High-throughput developability assays enable library-scale identification of producible protein scaffold variants

Abstract

Proteins require high developability—quantified by expression, solubility, and stability—for robust utility as therapeutics, diagnostics, and in other biotechnological applications. Measuring traditional developability metrics is low throughput in nature, often slowing the developmental pipeline. We evaluated the ability of 10 variations of three high-throughput developability assays to predict the bacterial recombinant expression of paratope variants of the protein scaffold Gp2. Enabled by a phenotype/genotype linkage, assay performance for 105 variants was calculated via deep sequencing of populations sorted by proxied developability. We identified the most informative assay combination via cross-validation accuracy and correlation feature selection and demonstrated the ability of machine learning models to exploit nonlinear mutual information to increase the assays’ predictive utility. We trained a random forest model that predicts expression from assay performance that is 35% closer to the experimental variance and trains 80% more efficiently than a model predicting from sequence information alone. Utilizing the predicted expression, we performed a site-wise analysis and predicted mutations consistent with enhanced developability. The validated assays offer the ability to identify developable proteins at unprecedented scales, reducing the bottleneck of protein commercialization.

Mega-scale experimental analysis of protein folding stability in biology and protein design

“…Proteases typically cleave unfolded proteins more quickly than folded ones, and proteolysis assays have been used for decades to measure folding stability (37) and select for high stability proteins (38,39). In 2017, we introduced the high-throughput yeast display proteolysis method for measuring folding stability using next generation sequencing (29,(40)(41)(42)(43)(44)(45)(46). To improve the scale, precision, speed, and cost of stability measurements, we developed cDNA display proteolysis.…”

Deep Antimicrobial Activity and Stability Analysis Inform Lysin Sequence–Function Mapping

“…In our previous work, V4 and V14 were the only tested variants with expression yields <0.2 mg/L (unpublished data affiliated with ref ). Additionally, while neither of these variants were tested for thermal stability, previous literature has demonstrated a correlation between stability and producibility . Here, a correlation is observed between lysin CDs of comparable producibility; so less stable or less producible molecules may inhibit growth when expressed in this assay via mechanisms unrelated to catalytic activity.…”

“…Naturally occurring antimicrobial peptides and proteins offer compelling starting points for developing therapies, but they often require engineering to improve activity, stability, solubility, or other molecular properties to ultimately be translatable . While the field has developed HT assays for screening large (10 6 –10 9 variants) protein libraries for stability, , developability, and other functions of interest, − screening of antimicrobial activity has remained mostly limited given the complex and diverse modes of antimicrobial activity. − The complex nature of antimicrobial activity and the limited quantity of activity data have further limited the development of statistical models to predict activity and inform the design of improved variants.…”

A Platform for Deep Sequence–Activity Mapping and Engineering Antimicrobial Peptides

“…Simultaneous evaluation of these mutually important elements aids efficiency yet also requires deconvolution if a particular enhanced mechanism is desired. Other techniques could assess biophysical metrics including stability via deep sequencing of flow cytometrically sorted yeast displayed protein exposed to protease , and expression via a library-scale split green fluorescent protein assay or dot blots for midthroughput.…”