In many cases, exploration of orthologous protein sequences can help identify amino acid residues or regions which adversely affect protein expression, purification, or stability. By generating chimeras of orthologous proteins, some of these challenges can be overcome while minimizing the dramatic impact of using a complete orthologous sequence. Case studies will be presented of very limited amino acid changes leading to significant improvements in protein yield and quality.

LifeArc’s therapeutic platforms include antibody discovery, humanization, and small molecule and PROTAC development. Protein production can be a key challenge, with intractable targets posing significant risks to project progression. To improve success rates for a challenging target class, we implemented a funnelled strategy taking multiple potential candidate proteins into expression and purification in parallel. This approach mitigated target-specific risks and enabled early identification of tractable proteins.

Complex membrane protein lead generation is challenging due to the inherent characteristics of the target, including low expression, conformational flexibility and dependence on lipids for solubility. In response, we have deployed unilamellar vesicles and nanodiscs as antigen presenting formats for multi-pass transmembrane targets. Herein, we show that these formats preserve relevant complex membrane protein features and are amenable to in vitro and in vivo lead generation workflows.

GPCRs are highly druggable but technically challenging targets for recombinant protein production and binding-first hit discovery. DEL screening requires GPCR reagents with appropriate stability, ligand accessibility, and conformational control. In this talk, I will describe some of our successful strategies for enabling GPCR-focused DEL campaigns through construct design, expression testing, purification, stabilization, and validation, with emphasis on membrane mimetics such as SMALPs and other detergent-free formulations for screening and hit follow-up.

Membrane scaffold protein (MSP) production is a key determinant of successful lipid nanodiscs assembly but is often limited by proteolysis and polymerization. We present an optimized MSP expression and purification protocol coupled with a rapid PD-10 desalting column-based reconstitution method that streamlines nanodiscs assembly, improves sample yield and quality, and enables efficient reconstitution of the two prototypical membrane proteins, KcsA and ABCB10, for structural and biophysical applications.

Antibody and binder discovery scales only as fast as the antigens feeding it, yet construct design stays manual and per-target. Across the antibody-accessible human surfaceome and secretome, more than 5,000 targets, OpenAntigens.org computes construct boundaries from topology, domains, and deposited PDB precedent, guided by AlphaFold structure. It flags developability liabilities like free cysteines, furin sites, and glycosylation. It scores cross-reactivity both ways: mouse and cynomolgus orthologs you want for preclinical work, paralogs you usually do not. Disease associations add context, and every design exports as copy-ready FASTA and TSV. All open and free.

The RAS-binding domain (RBD) and cysteine-rich domain (CRD) of RAF1 kinases are key to RAS GTPases, and activation of the mitogen-activated protein kinase (MAPK) pathway. This activation occurs via interaction of a GTP-bound RAS with the RAF1 RBDCRD domains. This interaction is critical to signaling, which controls a host of basic cell function and whose mutation drives many cancers. Here, we report improvements to both yield via expression medium, and protein solubility via buffers of 15N labeled and unlabeled RAF1 RBDCRD that give significantly higher yields and highly soluble protein. These protocols will be helpful for drug discovery efforts.

Engineering lentiviral payloads for solid tumors requires balancing antigen heterogeneity, tumor microenvironments, multicistronic design tradeoffs, and delivery constraints. This talk presents a framework for designing and testing multispecific transfer plasmids in primary T cells, emphasizing how payload architecture (e.g. promoters, codon optimization) influences recombinant protein expression, and viral construct characteristics. Our results argue for integrated datasets to enable and accelerate the development of machine learning-guided design of cell therapy payloads.