Abstract 2354 Engineering CRISPR-Associated Transposons for Facile Tagging of Endogenous Proteins

CRISPR-associated transposases (CASTs) are promising tools for the programmable and efficient genomic insertion of multi-kilobase genetic payloads. These systems direct nuclease-deficient CRISPR-Cas proteins to loci of interest and subsequently recruit transposase enzymes for downstream DNA integration. Payloads must be flanked by transposon end sequences to be recognized and mobilized by transposases, meaning that CAST-mediated insertions leave behind genetic "scars." Within a broader effort to uncover the molecular requirements of obligate transposase-DNA interactions, we aimed to repurpose these "scars" as functional sequences encoding peptide linkers for protein tagging. We leveraged pooled DNA libraries and high-throughput sequencing to test hundreds of rationally designed transposon variants in which stop codons and codons encoding bulky and/or charged amino acids were substituted to encode small and non-polar ones. Our library approach revealed several linker-encoding transposon variants that retained near-wild-type integration efficiencies, which we validated by testing individual candidates with quantitative PCR. In a proof-of-concept experiment, we showcased an engineered transposon variant by efficiently and stably tagging an endogenous E. coli protein with GFP, which we visualized by fluorescence microscopy and validated by Western blot. While these experiments were all performed in bacteria on one of the first discovered CAST systems (VchCAST), our recent data have further unveiled an improved CAST homolog (PseCAST) that is more active in both bacteria and human cells. Using a similar library approach, we are engineering this system to tag human genes, thereby expanding beyond the bacterial proteome. Our results collectively demonstrate that CASTs can be harnessed for facile tagging of endogenous proteins without a need for homology arms or recombination, unlocking a novel biological toolkit to study proteins in their native regulatory environments. More broadly, our work underscores the engineering potential of combining rationally designed libraries with high-throughput sequencing-based approaches. This work was supported by National Institutes of Health [DP2HG011650, R21AI168976 to S.H.S.]; Pew Biomedical Scholars Program (to S.H.S.); Alfred SloanFoundation Research Fellowship (to S.H.S.); Irma T. Hirschl Career Scientist Award (to S.H.S.); National Science Foundation (GRFP to M.W.W.).