IgG-like asymmetric bispecific antibodies are one of the challenging biotherapeutics—and successful scalable production might demand engineering of both heavy and light chains to facilitate correct pairing. Multiple engineering possibilities have been developed to increase the quality, quantity, and stability of bispecifics. In this case study we will be presenting the challenges encountered and the analytical strategies implemented to check product quality during expression, purification and formulation.

AI protein design strategies show promise for drug design—from de novo binder discovery to redesign of antibodies and antigens with improved properties. Even with best-in-class computational design, high-throughput laboratory methods are required to identify successful designs at scale. Here, we describe our experience producing and characterizing AI-designed proteins, including antibodies and solubilized multi-pass membrane protein proxies ("solMPMPs"), highlighting benefits, challenges, and the integration of computational and experimental methods.

Polymeric IgMs are abundantly secreted from plasma cells despite their structural complexity and intricate polymerization steps. To gain insights into IgM’s assembly mechanics that underwrite high-level secretion, we characterized IgM’s biosynthetic process by testing a series of mutant subunits that differentially disrupt secretion, folding, and specific inter-chain disulfide bond formation. The findings demonstrate the crucial role of underlying non-covalent protein-protein interactions in orchestrating the initial subunit interactions and maintaining the polymeric IgM product integrity during ER quality control steps, secretory pathway trafficking, and secretion. Insights obtained from this study provide a foundation for designing IgM-like multivalent binders.

Protein lipidation remains a largely untapped resource in therapeutic development, despite its prevalence regulating cell biology. Traditional semi-synthetic methods for protein lipidation often suffer from low yields, harsh reaction conditions, and can induce protein misfolding, complicating purification processes and limiting therapeutic potential. Our work addresses these challenges by genetically engineering prokaryotes to rapidly produce diverse libraries of lipidated proteins, enabling systematic investigation of lipid-driven protein behavior.

mRNA therapeutics is rapidly emerging as a groundbreaking strategy for treating cardiovascular diseases (CVD), which affect 650 million people. Despite advances in medicine, the need for curative therapies remains urgent. I will share our decade of work on mRNA therapies that promote cardiac regeneration—and combat fibrosis, cell death, and hypertrophy in CVD animal models. Additionally, we introduce novel cell-specific mRNA expression platforms, advancing the field of CVD therapeutics.

Unlike other biopolymers, central dogma describes glycosylation as non-template biosynthesis. Without a genetic encoding, glycan impact on biology is opaque. Here, we challenge template-free glycosylation and describe protein-encoded rules for glycan biosynthesis. We quantified glycan-protein associations and used these associations to predict glycosylation in HIV, SARS-CoV-2, and multiple immunoglobulins. Next, we reformulated these associations as an engineering strategy, leveraging many amino acid substitutions that minimally change protein structure but significantly impact glycosylation.  With this genetic encoding for glycans, we can integrate the siloed practices of glycan and protein engineering into a unified process of glycoprotein engineering.

We applied AI-assisted, iterative in silico approaches to stabilize RSV and influenza antigens for mRNA vaccines. Abrogated protein expression and secretion in early design rounds motivated adaptations to in silico design methodology. Within two iterations and twenty variants, four sets of unique mutations demonstrated RSV expression and stability enhancement comparable to established stabilizing mutations. Similar success with influenza B optimization supports broader applicability of iterative in silico design methods to streamline antigen engineering.

Despite minimal sample preparation of cryo-EM results in thousands of GPCR structures resolved, we still face challenges to conduct conformational transitions and dynamics study using 19F-qNMR, a super tool in quantifying conformational states because of its ultra-sensitivity to the microenvironmental changes compared to other nuclei. To address this, my lab developed two protocols driving the field forward, allowing us to produce 5-10 mg. of functional receptors/1L cell culture.

Cell-surface receptors pose challenges in expression and purification due to low levels, misfolding, and instability. We introduce a high-throughput ELISA fluorescence approach to rapidly assess multiple recombinant constructs. Utilizing small-scale expression, enzymatic biotinylation, and C-terminal His-tag capture, this approach efficiently prioritizes constructs for large-scale production. We also tested several codon optimization schemes using a minimally designed expression vector. Testing truncation constructs across various protein families demonstrated its effectiveness, significantly saving time in identifying optimal candidates for downstream applications

Based on fully-automated, well-controlled, parallel fed-batch cultivations with integrated analytics and model-based DoEs/Machine learning, the KIWI biolab allows a fast selection of best clones and optimization of process parameters in a single experiment. The power of the established self-driven lab is demonstrated with difficult-to-express protein processes, including hydrogenase, Fabs and elastin like proteins.

In this talk, I will share with the audience a new protein tag, designated EZ-tag, as a solution for difficult-to-expression proteins. Our EZ-tag platform is based on the unique behavior of an endogenous calcium-storage protein in cells. Our EZ-tag demonstrates notable increase in the expression and solubility of various recombinant proteins compared with other conventional protein tags. Moreover, the fusion protein can be facilely purified using a simple calcium-ion-mediated precipitation method. Our EZ-tag will benefit researchers in related industry and academia who are seeking an efficient protein tag for expression and purification of difficult-to-express proteins of interest.

A versatile and highly-efficient bioproduction platform to generate various forms of disulfide-constrained peptides (DCPs) has been developed as an environmentally sustainable alternative to SPPS. This platform can be used to generate: (1) multivalent DCPs with different geometries, (2) DCPs with functional chemical groups such as biotin, (3) DCPs with unnatural amino acids through amber codon suppression, and (4) isotope-labeled DCPs.