Engineered transcription-associated Cas9 targeting in eukaryotic cells

Abstract

DNA targeting Class 2 CRISPR-Cas effector nucleases, including the well-studied Cas9 proteins, evolved protospacer-adjacent motif (PAM) and guide RNA interactions that sequentially license their binding and cleavage activities at protospacer target sites. Both interactions are nucleic acid sequence specific but function constitutively; thus, they provide intrinsic spatial control over DNA targeting activities but naturally lack temporal control. Here we show that engineered Cas9 fusion proteins which bind to nascent RNAs near a protospacer can facilitate spatiotemporal coupling between transcription and DNA targeting at that protospacer: Transcription-associated Cas9 Targeting (TraCT). Engineered TraCT is enabled in eukaryotic yeast or human cells when suboptimal PAM interactions limit basal activity and when one or more nascent RNA substrates are still tethered to the actively transcribed target DNA in cis. Using yeast, we further show that this phenomenon can be applied for selective editing at one of two identical targets in distinct gene loci, or, in diploid allelic loci that are differentially transcribed. Our work demonstrates that temporal control over Cas9's targeting activity at specific DNA sites may be engineered without modifying Cas9's core domains and guide RNA components or their expression levels. More broadly, it establishes co-transcriptional RNA binding as a cis-acting mechanism that can conditionally stimulate CRISPR-Cas DNA targeting in eukaryotic cells.

Fig. 1. Cas9 activity can be controlled by conditional interactions between auxiliary fusion domains and protospacer-proximal cis elements.

a General schematic of the LYS2 genetic reporter system, engineered with a heterologous protospacer (PS) insertion between 100 bp direct repeats. Co-expression of Cas9 with a targeting sgRNA licenses cleavage and deletion of the repeat-intervening sequences through SSA-based recombination that restores LYS2⁺. The native pLYS2 promoter (bent arrow) and an NGG PAM are used throughout the work, except where stated otherwise. b Transformation-associated editing with the Cas9-scTetR fusion or native Cas9 control (Ø) in yeast containing a tetO2 element 13 bp downstream of the PS. The edited fraction of LYS2⁺ transformants (out of total transformants) was measured for Wt, MT3, and NG PID variants using synthetic dropout plates (see “Methods”) that lack (–) or contain (+) Dox. Inset schematic illustrates Dox-induced dissociation of the scTetR domain. Error bars, mean ± s.d. (n = 6 or n = 3, biological replicates). c Same as in b but with Cas9-scRevTetR fusions as indicated (V10 or V16). Inset schematic illustrates Dox-induced binding of the scRevTetR domain. d Simplified schematic illustration of the Cas9-scMCP fusion interacting with a nascent MS2^U/C RNA hairpin, generated from a single Hairpin Template (HT) upstream of the PS. The scMCP domain can be replaced with a 2xN domain that in principle allows binding of two boxB hairpins simultaneously. The nascent transcript is elongated at its 3’ end by an RNA polymerase (RNAP) transcribing left to right. e Transformation-associated editing fractions measured for MT3 and NG variants of Cas9-scMCP or native Cas9 in strains containing the indicated HTs. Error bars, mean ± s.d. (n = 3, biological replicates). p values were calculated using unpaired, two-tailed t tests with Welch’s correction, Gaussian distributions assumed, and no correction for multiple comparisons. f Same experimental setup as in e but performed with the 2xN domain in place of scMCP and different HT strains as indicated. Key elements used to create images in this figure were adapted from “Engineered dual selection for directed evolution of SpCas9 PAM specificity.” Nat Commun 12, 349 (2021) by Goldberg, G. W. et al., under CC BY 4.0. Source data are provided as a Source Data file.

Fig. 2. Nascent RNA binding is insufficient to compensate for subPAM interactions in trans.

a Schematic summary of the LYS2 and CAN1 dual-reporter system, engineered with identical protospacer and PAM insertions between direct repeats at each locus. SSA-based editing can occur via cleavage at either or both loci. Reconstitution of the LYS2 gene confers lysine prototrophy as in Fig. 1a, whereas reconstitution of the CAN1 gene confers canavanine sensitivity for counterselection; only cells that were unedited at their CAN1 locus can form canavanine-resistant (Can^R) colonies. Strains contained either no HTs, or two upstream MS2^U HTs inserted at one of their loci. The native constitutive promoters (bent arrows) were left unmodified at each locus. b Transformation-associated editing fractions measured for Cas9^NG-scMCP or native Cas9^NG (Ø) in dual-reporter strains harboring (::MS2) or lacking (::ΔHT) the HT insertions at each locus. Measurements were performed as in Fig. 1e. Error bars, mean ± s.d. (n = 3, biological replicates). c Editing measurements obtained with the same transformation samples tested in panel (b) except using canavanine-supplemented synthetic dropout plates that select only for Can^R, LYS2⁺ colonies. Error bars, mean ± s.d. (n = 3, biological replicates). Key elements used to create images in this figure were adapted from “Engineered dual selection for directed evolution of SpCas9 PAM specificity.” Nat Commun 12, 349 (2021) by Goldberg, G. W. et al., under CC BY 4.0. Source data are provided as a Source Data file.

Fig. 3. TraCT can be exploited for preferential editing at differentially transcribed alleles with identical target sequences.

a Schematic summary of the diploid SSA reporter system. Each allele differs only in its promoter sequences and a single synonymous point mutation (red asterisk) downstream of identical direct repeat and intervening sequences that disrupt LYS2 as shown. Allele 1’s pZEVi promoter can be highly induced with β-Estradiol (β-E) but is repressed in its absence; hence, Allele 1 cannot confer a LYS2⁺ phenotype in the absence of β-E, even if reconstituted via SSA. b Transformation-associated editing fractions measured for Cas9^MT3-scMCP or native Cas9^MT3 (Ø) in haploid strains containing only Allele 1 with the pZEVi or native pLYS2 promoter. Measurements were performed as in Fig. 1e except that transformants were also plated in parallel on synthetic dropout plates supplemented with β-E at the indicated concentrations. Error bars, mean ± s.d. (n = 3, biological replicates). Zero (0) LYS2⁺ colonies of the pZEVi strain were detected in the absence of β-E at the time of scoring. p values were calculated using unpaired, two-tailed t tests with Welch’s correction, Gaussian distributions assumed, and no correction for multiple comparisons. c Allelic editing in diploid transformants with Cas9^MT3-scMCP or native Cas9^MT3, determined by deep sequencing PCR amplicons to distinguish between alleles via the point mutation shown in panel (a). Amplicons were generated with a primer pair that specifically amplifies unedited alleles (depicted as small red half arrows in panel (a)). Data are plotted as the ratio of unedited Allele 2:Allele 1 genotypes measured at each β-E concentration (color coded as in panel (b)). Error bars, mean ± s.d. (n = 3, biological replicates). p values were calculated using unpaired, two-tailed t tests with Welch’s correction, Gaussian distributions assumed, and no correction for multiple comparisons. Key elements used to create images in this figure were adapted from “Engineered dual selection for directed evolution of SpCas9 PAM specificity.” Nat Commun 12, 349 (2021) by Goldberg, G. W. et al., under CC BY 4.0. Source data are provided as a Source Data file.

Fig. 4. 2xN peptides expressed independently of Cas9 or Cas9-2xN fusions do not compensate for boxB-proximal subPAM interactions and can limit TraCT.

a Simplified schematic illustration similar to Fig. 1d, except that a Cas9-2xN fusion is shown co-expressed with an mCherry−2xN (mC−2xN) fusion in the same strain; either can interact with a nascent boxB RNA hairpin. b Transformation-associated editing fractions measured for Cas9^MT3−2xN or native Cas9^MT3 in SSA reporter strains with one boxB HT upstream of their protospacer and mC−2xN fusion proteins expressed in trans at the indicated levels (+/++/+++), or not at all (–). Measurements were performed as in Fig. 1f; data shown for the strain lacking mC−2xN (–) are replotted from Fig. 1f (1x boxB, MT3) to facilitate comparisons. Error bars, mean ± s.d. (n = 3, biological replicates). p values were calculated using unpaired, two-tailed t tests with Welch’s correction, Gaussian distributions assumed, and no correction for multiple comparisons. Key elements used to create images in this figure were adapted from “Engineered dual selection for directed evolution of SpCas9 PAM specificity.” Nat Commun 12, 349 (2021) by Goldberg, G. W. et al., under CC BY 4.0. Source data are provided as a Source Data file.

Fig. 5. Transcription of boxB stimulates Cas9−2xN accumulation at downstream DNA sequences within the transcribed locus.

a ChIP-qPCR for yeast strains carrying a non-targeting sgRNA and Cas9^NG (Ø) or Cas9^NG-2xN (both of which bear a 3xHA tag, see Supplementary Note 1), either untreated (0 nM) or treated (15 and 1000 nM) with β-E for transcriptional induction of the boxB-tagged pZEVi-LYS2 locus. The percentage of chromatin captured by immunoprecipitation (IP) with anti-HA antibody out of total input chromatin was quantified using LYS2-specific primers for an amplicon within 250 bp downstream of boxB. Error bars, mean ± s.d. (n = 3, biological replicates). p values were calculated using unpaired, two-tailed t tests with Welch’s correction, Gaussian distributions assumed, and no correction for multiple comparisons. b Background control reactions performed in parallel for all panel (a) samples, with qPCR primers amplifying the TAF10 locus on a separate chromosome that lacks boxB. Error bars, mean ± s.d. (n = 3, biological replicates). p values were calculated using unpaired, two-tailed t tests with Welch’s correction, Gaussian distributions assumed, and no correction for multiple comparisons. c ChIP-seq enrichments across the boxB-tagged pZEVi-LYS2 locus, determined with anti-HA antibody for Cas9^NG and Cas9^NG-2xN at each β-E concentration. Data were generated from a single batch of samples tested in panel (a) and plotted as the ratio of IP coverage over input coverage after normalizing each sample for read depth (n = 1, biological replicates). The region downstream of boxB that was probed by qPCR is delineated with purple shading and a purple line above LYS2 in the scaled linear map atop the plots. Boxed insets are rescaled to dimensions that emphasize relative differences, or lack thereof, between the maximum values near boxB and other local maxima further downstream. TSS transcriptional start site, PS protospacer. Source data are provided as a Source Data file.

Fig. 6. Constitutive R-loop formation with boxB-fused sgRNAs that bind DNA but do not license cleavage can stimulate Cas9^Wt-2xN targeting at nearby subPAMs.

a Schematic summary of the paired-protospacer SSA reporter system with dual sgRNA programming for Cas9^Wt-2xN. Three boxB hairpins are fused to the Spacer 1 sgRNA targeting Protospacer 1 (PS1) but not fused to the Spacer 2 sgRNA targeting Protospacer 2 (PS2). Mismatches (Mm) at the 5’ end of Spacer 1 (depicted as the red segment) do not prevent binding but prevent cleavage at PS1. The PAM sequences at each target are denoted 5’ to 3’ irrespective of their orientation. b Transformation-associated editing fractions measured for Cas9^Wt-2xN or native Cas9^Wt in the SSA reporter strain summarized in panel (a). Various combinations of dual sgRNAs with (+) or without (–) the 3x boxB fusion were delivered via plasmid expression vectors. The number of Mm present in each spacer, relative to their intended protospacer targets, is indicated. Error bars, mean ± s.d. (n = 3, biological replicates). p values were calculated using unpaired, two-tailed t tests with Welch’s correction, Gaussian distributions assumed, and no correction for multiple comparisons. Key elements used to create images in this figure were adapted from “Engineered dual selection for directed evolution of SpCas9 PAM specificity.” Nat Commun 12, 349 (2021) by Goldberg, G. W. et al., under CC BY 4.0. Source data are provided as a Source Data file.

Fig. 7. Engineered TraCT can stimulate indel formation in human cells.

a Schematic summary of the non-essential EGFP-PEST ORF targeted in human cell culture experiments with the U2OS.EGFP (ΔHT) parent line and its U2OS.EGFP::5’UTR-boxB^2x or U2OS.EGFP::Intron-boxB^2x(V1.0) derivatives. Relative positions of the four sgRNA targets (purple half arrows with numbering) and two upstream boxB HT insertions within the 5’ UTR or 194 bp synthetic intron are indicated. Transcription is driven by the strong constitutive pCMV promoter in all cell lines. b Deep sequencing of protospacer-spanning amplicons to quantify the percentage of reads that contain indels, inferred to result from donor-independent NHEJ. gDNA templates were extracted 48 h after co-delivering one of four targeting sgRNA plasmids (1 through 4) and the Cas9-2xN^Wt-alt fusion or control Cas9^Wt-alt plasmid into the indicated cell lines. The PAMs CGAGG, CGAGC, GAGCT, and TGACC license targeting for sgRNAs 1 through 4, respectively. Error bars, mean ± s.d. (n = 3, biological replicates). p values were calculated using unpaired, two-tailed t tests with Welch’s correction, Gaussian distributions assumed, and no correction for multiple comparisons. Source data are provided as a Source Data file.