(H-301) High-Throughput Characterization of Heterologously-Expressed Membrane Transporter Proteins: A Snorkel Tag Approach

Designing membrane transport proteins is an exciting approach to addressing aberrant metabolic pathways and engineering targeted drug delivery systems. One main aim of the Woldring Lab is the application of high-throughput cytometry of diverse ( >106 variants) libraries of proteins to generate rich, quantitative datasets from which deep learning models can capture latent protein design “rules.” With these models, we hope to rapidly advance the pace of designing transporters with novel functions. Organic Anionic Transporting Proteins (OATPs) is one class of transporters with extensive involvement in many metabolism pathways yet hold untapped potential due to their understudied nature. Previous studies exploring the ligands transported by OATPs have generally relied on low throughput methodologies (e.g., microscopy, liquid scintillation counting). This motivates the development of new approaches to gain a deeper understanding of these essential proteins.

This project addresses the dearth of experimental data through the high-throughput characterization of diverse libraries of OATPs generated by ancestral sequence reconstruction (ASR). Since thermally stable ancestral proteins can be inferred from phylogenetic relationships, ASR facilitates the efficient design and production of large libraries. The diversity of this library is enhanced by sampling across probable amino acid substitutions of each inferred ancestral protein. To enrich the data observed from flow cytometry and determine the surface expression of OATPs, an extracellular epitope tag (snorkel) will be fused to the N- or C- terminus of OATPs. In addition to heterologous expression in mammalian cell culture, we will functionally characterize these libraries in yeast.

Materials and Methods:

Ancestral sequence reconstruction was performed on OATPs to generate a diverse portfolio of ancestral protein sequences. To demonstrate the functionality of these ancestors, several ancestor sequences from this library were synthesized and cloned into pLenti-based constructs for EF‐1α promoter-driven expression in HEK-293 cell culture following transient transfection. The snorkel tag, consisting of an extracellular FLAG-tag, a transmembrane domain, and an intracellular linker, will be fused to either the N- or C-terminus to determine the surface expression of OATPs. Similarly, a pCT-based vector for yeast will be used to facilitate the galactose-inducible expression of the OATP-Snorkel open reading frame described above. For the yeast expression vectors with the N-terminus snorkel tag, the putative plasma membrane association sequence (PMasseq) will be fused to the C-terminus to facilitate localization to the plasma membrane. In contrast, the prepro signal leader sequence from the ɑ-Mating factor will be fused to the N-terminus of OATP constructs with C-terminus snorkel tags. Both mammalian and yeast expression constructs will be tagged to eGFP between the snorkel tag and OATP sequences.

Through confocal fluorescence microscopy and flow cytometry, confirmation of surface expression and function of OATPs will be accomplished by secondary labeling of the extracellular FLAG-tag and the uptake of various fluorescent ligands (e.g., indocyanine green), respectively. Once surface expressions and functional uptake are observed, libraries of OATPs will be sorted based on expression strength and substrate uptake and subsequently deep sequenced, resulting in a rich dataset of labeled OATP sequences.

Several ancestral OATP sequences generated from our automated ASR workflow were synthesized and cloned into pLenti expression vectors. Despite having up to 170 mutations from modern canonical OATPs, these ancestors demonstrate efficient transport of substrates in HEK-293 cell culture. Furthermore, the addition of the snorkel tag and eGFP does not abolish transport activity. While preliminary experiments indicate efficient induction of OATP-SNKL constructs via GFP expression in yeast, further experiments are needed to sufficiently characterize the surface expression and transport function of these constructs.

The use of ASR to generate libraries is a novel approach that is particularly relevant for OATPs. Modern, canonical members of the OATP family have diverse capabilities in transporting various substrates. Characterizing the evolutionary path by which certain modern members have gained or lost the capacity to transport various substrates is key to elucidating underlying protein design rules

Current methods of characterizing OATP expression in flow cytometry rely on the crude estimation of total OATP expression by harvesting protein after sorting cells based on substrate uptake. By displaying an epitope tag extracellularly with the snorkel tagging strategy, we can generate diverse libraries that are not restricted to accommodate potentially interfering tags within extracellular loops. 

To the best of our knowledge, functional display of OATPs with yeast surface display (YSD) has not been demonstrated. In addition to the relative ease of generating and maintaining YSD libraries compared to mammalian display systems, YSD offers benefits in terms of designing and executing flow cytometric workflows. Due to the inherent difficulty of heterologously expressing and shuttling complex membrane proteins to the plasma membrane, more extensive engineering measures (e.g., evaluating alternative plasma membrane signal sequences and implementing glycoengineered strains of yeast to optimize OATP glycosylation) may be required to efficiently express OATPs on the plasma membrane of yeast.

The work in this presentation represents significant progress on preliminary steps toward generating large datasets of labeled sequences needed to develop machine learning models that can uncover latent protein design rules.