Computational design of PAM-relaxed Staphylococcus aureus Cas9 variants.

(A) Improved UniDesign workflow for point-mutation generation. (B) Specific recognition of the third guanine in the NNGRRT PAM by Arg1015, positioned through salt-bridge interaction with Glu993. Non-target strand (NTS) nucleotides are indicated by asterisks (as in subsequent panels). (C) Mean absolute deviation (MAD) of binding energies across four PAMs (TTAGGT, TTCGGT, TTGGGT, and TTTGGT) for substitutions of Arg1015 with polar or positively charged residues. (D) UniDesign model of the R1015H mutant, showing His1015 forming a hydrogen bond with Glu993. (E) Mean binding energy across the four PAMs for substitutions of Arg1015 with polar or positively charged residues; error bars represent MADs. (F–L) UniDesign models of mutations: N968R (F), N968K (G), K37R (H), E782K (I), N803K (J), Y882R (K), and K822R (L). (M) Mean binding energies of the top triple-mutant variants across the four PAMs; error bars represent MADs.

KRH expands the targeting range of SaCas9 across diverse cell types.

(A) Editing efficiencies of wild-type (WT) SaCas9 and the KRH variant at different PAM sites in HEK293T cells. Bars represent the mean of n = 3 independent biological replicates; error bars indicate s.d. (B) Editing efficiencies of WT SaCas9 and the KRH variant at different PAM sites in A549 cells. Bars represent the mean of n = 3 independent biological replicates; error bars indicate s.d. (C) Editing efficiencies of WT SaCas9 and the KRH variant at different PAM sites in HeLa cells. Bars represent the mean of n = 3 independent biological replicates; error bars indicate s.d. (D) Editing efficiencies of WT SaCas9 and the KRH variant at different PAM sites in U2OS cells. Bars represent the mean of n = 3 independent biological replicates; error bars indicate s.d. (E) Statistical comparison of editing efficiencies between WT SaCas9 and the KRH variant at non-canonical PAM sites across HEK293T, A549, HeLa, and U2OS cells. Data are shown as mean ± s.d.; statistical significance was assessed using a two-tailed unpaired Student’s t-test. (F) Statistical comparison of editing efficiencies between WT SaCas9 and the KRH variant across different PAM classes (NNARRT, NNCRRT, NNTRRT, and NNGRRT). Data are shown as mean ± s.d.; statistical significance was assessed using a two-tailed unpaired Student’s t-test.

KRH-based ABE further broadens base-editing capabilities.

(A) Base-editing efficiencies of WT-ABE and KRH-ABE at different PAM sites in HEK293T cells. Bars represent the mean of n = 3 independent biological replicates; error bars indicate s.d. (B) Base-editing efficiencies of WT-ABE and KRH-ABE at different PAM sites in A549 cells. Bars represent the mean of n = 3 independent biological replicates; error bars indicate s.d. (C) Base-editing efficiencies of WT-ABE and KRH-ABE at different PAM sites in HeLa cells. Bars represent the mean of n = 3 independent biological replicates; error bars indicate s.d. (D) Base-editing efficiencies of WT-ABE and KRH-ABE at different PAM sites in U2OS cells. Bars represent the mean of n = 3 independent biological replicates; error bars indicate s.d. (E) Statistical comparison of base-editing efficiencies between WT-ABE and KRH-ABE at non-canonical PAM sites across HEK293T, A549, HeLa, and U2OS cells. Data are shown as mean ± s.d.; statistical significance was assessed using a two-tailed unpaired Student’s t-test. (F) Statistical comparison of base-editing efficiencies between WT-ABE and KRH-ABE across different PAM classes (NNARRT, NNCRRT, NNTRRT, and NNGRRT). Data are shown as mean ± s.d.; statistical significance was assessed using a two-tailed unpaired Student’s t-test.

Comparison of KKH and KRH Staphylococcus aureus Cas9 variants.

(A) UniDesign models of the KKH and KRH variants. Non-target strand (NTS) nucleotides are indicated by asterisks. (B) Editing efficiencies of the KKH and KRH variants at different PAM sites in HEK293T cells. Bars represent the mean of n = 3 independent biological replicates; error bars indicate s.d. (C) Statistical comparison of editing efficiencies between the KKH and KRH variants in HEK293T cells. Data are shown as mean ± s.d.; statistical significance was assessed using a two-tailed unpaired Student’s t-test. (D) Base-editing efficiencies of KKH-ABE and KRH-ABE in HEK293T cells. Bars represent the mean of n = 3 independent biological replicates; error bars indicate s.d. (E) Statistical comparison of base-editing efficiencies between KKH-ABE and KRH-ABE in HEK293T cells. Data are shown as mean ± s.d.; statistical significance was assessed using a two-tailed unpaired Student’s t-test.

UniDesign models of mutations introducing bulky hydrophobic interactions with the DNA backbone.

(A) T787W. (B) T787M. (C) T787F. (D) N803W.

Evaluation of off-target (OT) effects of WT-ABE and KRH-ABE at predicted OT sites.

Bars represent the mean of n = 3 independent biological replicates; error bars indicate s.d.

UniDesign RESFILE contents used for SaCas9 redesign.

(A) First iteration. (B) Second iteration. (C) Third iteration.

DNA sequences of constructs used in this study.

List of sgRNA sequences used in this study.

List of PCR primers used in this study.

Predicted off-target site sequences.