Diversifying Design of Nucleic Acid Aptamers Using Unsupervised Machine Learning.

Inverse design of short single-stranded RNA and DNA sequences (aptamers) is the task of finding sequences that satisfy a set of desired criteria. Relevant criteria may be, for example, the presence of specific folding motifs, binding to molecular ligands, sensing properties, and so on. Most practical approaches to aptamer design identify a small set of promising candidate sequences using high-throughput experiments (e.g., SELEX) and then optimize performance by introducing only minor modifications to the empirically found candidates. Sequences that possess the desired properties but differ drastically in chemical composition will add diversity to the search space and facilitate the discovery of useful nucleic acid aptamers. Systematic diversification protocols are needed. Here we propose to use an unsupervised machine learning model known as the Potts model to discover new, useful sequences with controllable sequence diversity. We start by training a Potts model using the maximum entropy principle on a small set of empirically identified sequences unified by a common feature. To generate new candidate sequences with a controllable degree of diversity, we take advantage of the model's spectral feature: an "energy" bandgap separating sequences that are similar to the training set from those that are distinct. By controlling the Potts energy range that is sampled, we generate sequences that are distinct from the training set yet still likely to have the encoded features. To demonstrate performance, we apply our approach to design diverse pools of sequences with specified secondary structure motifs in 30-mer RNA and DNA aptamers.

High-Throughput Strategy for Glycine Oxidase Biosensor Development Reveals Glycine Release from Cultured Cells.

Glycine is an important biomarker in clinical analysis due to its involvement in multiple physiological processes. As such, the need for low-cost analytical tools for glycine detection is growing. As a neurotransmitter, glycine is involved in inhibitory and excitatory neurochemical transmission in the central nervous system. In this work, we present a 10 µM Pt-based electrochemical enzymatic biosensor based on the flavoenzyme glycine oxidase (GO) for localized real-time measurements of glycine. Among GO variants at position 244, the H244K variant with increased glycine turnover was selected to develop a functional biosensor. This biosensor relies on amperometric readouts and does not require additional redox mediators. The biosensor was characterized and applied for glycine detection from cells, mainly HEK 293 cells and primary rat astrocytes. We have identified an enzyme, GO H244K, with increased glycine turnover using mutagenesis but which can be developed into a functional biosensor. Noteworthy, a glycine release of 395.7 ± 123 µM from primary astrocytes was measured, which is ∼fivefold higher than glycine release from HEK 293 cells (75.4 ± 3.91 µM) using the GO H244K biosensor.

Quantitative measurements of free and immobilized RgDAAO Michaelis-Menten constant using an electrochemical assay reveal the impact of covalent cross-linking on substrate specificity.

Challenges facing enzyme-based electrochemical sensors include substrate specificity, batch to batch reproducibility, and lack of quantitative metrics related to the effect of enzyme immobilization. We present a quick, simple, and general approach for measuring the effect of immobilization and cross-linking on enzyme activity and substrate specificity. The method can be generalized for electrochemical biosensors using an enzyme that releases hydrogen peroxide during its catalytic cycle. Using as proof of concept RgDAAO-based electrochemical biosensors, we found that the Michaelis-Menten constant (Km) decreases post immobilization, hinting at alterations in the enzyme kinetic properties and thus substrate specificity. We confirm the decrease in Km electrochemically by characterizing the substrate specificity of the immobilized RgDAAO using chronoamperometry. Our results demonstrate that enzyme immobilization affects enzyme substrate specificity and this must be carefully evaluated during biosensor development.

Enhancing Electrochemical Biosensor Selectivity with Engineered d-Amino Acid Oxidase Enzymes for d-Serine and d-Alanine Quantification.

d-Amino acid oxidase (DAAO) enzymes bind a range of d-amino acids with variable affinity. As such, the design of selective DAAO-based enzymatic biosensors remains a challenge for real-world biosensor application. Herein, a methodology for developing biosensors with varying substrate selectivity is presented. First, we address DAAO-based biosensor selectivity toward d-serine by introducing point mutations into DAAO using rational design. Next, the wild-type yeast DAAO (RgDAAO WT) and variants human DAAO W209R and yeast M213G are characterized for their selectivity and activity toward d-serine and d-alanine, the preferred DAAO substrates. The DAAO enzymes have been immobilized for final biosensor design, where they demonstrate selectivity comparable to free DAAO. The cross-linking procedure impacts on DAAO structure and function and the use of a regeneration strategy allows the biosensor response to be improved.