High-throughput targeted gene deletion in the model mushroom Schizophyllum commune using pre-assembled Cas9 ribonucleoproteins

Efficient gene deletion methods are essential for the high-throughput study of gene function. Compared to most ascomycete model systems, gene deletion is more laborious in mushroom-forming basidiomycetes due to the relatively low incidence of homologous recombination (HR) and relatively high incidence of non-homologous end-joining (NHEJ). Here, we describe the use of pre-assembled Cas9-sgRNA ribonucleoproteins (RNPs) to efficiently delete the homeodomain transcription factor gene hom2 in the mushroom-forming basidiomycete Schizophyllum commune by replacing it with a selectable marker. All components (Cas9 protein, sgRNA, and repair template with selectable marker) were supplied to wild type protoplasts by PEG-mediated transformation, abolishing the need to optimize the expression of cas9 and sgRNAs. A Δku80 background further increased the efficiency of gene deletion. A repair template with homology arms of 250 bp was sufficient to efficiently induce homologous recombination. This is the first report of the use of pre-assembled Cas9 RNPs in a mushroom-forming basidiomycete and this approach may also improve the genetic accessibility of non-model species.


Cas9 RNPs are not sufficient for efficient gene disruption without a positive selection marker.
The minimal form of gene inactivation using CRISPR-Cas9 is the introduction of errors by NHEJ-mediated repair after the introduction of a double-stranded break 37 . We attempted this approach by disrupting a previously introduced nourseothricin resistance cassette in the genome of strain H4.8 A Δhom2 6 . Two sgRNAs targeting the first 100 bp of the nourseothricin resistance gene were designed and transcribed in vitro. The efficacy of these sgRNAs was confirmed by in vitro digestion of the nourseothricin gene. The sgRNAs were assembled with Cas9 and introduced in H4.8 A Δhom2 protoplasts using PEG-mediated transformation. After regeneration, the protoplasts were plated on non-selective medium. The resulting colonies were tested for a loss of nourseothricin resistance, which would indicate a disruption of the nourseothricin resistance cassette by the RNPs. However, despite repeated attempts no nourseothricin sensitive colonies were identified among over 100 potential mutants, indicating a lack of gene-inactivating mutations. This means that the efficiency of the procedure is less than 1%, which strongly reduces its use in high-throughput gene deletion.

Co-transformation of Cas9
RNPs with a repair template greatly increases the rate of gene inactivation. Next, we determined whether Cas9-sgRNA RNPs were able to increase the rate of gene deletions, when used in combination with a repair template containing a selection marker. We used the previously published hom2 deletion vector, which is comprised of a nourseothricin antibiotic resistance cassette flanked by approximately 1200 bp flanks of the hom2 locus (Fig. 1). Furthermore, it contains a phleomycin antibiotic resistance cassette, allowing us to distinguish between the desired gene deletion by homologous recombination (nourseothricin resistant and phleomycin sensitive transformants) and the undesired ectopic integration of the Figure 1. Schematic overview of the approach for targeted deletion of the hom2 gene. Two single guide RNAs (sgRNAs) were designed that target the 5′ region of the hom2 gene. A circular deletion vector was used as a repair template, comprised of homology arms of approximately 1200 bp flanking a nourseothricin resistance cassette. It also contains a phleomycin resistance cassette that was used to distinguish between the desired gene deletion by homologous recombination (nourseothricin resistant and phleomycin sensitive transformants) and the undesired ectopic integration of the full construct (nourseothricin and phleomycin resistant transformants). Moreover, linear PCR products with shorter homology arms were used as repair template, as indicated. The Cas9 protein, sgRNAs and repair template were introduced into protoplasts by PEG-mediated transformation. All repair templates resulted in confirmed gene deletions.
www.nature.com/scientificreports www.nature.com/scientificreports/ full construct (nourseothricin and phleomycin resistant transformants). This approach was used previously to generate a single deletion strain of the hom2 gene that grows faster and is not capable of forming mushrooms as a dikaryon 6 . Two sgRNAs were designed that target the hom2 gene 36 and 57 bp downstream of the translation start site, introducing double-stranded DNA breaks (Fig. 1). The efficacy of the sgRNAs was confirmed by in vitro digestion of the hom2 locus. Wild type H4-8 protoplasts were transformed with the hom2 deletion vector with or without Cas9-sgRNA RNP complexes. After regeneration, the protoplasts were plated on medium containing nourseothricin and the resulting colonies were screened for phleomycin resistance. Similar numbers of nourseothricin resistant and phleomycin sensitive transformants were found (Table S1). However, seven hom2 gene deletions were observed with Cas9 RNPs as determined by PCR, while no gene deletions were observed without Cas9 in three replicate experiments ( Fig. 2; Table S1). Sequencing the genomic locus of these potential gene deletion mutants indeed showed replacement of the hom2 gene with the nourseothricin resistance cassette in all cases (Fig. 3). This indicates that the double-stranded DNA breaks introduced by the Cas9 RNPs increase the rate of gene deletion by homologous recombination. The phenotype of the hom2 deletion strains was the same phenotype as previously reported (Fig. 4) 6 .
A Δku80 background further improves gene deletion efficiency. Current gene deletion methods in Basidiomycota often rely on a Δku80 background, which inactivates the NHEJ pathway and greatly reduces the number of transformants with ectopic integrations. Therefore, we attempted the same transformation method in S. commune Δku80. As expected, few nourseothricin resistant transformants were observed without the addition of Cas9 RNPs and no confirmed gene deletions were identified among these ( Fig. 2; Table S1). In the Δku80 strain co-transformed with Cas9 RNPs, a similar number of nourseothricin resistant transformants was observed as in the wild type strain with Cas9 (Table S1). However, both the number of phleomycin sensitive transformants and  . Sanger sequencing confirms the replacement of the hom2 gene with the nourseothricin resistance cassette for 17 tested deletion strains. The sequence alignments between the left and right homology arms (shaded in blue) are partially shown. As expected, in the wild type the sequence between the homology arms corresponds to the wild type hom2 locus (shaded in green). In contrast, in the 17 deletion strains this sequence corresponds to the nourseothricin resistance cassette (shaded in orange) of the repair template, confirming that the hom2 gene was replaced in these strains.
www.nature.com/scientificreports www.nature.com/scientificreports/ the number of transformants with a confirmed gene deletion were significantly higher ( Fig. 2; Table S1), with a total of 30 confirmed gene deletions. Sequencing the genomic locus of 10 hom2 deletion strains confirmed the expected deletion of the hom2 gene and insertion of the nourseothricin resistance cassette (Fig. 3). This indicates that using Cas9 RNPs in a Δku80 background increases the rate of homologous recombination and results in a higher rate of gene deletion than in the wild type.

Homology arms of 250 bp are sufficient for consistent recruitment of homology-directed repair.
Current gene deletion methods in mushroom-forming Basidiomycota generally rely on repair templates comprised of homology arms of approximately 1000 bp that flank a selectable marker. Indeed, the hom2 deletion vector described above contains homology arms of 1190 and 1147 bp. We determined whether repair templates with shorter homology arms can also assist in gene deletion when used in combination with Cas9-sgRNA RNPs. To generate the repair templates, PCRs were performed on the hom2 deletion vector resulting in linear double-stranded DNA fragments with reduced homology arm lengths of approximately 1000, 750, 500, 250 and 100 bp (Fig. 1). In all cases deletion mutants were obtained, ranging from 3.6 deletion mutants per 2 × 10 7 protoplasts with 1000 bp homology arms to 0.3 deletion mutants with 100 bp homology arms (Table S2). For all flank lengths gene deletions were consistently observed in all replicas, except for 100 bp homology arms where only 1 gene deletion was identified after multiple attempts. This indicates that 250 bp of homology is sufficient to efficiently recruit homology-directed repair, resulting in a gene deletion.

Discussion
Efficient gene deletion methods are essential for the high-throughput study of gene function. Here, we described a method based on a pre-assembled ribonucleoprotein (RNP) complex of Cas9 protein and in vitro transcribed short guide RNAs. This method was used to create a total of 65 deletion strains of the hom2 gene by replacing it with a selectable marker. Moreover, we determined that homology arms of 250 bp are sufficient to efficiently induce homologous recombination. This method yields considerably more gene deletion strains than our previously published methods 5,15 , in which usually only 1 correct deletion strain was obtained and only after repeated attempts.
Various efficient approaches of genome editing using CRISPR/Cas9 have been reported in the model fungus Saccharomyces cerevisiae. For example, cas9 and sgRNAs may be expressed from a plasmid and cause genome edits with such high efficiency that the use of a selectable marker is not necessary 38 . We have attempted this approach in S. commune, but were unsuccessful in producing strains with a targeted gene deletion (data not shown). Nor were we successful with in vitro produced Cas9 RNPs without the use of a selectable marker, since no strains with a gene deletion were identified among >100 colonies. This suggests that the efficiency of the Cas9 RNP approach www.nature.com/scientificreports www.nature.com/scientificreports/ is much lower in S. commune (and possibly other mushroom-forming basidiomycetes) than in Ascomycota. This relatively low efficiency is also observed in Coprinopsis cinerea 28 . In contrast, many correct gene deletion strains were obtained when we attempted to replace the target gene with a selectable marker. We hypothesize that the Cas9 RNPs cause double strand breaks in the target gene, after which the DNA repair machinery is recruited to the site and repairs the damage using the provided repair template. Although this is a relatively rare event, positive selection subsequently allows the identification of the (relatively small number of) correct transformants.
Current gene deletion methods in mushroom-forming Basidiomycota generally rely on templates comprised of homology arms of approximately 1000 bp that flank a selectable marker 5,16 . The creation of this gene deletion template is currently a bottleneck in high throughput gene deletion, although recent approaches for the assembly of multiple DNA fragments (e.g. Gibson assembly) have considerably improved this 39 . Our results show that homology arms as short as 250 bp can efficiently induce homologous recombination when used in combination with RNPs. This relatively short length of the template is well within the range of cost-effective chemical synthesis, which will further facilitate high-throughput gene deletion by allowing in vitro construction of this template.
The methods described in this study will remove several bottlenecks for efficient genome editing in non-model mushroom-forming fungi, as well as wild isolate strains of S. commune. Firstly, our approach works efficiently not only in a Δku80 background, but also in the wild type. This will simplify the gene deletion procedure for non-model system species for which no Δku80 strains are available. Secondly, the expression levels of cas9 and the sgRNAs do not need to be optimized using organism-specific promoters, since both components are supplied externally. Previous studies have shown that the selection markers needed for the repair template are functional in other mushroom-forming fungi 40 .
In addition to gene deletion, these techniques can be modified to produce other genome edits, including promoter replacement and targeted gene integration (knock-in). Recent efforts in genome and transcriptome sequencing of mushroom-forming species have resulted in a large number of candidate gene families involved in mushroom development and other processes 11,14,[41][42][43][44][45][46][47] . The important addition of Cas9 RNP-mediated genome editing to the molecular toolkit of S. commune will facilitate the functional characterization of these gene families, leading to important new insights into the biology of this important group of fungi.
Cas9 expression and purification. pET-NLS-Cas9-6xHis was a gift from David Liu (Addgene plasmid # 62934; http://n2t.net/addgene:62934; RRID:Addgene_62934) 49 and contains a NLS-Cas9-6xhis fusion sequence under control of the T7 promotor. The plasmid was introduced into Escherichia coli BL21 Star (DE3) (Life Technologies, USA). The resulting expression strain was pre-grown for 16 hours in liquid culture at 37 °C in LB supplemented with 50 μg ml −1 ampicillin. This pre-culture was diluted 1:100 in fresh LB with ampicillin and grown in liquid culture at 22.5 °C to an OD 600 of 0.6. Cas9 expression was induced with IPTG at a final concentration of 0.4 mM and incubation was continued for an additional 18 hours. Cells were harvested by centrifugation and resuspended in 80 ml wash buffer (20 mM Tris-HCl pH 8.0, 30 mM imidazole, 500 mM NaCl) supplemented with 1 mM PMFS, 1 mM MgCl 2 and 6.25 U ml −1 DNase I (ThermoFisher Scientific, USA). Cells were lysed three times with a French press (AMINCO, USA) at 500 psi on high setting in a pre-chilled cylinder (4 °C). After lysis, all steps were performed at 4 °C. The lysate was cleared of cellular debris by centrifugation for 30 minutes at 15,000 g. Per liter of culture, 8 ml of bed volume of Ni-NTA resin (Qiagen, Germany) was added and incubated for one hour with gentle agitation. Column purification was performed according to the manufacturer's recommendations and Cas9 was eluted in 24 ml wash buffer with 500 mM imidazole. The protein purity was assessed on SDS-PAGE with colloidal coomassie staining 50 and protein yield was quantified by BCA-assay (ThermoFisher Scientific, USA). Imidazole and salts were removed by three washing steps with 21 ml dialysis buffer (20 mM Tris-HCl pH 8.0, 200 mM KCl, 10 mM MgCl 2 ) on Amicon 30 kDa filters (Merck Millipore, USA). Cas9 protein was diluted to 10 mg ml −1 in dialysis buffer, frozen in liquid nitrogen and stored at −80 °C until use.
Design and synthesis of gene-specific sgRNAs for gene disruption. Candidate protospacers (including PAM site) were identified in the coding region of the nourseothricin resistance cassette 48 and hom2 (protein ID 257987 in version Schco3 of the genome of S. commune 10 ) with CCtop 51 and checked against the full genome to identify potential off-target regions (defined here as regions with fewer than 5 bp differences with the actual target region as identified by Bowtie2) 52 . Two sgRNAs were selected per gene based on fewest off-targets and the presence of one or more guanines at the start of the sgRNA as this promotes a high yield of in vitro T7 transcription. For the nourseothricin resistance gene, sgRNA 1 and 2 were located 57 and 22 bp downstream of the translation start site, respectively, and for hom2, sgRNA 1 and 2 were located 57 and 36 bp downstream of the translation start site of hom2, respectively (Fig. 1). Oligos were designed (p1_sgRNA and p2_sgRNA; Table 1) and the sgRNAs were synthesized in vitro according to the specifications of the GeneArt Precision sgRNA Synthesis Kit (ThermoFisher Scientific, USA).
In vitro Cas9 digestion assay. The efficacy of the RNPs was tested by an in vitro digestion of a template containing the target regions. For the nourseothricin resistance cassette, the resistance gene was cut from the pDELCAS deletion vector 5 with EcoRI, resulting in a 3759 bp fragment. In vitro digestion by the Cas9 sgRNA www.nature.com/scientificreports www.nature.com/scientificreports/ RNPs results in two fragments of 2719 bp and 1040 bp for sgRNA 1 and 2684 bp and 1075 bp for sgRNA 2. The template for hom2 was generated by PCR with primers annealing to the promotor and terminator region of hom2 (hom2_250_fw and hom2_250_rv). This resulted in a 3656 bp fragment that can be digested by Cas9 with either sgRNA to two fragments of 433 bp and 3223 bp for sgRNA 1 and 412 bp and 3244 bp for sgRNA 2. For digestion, 1 μg of Cas9, 200 ng sgRNA (which is approximately a 1:1 molar ratio) and 1x Cas9 reaction buffer (20 mM HEPES pH 6.5, 100 mM NaCl, 5 mM MgCl 2 0.1 mM EDTA) were mixed in 25 μl final volume and assembled at 37 °C for 10 minutes. After assembly, 500 ng of template (100 ng μl −1 ) was added to a final volume of 30 μl and the sample was incubated at 37 °C for 15 minutes. Cas9 was degraded by the addition of 1 U Proteinase K (ThermoFisher Scientific, USA) and incubated at room temperature for 15 minutes, after which DNA digestion was verified on agarose gel.
Design of the repair templates for homologous recombination. The previously described hom2 deletion vector was used as a template for homologous recombination 6 . This plasmid contains a nourseothricin resistance cassette flanked by 1200 bp homology arms outside hom2. Moreover, the plasmid harbors a phleomycin resistance cassette that is only integrated if the plasmid integrates through a single cross-over (i.e. ectopically). Linear templates with reduced homology arm lengths (approximately 1000 bp, 750 bp, 500 bp, 250 bp and 100 bp) were made by PCR on the full vector. The primers were designed to bind 1000 bp (hom2_1000_ fw and hom2_1000_rv;  Table 1) outside of the nourseothricin resistance cassette (Fig. 1).
In the case of the nourseothricin gene inactivation, 4 × 10 3 protoplasts in 200 μl were added. Subsequent transformation was performed as previously described, but without selection 6 . The regenerated protoplasts were plated on non-selective medium and grown for two days.
In the case of hom2 deletions, the number of protoplasts was changed to 2 × 10 7 in 200 μl and 12.5 μg repair template was added. The regenerated protoplasts were plated on medium supplemented with nourseothricin and grown for 3 days. Next, 48 colonies were randomly selected per transformation after three days and transferred to fresh selection media for a second round of selection. In the case of transformation with a full hom2 deletion plasmid, nourseothricin-resistant colonies were then transferred to phleomycin selection medium and their resistance was scored. DNA was isolated from agar plugs by microwaving samples for 120 seconds in 50 μl TE buffer (10 mM Tris pH 8.0, 1 mM EDTA) from candidates that were resistant to nourseothricin and sensitive to phleomycin for PCR verification with primers before the upstream homology arm (hom2_ChkA; Table 1) and after the downstream homology arm (hom2_ChkD; Table 1) 6,53 . The size of the DNA band was used to determine whether the gene of interest was replaced with the nourseothricin resistance cassette. Primers hom2_100_fw and hom2_100_rv (Table 1) were used to sequence the PCR fragment to confirm the correct integration.