Multi-Knock-a multi-targeted genome-scale CRISPR toolbox to overcome functional redundancy in plants

Abstract

Plant genomes are characterized by large and complex gene families that often result in similar and partially overlapping functions. This genetic redundancy severely hampers current efforts to uncover novel phenotypes, delaying basic genetic research and breeding programmes. Here we describe the development and validation of Multi-Knock, a genome-scale clustered regularly interspaced short palindromic repeat toolbox that overcomes functional redundancy in Arabidopsis by simultaneously targeting multiple gene-family members, thus identifying genetically hidden components. We computationally designed 59,129 optimal single-guide RNAs that each target two to ten genes within a family at once. Furthermore, partitioning the library into ten sublibraries directed towards a different functional group allows flexible and targeted genetic screens. From the 5,635 single-guide RNAs targeting the plant transportome, we generated over 3,500 independent Arabidopsis lines that allowed us to identify and characterize the first known cytokinin tonoplast-localized transporters in plants. With the ability to overcome functional redundancy in plants at the genome-scale level, the developed strategy can be readily deployed by scientists and breeders for basic research and to expedite breeding efforts.

Fig. 1. Overview of the Multi-Knock, genome-scale, multi-targeted CRISPR platform.

Stage 1: Multi-targeted sgRNAs were designed to target multiple genes (coding sequences) from the same family. The Arabidopsis genome was clustered into gene families and multiple sgRNAs were designed to target each node using the CRISPys algorithm. Stages 2 and 3: sgRNA sub-library sequences were synthesized, amplified, and cloned into CRISPR/Cas9 vectors. Stage 4: The pooled CRISPR library was introduced into Agrobacterium and transformed into Arabidopsis to generate stable lines. Each plant expresses a single sgRNA, targeting a clade of 2 to 10 genes from the same family. Stage 5: A phenotypic forward genetic screen was conducted. Candidate lines were genotyped for sgRNAs and targets.

Fig. 2. An overview of sgRNA design strategy for gene families.

a, For each gene family, the multiple alignments of the respective protein sequences is computed. P stands for protein, and letters indicate amino acids. b, A phylogenetic tree is constructed based on the sequence similarity of the protein sequences. Optimal sgRNAs for each subgroup of genes, which are induced by internal nodes in the tree (marked by lowercase letters a-c), are then designed. c, For each subfamily of genes, and illustrated here for node a, all potential CRISPR target sites are extracted. In this case, the subfamily induced by node a includes two genes (g₂ and g₄, encoding for proteins P₂ and P₄, respectively). Typically, each gene contains dozens of possible targets. For simplicity, only five targets are presented. Nucleotide positions that are identical in all targets are colored in black, while the polymorphic sites are colored in red and are in bold type. d, A tree of the target sites is constructed based on sequence similarity among the targets while accounting for CRISPR-specific characteristics. sgRNA candidates are constructed for each internal node, where all combinations of the polymorphic sites are considered (marked in red), and the ones with the highest editing efficacy to target the considered subgroup of genes are chosen. For simplicity, only a few candidates (denoted by s_i) are shown for each internal node. Assuming that the cutoff of the number of polymorphic sites k is 4, the search of sgRNA candidates stops at node z. In practice, k was set to 12 polymorphic sites.

Fig. 3. Multi-targeted genome-scale sgRNA design and construction.

a, Schematic illustration of the computational workflow used to design the Multi-Knock sgRNA library. A filtering process yielded a selection of 59,129 sgRNAs targeting 16,152 genes (~74% of all coding genes belonging to families). The red color represents coding genes, blue color represents sgRNAs, and curved arrows represent filtering steps. Abbreviations: Mt-genes, mitochondrial genes; Cp-genes, chloroplast genes; Singletons, genes that do not belong to a family. b, Histogram showing the number of genes targeted by individual sgRNAs. c, Representative sgRNA-target network in the CRISPR library. Genes are targeted by multiple sgRNAs, and sgRNAs target multiple genes. d, Total number of sgRNAs and target genes in each functional sub-library. e, Deep-sequencing data of sgRNAs in individual sub-libraries. Columns indicate the distribution of sgRNAs. Coverage is indicated for each group. TRP (transporters); PKR (protein kinases, protein phosphatases, receptors, and their ligands); TFB (transcription factors and other RNA and DNA binding proteins); BNO (proteins binding small molecules); CSI (proteins that form or interact with protein complexes including stabilizing factors); HEC (hydrolytic enzymes); TEC (metabolic enzymes and enzymes that catalyze transfer reactions); PEC (catalytically active proteins, mainly enzymes); DMF (proteins with diverse functional annotations not found in the other categories); and UNC (proteins of unknown function or cannot be inferred).

Fig. 4. Transportome-specific Multi-Knock screen.

a, To create independent sub-libraries, 5,635 sgRNAs, each targeting 2 to 10 transporters from the same family, were amplified and cloned into four different Cas9 vectors to create pRPS5A:Cas9 (OLE:CITRINE), pUBI:Cas9, pEC:Cas9, and pRPS5A:zCas9i sub-libraries. Graphs show coverage and frequency based on next-generation sequencing of the four sub-libraries. The four libraries were transformed into Col-0 plants yielding 3,500 transgenic T1 plants. b, Photographs show representative phenotypes of TRP Multi-Knock proof-of-concept lines. From top to bottom are Col-0 and plant expressing sgRNA targeting toc120 and toc132 (scale bar = 2 cm), Col-0 and plant expressing sgRNA targeting mex1 and mex1l (scale bar = 1 cm), and control DR5:VENUS plant and the T1 plant harboring sgRNA targeting bor1 and bor2 (scale bar = 4 cm). Chromatograms show the types of mutations. Red arrows indicate the mismatches between sgRNA and target sequence. PAM is marked with a black underline. c, Images show lines with abnormal phenotypes that had not previously been described: from top to bottom adjacent to Col-0 control are plants expressing sgRNA targeting clc-a and clc-b (scale bar = 2 cm), vha-d1 and vhad-2 (scale bar = 2 cm), and pup8 and pup21 (scale bar = 1 cm). Chromatograms show the type of mutations. Red arrows indicate the mismatches between sgRNA and target sequence. PAM is marked with a black underline.

Fig. 5. PUP7, 8, and 21 redundantly regulate shoot growth and phyllotaxis.

a, Phylogenetic tree of Arabidopsis PUP family based on amino acid sequences. Red dots indicate proteins coded by putative CR7/8/21 target genes. b, Chromatograms showing the types of mutations in the CR8/21 (T3 generation), CR7/21 (T3 generation) and CR7/8/21 (T4 generation) lines as identified by sequencing. CR8/21 and CR7/21 stand for CRISPR double mutant pup8/21 and pup7/21, respectively; CR7/8/21 stands for CRISPR triple mutant pup7/8/21. PAM is underlined in black; 20-bp sgRNA is underlined in green. The sgRNA in line CR8/21 was not designed to target PUP7 as it does not contain a respective PAM sequence. c, Shown are representative images of 18-day-old WT (Col-0), the T-DNA single pup7, pup8, pup21 mutants, the double pup8pup21 (CR8/21) (T3 generation), double pup7, pup21 (CR7/21) (T3 generation), and the triple pup7pup8pup21 CRISPR mutant (CR7/8/21) (T4 generation). Scale bar = 1 cm. d, Quantification of genotypes shown in (c). Shown are means (±SE). p value in ordinary one-way ANOVA is indicated for each analysis, Col-0, pup7. pup21, CR7/21 and CR7/8/21: n=16; pup8 and CR8/21: n=15. e, Shown are representative images of 18-day-old Control (TCS:VENUS) and amiRNA7/8/21 (TCS:VENUS background). amiRNA7/8/21 stands for amiRNA knockdown PUP7/8/21. Scale bar = 1 cm. f, Quantification of genotypes shown in (e). Shown are means (±SE). p value two tailed t test is indicated. n = 15 (Control); n = 16 (amiRNA7/8/21). g, Phyllotaxis patterns in inflorescences stem of wild-type (Col-0), single T-DNA insertion mutants, CR8/21, CR7/21 and CR7/8/21. Scale bar = 5 cm. h, Silique divergence angle distribution in inflorescences of Col-0, pup single mutants, CR8/21. CR7/21 and CR7/8/21. P-value, n number and standard deviation (sd) are indicated for each analysis. P-value was extracted using Fligner-Killeen test for equality of variance. i, Phyllotaxis patterns in inflorescence stem of control (TCS:VENUS) and amiRNA7/8/21 mutant. amiRNA7/8/21 stands for amiRNA triple PUP7/8/21 knockdown. Scale bar = 5 cm. j, Distribution of divergence angle frequencies between successive siliques in control and amiRNA7/8/21 stems. p value Fligner-Killeen test for equality of variance is indicated for each analysis.

Fig. 6. PUP7, PUP8, and PUP21 regulate cytokinin transport and shoot meristem.

a, Root meristem confocal microscopy images of 35S:YFP-PUPs localization. YFP (yellow) was used to label PUPs, Vac-ck (cyan) was used as a tonoplast marker, and propidium iodide (PI) (red) was used to stain the cell wall. Scale bars = 10 μm. The experiments were independently repeated three times with similar results. b, Standard import Xenopus laevis oocytes assays using the indicated cytokinin froms. 60 min transport assay in 100 μM cytokinins at pH = 5.5. n = 7 (Mock), n = 6 (PUP8 and PUP7), n = 7 (PUP21 tZ and iPR), n = 6 (PUP21 tZR). Shown are means (±SE). p values were determined by One-way ANOVA. c, Injection-based export assay of PUPs in Xenopus laevis oocytes. n = 6 ± SE, p value two-tailed t-test is indicated for each analysis, d, e, tZ export from tobacco protoplasts as percentage of initial export as a function of time (d) and at 20 minutes (e). n = 14 (Control), n = 4 (PUP8 and PUP7), n = 5 (PUP21). Shown are means (±SE). p values were determined by One-way ANOVA. Box plots represent 25^th-75^th percentiles; whiskers represent min-max (with all points), central bands in the box plots show the medians. f-i, TCS-Venus intensity in control plants (TCS:VENUS) compared to amiR7/8/21 at (f) vegetative and (g) inflorescence stages. amiRNA7/8/21 stands for amiRNA triple knockdown PUP7/8/21. Optical longitudinal sections of the meristem (YZ direction) are shown. LP: leave primordia, FP: floral primordia. Scale bars = 100 μm. The color scale at the right indicates TCS expression. Quantification of TCS-Venus intensity in vegetative (h) and inflorescence (i) stages. Control: n = 5 in (h, i); amiRNA7/8/21: n = 7 in (h), n = 11 in (i). Shown are means (±SE). p value two tailed t test is indicated for each analysis.