Use of RNA-Protein Complexes for Genome Editing in Non-albicans Candida Species

Existing CRISPR-Cas9 genome modification systems for use in Candida albicans, which rely on constructs to endogenously express the Cas9 protein and guide RNA, do not work efficiently in other Candida species due to inefficient promoter activity. Here, we present an expression-free method that uses RNA-protein complexes and demonstrate its use in three Candida species known for their drug resistance profiles. We propose that this system will aid the genetic analysis of fungi that lack established genetic systems.

I n the past few years, clustered regularly interspaced short palindromic repeat (CRISPR) genome editing, a process first discovered through the study of bacteriumbacteriophage interactions (1), has emerged as a major strategy for genome editing in eukaryotes (2,3). CRISPR-Cas9 genome modification methods rely on a nuclease, Cas9, to make a precise double-stranded DNA break which can then be repaired by nonhomologous end joining, causing a potential insertion or deletion, or homologous recombination using a construct with homology both up-and downstream of the cut site. Repair constructs contain selectable markers and can be engineered to introduce genome changes such as gene replacements, protein tags, or other directed mutations (4)(5)(6). Cas9 interacts with the guide RNA which directs Cas9 to the cut site by hybridizing to the gene of interest with a 20-bp target-specific guide or protospacer sequence. Cas9-mediated cutting occurs only if the protospacer sequence is directly followed by a PAM (protospacer adjacent motif) site, an essential targeting component of the CRISPR-Cas9 system, which distinguishes bacterial self from nonself DNA in phage resistance (1)(2)(3).
The study of pathogenic fungi, including Candida albicans, has been enhanced by the introduction of strategies for genome modification via CRISPR-Cas9 systems. CRISPR-Cas9-mediated modification of C. albicans was first reported by Vyas et al. (7), using a method that employed the stable expression of the Cas9-encoding gene from a construct located on the chromosome. Later iterations of CRISPR-Cas9 modification of C. albicans demonstrated the use of transient gene expression systems wherein DNAs encoding Cas9 and guide RNAs could be cotransformed with the repair construct (8). In these highly effective published systems, designed for use in C. albicans, the Cas9 protein is derived from DNA constructs in which the CAS9 gene is under the control of a promoter known to be well expressed in C. albicans. Recent studies performed by Norton et al. (see reference 9) have shown that constructs designed for use in C. albicans do not lead to efficient transformation in Candida (Clavispora) lusitaniae due to differences in promoter requirements. A CRISPR-Cas9 system has also been recently established for Candida glabrata, but again, this system relies on the use of speciesspecific promoters from either Saccharomyces cerevisiae or C. glabrata (10). A strategy that does not require the availability of established promoters and terminators for each species could enhance research in diverse fungal pathogens.
Genome editing research in human cells has demonstrated the successful use of a CRISPR-Cas9 method utilizing purified Cas9 protein and CRISPR RNAs, rather than CAS9and guide RNA-expressing constructs (11). Expression-free CRISPR-Cas9 genome editing has also recently been demonstrated in plants (12), algae (13), and the filamentous fungus Penicillium chrysogenum after protoplast formation (14). This method has not yet been applied to pathogenic yeast.
Here, we present genomic modification of C. lusitaniae, Candida auris, and C. glabrata, three species that are notorious for innate or acquired antifungal resistance (15)(16)(17)(18)(19)(20), using CRISPR RNA-Cas9 protein complexes (RNPs) along with a repair construct that contains the desired genome modification. The inclusion of RNPs increased both the number of transformants and the percentage of transformants with the desired mutation in all three Candida species. Using RNP-mediated genome editing, we constructed mutants lacking either a known or a putative catalase gene(s), and in all three species, the deletion of the selected catalase gene led to increased sensitivity to hydrogen peroxide as expected. We expect that this method will be useful for the genetic analysis of diverse fungal species that lack established gene expression systems.

RESULTS
Strategy for RNP-mediated deletion of known or putative catalase genes in diverse Candida spp. The lack of an established gene expression system for CRISPR-Cas9-mediated genome editing slows research on emerging pathogens. In this study, we sought to determine if RNPs, which contain the protein and RNA components of the CRISPR-Cas9 system, could be used to make genetic alterations in three diverse fungal pathogens without the need for defined promoters for heterologous gene expression. To do this, we designed a strategy to knock out proposed or validated catalaseencoding genes in C. lusitaniae, C. glabrata, and C. auris. The gene encoding catalase was targeted because the null mutant phenotype is easily assayed, thus facilitating these proof-of-principle experiments, and because the catalase mutants may be useful in future studies on the role of reactive oxygen species (ROS) in killing by antifungal compounds or through interactions with the host immune system (21,22). Genes in C. lusitaniae, C. glabrata, and C. auris with the highest level of homology to the gene encoding C. albicans catalase, CAT1 (C1_06810W_A), were identified by BLAST (23). C. lusitaniae CLUG_04072, an uncharacterized gene, and C. glabrata CTA1 (24), which has been previously characterized as catalase (25), were found. In C. auris, the gene with the highest homology to C. albicans CAT1 (CaCAT1) was hypothetical gene QG37_05842, but subsequent protein sequence alignment of this open reading frame (ORF) to the known and putative catalases of C. albicans, C. glabrata, and C. lusitaniae, performed using Clustal Omega (26), suggested that this initial annotation did not encompass the full-length protein. This analysis strongly suggested that the ORFs annotated as QG37_05842 and QG37_05843 are actually one gene that encodes a catalase (see Fig. S1 in the supplemental material). To create the constructs necessary to delete the catalase-encoding genes, PCR was used to fuse a selectable nourseothricin resistance marker, NAT1, with DNA sequence from upstream and downstream of the indicated catalase genes. The NAT1 gene was codon optimized for use in Candida spp. (27). The amount of homologous sequence used varied slightly between constructs, ranging from 500 to 1,000 bp. The final constructs contained 896 and 984 bp for C. lusitaniae, 977 and 968 bp for C. auris, and 818 and 531 bp for C. glabrata for the left and right flanks, respectively. The methods for generating these gene deletion constructs are outlined in Fig. 1A.
In the CRISPR genome modification method tested here, the CRISPR machinery that is cotransformed with the gene deletion construct cleaves the chromosome at a site between the regions where homologous recombination is desired. The CRISPR machinery includes purified Cas9 protein and two RNAs: the CRISPR guide RNA (crRNA), which contains 20 bp homologous to the target gene fused to the scaffold sequence, and a universal transactivating CRISPR RNA (tracrRNA), which forms an RNA duplex with ) were amplified using the primers shown in PCRs 1 and 3. The NAT1 cassette was amplified in PCR 2 with primers 5 and 6. Primers 2 and 3 contained reverse-complemented sequences to primers 5 and 6 that were used to fuse the NAT cassette to the LF and RF amplicons in the PCR 4 stitching reaction with nested primers 7 and 8. The resulting gene deletion construct was used for transformation. (B) The gene-specific part of the crRNA is a 20-bp sequence (underlined) that ends with a CRISPR-Cas9 PAM site and is directly adjacent to an additional PAM site (PAM sites shown in bold).
Expression-Independent CRISPR-Cas9 Editing in Fungi the gene-specific crRNA and subsequently complexes with the Cas9 nuclease. To ensure that the protein and RNA components are properly assembled and transformed together, the crRNA and tracrRNA are coincubated and then added to purified Cas9 protein, allowing formation of the RNA-protein complex prior to electroporation. Typically, crRNAs are designed with homology to 20 bp adjacent to a PAM site (NGG). However, to maximize knockout efficiency, we designed the crRNA used in this study to target adjacent PAM sites (NGGNGG) as this dramatically increases CRISPR efficiency in other species (28). Thus, the gene-specific crRNAs used below recognize a 20-bp sequence that ends with a CRISPR-Cas9 PAM site (7) and is directly adjacent to an additional PAM site (Fig. 1B). NCBI BLAST was used to verify that the chosen gene-specific 20 bp had no off-site targets in the genome of the respective fungal pathogen.
RNPs increase the efficiency of C. lusitaniae CLUG_04072 gene replacement. Recent studies have indicated that C. lusitaniae is refractory to the CRISPR method developed for C. albicans (9); thus, we sought to test whether RNPs could increase accurate transformation efficiency. The haploid clinical C. lusitaniae strain A04 was transformed with the CLUG_04072 knockout construct in the presence and absence of CLUG_04072-targeted RNPs (prepared as described in Materials and Methods). As shown in Table 1, inclusion of CLUG_04072-targeted RNPs in the transformation significantly increased the number of nourseothricin-resistant (NAT r ) C. lusitaniae transformants, resulting in approximately 10 times more transformants than transformation with the deletion construct alone. Ten NAT r colonies were randomly chosen for genotype assessment from each transformation reaction using a PCR-based strategy outlined in Fig. 2A. Using the presence of the correct-size band for both the left and right flanking regions of the knockout construct as an indication of correct construct integration, we determined that only 1 out of 10 transformants obtained in the absence of RNPs was correct, whereas 7 out of 10 transformants obtained with RNPs were correct (Fig. 2B). Thus, in C. lusitaniae the addition of gene-specific RNPs to a transformation drastically increases the number of accurate transformants obtained.
RNP-mediated CRISPR-Cas9 gene editing increases the efficiency of gene replacement in C. glabrata and C. auris. To determine if RNPs also increase the efficiency of CRISPR-Cas9-mediated gene replacement in clinical strains of C. auris and C. glabrata, we performed a similar analysis in strains of these species. As for the transformation of C. lusitaniae, the use of RNPs significantly increased the total number of transformants in both species (Table 1). Again, using the presence of the correct-size band for both the left and right flanking regions of the knockout construct as an indication of correct construct integration, we determined that out of 10 randomly selected NAT r transformants, the addition of gene-specific RNPs increased the number of accurate transformants from two to six in C. glabrata and five to seven in C. auris (Fig. 3). Therefore, not only is the number of transformants higher but also the percentage of transformants with the marker integrated at the correct locus is increased by the addition of CRISPR-Cas9 RNPs.
Mutants lacking known or putative catalase-encoding genes have increased hydrogen peroxide sensitivity. We analyzed the catalase phenotypes for four confirmed mutants in each of the three Candida species included in these studies by  (25), and in this strain, deletion of CTA1 had a more modest effect on H 2 O 2 resistance, though the effects were consistent across mutants and reproducible across experiments. The H 2 O 2 susceptibility phenotypes of the C. glabrata cta1⌬ strains are consistent with results reported for a cta1⌬ mutant constructed as described in the work of Cuéllar-Cruz et al. (25).

DISCUSSION
We propose that the methods described here will advance our ability to study pathogens of medical importance that currently lack well-developed systems for their genetic manipulation. One of the benefits of this approach is the ability to use commercially available Cas9 protein and custom-synthesized RNAs; thus, no additional lab equipment or techniques are required, and only the deletion construct needs to be synthesized by the investigator. Both design of the RNA and the deletion construct can be accomplished with minimal knowledge about an organism's biology; the only data necessary for these steps are the gene sequence plus and minus approximately 1,000 bp. We have demonstrated that the delivery of the mixture of Cas9 protein and guide RNAs with the deletion construct enhances the number of total NAT r transformants as well as the percentage of transformants with the correctly integrated marker, Expression-Independent CRISPR-Cas9 Editing in Fungi even in species like C. lusitaniae, which we, and others, have had extreme difficulties mutagenizing using non-CRISPR-Cas9-based methods. Parallel studies by Bennett and colleagues (see reference 9) highlight the inefficiency of a transient expression system developed for use in C. albicans when used in C. lusitaniae and show that efficiency   Fig. 2 and 3. Images were captured after 24 h of growth on either YPD or YPD plus 3 mM H 2 O 2 as indicated.
increases when a species-specific promoter is used to drive the expression of the Cas9-encoding gene. Thus, the RNP-based method may be particularly impactful for species for which promoters that can drive the expression of the CAS9 gene have not been constructed, including C. auris, in which multidrug-resistant strains have emerged multiple times and caused hospital-associated outbreaks (15,17,29,30), and C. glabrata, known for its innate resistance to the most commonly used antifungals (16). In this report, we demonstrate, for the first time, that C. auris is naturally capable of efficient recombination. However, even in C. auris, the inclusion of Cas9 and guide RNAs increased the percentage of transformants with the correct gene replacement, thereby reducing the number of transformants that need to be screened. These studies have built upon the prior work by Vyas and colleagues (7), who developed the guide RNA selection strategy; Min and colleagues, who further characterized the use of CRISPR-Cas9 in C. albicans (8); and Shen and colleagues, who developed the optimized nourseothricin resistance marker for use in fungal species that use the CUG codon to encode serine instead of leucine (27).
To demonstrate the effectiveness of this method, we chose to delete catalaseencoding genes because although catalase-encoding genes are not essential to cellular growth under standard laboratory conditions, they are found among many diverse organisms; thus, we expected to identify putative catalase genes in C. lusitaniae and C. auris. In addition, not only is the loss of catalase activity easily assessed phenotypically, allowing us to show that the genetic evidence of gene deletion matches an expected phenotype, but these enzymes are important for resistance to oxidative damage and associated with virulence; thus, they may be of interest for further study among other labs. Interestingly, we identified another locus in C. lusitaniae with homology to ClCAT1 (CLUG_04072), CLUG_05766; therefore, future studies will be necessary to determine its contribution to oxidative stress resistance in this species. While we chose to knock out catalase-encoding genes to demonstrate the power of the expression-free CRISPR-Cas9 system, the possibilities are endless as to what modification could be made in a diverse array of fungi with this system with minimal tool development.
Many other studies have documented the need for both Cas9 and guide RNA for CRISPR-mediated genome editing. Further, we show that protein, RNA, and DNA can be transformed into cells of Candida species via electroporation without prior formation of protoplasts. While we imagine that the protein and RNA are in complex at the time of transformation, this has not been demonstrated in these studies. The ability to transform protein and RNA into Candida spp. opens up the possibility of transforming other proteins such as biosensors, enzymes, or inhibitory proteins (31).

MATERIALS AND METHODS
Strains and culture conditions. Strains used in this study, listed in Table S1 in the supplemental material, were grown at 30°C in YPD (2% Bacto peptone, 2% dextrose, 1% yeast extract) with shaking. Transformants were selected on YPD with 200 g/ml nourseothricin. Cultures were streaked from frozen stocks stored at Ϫ80°C in 15% glycerol on a weekly basis.
Plasmids and DNA. Gene replacement constructs for transformation were generated by fusion PCR. Briefly, 0.5 to 1.0 kb of the 5= and 3= regions flanking the gene of interest was amplified from genomic DNA (isolated using the MasterPure yeast DNA purification kit [Epicentre]). The nourseothricin resistance (NAT1) cassette was amplified from plasmid pNAT (8). PCR primers used to amplify the three fragments (two flanking regions and NAT1 gene) contained overlapping homologous sequences as necessary for the subsequent fusion PCR using nested primers (Fig. 1A). The fusion PCR resulted in the NAT1 cassette flanked by the 5= and 3= regions of the gene of interest. PCR products for transformations were purified and concentrated with the Zymo DNA Clean & Concentrator kit (Zymo Research) with a final elution of the construct in molecular-biology-grade water (Corning). One microgram of DNA repair construct was used in each transformation (not including negative controls). All oligonucleotides used in this study can be found in Table S2; primer type corresponds with primer numbers indicated in Fig. 1 to 3.
Transformation with RNPs. All transformations were performed by electroporation of competent cells prepared using lithium acetate (LiAc) (32). Overnight cultures with an optical density value at 600 nm (OD 600 ) of approximately 1.6 to 2.2 were pelleted and resuspended in 10 ml of transformation buffer (100 mM LiAc, 10 mM Tris-HCl, 1 mM EDTA). These were incubated with shaking for 1 h, before addition of 100 mM dithiothreitol (DTT) for an additional 30 min. After incubation, cells were washed twice with ice-cold water and once with ice-cold 1 M sorbitol before resuspension in approximately 200 l of ice-cold 1 M sorbitol. Forty microliters of this cell slurry was used per transformation.
Expression-Independent CRISPR-Cas9 Editing in Fungi RNPs were created using the Alt-R CRISPR-Cas9 system from IDT (Integrated DNA Technologies, Inc.) and assembled for usage during the final washing steps to generate competent cells. Stocks of crRNA (gene specific) and tracrRNA (universal) were dissolved in RNase-free distilled water (dH 2 O; 100 M) and stored at Ϫ20°C. To generate the complete guide RNA, equimolar concentrations (4 M final) of the gene-specific crRNA and tracrRNA were mixed in water, with a final volume of 3.6 l per transformation required, and incubated at 95°C for 5 min. Alt-R S.p. Cas9 nuclease 3NLS (60 M stock from IDT) was diluted to 4 M in RNase-free dH 2 O, with a final volume of 3 l per transformation. After the guide RNA (crRNA plus tracrRNA) was allowed to cool to room temperature, it was mixed with diluted Cas9 protein (4 M) in a 1.2-to-1 ratio (3.6 l of guide RNA to 3 l of Cas9 protein) and incubated at room temperature for at least 5 min to assemble the RNP complex. RNPs were used at 6.6 l per transformation.
Electroporation was performed using an 0.2-cm electroporation cuvette and electroporated with a manual 1.8 pulse (Bio-Rad MicroPulser). Immediately following transformation, cells were resuspended in 1 ml ice-cold 1 M sorbitol and then gently pelleted (3 min, 3,000 rpm) before resuspension in 1 ml liquid YPD. Cells were recovered for 2 to 4 h at 30°C while gently shaking. After recovery, cells were pelleted and resuspended in 200 l liquid YPD before aliquots were spread plated onto YPD plates with 200 g/ml nourseothricin and incubated at 30°C for 2 days. To verify nourseothricin resistance, transformants were patched onto selective plates an additional time followed by genotypic and phenotypic characterization.
Negative-control transformation mixtures contained 40 l of cell slurry and 10 l of ice-cold 1 M sorbitol (no RNPs, no DNA repair construct). Negative controls did not yield any colonies when plated on YPD plus nourseothricin but yielded robust growth on YPD alone. Minus-RNP (ϪRNP) transformation mixtures contained 40 l of cell slurry, 1 g of DNA repair construct, and 10 l of ice-cold 1 M sorbitol (no RNPs). Plus-RNP (ϩRNP) transformation mixtures contained 40 l of cell slurry, 1 g of DNA repair construct, 6.6 l of RNPs, and 10 l of ice-cold 1 M sorbitol.
The number of NAT r colonies reported in Table 1 represents an average from 3 or 4 independent transformations. Statistics representing the relationship between ϪRNP and ϩRNP NAT r colonies were generated using a ratio-paired t test.
Phenotypic assessment. To assess H 2 O 2 susceptibility of various Candida species with or without deletions of catalase-encoding genes, strains were grown in YPD medium overnight, with aeration by rotating, at 30°C. These cultures were then diluted with fresh medium to an OD 600 of 1. Serial dilutions of 10-fold were carried out in a microtiter plate to yield seven concentrations ranging from approximately 10 7 cells/ml (for an OD 600 of 1) to approximately 10 1 cells/ml. Ten microliters of each dilution for parental Candida strains and catalase null derivatives was applied to YPD plates containing 0 or 3 mM H 2 O 2 . Images were captured after incubation at 37°C for 24 h.