Figure 1. A platform for deep mutational scanning of GPCRs.

(A) Overview of the multiplexed GPCR activity assay. Plasmids encoding ADRB2 variants, a transcriptional CRE reporter of signaling activity, and 15 nucleotide barcode sequences that identify the variant are integrated into a defined genomic locus such that one variant is present per cell. Upon stimulation by isoproterenol, G-protein signaling induces transcription of the CRE genetic reporter and the barcode. Thus, the activity of a given variant is proportional to the amount of barcode mRNA which can be read out in multiplex by RNA-seq. (B) Schematic detailing the recombination of the reporter-receptor expression plasmid into the landing pad locus. Top right: activation of the CRE reporter integrated with (purple) or without (grey) exogenous ADRB2 into the landing pad when stimulated with isoproterenol in ΔADRB2 cells via a luciferase CRE reporter gene assay. (C) Overview of library generation and functional assay. Missense variants are synthesized on an oligonucleotide microarray, the oligos are amplified with random DNA barcode sequences appended, and the variants are cloned into wild-type background vectors. Barcode-variant pairs are mapped with next-generation sequencing and the remaining wild-type receptor and CRE reporter sequences are cloned into the vector. Next, the variant library is integrated en masse into the serine recombinase (Bxb1) landing pad engineered at the H11 locus of ΔADRB2 HEK293T cells. This integration strategy ensures a single pair of receptor variant and barcoded CRE reporter is integrated per cell and avoids crosstalk. After selection, the library is stimulated with various concentrations of the β₂AR agonist, isoproterenol. Finally, mutant activity is determined by measuring the relative abundance of each variant’s barcoded reporter transcript with RNA-seq.

Figure 1—figure supplement 1. Cellular engineering and reporter optimization for multiplexed assay.

(A) Schematic of experiment to ensure the landing pad is present at single copy in the genome and thus recombine a single donor plasmid per cell. Single copy integration is essential to prevent receptors of variable functionality to activate barcoded reporters mapped to other variants. Upon co-expression of the promoterless GFP and mCherry plasmids with bxb1 recombinase sites, a cell line with a single landing pad will exclusively integrate one cassette. Therefore, cells will be either GFP⁺ or mCherry⁺ but never both. (B) Flow cytometry plots detailing the percentage of GFP⁺ and mCherry⁺ cells when transfected with an equimolar ratio of promoterless GFP and mCherry expression cassettes with or without Bxb1 recombinase expression. (C) Activation of a cAMP-responsive luciferase reporter gene integrated in the landing pad when stimulated with isoproterenol in a WT or ΔADRB2 background. Activation of the reporter in the WT background emphasizes the importance of generating a ΔADRB2 cell line for the multiplexed assay. (D) Activation of the genetic reporter/ADRB2 expression cassette with or without a CHS4 DNA insulator upstream of the reporter gene integrated in the landing pad when stimulated with isoproterenol. (E) Fold activation of an integrated genetic reporter/ADRB2 expression cassette with a FLAG-tag or 3x-FLAG tag fused to the N-terminus of ADRB2. (F) Alterations to the processes in the multiplexed assay that improve barcode abundance estimate. Initially, we seeded 2,300,000 cells/replicate and processed ~25 µg of total RNA per replicate and observed modest correlation (r = 0.54). We then scaled up to seeding 32,000,000 cells/replicate and correlation improved (r = 0.85). Finally, we began processing ~650 µg of total RNA/replicate and noticed further improvement (r = 0.90). (G) Heatmap displaying the barcodes per variant across the genomically integrated library. (H) Histogram displaying the frequency barcodes per variant across the genomically integrated library.