Single crossover-mediated targeted nucleotide substitution and knock-in strategies with CRISPR/Cas9 system in the rice blast fungus

Clustered regularly interspaced short palindromic repeats (CRISPR)/CRISPR-associated protein 9 (Cas9)-mediated genome editing has become a promising approach for efficient and versatile genetic engineering in various organisms; however, simple and precise nucleotide modification methods in filamentous fungi have been restricted to double crossover type homologous recombination (HR). In this study, we developed a novel genome editing strategy via single crossover-mediated HR in the model filamentous fungus Pyricularia (Magnaporthe) oryzae. This method includes the CRISPR/Cas9 system and a donor vector harboring a single homology arm with point mutations at the CRISPR/Cas9 cleavage site. Using this strategy, we demonstrated highly efficient and freely programmable base substitutions within the desired genomic locus, and target gene disrupted mutants were also obtained via a shortened (100–1000 bp) single homology arm. We further demonstrated that this method allowed a one-step GFP gene knock-in at the C-terminus of the targeted gene. Since the genomic recombination does not require an intact protospacer-adjacent motif within the donor construct and any additional modifications of host components, this method can be used in various filamentous fungi for CRISPR/Cas9-based basic and applied biological analyses.

reporter gene knock-ins 6 . This strategy enabled efficient and precise targeted gene disruption and knock-in with short homology arms; however, the desired alteration of the nucleotides cannot be performed.
Filamentous fungi comprise a large and diverse group of eukaryotic organisms and are associated with various food, agricultural, and pharmaceutical industries. In contrast, this group contains human and plant pathogens including strains that can produce toxic substances, such as aflatoxin 7 . Therefore, developments in genetic engineering of filamentous fungi can directly affect various fields of basic and applied biological research. CRISPR/ Cas9 technologies have been established as one of the simplest genome editing tools in several filamentous fungi 8 and are still being used in the non-model fungal strains [9][10][11][12] . In contrast to other eukaryotes, in a few filamentous fungi, the introduction of DSBs without homologous templates tends to induce large deletions [13][14][15] ; whereas, HR-mediated gene modifications including the MMEJ-mediated knock-in have been demonstrated in most filamentous fungi with higher efficiency than that of the other eukaryotes, excluding yeast 8,15 . The HR repair can be further divided into two sub-pathways, noncrossover (gene conversion) and crossover type recombination 16 . In the gene conversion process, the cleaved DNA site searches homologous or identical sequences as a donor template, and copies their sequence information (Fig. 1a,b). In the crossover process, the DNA strand exchanges occur between the cleaved DNA and donor DNA strands through the resolution of a double holiday junction by a cut at the crossover sites via structure-selective nucleases (Fig. 1a,b) 17,18 . Although crossover type repairs seem to be limited in mitotic DSB repair in mouse embryonic stem cells and human lymphoblastoid cells 19,20 , this has not been validated in the filamentous fungi.
Pyricularia oryzae (Magnaporthe oryzae) is a model filamentous fungus, which causes destructive fungal diseases in rice. This fungus comprises mononuclear haploid cells and has been established as a basic tool for the analysis of DSB repair and genome editing 13,[21][22][23] . In this study, we used HR detection/selection system to survey whether the crossover type HR has a potential to be applied for the genome editing in P. oryzae. Using the obtained data, we established simplified and free genome editing strategies for targeted gene disruption, base substitution, and knock-in in the filamentous fungus.

Results
Detection and validation of the crossover type HR in the genome of P. oryzae using reporter gene constructs. We previously established an HR detection/selection system for evaluating the DSBmediated HR in P. oryzae 21 . This system comprises two nonfunctional yellow fluorescent protein (YFP) and blasticidin S (BS) deaminase (BSD) fusion genes (ISTG-YFP::BSD and RS-YFP::BSD) and a rare-cutting endonuclease I-SceI. To validate whether the crossover type HR can be induced by targeted DSBs in the genome of P. oryzae, we further developed this system by modifying pRS-YFP::BSD, for detecting the crossover type HR. The bialaphos resistant (bar) gene cassette was inserted into the pRS-YFP::BSD vector, which resulted in pRS-YFP::BSD-bar (donor vector) (Fig. 1a). The donor and I-SceI expression (DSB inducer) vectors were simultaneously introduced into the protoplasts of P. oryzae transformant ISTG6 that has a single copy of ISTG-YFP::BSD (recipient gene) in the genome 22 . Since the recipient gene contains an 18-bp I-SceI recognition site between 327 and 328 bp of the YFP::BSD open reading frame, co-introduction with the DSB inducer and donor vector can induce the recipient gene specific DSB and HR between the donor and recipient YFP::BSD gene regions (Fig. 1a). The HR occurredcells would include functionally restored YFP::BSD gene. When the gene conversion type HR occurred between the donor and recipient genes, the cells would reveal only blasticidin S resistance. In contrast, the crossover type HR would induce the integration of the whole donor vector sequence into the genome, which would result in double-resistant colonies for BS and bialaphos (Fig. 1a). The transformed protoplasts were selected by BS, and then the obtained colonies were counted and transferred onto PSA medium containing bialaphos (30 µg/mL). Consistent with the previous study, co-introduction with the DSB inducer and donor template generated BS resistant colonies; however, no drug-resistant colonies were obtained when only the donor vector was introduced into protoplasts (Fig. 1b). About a half of the BS resistant colonies showed resistance against bialaphos (Fig. 1b). PCR and sequencing analysis showed that randomly selected double drug resistant colonies (5/5) were produced by crossover type HR. Similar results were observed when the protoplasts were initially selected by bialaphos, and then the colonies were transferred onto the PSA plates with BS. Nevertheless, the double-resistant colonies were fewer than those initially selected by BS (Fig. 1b), suggesting that the initially selected bialaphos resistant colonies comprised the transformants produced by the ectopic insertion of bar cassette. These results indicated that introducing site-specific DSB could efficiently induce crossover type HR at the recipient gene locus in P. oryzae.
Single crossover-mediated targeted nucleotide substitution with CRISPR/Cas9 system. We previously succeeded in increasing the HR-mediated targeted gene replacement at the scytalone dehydratase (SDH) www.nature.com/scientificreports www.nature.com/scientificreports/ gene (MGG_05059) locus via cotransfromation with the fungal CRISPR/Cas9 and targeting vectors 23 . The SDH gene disrupted mutants can be detected by the loss of melanin deposition (white phenotype). To induce and evaluate the single crossover-mediated target base substitutions, we constructed a donor vector optimized for the induction of single crossover type HR. The 1335-bp SDH region containing point mutations at the CRISPR/Cas9 target site was obtained by fusion PCR, and was cloned into the vector harboring the hygromycin B phosphotransferase (hph) gene cassette (pMK-PSDH). Because the DNA strand exchanges occur after the DSB repair of cleaved DNA by the copy of the donor DNA sequence, specific DSB-mediated single crossover would induce a pin-point base substitution (stop codon insertion) at the desired SDH locus and integrate whole vector sequence containing hph gene cassette into the genome (Fig. 2b). In contrast, the gene conversion type HR would occur only in the base substitutions (Fig. 2b). The donor vector and each pCRISPR/Cas-U6-1 site2, pCRISPR/Cas-U6-2 site2, or pCRISPR/Cas-TrpC site2, targeting SDH gene 23 , were simultaneously introduced into the wild-type protoplasts and we obtained the hygromycin B resistant colonies. By introducing with the donor and CRISPR/Cas9 vector, the white colony count was dramatically increased, compared with that of only the donor vector (Fig. 2c,d). Sequencing analysis showed that all the white colonies (5/5) comprised desired point mutations (Fig. 2e) and the whole vector sequences were integrated into the SDH locus. Consistent with the previous study 23 , in these white colonies, the Cas9 gene was not detected by the PCR analysis using a specific primer set (Table S1). These results indicated that a single crossover-mediated HR was induced by CRISPR-based DSB and this strategy enables targeted base substitutions with high efficiency.
To assess the feasibility of this strategy, the donor vectors harboring very short homologous sequence (1000, 750, 500, 250, or 100 bp) with stop codon was constructed (Fig. 3a). Both donor and CRISPR/Cas9 vectors were co-introduced into the wild-type protoplasts, and the efficacies of the SDH disruption were evaluated by the depleting melanin deposition. White colonies were obtained with high efficiencies (10-25.58%) by co-introduction with pCRISPR/Cas-U6-1 site2 and each donor vector comprising 250-1000 bp homologous sequences (Fig. 3b). The pCRISPR/Cas9-U6-2 could produce more white colonies than pCRISPR/Cas9-U6-1, and it generated the mutant using only 100-bp homologous sequence (Fig. 3b). These results indicated that a single crossover-mediated gene disruption can be performed by using the shortened homology arms in P. oryzae.
Single crossover-mediated reporter gene knock-in with CRISPR/Cas9 system. To investigate whether the single crossover type HR can be applied for the reporter gene knock-in, we constructed a donor vector comprising a GFP fused SDH homologous sequence. In addition, we introduced silent mutations at the CRISPR/Cas9 target site to avoid the cleavage of the donor vector. Precisely, a single crossover recombination would introduce GFP at the C-terminus of the SDH gene (Fig. 4a). The hygromycin B resistant colonies with the GFP fluorescence were not obtained via introduction with the only donor vector, whereas co-introduction with the CRISPR/Cas9 and donor vectors helped in obtaining the transformants with the GFP fluorescence ( Fig. 4b-d).
Randomly selected nine fluorescent colonies comprised the desired silent mutations at the CRISPR/Cas9 target site of the SDH gene (Fig. 4e). These results indicated that the CRISPR/Cas9-induced single crossover-type recombination can be applied for the one-step reporter gene knock-in at the targeted genomic locus.
Functional and expression analysis of Spo11 using single crossover-mediated gene disruption and reporter gene knock-in. To assess whether the single crossover-mediated genome editing is effective for other genomic loci, we analyzed the functions and expression of SPO11 (MGG_10666) in P. oryzae. In Saccharomyces cerevisiae, Spo11 catalyzes DSB formation to act via topoisomerase-like reaction in meiosis 24 . Since asexual reproduction is predominantly observed in P. oryzae, the function of genes related to the meiotic recombination in asexual life cycle has not been characterized. We created two donor vectors for the functional disruption (Fig. S1) and GFP tagging of Spo11 (Fig. S2). When seven independent well-grown colonies obtained via co-introduction with CRISPR/Cas9 and donor vectors for functional disruption (pMK-PSPO) were subjected to sequencing analysis, three out of seven transformants had desired stop codons at the SPO11 gene (ca. 43%); however, the three mutants indicated no significant differences in the growth, conidiation, germination, www.nature.com/scientificreports www.nature.com/scientificreports/ and appressorium formation rates, compared with those of the wild-type strain. For the Spo11-GFP plus CRISPR/Cas9 co-transformation, ten hygromycin B-resistant colonies were randomly selected and two (20%) contained the desired allelic replacement based on PCR and sequencing analysis; however, the GFP fluorescence was not observed during the vegetative growth and appressorium formation. These results suggest that the single-crossover strategy could be applied universally across genomic loci. Furthermore, we have employed this strategy to functionally demonstrate that Spo11 does not play a role in the asexual life cycle of P. oryzae.

Discussion
We demonstrated that a single crossover type HR could be induced by the site-specific DSB, which allows free and efficient genome editing including targeted nucleotide substitutions and reporter gene knock-ins in the model filamentous fungus P. oryzae. Because the CRISPR/Cas9 system derived from S. pyogenes is considerably simpler than the other CRISPR/Cas systems, this strategy would be useful for various types of research and can be applied to a majority of filamentous fungi. To the best our knowledge, this is the first demonstration of efficient and desired base editing without additional modifications of host components, such as NHEJ related genes disruptions, in filamentous fungi.
The easiest functional gene disruption method with the CRISPR/Cas system is the NHEJ-mediated mutagenesis because this strategy only needs to integrate the CRISPR/Cas system into the genome. Nevertheless, in filamentous fungi, the repair of DSBs without donor templates often induces undesired large deletions or mutations [13][14][15] . Additionally, the constitutive expression of the CRISPR/Cas9 system would trigger mutations outside of the target site (off-target effect), which might affect the phenotypes of interest 25 . Using the CRISPR/Cas9 www.nature.com/scientificreports www.nature.com/scientificreports/ system and the other nucleases, HR could be efficiently induced via the transient expression of the nucleases in fungal protoplasts 8 , which would allow the minimization of the off-target effects. The single crossover-mediated genome editing is easier than the double crossover-mediated gene replacement because the donor vector can be constructed via one-step cloning. Moreover, as the single crossover-mediated recombination enables the integration of the whole plasmid structure into the targeted genomic locus via DNA strand exchange, a large genomic region, such as a metabolic gene cluster, could be efficiently knocked-in at the desired locus.
The most significant limitation of the CRISPR/Cas system is the requirement of the PAM sequence for effective DNA cleavage. Recently, various CRISPR/Cas homologs and orthologs, with various types of the PAM, have been identified from a broad range of bacteria and archaea, and their application in flexible genome editing tools is progressing 26 . The protein engineering of Cas9 also enabled modifications in the PAM recognition 27 . Nevertheless, for pin-point genome and base editing, it is necessary to use the appropriate CRISPR/Cas homolog, ortholog, or engineered Cas protein every time. One of the advantages of our strategy is that its target bases can be freely modified independent of the PAM sequence. We anticipate that the single crossover-mediated genome editing strategy can expand the scope of the genome editing technology, and aid a variety of basic and applied biological research in filamentous fungi.

Methods
Fungal strain, growth conditions, and DNA analysis. The P. oryzae strain Hoku-1 was used as a wild-type strain in this study. The media composition and genomic DNA extraction were as described previously 21 . PCR and sequencing analysis were performed via standard procedures 22 . BLAST search was performed using http://blast.be-md.ncbi.nlm.nih.gov/Blast.cgi. Primers and oligonucleotides used in this study are listed in Table S1.

Construction of a single crossover detection/selection system and its validation.
For detecting the single crossover events in the genome of P. oryzae, the YFP::BSD-based HR detection/selection system 22 was modified. To construct pRS-YFP::BSD-bar, as a donor vector for the single crossover, the bar gene cassette was excited from pGEM-bar 22 via ApaI digestion and was inserted into the ApaI site of pRS-YFP::BSD 22 which resulted in pRS-YFP::BSD-bar. The pGEM-I-SceI::SvNLS and ISTG6 strains were used as DSB inducers and recipient host cells, respectively 22 . The preparation of ISTG protoplasts and PEG transformation were described previously 22 , and the procedures were slightly modified. Briefly, 2.5 µg of each pRS-YFP::BSD-bar and pGEM-I-SceI::SvNLS vectors was mixed and added to the protoplasts. About 24 h after transformation, the protoplasts were transferred onto a PSA plate with BS (200 µg/mL) or bialaphos sodium (30 µg/mL). For the validation of single crossover-mediated HR between the ISTG region and pRS-YFP::BSD-bar, the PCR analysis was performed by using a YFP-seq-1/TTrpC-1 primer set. The PCR products were sequenced by using the YFP-seq-1.

Construction of CRISPR/Cas9 expression vector targeting endogenous genes. Sense
(SPO-gRNA-1) and antisense (SPO-gRNA-2) oligonucleotides for the CRISPR/Cas9 targeting SPO11 were designed using the website service, CRISPR direct (http://crispr.dbcls.jp), and were annealed according to the procedures of the previous report 28 (Table S1). The annealed oligonucleotides were inserted into pCRISPR/Cas-U6-2 via the Golden Gate cloning method according to the previous report 23 with certain modifications. Briefly, 0.8 µL 50-fold-diluted oligonucleotides, 0.3 µL pCRISPR/Cas-U6-2, 0.2 µL Esp3I enzyme (Thermo Scientific), 0.2 µL Quick ligase (New England Biolabs), and 0.4 µL 10 × T4 DNA ligase buffer (New England Biolabs) were mixed in a single tube and the volume was made up to 5 µL by adding sterilized water. The mixtures were then subjected to a thermal cycling reaction as follows: 6 cycles of 37 °C for 5 min and 16 °C for 10 min. After the cycling reaction, 1 µL of the reaction mixtures was used for transformation of Escherichia coli DH5α chemical competent cells.

Construction of donor vectors for single crossover recombination.
The SDH gene region comprising point mutations was amplified via fusion PCR methods. For the first PCR, the SDH gene regions were amplified from the genome of P. oryzae using a PSDH-1/-2 primer set and PSDH-3/-4 set, respectively. The PCR products obtained from the first PCR were mixed and used for the second PCR. The fusion PCR product was generated by a PSDH-1/-4 primer set and cloned between the HindIII and SpeI sites of pMK412-dGFP 21 , which resulted in pMK-PSDH. The donor vectors with short homology regions (pMK-PSDH1000, pMK-PSDH750, pMK-PSDH500, pMK-PSDH250, and pMK-PSDH100) were constructed from pMK-PSDH via PCR-based methods and cloned between HindIII and SpeI sites of pMK412-dGFP. The PSDH-1/PSDH1000-2, PSDH-1/ PSDH750-2, PSDH500-1/-2, PSDH250-1/-2, and PSDH100-1/-2 were used for the amplification of 1000, 750, 500, 250, and 100 bp of SDH region with mutations, respectively. The donor vector for GFP knock-in at the C-terminus of the SDH gene was constructed via fusion PCR methods. For the first PCR, the SDH gene regions with silent mutations, which removed the start and stop codons, were amplified from pMK-PSDH using a KSDH-1/-2 primer set and KSDH-3/-4 primer set, respectively. The PCR products obtained from the first PCR were mixed and used for the second PCR. The fusion PCR product was generated by a KSDH-1/-4 primer set. The GFP gene and putative terminator region of the SDH gene were also amplified from pMK412 21 and pMK-PSDH using KGFP-1/-2 and KSDH-5/PSDH-4 primer sets, respectively. The PCR products were mixed and used for second PCR, and the fusion PCR product was generated by a KGFP-1/PSDH-4 primer set. The fusion PCR products generated by the KSDH-1/-4 and KGFP-1/KSDH-6 primer sets were mixed and used for the third PCR. The third PCR product was generated by a KSDH-1/PSDH-4 primer set and cloned between the HindIII and SpeI sites of pMK-dGFP, which resulted in pMK-KSDH-GFP. The SPO11 gene region with point mutations was constructed via fusion PCR methods. For the first PCR, the SPO11 gene regions were amplified from the genome of P. oryzae using PSPO-1/-2 and PSPO-3/4 primer sets, respectively. The PCR products obtained from the first PCR were mixed and used for the second PCR. The fusion PCR product was generated by a PSPO-1/-4 primer set and was www.nature.com/scientificreports www.nature.com/scientificreports/ cloned between the HindIII and SpeI sites of pMK412-dGFP, which resulted in pMK412-PSPO. The donor vector for GFP knock-in at the C-terminus of the SPO11 gene was constructed via fusion PCR methods. For the first PCR, the SPO11 gene regions with silent mutations, which removed the start and stop codons, were amplified from the pMK-KSPO and genome using a KSPO-1/-2 and KSPO-3/-4 primer sets, respectively. The PCR products obtained from the first PCR were mixed and used for the second PCR. The fusion PCR product was generated by a KSPO-1/KSPO-4 primer set and was cloned between the SpeI and SacI sites of pMK412-dGFP, which resulted in pMK-KSPO. The GFP gene and terminator region of SPO11 were amplified from the pMK412 and genome using the KGFP-3/-4 and KSPO-5/PSPO-4 primer sets, respectively. The PCR products obtained from the first PCR were mixed and used for the second PCR. The fusion PCR product was generated by a KGFP-1/PSPO-4 primer set and was cloned between the AscI and PacI sites of pMK-KSPO, which resulted in pMK-KSPO-GFP.
Transformation of P. oryzae with CRISPR/Cas9 and donor vectors. The 2.5 µg of each CRISPR/Cas9 vector and donor vector was mixed and added to the protoplasts. About 24 h after transformation, the protoplasts were transferred onto a PSA plate with hygromycin B (200 µg/mL).