AlphaFold predictions of fold-switched conformations are driven by structure memorization.

Recent work suggests that AlphaFold (AF)-a deep learning-based model that can accurately infer protein structure from sequence-may discern important features of folded protein energy landscapes, defined by the diversity and frequency of different conformations in the folded state. Here, we test the limits of its predictive power on fold-switching proteins, which assume two structures with regions of distinct secondary and/or tertiary structure. We find that (1) AF is a weak predictor of fold switching and (2) some of its successes result from memorization of training-set structures rather than learned protein energetics. Combining >280,000 models from several implementations of AF2 and AF3, a 35% success rate was achieved for fold switchers likely in AF's training sets. AF2's confidence metrics selected against models consistent with experimentally determined fold-switching structures and failed to discriminate between low and high energy conformations. Further, AF captured only one out of seven experimentally confirmed fold switchers outside of its training sets despite extensive sampling of an additional ~280,000 models. Several observations indicate that AF2 has memorized structural information during training, and AF3 misassigns coevolutionary restraints. These limitations constrain the scope of successful predictions, highlighting the need for physically based methods that readily predict multiple protein conformations.

Disordered proteins mitigate the temperature dependence of site-specific binding free energies.

Biophysical characterization of protein-protein interactions involving disordered proteins is challenging. A common simplification is to measure the thermodynamics and kinetics of disordered site binding using peptides containing only the minimum residues necessary. We should not assume, however, that these few residues tell the whole story. Son of sevenless, a multidomain signaling protein from Drosophila melanogaster, is critical to the mitogen-activated protein kinase pathway, passing an external signal to Ras, which leads to cellular responses. The disordered 55 kDa C-terminal domain of Son of sevenless is an autoinhibitor that blocks guanidine exchange factor activity. Activation requires another protein, Downstream of receptor kinase (Drk), which contains two Src homology 3 domains. Here, we utilized NMR spectroscopy and isothermal titration calorimetry to quantify the thermodynamics and kinetics of the N-terminal Src homology 3 domain binding to the strongest sites incorporated into the flanking disordered sequences. Comparing these results to those for isolated peptides provides information about how the larger domain affects binding. The affinities of sites on the disordered domain are like those of the peptides at low temperatures but less sensitive to temperature. Our results, combined with observations showing that intrinsically disordered proteins become more compact with increasing temperature, suggest a mechanism for this effect.

Development and optimization of an in vivo electrocardiogram recording method and analysis program for adult zebrafish.

Clinically pertinent electrocardiogram (ECG) data from model systems, such as zebrafish, are crucial for illuminating factors contributing to human cardiac electrophysiological abnormalities and disease. Current zebrafish ECG collection strategies have not adequately addressed the consistent acquisition of high-quality traces or sources of phenotypic variation that could obscure data interpretation. Thus, we developed a novel platform to ensure high-quality recording of in vivo subdermal adult zebrafish ECGs and zebrafish ECG reading GUI (zERG), a program to acquire measurements from traces that commercial software cannot examine owing to erroneous peak calling. We evaluate normal ECG trait variation, revealing highly reproducible intervals and wave amplitude variation largely driven by recording artifacts, and identify sex and body size as potential confounders to PR, QRS and QT intervals. With this framework, we characterize the effect of the class I anti-arrhythmic drug flecainide acetate on adults, provide support for the impact of a Long QT syndrome model, and establish power calculations for this and other studies. These results highlight our pipeline as a robust approach to evaluate zebrafish models of human cardiac electrophysiological phenotypes.

Danio rerio Oocytes for Eukaryotic In-Cell NMR.

Understanding how the crowded and complex cellular milieu affects protein stability and dynamics has only recently become possible by using techniques such as in-cell nuclear magnetic resonance. However, the combination of stabilizing and destabilizing interactions makes simple predictions difficult. Here we show the potential of Danio rerio oocytes as an in-cell nuclear magnetic resonance model that can be widely used to measure protein stability and dynamics. We demonstrate that in eukaryotic oocytes, which are 3-6-fold less crowded than other cell types, attractive chemical interactions still dominate effects on protein stability and slow tumbling times, compared to the effects of dilute buffer.