Two stones, one bird: co-targeting facilitates precise gene editing of non-selectable genes in Chlamydomonas

Chlamydomonas (Chlamydomonas reinhardtii) has long been one of the leading unicellular reference systems for the plant kingdom. Its excellent classical genetic properties and haploid lifestyle has facilitated genetic analyses of a broad range of processes, from cilia biology to photosynthesis (Salomé and Merchant, 2019). CRISPR-Cas technology now also enables routine generation of targeted knock-out mutants in any gene of interest for functional study (Greiner et al., 2017; Picariello et al., 2020). These knock-out events generally result from error-prone repair of a Cas-induced double-stranded break (DSB) via the non-homologous end-joining (NHEJ) pathway (Sizova et al., 2021). However, the reluctance of Chlamydomonas to engage in homology-directed repair using exogenously provided DNA templates (Sodeinde and Kindle, 1993) has remained a severe limitation. Without a method to modify genes at their native locus, reverse genetic studies will continue to rely on introduction of transgenes into random places in the genome, a process that is fraught with problems such as unpredictable expression levels, gene silencing, and other positional effects.

In this issue of Plant Physiology, Akella and co-workers take an important step toward a practical method for precise, homology-dependent gene editing in Chlamydomonas by adopting a CRISPR/Cas9-based co-targeting approach. Their method builds on previous work demonstrating that single-stranded oligodeoxynucleotides (ssODNs) can direct precise homology-dependent repair of DSBs at a high rate (Ferenczi et al., 2017) with much less likelihood to engage in unwanted NHEJ (Zorin et al., 2005). The authors used in vitro-assembled Cas9-guide RNA complexes (referred to as ribonucleoproteins, or RNPs in the paper) together with ssODNs to guide DSB repair at specific sites (Akella et al.). One of the ssODNs converts an endogenous gene into a selectable marker, and the other introduces the desired edit in a non-selectable gene of interest. With this approach, they achieved a precise editing frequency of ∼1%, a level that is high enough for identifying positive transformants by PCR-based screening without too much effort.

To develop their method, the authors chose as a test case the FTSY gene, which is involved in assembly of the photosynthetic apparatus. Cells with a mutated FTSY gene are easily recognizable because they form pale green, rather than dark green colonies. They designed an ssODN template spanning the CRISPR-Cas9 cut site in the FTSY gene with ∼40 bp homology arms extending to either side. The ssODN, if used for repair, would introduce STOP codons disrupting the coding sequence and new enzyme restriction sites to facilitate analysis of the result. The authors transformed the RNP complex targeting FTSY together with the ssODN repair template. They also included a linear double-stranded DNA fragment containing an antibiotic resistance gene to select transformed cells. However, when they analyzed the pale green mutants from this experiment, they found that although many of them had cut the FTSY gene at the expected position, none of them had used the ssODN template for repair. Instead, the genes displayed insertions and deletions of varying size and complexity at the cut site, the typical result of NHEJ (Sizova et al., 2021).

Vegetative Chlamydomonas cells contain an active machinery for homology-dependent repair. Two non-functional halves of a selectable gene can be transformed together and recombine inside the cell to reconstitute a functional unit (Sodeinde and Kindle, 1993), and pre-installed defective resistance genes can be repaired using a homologous temple (Greiner et al., 2017). To fish out cells with an active recombination machinery by positive selection, the authors designed a second RNP + ssODN pair, this time against the protoporphyrinogen IX oxidase gene, PPX1. In this case, successful repair using a ssODN template would introduce a single nucleotide (G→A) substitution that would render the cell resistant to the herbicide oxyfluorfen. They also added a second, silent mutation on the same template to distinguish the edited cells from potential spontaneous oxyfluorfen-resistant mutants. When they analyzed oxyfluorfen-resistant clones after transformation, they found that 60% (9/15 clones) had incorporated both the G→A mutation conferring resistance and the silent control mutation, proving that these clones came from true editing events.

For the real test, the authors co-transformed the cells with both sets of editing components at the same time, targeting both FTSY and PPX1. In this case, roughly 1% of the resulting oxyfluorfen-resistant colonies also contained a precisely edited FTSY gene (Figure  1), indicating that once repair-proficient cells were selected through successful editing of PPX1, these cells were rather likely to also contain the edit at the non-selectable FTSY locus. The frequency (∼1%) was low but well within reach for PCR-based screening for positive clones, which should make this method an attractive option for many Chlamydomonas laboratories and applicable to basically any strain background, since it uses the native PPX1 gene to generate a selectable marker. The authors further demonstrated the utility of the co-targeting method by generating precise edits in a gene involved in lipid metabolism.

Figure 1.

Co-editing of a non-selectable (FTSY) and selectable (PPX1) gene in Chlamydomonas. The single-stranded ssODN template for FTSY inactivates the gene and turns edited cells pale green. The ssODN template for PPX1 provides resistance against oxyfluorfen. After electroporation of both editing constructs, together with their cognate Cas9 RNPs, and selection for survivors on oxyfluorfen, ∼1% of precisely edited FTSY mutants are recovered.

Two main pathways have been proposed to explain how single-stranded DNA donors direct homologous repair at DSBs. In the single-stranded template repair pathway, the repair begins with 5′-to-3′ end resection of the DSB to generate two free 3′ ends of single-stranded DNA. One of these ends anneals to and copies information from the complementary ssODN template (Gallagher et al., 2020). In the other pathway, called single-stranded DNA incorporation (Kan et al., 2017), the ssODN is physically incorporated into the final product. Future studies may clarify if ssODN-templated repair in Chlamydomonas follows one of these pathways or perhaps yet another pathway. It will also be interesting to discover how the efficiency of precise editing depends on the number and complexity of mismatches, and what factors currently limit the co-editing efficiency to one percent. With this proof-of-principle publication, many laboratories may use the method and begin to answer some of these questions.

Funding

This project has received funding from the European Union’s Horizon 2020 research and innovation programme under the Marie Skłodowska-Curie grant agreement No 798198.

Conflict of interest statement. None declared.