|Home | About | Journals | Submit | Contact Us | Français|
Forward genetic screens have been instrumental in deciphering the molecular mechanisms that control agriculturally important traits. Traditionally, crop mutant collections are generated by chemical mutagenesis such as ethyl methanesulfonate (EMS) treatment, physical irradiation, and insertional agents including T-DNAs and transposons (Wei et al., 2013; Li et al., 2016). Because EMS and irradiation-generated mutations are mosaic at the first generation (M1), forward genetic screens for any desirable phenotypes have to be conducted at the M2 generation. T-DNA/transposon insertional mutations at T0 generation usually are heterozygous, making it impossible to screen for phenotypes associated with loss-of-function mutations at the T0 generation. Overall, all previous forward genetic screens except RNA interference-based screens need at least two generations in order to identify stable loss-of-function mutants with a particular phenotype. However, RNAi-based screens often encounter off-target effects and stability issues. Given that a complete life cycle from seed to seed for most major crops takes at least several months, development of technologies that enable the completion of both constructing a mutant library and screening for mutant phenotypes within a single generation will revolutionize plant genetics and crop improvement. Another challenge in traditional forward genetic screens is that identification of the causal mutations for the observed phenotypes is often labor-intensive and time-consuming. Identification of the causal mutation in an EMS mutant or irradiation mutant often requires the construction of a mapping population. T-DNA/transposon mutants sometimes are not tagged, requiring the employment of map-based cloning as well. For tagged T-DNA mutants, it is sometimes difficult to identify the sequences flanking the T-DNA insertion because of the truncation or tandem repeats of the T-DNA fragments. Overall, determination of the relationship between a genotype and a phenotype among mutants is a lengthy process because it often requires crosses/backcrosses, genetic complementation, and growth of several generations of plants.
CRISPR/Cas9 (Clustered Regulatory Interspaced Short Palindromic Repeat/CRISPR associated protein 9) gene editing technology has made targeted mutagenesis easy, and the technology has successfully edited various targets in many organisms including all major crops (Ma et al., 2016). Cas9 binds to the DNA target, which is complementary to the first 20-bp sequence of the guide RNA (gRNA), and generates a double strand break (DSB) within the target site. A small insertion or a deletion, which often causes frameshift of an open reading frame, is generated when the DSB is repaired by a non-homologous end-joining DNA repair mechanism. In rice, targeted mutagenesis by CRISPR/Cas9 is extremely efficient. It was reported that over 80% of the T0 rice plants transformed with a CRISPR/Cas9 construct contain mutations at the intended target site (Ma et al., 2016). More importantly, the percentage of T0 mutants that are either homozygous or bi-allelic can be as high as 90% (Ma et al., 2016). The observed high mutagenic efficiency and high percentage of homozygous/bi-allelic mutations makes it feasible to complete three key genetic tasks within a single generation in rice: (1) generate a CRISPR/Cas9 mutant library; (2) conduct high-throughput forward genetic screens for desired traits; (3) identify and confirm the causal mutations for the observed phenotypes.
Two recent studies in rice demonstrated that it is technically and logistically feasible to generate large-scale CRISPR/Cas9 mutant libraries that cover most of the genes in rice for the purpose of defining gene functions and improving crops (Lu et al., 2017; Meng et al., 2017). Designing and constructing genome-scale CRISPR/Cas9 plasmids are straightforward and relatively inexpensive because the targeting specificity of CRISPR/Cas9 is determined by the first 20-bp sequence in a gRNA molecule, which can be easily synthesized using array-based oligonucleotide synthesis (Figure 1). Several large-scale genome-wide CRISPR-based genetic screens have been successfully conducted in mammalian cells (Shalem et al., 2015). The same strategies can be used to generate CRISPR/Cas9 plasmid libraries for crops. Compared with the mammalian cell system, we believe that the major challenge in constructing a CRISPR/ Cas9 mutant library in plants is the delivery of the CRISPR plasmids into plant cells and to regenerate transgenic plants. The two large-scale and high-throughput studies in rice discussed here used pooled gRNA-producing binary plasmids and Agrobacterium-mediated transformation to generate a collection of transgenic T0 plants (Figure 1). Both studies reported efficient editing of the target sites and very few off-target events. Together the two studies generated almost 100,000 independent CRISPR/Cas9 T0 rice lines, providing tremendous resources for the plant biology community (Lu et al., 2017; Meng et al., 2017).
A major advantage of CRISPR/Cas9 mutagenesis over traditional T-DNA/transposon/EMS/irradiation mutagenesis is that CRISPR/Cas9 allows quick identification of the potential causal mutations for a phenotype by identifying the gRNA sequence followed by Sanger-sequencing of the corresponding target. Lu et al. (2017) creatively used barcodes and next-generation sequencing (NGS) to identify the gRNAs for each transgenic plant in a CRISPR/Cas9 mutant collection that they generated, enabling high-throughput analysis of the relationship between genotypes and phenotypes. As a pilot analysis, Lu et al. (2017) conducted 96 × 96 PCR reactions with a unique pair of barcoded PCR primers for each sample to amplify gRNAs from 9216 transgenic plants, resulting in the identification of gRNAs in more than 5500 plants (some of the PCR or sequencing reactions did not work) (Figure 1). This study demonstrates that genotyping CRISPR/Cas9 mutant library has become efficient and inexpensive, greatly accelerating the identification of the causal mutations for a given phenotype (Lu et al., 2017).
Forward genetic screens using CRISPR/Cas9 mutagenesis may allow the elucidation of a gene function within a single generation if the CRISPR/Cas9 mutant collection is large enough to contain several allelic mutations. Indeed, Lu et al. (2017) reported that multiple T0 plants had obvious developmental phenotypes, including spotted leaves, altered tiller number, altered tiller angle, dwarf, and sterility. The genes responsible for the phenotypes at T0 generation potentially can be inferred from the gRNA sequences, which can be obtained through barcoded/NGS analysis or simple PCR/sequencing (Figure 1).
Once the library is constructed, it is quite easy to conduct genetic screens at the T1 generation with the added benefit that multiple plants with the same phenotypes are available (Figure 1). For example, Meng et al. (2017) discovered that 11 T1 plants that were offspring of a single T0 plant had altered tiller number and reduced height. They were able to show that the phenotypes were caused by loss-of-function mutations in the TAD1 gene (Meng et al., 2017). Large-scale barcoded/NGS analysis will establish sequence-indexed CRISPR/Cas9 mutant collections similar to sequence-indexed Arabidopsis mutant collections (Alonso et al., 2003), which have greatly benefited the Arabidopsis community.
High-throughput and large-scale CRISPR/Cas9 mutagenesis can be applied to other crops as well. In a pilot study in tomato, Jacobs et al. (2017) developed a CRISPR/Cas9 plasmid library that targets 54 immunity-associated leucine-rich repeat subfamily XII genes. From a single transformation with this CRISPR/Cas9 library, they were able to identify heritable mutations in 15 out of 54 genes. Jacobs et al. (2017) generated a second CRISPR/Cas9 library, which is composed of plasmids that express three gRNAs in each plasmid. Some of the T0 plants had mutations in multiple targeted genes. More importantly, Jacobs et al. (2017) were able to establish an association between the developmental phenotypes observed in T0 plants and the mutations in the targeted genes. They demonstrated that mutations in a homolog of an Arabidopsis boron efflux transporter caused boron-deficiency phenotypes in tomato, defining a gene function in a single generation (Jacobs et al., 2017). Such a task would not be possible to complete in a generation using traditional mutagenesis approaches.
Although still limited in scale, the studies in rice and tomato discussed here demonstrated that the CRISPR/Cas9-mediated targeted mutagenesis is reliable, efficient, and high-throughput for generating mutant collections that cover most genes in a genome (Jacobs et al., 2017; Lu et al., 2017; Meng et al., 2017). The CRISPR technology has revolutionized genetic studies in crops because of its unprecedented scale, versatility, and speed. CRISPR technology is not only ideal for producing knockout mutants, it can also be adapted for generating gain-of-function mutant libraries if the nuclease-dead Cas9 (dCas9) fused with a transcriptional activation domain is used (Tang et al., 2017). Moreover, the technology can be used to generate multiple mutants with a single plasmid that produces several gRNAs as shown in the tomato study (Jacobs et al., 2017). Undoubtedly, more features such as fluorescence markers for transgene identification (Gao et al., 2016), different types of nucleases including Cpf1 (Tang et al., 2017), and diversified gRNA production methods (Gao and Zhao, 2014) will further improve CRISPR-mediated mutagenesis and broaden the applications of CRISPR gene editing technologies in genetic studies and crop improvement.
Our research is supported by the National Transgenic Science and Technology Program of China (2016ZX08010002) to R.W. and an NIH grant R01GM114660 to Y.Z.
No conflict of interest declared.