Fig. 3. Stacking multiple domain deletions on Cas9 results in defective DNA-binding activity.

A In vivo transcription repression activity of MISER CRISPR effectors containing triple (Δ3CE) and quadruple (Δ4CE) deletion variants. Sublibraries of REC2, REC3, HNH, and RuvC were combined to build a library of stacked deletions, and the resulting library was assayed for high-performing variants using FACS (light blue bars). As none of the variants contained a REC2 deletion (~Δ167–307), we named the highest-performing triple-deletion variant in this library (Library 2; see Supplementary Fig. 6) Δ3CE. To force a library containing REC2 deletions, a sublibrary of REC2 deletions was added to Δ3CE, resulting in a library of quadruple deletion variants that contain Δ3CE and a REC2 deletion (dark blue bars). Data are plotted as mean ± SD from biological triplicates. B Expression constructs for Δ3CE and Δ4CE, with specified deletions manually cloned in. C BLI assay of CE constructs. Δ3CE and Δ4CE exhibit almost no binding against a fully complementary dsDNA target at 300 nM RNP (see Supplementary Fig. 10); and weak binding at 1000 nM RNP. Binding is rescued to near-WT levels when RNP concentration is 3.3× that of dCas9 if the dsDNA contains a 3-bp bubble in the PAM-proximal seed region. Data are normalized to 300 nM dCas9 binding to fully complementary dsDNA. D Measurement of CRISPRi efficacy of Δ3CE and Δ4CE in U-251 cells using RT-qPCR. Fold change in PCNA expression levels is measured by RT-qPCR, 2 and 5 days after KRAB-Δ3CE and KRAB-Δ4CE expressing cell lines are transduced with a sgRNA targeting PCNA. Δ3CE and Δ4CE exhibit weak DNA binding and transcriptional repression activity compared to dCas9. Bars represent the fold change of PCNA expression relative to a nontargeting sgRNA. Data are presented as mean and SD (for triplicates where shown). Source data are provided as a Source Data file.