granule. Moreover, by tracking puncta fluorescence signals over time, we detected addition or shedding events of slncRNA-CP nucleoprotein complexes. Consequently, our granules constitute a genetically encoded storage compartment for protein and RNA with a programmable controlled release profile that is determined by the number of hairpins encoded into the RNA. Our findings have important implications for the potential regulatory role of naturally occurring granules and for the broader biotechnology field.

A Cell-Free Assay for Rapid Screening of Inhibitors of hACE2-Receptor-SARS-CoV-2-Spike Binding.

We present a cell-free assay for rapid screening of candidate inhibitors of protein binding, focusing on inhibition of the interaction between the SARS-CoV-2 Spike receptor binding domain (RBD) and human angiotensin-converting enzyme 2 (hACE2). The assay has two components: fluorescent polystyrene particles covalently coated with RBD, termed virion-particles (v-particles), and fluorescently labeled hACE2 (hACE2F) that binds the v-particles. When incubated with an inhibitor, v-particle-hACE2F binding is diminished, resulting in a reduction in the fluorescent signal of bound hACE2F relative to the noninhibitor control, which can be measured via flow cytometry or fluorescence microscopy. We determine the amount of RBD needed for v-particle preparation, v-particle incubation time with hACE2F, hACE2F detection limit, and specificity of v-particle binding to hACE2F. We measure the dose response of the v-particles to known inhibitors. Finally, utilizing an RNA-binding protein tdPP7 incorporated into hACE2F, we demonstrate that RNA-hACE2F granules trap v-particles effectively, providing a basis for potential RNA-hACE2F therapeutics.

Overcoming the design, build, test bottleneck for synthesis of nonrepetitive protein-RNA cassettes.

We apply an oligo-library and machine learning-approach to characterize the sequence and structural determinants of binding of the phage coat proteins (CPs) of bacteriophages MS2 (MCP), PP7 (PCP), and Qß (QCP) to RNA. Using the oligo library, we generate thousands of candidate binding sites for each CP, and screen for binding using a high-throughput dose-response Sort-seq assay (iSort-seq). We then apply a neural network to expand this space of binding sites, which allowed us to identify the critical structural and sequence features for binding of each CP. To verify our model and experimental findings, we design several non-repetitive binding site cassettes and validate their functionality in mammalian cells. We find that the binding of each CP to RNA is characterized by a unique space of sequence and structural determinants, thus providing a more complete description of CP-RNA interaction as compared with previous low-throughput findings. Finally, based on the binding spaces we demonstrate a computational tool for the successful design and rapid synthesis of functional non-repetitive binding-site cassettes.

Single-Particle Diffusion Characterization by Deep Learning.

Diffusion plays a crucial role in many biological processes including signaling, cellular organization, transport mechanisms, and more. Direct observation of molecular movement by single-particle-tracking experiments has contributed to a growing body of evidence that many cellular systems do not exhibit classical Brownian motion but rather anomalous diffusion. Despite this evidence, characterization of the physical process underlying anomalous diffusion remains a challenging problem for several reasons. First, different physical processes can exist simultaneously in a system. Second, commonly used tools for distinguishing between these processes are based on asymptotic behavior, which is experimentally inaccessible in most cases. Finally, an accurate analysis of the diffusion model requires the calculation of many observables because different transport modes can result in the same diffusion power-law α, which is typically obtained from the mean-square displacements (MSDs). The outstanding challenge in the field is to develop a method to extract an accurate assessment of the diffusion process using many short trajectories with a simple scheme that is applicable at the nonexpert level. Here, we use deep learning to infer the underlying process resulting in anomalous diffusion. We implement a neural network to classify single-particle trajectories by diffusion type: Brownian motion, fractional Brownian motion and continuous time random walk. Further, we demonstrate the applicability of our network architecture for estimating the Hurst exponent for fractional Brownian motion and the diffusion coefficient for Brownian motion on both simulated and experimental data. These networks achieve greater accuracy than time-averaged MSD analysis on simulated trajectories while only requiring as few as 25 steps. When tested on experimental data, both net and ensemble MSD analysis converge to similar values; however, the net needs only half the number of trajectories required for ensemble MSD to achieve the same confidence interval. Finally, we extract diffusion parameters from multiple extremely short trajectories (10 steps) using our approach.

Designing Bacterial Chemotactic Receptors Guided by Photonic Femtoliter Well Arrays for Quantifiable, Label-Free Measurement of Bacterial Chemotaxis.

Whole cell bioreporters, such as bacterial cells, can be used for environmental and clinical sensing of specific analytes. However, the current methods implemented to observe such bioreporters in the form of chemotactic responses heavily rely on microscope analysis, fluorescent labels, and hard-to-scale microfluidic devices. Herein, we demonstrate that chemotaxis can be detected within minutes using intrinsic optical measurements of silicon femtoliter well arrays (FMAs). This is done via phase-shift reflectometric interference spectroscopic measurements (PRISM) of the wells, which act as silicon diffraction gratings, enabling label-free, real-time quantification of the number of trapped bacteria cells in the optical readout. By generating unsteady chemical gradients over the wells, we first demonstrate that chemotaxis toward attractants and away from repellents can be easily differentiated based on the signal response of PRISM. The lowest concentration of chemorepellent to elicit an observed bacterial response was 50 mM, whereas the lowest concentration of chemoattractant to elicit a response was 10 mM. Second, we employed PRISM, in combination with a computational approach, to rapidly scan for and identify a novel synthetic histamine chemoreceptor strain. Consequently, we show that by using a combined computational design approach, together with a quantitative, real-time, and label-free detection method, it is possible to manufacture and characterize novel synthetic chemoreceptors in Escherichia coli (E. coli).