CRISPR/Cas9‐mediated knockout of an oil palm defense‐related gene to the pathogenic fungus Ganoderma boninense

Oil palm plantation in Indonesia is significantly affected by basal stem rot disease caused by the pathogenic fungus Ganoderma boninense. Tolerant oil palm cultivars toward G. boninense have been developed through a breeding program accelerated by the implementation of the CRISPR/Cas9 technology. This study was conducted to perform a gene knockout (KO) of oil palm that confers a putative defense‐related trait toward G. boninense. A plasmid pCRISPR_EMLP containing modules, i.e., 35S‐CaMV‐promoter‐driven CRISPR/Cas9, U6‐promoter‐driven sgRNA to the target EgEMLP gene (EL695076), and hygromycin resistance gene as the selectable marker, was established for Agrobacterium‐mediated delivery into oil palm calli (OPC). The transformed OPCs were regenerated and screened in DF (de Fossard) media containing hygromycin. The working concentration of hygromycin was successfully optimized for selection at 20 ppm. Through PCR‐based selection using HYG primers, we succeeded in discerning positive transformed OPC clones. The sequenced PCR products of genomic DNA as the template amplified using EMLP1 primers showed a point mutation, causing a frameshift in the edited EgEMLP and premature stop codon. Furthermore, in silico modeling demonstrated that the mutation resulted in a change in the C‐terminal region, affecting the tertiary protein structure. Moreover, electrophoresis analysis of PCR products of cDNA as the template from transformed OPC clones showed several samples with faint or undetected bands. This indicated that the CRISPR/Cas9 module induced a mutation that could destabilize the transcribed mRNA, e.g., premature degradation. Altogether, this study has successfully implemented CRISPR/Cas9 gene editing in oil palm in a model gene that is responsible for putative defense‐related traits toward the pathogenic fungus G. boninense.


Introduction
Basal stem rot (BSR) disease that affects the oil palm is caused by the pathogenic fungus Ganoderma boninense (Ho and Tan 2015). This disease affects a wide area of oil palm plantation in Indonesia, causing devastating eco nomic losses amounting to 500 million USD every year (Ommelna et al. 2012; Hushiarian et al. 2013. Recently, several efforts have been made to overcome the BSR dis ease. However, reported infections, casualties, and losses still appear to increase every year.
Currently, the development of tolerant oil palm culti vars against G. boninense is considered as the best solu tion. An engineered tolerant cultivar can be established using a breeding program from parent stocks with known traits resisting G. boninense. However, this traditional program is tedious and requires a long time to generate an improved trait.
An accelerated breeding program can be achieved by the implementation of cuttingedge technology, e.g., CRISPR/Cas9mediated gene editing (Jaganathan et al. 2018). It has been reported that the CRISPR/Cas9 system naturally occurs in bacteria, which functions as an immu nity element against virus (phage) invasion by incising the foreign DNA fragments (Horvath and Barrangou 2010). This system was repurposed and engineered to break the DNA target at a specific site using programmable guide RNA (sgRNA) (Jinek et al. 2012). The programed sgR NAs are generally synthetically built adjoining the Cas9 module in a suitable expression plasmid. When expressed, the Cas9 enzyme acts as a homing endonuclease directed by the sgRNA, which can hybridize to a particular site within the targeted DNA (Mali et al. 2013). Cas9 in duces doublestrand breaks (Shen et al. 2017), which are later repaired by the intracellular machinery. Edits are introduced in this manner in the form of point mutation, insertion, and deletion through the nonhomologous end joining (NHEJ) pathway (Chiruvella et al. 2013) or by gene replacement through homologydirected recombina tion (Hahn et al. 2018). Therefore, the trait of interest can be generated using CRISPR/Cas9 using advantageous cis genesis (Hou et al. 2014), avoiding the introduction of any transgenes (Telem et al. 2013).
In this study, we aimed at improving a trait of oil palm against infection with G. boninense. This was achieved by conducting a gene knockout (KO) experiment on the defenserelated gene against the pathogenic fungus. The target gene was determined from another study that had demonstrated that the expression of EgEMLP was elevated during infection with G. boninense in oil palm. We hy pothesized that EgEMLP is a marker gene that, if knocked out, will alter the oil palm trait, preferably increasing its tolerance toward G. boninense. In this study, we suc ceeded in knocking out EgEMLP by demonstrating several postediting effects at the genomic (DNA) and transcrip tomic (RNA) levels and through protein modeling. Further study is important to examine the growth of the edited oil palm clones under G. boninense infection in the field.

CRISPR/Cas9 editing module design
CRISPR/Cas9 modules were designed according to an earlier pipeline study (Budiani et al. 2018). The edit ing module was constructed to the target EgEMLP gene (EL695076). It was selected as described in a previous study that demonstrated its elevated expression in oil palm leaves upon G. boninense infection (Tan et al. 2013). We hypothesized that the CRISPR/Cas9mediated knockout on EgEMLP would repress its expression, thereby alter ing the trait of the edited oil palm.
Optimum designed sgRNAs were assessed to examine the ontarget specificity toward the target gene and the off target probability toward nontarget sites. One optimum sgRNA was selected for synthesis and assembly into the expression plasmid ( Figure 1) using the services of Sigma Aldrich.

Agrobacterium-mediated transformation, calli regeneration, and screening
The transformation was delivered using Agrobacterium mediated protocols as optimized in previous research (Bu diani et al. 2019). The transformed oil palm calli (OPC) were regenerated in de Fossard (DF) media supplemented FIGURE 1 CRISPR_EMPLP construct containing Cas9, sgRNA for EgEMLP, and hygromycin selectable marker expression modules.
with hygromycin. Optimization of hygromycin for the se lective marker was conducted using working concentra tions of 10, 20, and 30 ppm. The working concentration for supplementation was considered to be lethal if it can in duce calli browning. The optimal working concentration was deduced as the concentration at one level lower below the lethal concentration. Screenings were conducted to se lect the population of calli, which can survive under the supplementation of hygromycin at the optimum working concentration. They were then subjected to further screen ing.

DNA and RNA isolation, cDNA synthesis, and electrophoresis
All samples were ground using mortar and pestle in liquid nitrogen. Genomic DNA from calli was extracted using the Genomic DNA Mini Kit Plant (Geneaid) according to the manufacturer's instruction. Total RNA was extracted using the RNAEasy Plant Mini Kit (Qiagen). Firststrand cDNA synthesis was conducted using the iScript cDNA Synthesis Kit (BIORAD). The electrophoresis was per formed in an agarose gel (0.5% for RNA or 0.8% for DNA) in 1 × TBE buffer in DEPCtreated or nucleasefree water.

PCR-based selection, sequencing, and sequence analysis
PCR experiments were conducted to screen for positive transformed OPC clones using HYG and EMLP1 primers ( Table 1). The HYG primers were designed to amplify the hygromycin resistance gene. Meanwhile, the EMLP1 primers were designed to flank the regions targeted by the sgRNA where the Cas9 protein will cut and cause a doublestrand break.
The PCR experiments using EMLP1 were conducted using genomic DNA and cDNA as templates to observe

In silico modeling of partial EgEMLP protein structure
The amino acid sequences of EgEMLP exon of wildtype (nonedited) and transformed (edited) were translated using a standard genetic code in frame with the reference gene in Geneious Prime suite (Biomatters, Ltd.). The amino acid sequences were submitted to the ITASSER server (Yang 2007) to yield the protein structure model. Ideal ized models were retrieved separately and saved as. PDB file. Both models were superimposed in the PyMOL suite (Schrödinger LLC). A descriptive analysis was conducted on the superimposed models to evaluate the differences between edited and nonedited models.

Transformation and optimization of hygromycin working concentration for calli screening
Screened calli were further selected by PCRbased selec tion using HYG and EMLP1 primers to confirm the trans formations of the pCRISPR_EMLP construct. The elec trophoresis profile obtained by PCRbased screening us ing HYG primers demonstrated successful OPC transfor mation using the pCRISPR_EMLP construct ( Figure 3, upper panel). Meanwhile, the application of 10 ppm hy gromycin was considered to produce falsepositive results that allowed the growth of most of the calli. Therefore, the optimum working concentration of hygromycin was determined as 20 ppm, even if it could induce calli brown ing after 4 and 8 weeks of culture ( Figure 2 D and E). In this case, the results demonstrated that a gradually lowered hygromycin working concentration should be used dur ing calli subcultures. Further screening procedures should also be implemented, e.g., PCRbased selection.

PCR-based screening of CRISPR/Cas9-edited calli
Screened calli were further selected by PCRbased selec tion using HYG and EMLP1 primers to confirm the trans formations of the pCRISPR_EMLP construct. The elec trophoresis profile obtained by PCRbased screening using HYG primers demonstrated successful OPC transforma tion using the pCRISPR_EMLP construct (Figure 3, upper panel). All the examined calli harbored the hygromycin resistance gene with an expected amplicon size of 700 FIGURE 2 Calli selection in DF media supplemented with hygromicin. All calli were grown with 250 ppm cefotaxime to reduce Agrobacterium tumefaciens over growth. Based on this result, the optimum hygromycin concentration for calli selection was deduced at 20 ppm, as higher concentration (30 ppm) will induce calli death (browning) while lower concentration at 10 ppm will produce nontransformed (escapee) calli thus allowing for false positives.
bp, which conferred the ability to grow in media supple mented with hygromycin. PCRbased experiments were also established using EMLP1 primers to examine the edit ing events at the genomic level. The PCR was able to pro duce the expected distinct amplicon at the size of 300 bp using the genomic DNA as the template (Figure 3A.). Fur ther examinations were conducted to confirm the editing events using Sanger sequencing of these amplicons. PCR experiments were also conducted using EMLP1 primers on cDNA templates to examine whether the edit ing events can be discerned at the transcriptomic (mRNA) level. The electrophoresis profile revealed that some calli produced faint or none of the expected bands (200 bp) ( Figure 3B.), which indicated that editing events could be discerned at the transcriptomic level.
In this study, we established a procedure to screen OPC by the CRISPR/Cas9mediated editing technology using PCR and electrophoresis on the cDNA template. We succeeded in demonstrating the CRISPR/Cas9induced knockout (KO), which can later change the level of tran scribed mRNA as shown by the faint or none of the ex pected bands. Therefore, the edited and nonedited calli obtained by the CRISPR/Cas9mediated KO can be dif ferentiated using routine PCR.

Sequence analysis of CRISPR/Cas9-edited EgEMLP gene
Sequence analysis of the PCR amplicons produced using the genomic DNA template was conducted using align ment against an identical genomic DNA template from wildtype sequences. This analysis demonstrated that one of the edited calli harbored a mutation (Figure 4) located at

FIGURE 4
The pairwise sequence alignment of the edited calli against wild-type EgEMLP sequence produced in Geneious Prime suite. Highlighted nucleotide by red triangles showed where the change made by CRISPR/Cas9 creating SNP (single nucleotide polymorphism) and changing the translated amino acid sequence.
the CRISPR/Cas9 target site for EgEMLP. This indicated that the mutation occurred as a base insertion due to inter nal DNA repair after the doublestrand breakinduced by CRISPR/Cas9. This mutation induced a frameshift on the translated amino acid sequence, causing a change in the translated protein starting at the point mutation onward. It also introduced a premature stop codon in the edited se quence.

Homology-based modeling for CRISPR/Cas9edited EgEMLP protein
In silico modeling was implemented to determine the ef fect of the mutation on the protein structure. As shown in Figure 5, the mutation distinctively changed the ter tiary structure of EgEMLP protein based on the super imposed structures between the edited and wildtype pro teins. The mutation resulted in the premature stop codon and deleted nine amino acid residues at the Cterminal re gion of EgEMLP protein. the adjacent alphahelix structure, indicating disruption of the tertiary structure, which may abolish the protein func tion.
Altogether, this study was successful in implement ing the CRISPR/Cas9 editing technology to the engineer ing of OPC. Moreover, the optimum working concentra tion of hygromycin used in calli transformation screening was established, and it was deduced that a gradually low ered hygromycin working concentration could avoid false positive results. Furthermore, successful implementation of CRISPR/Cas9mediated editing was demonstrated us ing results evidenced at multiple levels ranging from ge nomic, transcriptomic, and protein modeling. This study could open up the possibility of engineering crops, partic ularly with long growth periods, e.g., the oil palm, accord ing to various applications, through the implementation of the cuttingedge markerassisted molecular plant breeding program.

Conclusions
In this study, we had successfully implemented the CRISPR/Cas9 system to conduct a gene knockout in the OPC. We also successfully demonstrated the editing events using multiple pieces of evidence at the genomic and transcriptomic levels. Protein models of the edited and nonedited gene were also generated to predict the struc tural change that may cause disruption in its function.