Molecular wrench activity of DNA helicases: Keys to modulation of rapid kinetics in DNA repair.

DNA helicase activity is essential for the vital DNA metabolic processes of recombination, replication, transcription, translation, and repair. Recently, an unexpected, rapid exponential ATP-stimulated DNA unwinding rate was observed from an Archaeoglobus fulgidus helicase (AfXPB) as compared to the slower conventional helicases from Sulfolobus tokodaii, StXPB1 and StXPB2. This unusual rapid activity suggests a "molecular wrench" mechanism arising from the torque applied by AfXPB on the duplex structure in transitioning from open to closed conformations. However, much remains to be understood. Here, we investigate the concentration dependence of DNA helicase binding and ATP-stimulated kinetics of StXPB2 and AfXPB, as well as their binding and activity in Bax1 complexes, via an electrochemical assay with redox-active DNA monolayers. StXPB2 ATP-stimulated activity is concentration-independent from 8 to 200 nM. Unexpectedly, AfXPB activity is concentration-dependent in this range, with exponential rate constants varying from seconds at concentrations greater than 20 nM to thousands of seconds at lower concentrations. At 20 nM, rapid exponential signal decay ensues, linearly reverses, and resumes with a slower exponential decay. This change in AfXPB activity as a function of its concentration is rationalized as the crossover between the fast molecular wrench and slower conventional helicase modes. AfXPB-Bax1 inhibits rapid activity, whereas the StXPB2-Bax1 complex induces rapid kinetics at higher concentrations. This activity is rationalized with the crystal structures of these complexes. These findings illuminate the different physical models governing molecular wrench activity for improved biological insight into a key factor in DNA repair.

Machine Learning for Estimating Electron Transfer Rates From Square Wave Voltammetry.

Electrochemistry of surface-bound molecules is of high importance for numerous electronic and sensor applications. Extracting the electron transfer rate is beneficial for understanding surface-bound processes, but it requires experimental or computational rigor. We evaluate methods to determine electron transfer rates from large voltammetry sets from experiments via machine learning using decision tree ensembles, neural networks, and gaussian process regression models. We applied these to reproduce computational measures of electron transfer rates modeled by first principles. The best machine learning models were a random forest with 80 decision trees and a neural network with Bayesian regularization, producing root mean squared errors of 0.37 and 0.49 s-1 , respectively, corresponding to mean percent errors of 0.38 % and 0.52 %, respectively. This work establishes machine learning methods for rapidly acquiring electron transfer rates across large datasets for widespread applications.