Fig. 1

The development of oA3G-BE3. a Key editors tested in this study. Editors #1 [13], #2 [3, 4] #3 [26], #5 [14], and #7 [15] have been formally published whereas #6 posted in a preprint [16]. In contrast, editors #4 (oA3G-BE3) and #8 (oA3G-BE4max) are developed in this study, where oA3G denotes “optimized A3G” bearing 5 substitutions. The BE4max architecture differs from that of BE3 in that it has 2 copies of UGI, optimized codon usage, and nuclear localization, in addition to using optimized linker sequences between the fused proteins. NLS, nuclear localization signal; BPNLS, bipartite nuclear localization signal; UGI, uracil DNA glycosylase inhibitor. b A3G-BE3 efficiently edited CCC present at HEK293 site 3. Various editors and a gRNA for the target site were coexpressed in HEK293T cell and analyzed using Sanger sequencing. The target sequence is depicted. The third C at CCC was edited most efficiently by A3G-BE3 and highlighted in the bar graph. Values are mean ± SEM from triplicate transfections. c A3G-BE3 created low levels of off-target edits at the transcriptome. Editors were coexpressed with the gRNA targeting HEK293 site 3 as in a, but transfection was done in duplicates and at a larger scale. Cells with top 15% GFP signal were sorted and analyzed by RNA-seq 48 h later. The jitter plot shows the off-target edits in duplicate samples, with the total numbers of the edits indicated. d, e On-target (d) and RNA off-target (e) editing by A3G-BE3-4M and oA3G-BE3. The samples are from the same experiment as in b, but the cells with top 15% GFP signal were analyzed in parallel by Sanger sequencing (to measure on-target editing at HEK293 site 3; d) and RNA-seq (to determine RNA off-target effects; e). The on-target editing rates were higher than b because the cells analyzed here, with top 15% GFP fluorescence, expressed higher levels of editors and gRNA. The bar graph in d displays mean ± SEM from duplicate transfections, with the blue numbers being the ratios of the editing rates at the target C (C₅, red bar) over that at the bystander (C₄), which is a measure of the selectivity of the editors. The P value (0.07%) for the on-target editing rates has not reached significance presumably due to the small sample size (n = 2). f, g On-target editing by A3G-BE3-4M and oA3G-BE3 at two more sites. Editors and gRNAs were coexpressed as in a. Gene editing was then analyzed by targeted deep-sequencing instead of Sanger sequencing, in order to detect the low level editing at C₅ at EMX1-2