Marker-free genome editing in Ustilago trichophora with the CRISPR-Cas9 technology

ABSTRACT

In this communication, we report the adaptation of the CRISPR-Cas9 technology in Ustilago trichophora prototrophic wild-type isolate obtained from its natural host Echinochloa crus-galli. The established CRISPR vector and method enable a rapid and marker-free introduction of Cas9-induced non-homologous end-joining (NHEJ) dependent mutation at the targeted gene. Moreover, the method allows a specific modification of the chromosomal DNA sequence by Cas9-induced homologous recombination using short DNA repair templates. The results demonstrate the applicability of the CRISPR-Cas9 technology in U. trichophora for both gene knock-out by the NHEJ pathway and specific gene modification by templated genome editing, paving the way for rapid metabolic engineering of this Ustilago species for industrial applications.

Introduction

The development of CRISPR-Cas9 technology has facilitated the genetic modification of a plethora of eukaryotic cells including non-model and industrial relevant organisms [1,2]. The method is based on the expression of the DNA endonuclease Cas9, e.g. SpCas9 from Streptococcus pyogenes, and a synthetic non-coding RNA of ~ 100 nucleotides (nt) in length, known as single-guide RNA (sgRNA). [3,4] Loaded with the sgRNA, Cas9 screens the genome for the presence of a DNA sequence that is complementary to the first 20 nt guide sequence of the sgRNA. The Watson-Crick base pairing of the guide sequence with the complementary strand of the target DNA leads to the formation of an R-loop, which triggers the nuclease activity of Cas9 leading to double-strand DNA break (DSB). Typically the DSB is located within the targeted DNA site 3-bp upstream of the protospacer-adjacent motif (PAM). [4–7] By modulation of the guide sequence of the sgRNA, Cas9 can be reprogrammed to target nearly any desired DNA sequence within a genome. [3,4] The DSB is repaired by the cell´s endogenous repair pathways, which is primarily either the non-homologous end-joining (NHEJ) pathway or the homology directed recombination (HDR) [8]. The NHEJ pathway is error-prone and leaves short indel errors (insertion/deletion of few nucleotides), which can result in frame-shift mutation of the Cas9-targeted gene. [4,9,10] Alternatively, by providing DNA repair template with homologous sequences flanking the DSB, genomic DNA sequences can be specifically modified trough Cas9-induced homologous recombination. Both strategies enable genome modification without the use of selection markers (e.g. antibiotic resistance), which is an important issue in the industrial food or feed biotechnology.

Members of the fungal family Ustilaginaceae have been recognized as promising industrial production hosts due to their capability to produce polyols and organic acids through metabolizing various carbon sources including raw materials or waste streams like crude glycerol [11–13]. One example for a valuable organic acid naturally produced by Ustilaginaceae is malate, which is employed in a wide range of applications in the food and pharmaceutical industry and personal care products [14–16]. Recently, it has been reported that among more than 70 analyzed Ustilaginacea, U. trichophora is one of the most efficient dicarboxylic acid producers on glycerol medium [16]. U. trichophora is known for its pathogenicity towards Echinochloa crus-galli and a close relative to the well-studied fungal pathogen U. maydis [17,18]. Molecular tools established for U. maydis have been recently adapted in U. trichophora including site-specific integration of genes-of-interest by homologous recombination [14]. Chromosomal integration by homologous recombination was hitherto the most employed method for the genetic manipulation in U. maydis [11,19,20]. Recently, the CRISPR-Cas9 technology has been established in U. maydis and successfully used for a marker-free gene knock-out through Cas9-induced NHEJ-mediated mutation. [21]

Here, we demonstrate the adaptation of the CRISPR-Cas9 technology in a wild-type U. trichophora strain isolated from its host Echinochloa crus-galli. The method enables the rapid and marker-free generation of NHEJ-mediated mutations. Moreover, using short double-stranded DNA repair templates, it is possible to edit the genome of U. trichophora in a sequence-specific manner.

Results and discussion

Construction of a CRISPR-Cas9 vector for genome editing in Ustilago trichophora

The simplest use of the CRISPR-Cas9 technology for genome editing is based on the plasmid-based co-expression of the Cas9 protein and a sgRNA. In eukaryotic cells a nuclear localization sequence must be attached to the Cas9 protein [4]. The transcription of the sgRNA has to be driven by an RNA polymerase III-dependent promoter with a precisely mapped transcriptional start site in order to prevent a post-transcriptional modification of the sgRNA. Alternatively, the sgRNA can be flanked by self-splicing ribozymes sequences, which ensure the release of unmodified sgRNAs transcribed from RNA polymerase II promoters. [22] It has been shown that the human-codon optimized cas9 is functional in several yeast cells. [23] Moreover, in Pichia pastoris the expression of human-codon optimized cas9 gene was necessary to achieve efficient gene knock-out, likely due to the toxic effect of high amounts of Cas9 in the cell by using a Pichia codon-optimized cas9 gene [24].

To establish the technology in U. trichophora we constructed the vector pMF-hCas9-PU6-Tri-scaffold at which the human-codon optimized cas9 gene from S. pyogenes is under the control of the constitutive promoter Phsp70 and the transcription terminator Tnos (Figure 1(a)). For the expression of the corresponding sgRNA based on the S. pyogenes CRISPR system, we used the U6 gene sequence of U. maydis to identify the U6 promoter in the U. trichophora genome. [21] The U6 gene sequences of U. maydis and trichophora share a similarity of 80% with a highly conserved 5´-terminal domain (Supplemental Fig. S1A). The 5´-terminal domains of eukaryotic U6 RNAs are known to form a stem-loop structure. [25] On the basis of the predicted structure of the5´-terminus of U. trichophora U6 RNA (Supplemental Fig. S1B), we suggested that the transcriptional start site of the U6 gene is conserved in U. maydis and trichophora [21]. We used the 550 bp upstream DNA region of the proposed U6 transcription start site as promoter sequence for the transcription of the sgRNA scaffold (Figure 1(a) and (b)), which exhibited a DNA sequence similarity of only 44% to the corresponding region in the U. maydis genome (Supplemental Fig. S1A). The sgRNA scaffold contains five consecutive T´s at the 3´-end, which acts as termination signal for RNA polymerase III promoter. Two inversely oriented BbsI restriction sites were included between the transcription start site and the sgRNA-scaffold, which enables the insertion of customized guide sequences to target a specific chromosomal locus by Cas9 (Figure 1(b)).

Figure 1.

Constructed CRISPR-Cas9 vector for genome editing in U. trichophora (a) The vector map pMF-hCas9-U6-tri-scaff is shown. The vector contains: ColE1 (colicin E1) and ApR (beta-lactamase), HygR (hygromycin B phosphotransferase), hCas9 (human codon-optimized Cas9, NLS from SV40 on both ends and an N-terminal V5-tag and an OFP gene coupled to hCas9 with a 2A sequence), BbsI endonuclease restriction site for the insertion of the guide sequence in-between of the U6 promoter of U. trichophora and the sgRNA scaffold. (b) The cloning scheme for guide sequences is shown. Inversely oriented BbsI sites are indicated in red, the cleavage sites by dashed lines. The transcription start site is marked with + 1. The ligation of two hybridized oligonucleotides with appropriate 5´-overhangs reconstitutes the start nucleotide ‘G’ of the U6 promoter followed by the 19 to 20 nt guide sequence (indicated by N´s) of sgRNA.

Note that the constructed vector pMF-hCas9-PU6-Tri-scaffold contains two copies of the constitutive promoter Phsp70 and three copies of the terminator Tnos. Moreover, the vector lacks any sequence elements known to mediate an autonomous self-replication in U. trichophora, suggesting that the stability of the plasmid will be impaired in U. trichophora. However, the transient expression of Cas9 and the sgRNA immediately after the transformation will very likely be sufficient to obtain gene targeting, comparable with genome editing approaches through electroporation of Cas9-sgRNA ribonucleoproteins. [26] The plasmid stability in E. coli NEB10β was not affected and no recombination events were observed during cloning procedures of different guide-sequences.

Targeting U. trichophora genome with CRISPR-Cas9

To test the functionality of the constructed CRISPR-Cas plasmid we designed guide sequences against the malA gene, which encodes for the conserved malic enzyme (Malate dehydrogenase, oxaloacetate-decarboxylating, (NADP+) (EC 1.1.1.40)) that converts malate with the cofactor NAD(P) to pyruvate, CO₂ and NAD(P)H. Two different guide sequences were derived from the coding region of the malA gene of U. trichophora, both flanked by the PAM ‘NGG’ at their 3´-ends, which is required for the targeting by the SpCas9 (Figure 2(a)). The malA-targeting guide sequences were cloned into the pMF-hCas9-PU6-Tri-scaffold in a single-step restriction-ligation reaction as described in the Material and methods section, yielding the plasmids pMF-hCas9-PU6-Tri-malAg#1 or pMF-hCas9-PU6-Tri-malAg#2, respectively. U. trichophora wild-type protoplasts were transformed either with the malA-targeting vectors or the empty control vector pMF-hCas9-PU6-Tri-scaffold, respectively. After the cultivation of the transformants for three days on agar plates supplemented with hygromycin, the genomic DNA of single clones were prepared, the malA region was amplified by PCR and analyzed by Sanger sequencing. As shown in Figure 2(b) transformation of U. trichophora with the pMF-hCas9-PU6-Tri-malAg#1 led to modified malA gene in the majority of the analyzed clones with deletions of 6 to 18 nucleotides immediately upstream and downstream of the protospacer-adjacent motif (Figure 2(b)). Similar, the transformation with pMF-hCas9-PU6-Tri-malAg#2 resulted in short deletions at the corresponding downstream target site without affecting the target site of malAg#1 (Figure 2(c)). In contrast, the transformation with the control vector pMF-hCas9-PU6-Tri-scaff without a specific guide sequence resulted in wild-type sequences. To analyze the appearance of potential off-target effects, the genomic DNA of two malAg#1-modified clones were extracted and nine potential off-target sites predicted by CCTop were PCR amplified and analyzed by Sanger sequencing (Supplemental Table S1) [27]. The results show that none of the analyzed potential off-target sites were affected.

Figure 2.

Modification of the malA gene with CRISPR-Cas9 in U. trichophora by using the vectors and malAg#2. .(a) The two target sites of malAg#1 and malAg#2 with their corresponding protospacer adjacent motifs (PAMs) in the malA gene are shown. The predicted Cas9-cleavage site 3-bp upstream of PAM is indicated. (b) Sequences of the targeted malA region from 13 transformants of pMF-hCas9-PU6-Tri-malAg#1 (numbered with 1–13) and 12 transformants of the control vector pMF-hCas9-U6-tri-scaff (numbered with c1-c12) are aligned to the wild-type malA sequence. The vertical line indicates the predicted CRISPR-Cas9 induced DSBs. Clone 10 (indicated with an asterisk) contained wild-type malA sequence. (c) Sequences of the targeted malA region from 4 analyzed transformants of pMF-hCas9-PU6-Tri-malAg#2 are aligned to the wild-type malA sequence.

Next, two guide-sequences targeting two different sites within the coding region of the pyr6 gene have been designed in order to proof a phenotypic effect of CRISPR-Cas9 induced mutations in U. trichophora. The pyr6 gene is a homolog of the ura3 gene of S. cerevisiae and encodes for oritidine-5´-phosphate decarboxylase [28]. Similar to the ura3-based selection system, pyr6-deficiency has been shown to confer resistance of U. maydis to 5-fluoroorotic acid (5-FOA) and enables a positive selection of the mutated transformants [29,30]. As can be seen in Supplemental Fig. S2A, several 5-FOA-resistant clones were obtained for both targeted regions within the pyr6 gene. Sequencing of the targeted regions confirmed the successful introduction of NHEJ-mutations at the corresponding DNA sites (Supplemental Fig. S2).

In addition to the deletion of a few nucleotides, the deletion of larger DNA regions ranging from 508 to 1270 bp around the Cas9-induced cleavage site at the malA locus was also observed (Supplemental Table S2). This could indicate an ineffective NHEJ repair system or strong exonuclease activity in some U. trichophora transformants resulting in extended exonucleolytic degradation at the DNA ends. Therefore, the Cas9-mediated targeting of a specific gene could cause unintended loss of gene functions at the surrounding regions, at least in some of the obtained transformants. To overcome this limitation, we next tested a more specific gene modification in U. trichophora by Cas9-mediated homologous recombination using a synthetic DNA repair template.

Genome editing with short homologous-recombination templates

The deletion of large regions of Cas9-induced DNA suggests a variable NHEJ-repair capacity in U. trichophora. It is well known that Ustilago sp. has very efficient homologous recombination system [31,32], which is likely preferred over NHEJ when introducing DSB using Cas9. In Saccharomyces cerevisiae a specific modification of chromosomal DNA using CRISPR-Cas9 technology can easily be achieved by the co-transformation of short synthetic DNA-repair templates [10]. To analyze a homology directed repair (HDR) of the Cas9-induced DSB in U. trichophora, the pMF-hCas9-PU6-Tri-malAg#1 was co-transformed together with a HDR-template. The double-stranded HDR-template was generated by hybridization of two complementary 94-mer oligonucleotides containing an EcoRI restriction site flanked by ~ 45-bp sequences with homology to the upstream and downstream region of the DSB (Figure 3(a)). Genomic DNA of randomly picked transformants was isolated and the malA region was amplified by PCR. Indeed, the EcoRI cleavage pattern of the PCR fragments indicated a specific insertion of the repair template into the targeted malA gene (data not shown), which was verified by Sanger sequencing (Figure 3(b)).

Figure 3.

Insertion of an EcoRI restriction site into the chromosomal malA gene using Cas9-mediated homologous recombination. (a) The malAg#1 targeted region and the 94-mer double-stranded repair template is shown. The EcoRI restriction sites and the primers used for the amplification of the malA fragment (oligonucleotides 17 and 18, Supplemental Table S3) are indicated. (b) Sequences of the targeted malA-region from five transformants are aligned to the wild-type malA sequence.

In summary, the vector described in this communication can be applied for specific gene knock-out in U. trichophora. The U. trichophora U6 promoter used in this study is appropriate to express functional sgRNA. Moreover, despite the lack of an autonomous replication sequence, the vector enables the modification of chromosomal DNA in U. trichophora. Due to the unpredictable outcome of the NHEJ-mediated modification, we recommend a templated genome editing, which enables a fast, easy and marker-free specific editing using synthetic double-stranded oligonucleotides. The vector and methods described here complement the available genetic tools to further improve the production of bio-based chemicals by U. trichophora as a promising industrial platform microorganism.

Material and methods

Strains, media and growth conditions

The strain E. coli NEB10β (New England Biolabs) was used for routine cloning procedures and cultivated in Luria-Bertani (LB, 1% Tryptone, 0.5% NaCl, 0.5% Yeast Extract) medium supplemented with the appropriate antibiotic at 37°C and 250 rpm shaking. The U. trichophora strain, isolated from Echinochloa crus-galli in its unicellular haploid saprophytic stage, was cultivated in YEP liquid medium (1% yeast extract, 1% peptone) on a rotary shaker at 30°C and 250 rpm. For the selection of transformants the medium was supplemented with 400 µg/ml hygromycin B (Thermo Fisher Scientific).

Construction of the pMF-Cas9-U6tri-scaff

All oligonucleotides used in this study were obtained from biomers.net GmbH (Ulm, Germany) and are listed in the Supplemental Table S3. The plasmid pMF1-hs was kindly provided by Prof. Dr. M. Feldbrügge (University Düsseldorf, Germany). The plasmid is based on the vector pBluescript KS(+) and contains a hygromycin resistance cassette (hygR) [19,20]. Cas9 including the nuclear localization signal (NLS) of the simian virus 40 large T-antigen (SV40) was amplified using oligonucleotide primers 1 and 2 (Supplemental Table S3) from the CRISPR Nuclease OFP vector obtained from Thermo Fisher Scientific (Schwerte, Germany) [33]. The amplification of the pMF1-hs backbone was achieved by using primer 3 and 4 (Supplemental Table S3). The resulting double-stranded DNA fragments were assembled to the plasmid pMF-Cas9 using Gibson Assembly Cloning Kit from New England Biolabs (Frankfurt am Main, Germany). The U6 promoter region was amplified from U. trichophora genome using the oligonucleotide primers 5 and 6 (Supplemental Table S3). The pMF-Cas9 was amplified using the oligonucleotide primers 7 and 8, and both fragments were assembled by Gibson Assembly to obtain pMF-Cas9-U6tri. The hygR expression cassette was obtained by digestion of the pMF-1hs with the restriction enzyme StuI (NEB) and the resulting 1884 bp fragment was ligated into pMF-Cas9-U6tri linearized with PsiI (NEB). The scaffold of the sgRNA template was synthesized by GeneArt (Regensburg, Germany). After digestion of the GeneArt Vector with BsmBI, the resulting 399 bp fragment was ligated into the BbsI-digested pMF-Cas9-U6tri to obtain the final CRISPR-Cas9 plasmid pMF-Cas9-U6tri-scaff. The sequences of the Cas9 expression cassette and the sgRNA transcription unit are shown in Supplemental Table S4.

Cloning of the guide sequences

200 pmol forward and 200 pmol reverse guide oligonucleotides (9–14, Supplemental Table S3) were initially denatured by incubating 5 min at 98°C in the annealing buffer (end concentration: 10 mM Tris-HCl pH 7.5, 0.5 M NaCl, 50 mM EDTA, DEPC-treated H₂O) in a total volume of 20 µl followed by slow cooling down to room temperature. One-step restriction-ligation was performed by digestion of 1 µg pMF-Cas9-U6tri-scaff with 10 U FD-BpiI (Fast digest BpiI; Fisher Scientific) in the provided FD-buffer in a total volume of 10 µl and incubation at 37°C for 8 min in a water bath. Subsequently, 1 µl hybridized oligonucleotides were mixed with 1 µl QuickLigase (NEB) and 8 µl 2x T4 ligase buffer. The reaction mixture was incubated for 4 min at room temperature and for additional 4 min at 37°C. 5 µl of the ligation mixture was used for the transformation of 50 µl of E. coli NEB10β cells (New England Biolabs, C3019H) according to manufacturer´s protocol. The isolation of the plasmid DNA was achieved by using Gene Jet Plasmid Miniprep Kit (Thermo Fisher Scientific) and analyzed by Sanger sequencing (GATC Biotech) using the primer 15 or 16 (Supplemental Table S3).

Protoplast generation and transformation

The protoplast preparation was carried out as described previously. [34,35] A 5 ml preculture of U. trichophora was cultivated overnight at 30°C in YEP medium. The day after, a 25 ml YEPS (1% yeast extract, 1% peptone, 0.5 M sorbitol) medium was inoculated at OD₆₀₀ 0.05 using the preculture. After 5 h incubation at 30°C on a rotary shaker, the cells were harvested by centrifugation (21°C, 4000 rpm, 10 min) and washed once in SCS buffer (20 mM sodium citrate pH 5.8; 1.5 M sorbitol) followed by a second centrifugation. The harvested cells were resuspended in 1 ml of lysis buffer (20 mg/ml lysing enzymes from Trichoderma harzianum in SCS buffer). Protoplast formation was monitored with a phase-contrast microscope. 15 ml of SCS buffer were added and the protoplasts were collected by centrifugation (10°C, 2000 rpm, 10 min) and washed 3-times using 15 ml SCS buffer. The pellet of protoplasts was resuspended in 150 µl STC buffer (100 mM calcium chloride, 1 M sorbitol, 10 mM Tris-HCl buffer) and stored on ice. All following DNA transformation steps were performed on ice. 50 µl protoplasts were incubated with 1 µg plasmid DNA for 10 min. After the addition of 500 µl STC/PEG 4000 buffer (40% PEG 4000, 100 mM calcium chloride, 1 M sorbitol, 10 mM Tris-HCl buffer) the mixture was incubated for additional 10 min. The protoplast-DNA suspension was spread onto freshly made two-layer selective agar consisting of a 15 ml lower layer (YEP agar supplemented with 1 M sorbitol and 800 µg/ml Hygromycin B) and a 15 ml top layer (YEP agar supplemented with 1 M sorbitol). The plates were incubated 3 days at 30°C. Genomic DNA of U. trichophora was prepared according to the manufactory instruction of „MasterPure Yeast DNA Purification“ kit (Epicentre). For templated genome editing 2 µg HDR template DNA in a total volume of 4 µl was co-transformed together with 1 µg of the corresponding plasmid DNA as described above. The HDR template was prepared by incubation of 12.5 µg forward and reverse oligonucleotides (29 and 30, Supplemental Table S3) in the annealing buffer for 5 min at 95°C followed by slow cooling down to room temperature. The malA region was amplified by PCR using the oligonucleotide primers 17 and 18 (Supplemental Table S3).

Funding Statement

The results described were generated in one of the projects of the ‘Innovationsallianz ZeroCarbFP’ that was supported by the Bundesministerium für Bildung und Forschung within the ‘Innovationsinitiative industrielle Biotechnologie’ (FKZ 031B0181F);Bundesministerium für Bildung und Forschung [FKZ 031B0181F].

Acknowledgments

We would like to thank Prof. Dr. Michael Feldbrügge for kindly providing the plasmid pMF1-hs, the members of ZeroCarbFP for helpful discussions, Jan Ziegler and Zihni Arslan for technical support.

Disclosure statement

No potential conflict of interest was reported by the authors.

Supplementary materials

Supplementary data for this article can be accessed here.