HyperCas12a enables highly-multiplexed epigenome editing screens

Abstract

Interactions between multiple genes or cis-regulatory elements (CREs) underlie a wide range of biological processes in both health and disease. High-throughput screens using dCas9 fused to epigenome editing domains have allowed researchers to assess the impact of activation or repression of both coding and non-coding genomic regions on a phenotype of interest, but assessment of genetic interactions between those elements has been limited to pairs. Here, we combine a hyper-efficient version of Lachnospiraceae bacterium dCas12a (dHyperLbCas12a) with RNA Polymerase II expression of long CRISPR RNA (crRNA) arrays to enable efficient highly-multiplexed epigenome editing. We demonstrate that this system is compatible with several activation and repression domains, including the P300 histone acetyltransferase domain and SIN3A interacting domain (SID). We also show that the dCas12a platform can perform simultaneous activation and repression using a single crRNA array via co-expression of multiple dCas12a orthologues. Lastly, demonstrate that the dCas12a system is highly effective for high-throughput screens. We use dHyperLbCas12a-KRAB and a ~19,000-member barcoded library of crRNA arrays containing six crRNAs each to dissect the independent and combinatorial contributions of CREs to the dose-dependent control of gene expression at a glucocorticoid-responsive locus. The tools and methods introduced here create new possibilities for highly multiplexed control of gene expression in a wide variety of biological systems.

Figure 1.. Optimization of the dCas12a system for multiplexed epigenome editing

(a) Comparison of dCas12a repressors. Five different Cas12a variants were fused to the transcriptional repressor KRAB. Each dCas12a-KRAB construct was co-transfected into HEK293T cells with either a pre-crRNA of three crRNAs targeting the RAB11A promoter or a control pre-crRNA containing three non-targeting crRNAs. After two days, cells were sorted on Cas12a expression using a GFP marker, mRNA was harvested and relative RAB11A expression levels were quantified using RT-qPCR (two or more independent experiments per transfection condition) (b) Comparison of dCas12a activators. Five different Cas12a variants were fused to the transcriptional activator VPR. Each dCas12a-VPR construct was co-transfected into HEK293T cells with either a pre-crRNA of three crRNAs targeting the HBE1 promoter or a control pre-crRNA containing three non-targeting crRNAs. After 2 days, mRNA was harvested and relative HBE1 expression levels were quantified using RT-qPCR (n = 3 independent experiments per transfection condition) (c) Constructs used to assess different promoters for pre-crRNAs of length ten. Pre-crRNAs of 10 crRNAs, driven by CAG or U6, were designed to contain three crRNAs targeting the HBE1 promoter in either position 1–3 (Front of array) or 8–10 (Back of array). These constructs were then co-transfected into HEK293T cells with either (d) dCas12a-VPR or (e) dCas12a-P300. After 2 days, mRNA was harvested and relative HBE1 expression levels were quantified using RT-qPCR (two or more independent experiments per transfection condition)

Figure 2.. Multiplexed gene repression by dCas12a-KRAB and dCas12a-4xSID

(a) dCas12a repressor constructs (b) Experimental design for testing multiplexed repression of a single gene by dCas12a-KRAB and dCas12a-4xSID constructs. Repressors were co-transfected into HEK293T cells along with a pre-crRNA of three crRNAs targeting a gene promoter of interest or a control pre-crRNA containing three non-targeting crRNAs. Two days post-transfection, dCas12a-positive cells were sorted on EGFP expression, RNA was harvested, and gene expression changes were assessed via RT-qPCR. (C) Repression of the RAB11A and VEGFA promoters by dEnAsCas12a- KRAB and 4xSID fusions (n = 3 independent experiments per transfection condition) (d) Repression of the RAB11A and VEGFA promoters by dHyperLbCas12a KRAB and 4xSID fusions (at least two independent experiments per transfection condition) (e) Experimental design for testing highly multiplexed gene repression by dCas12a repressors. A pre-crRNA of ten crRNAs targeting four genes: VEGFA, MYC, RAB11A, and PIK3C3 was designed to test highly multiplexed repression. HEK293T cells were co-transfected with dCas12a repressors and either this array or a control pre-crRNA containing ten non-targeting crRNAs and relative gene expression was determined as in (b). (f) Repression of four target genes by dEnAsCas12a KRAB and 4xSID fusions using a pre-crRNA of length ten (n = 3 independent experiments per transfection condition) (g) Repression of four target genes by dHyperLbCas12a KRAB and 4xSID fusions using a a pre-crRNA of length ten (n = 3 independent experiments per transfection condition)

Figure 3.. Activation of single genes using dCas12a

(a) dCas12a activator constructs (b) Experimental design for testing multiplexed repression by dCas12a-VPR and P300 constructs. Activator constructs were co-transfected into HEK293T cells along with a pre-crRNA of three crRNAs targeting a gene of interest or a control pre-crRNA of three non-targeting crRNAs. Two days post-transfection, dCas12a-expressing cells were sorted on EGFP, RNA was harvested, and gene expression changes were assessed via RT-qPCR. (c,d) Activation of the HBE1 promoter, MYOD1 promoter, or MYOD1 distal regulatory region (DRR) by (c) dEnAsCas12a- VPR and P300 fusions (at least two independent experiments per transfection condition) or (d), dHyperLbCas12a-VPR and P300 fusions (at least two independent experiments per transfection condition)

Figure 4.. Activation of multiple genes using dCas12a

(a) Experimental design for testing highly multiplexed gene activation by dCas12a activators. A pre-crRNA containing ten crRNAs targeting three genes: AR, MYOD1, and HBE1, was designed to test highly multiplexed activation. HEK293T cells were co-transfected with dCas12a activators and either this pre-crRNA or a control pre-crRNA of ten non-targeting crRNAs and gene expression was determined as in Fig. 3. (b,c) Activation of three target genes using the multiplexed pre-crRNA by (b) dEnAsCas12a-VPR and P300 fusions and (C) dHyperLbCas12a- VPR and P300 fusions (n = 3 independent experiments per transfection condition)

Figure 5.. Programmed, simultaneous activation and repression within a single cell

(a) Experimental design for simultaneous activation and repression. A hybrid pre-crRNA composed of both Lb and As Cas12a crRNAs is co-transfected with a dEnAsCas12a activator and a dHyperLbCas12a repressor in order to achieve programmed activation and repression within a single cell. The hybrid pre-crRNA or a control pre-crRNA of ten non-targeting crRNAs were co-transfected into HEK293T cells with a dEnAsCas12a activator and a dHyperLbCas12a repressor. Cells were sorted and changes in target gene expression were assessed as in Fig 2–4. (b,c) Simultaneous activation and repression by (b) dEnAsCas12a-VPR and dHyperLbCas12a-KRAB or (c) dEnAsCas12a-P300 and dHyperLbCas12a-4xSID (n = 3 independent experiments per transfection condition).

Figure 6.. Construction of a crRNA library to target two PER1 enhancers

(a) two enhancers, labeled A and B, control the glucocorticoid-responsiveness of the gene PER1. These sites are bound by the glucocorticoid receptor after treatment with the synthetic glucocorticoid dexamethasone (b) dHyperLbCas12a-KRAB is capable of dampening the PER1 glucocorticoid response through dual targeting of the enhancers in (a). A549 cells stably expressing dHyperLbCas12a-KRAB and either a pre-crRNA of four crRNAs targeting the PER1 enhancers (2 crRNAs per enhancer) or a control pre-crRNA of four non-targeting crRNAs were created. These cell lines were treated with either 100nM dexamethasone or vehicle for three hours before RNA was harvested and changes in PER1 expression were determined using RT-qPCR. (c) Schematic showing makeup of the 8,836 member PER1-targeting pre-crRNA library. Pre-crRNAs of six crRNAs were created to target each of the PER1 enhancers either individually or in combination (d) schematic showing the completed library array constructs. Two sets of 3 crRNAs were cloned into a CAG expression vector to create dual-targeting pre-crRNAs of 6 crRNAs with a bipartite barcode flanked by TruSeq adapter sequences. The crRNAs are also located on the same transcript as the mCherry-P2A-puroR selection marker. (e,f) Density plots showing the distribution of crRNA counts for each library. The distribution of pre-crRNA barcode counts for (e) the 10,003 member non-targeting library. (f) the 8,836 member PER1-targeting library

Figure 7.. PER1 Screen Results

(a) Timeline of the PER1 screen. A549 cells stably expressing dHyperLbCas12a-KRAB were transduced with either the library of PER1-targeting arrays (n = 8,836) or the library of non-targeting arrays and promoter-targeting arrays (n=10,003). They were then selected with puromycin, after which the cell populations for the two different libraries were combined. Cells were treated with either an ethanol control (vehicle), 0.2nM dexamethasone (low Dex), or 100nM dexamethasone (high Dex) for three hours before being harvested for HCR. Cells were then sorted into high and low bins (top and bottom 12%) based on PER1 expression. gDNA was recovered and sequenced and array abundance in high and low bins was compared to input. (b,c,d) Enrichment in pre-crRNA barcode abundance vs. input for (b) vehicle (c) 0.2nM Dex and (d) 100nM Dex. Log₂ fold change is an average of enrichment in the high and low PER1-expressing populations