LbCas12a-D156R Efficiently Edits LOB1 Effector Binding Elements to Generate Canker-Resistant Citrus Plants

Citrus canker caused by Xanthomonas citri subsp. citri (Xcc) is an economically important disease in most citrus production regions worldwide. Xcc secretes a transcriptional activator like effector (TALE) PthA4 to bind to the effector binding elements (EBEs) in the promoter region of canker susceptibility gene LOB1 to activate its expression, which in turn causes canker symptoms. Editing the EBE region with Cas9/gRNA has been used to generate canker resistant citrus plants. However, most of the EBE-edited lines generated contain indels of 1–2 bp, which has higher possibility to be overcome by PthA4 adaptation. The adaptation capacity of TALEs inversely correlates with the number of mismatches with the EBE. LbCas12a/crRNA is known to generate longer deletion than Cas9. In this study, we used a temperature-tolerant and more efficient LbCas12a variant (ttLbCas12a), harboring the single substitution D156R, to modify the EBE region of LOB1. We first constructed GFP-p1380N-ttLbCas12a:LOBP, which was shown to be functional via Xcc-facilitated agroinfiltration in Pummelo (Citrus maxima) leaves. Subsequently, we stably expressed ttLbCas12a:LOBP in Pummelo. Eight transgenic lines were generated, with seven lines showing 100% mutations of the EBE, among which one line is homozygous. The EBE-edited lines had the ttLbCas12a-mediated deletions of up to 10 bp. Importantly, the seven lines were canker resistant and no off-targets were detected. In summary, ttLbCas12a can be used to efficiently generate biallelic/homozygous citrus mutant lines with short deletions, thus providing a useful tool for the functional study and breeding of citrus.


Introduction
Citrus is one of the top fruit crops worldwide. However, citrus production faces many biotic and abiotic challenges including citrus huanglongbing and bacterial canker, droughts, flooding, and freezes [1][2][3][4][5]. CRISPR-mediated genome editing is promising in dissecting the genetic determinants for improving citrus fruit quality and yield and resistance against biotic and abiotic stresses, and in precision breeding [6][7][8][9]. Jia and Wang [8] first adapted SpCas9/gRNA from Streptococcus pyogenes to edit citrus PHYTOENE DESATURASE (PDS) via transient expression in Citrus sinensis. Transgenic expression of SpCas9/gRNA [7,[9][10][11][12][13][14][15], SaCas9/gRNA from Staphylococcus aureus [16], and LbCas12a/crRNA from Lachnospiraceae bacterium [17] have been used for citrus genome editing. Both Huang et al. [7] and Dutt et al. [6] demonstrated the genome editing of citrus protoplasts using SpCas9/gRNA. CRISPR-mediated genome editing has been used to generate canker resistant citrus varieties. Citrus canker is an economically important citrus bacterial disease that is present in most citrus producing countries. Citrus canker is caused by Xanthomonas citri subsp. citri (Xcc). Xcc causes typical hypertrophy and hyperplasia symptoms by secreting a transcriptional activator like effector (TALE) PthA4, a key pathogenicity factor of Xcc, into the nucleus of plant cells. PthA4 binds to the effector binding elements (EBEs) in the promoter region of the canker susceptibility gene LOB1 to activate its expression, thus causing canker symptoms [18][19][20]. Genome editing of the coding region of LOB1 abolishes the canker symptoms caused by Xcc [12]. In addition, genome editing of the EBE region also confers citrus resistance to Xcc [10,13,14,21]. Editing the EBE region has advantages over editing the coding region by reducing the putative side effect of mutation of the coding region. In our previous studies, we have generated one homozygous Pummelo (Citrus maxima) mutant line containing an adenine deletion within EBE PthA4 -LOBP, and one biallelic Pummelo mutant line comprising a thymine insertion within one EBE PthA4 -LOBP allele and a two-adenine deletion within another allele [13]. We have also generated one biallelic 'Duncan' grapefruit (C. paradisi) mutant line containing a thymine insertion in one allele and an adenine insertion in another allele [21]. These EBE-edited lines are resistant to citrus canker. However, it was reported that TALE effectors are capable of overcoming the disease resistance caused by the mismatches between TALEs and the edited EBE regions [22]. Our data showed that the adaptation capacity of TALEs inversely correlates with the number of mismatches. TALEs harboring seven to nine mismatches were unable to adapt to overcome the incompatible interaction, whereas TALEs that harbored a small number of mismatches (≤5) to the EBE were able to adapt [22]. Thus, it is necessary to generate EBE-edited citrus plants with more mutations. To achieve this goal, we have been using multiple approaches, including increasing the efficacy of the Cas9/sgRNA-based genome editing [10] and using non-Cas9/sgRNA-based tools such as Cas12a, which is known to generate longer deletion than Cas9 [17,23].
In this study, we used ttLbCas12a to modify citrus EBE PthA4 -LOBP. ttLbCas12a function was first tested via Xcc-facilitated agroinfiltration in Pummelo leaf. Subsequently, we conducted stable expression of ttLbCas12a in Pummelo and EBE PthA4 -LOBP was successfully modified with multiple mutations up to 10 nucleotides deletion. Notably, the ttLbCas12a-mediated mutation rates were 100% in seven transgenic Pummelo lines, including one homozygous line, and the seven edited lines were resistant against citrus canker.

Figure 1.
Schematic representation of the binary vector GFP-p1380N-ttLbCas12a:LOBP used to modify Type II LOBP. (a) Pummelo Type II LOBP. Part of the Type II LOBP sequence and its chromatogram were presented, in which EBEPthA4 was highlighted by red rectangles. A crRNA was designed to target EBEPthA4-LOBP, which was indicated by blue. (b) Schematic diagram of GFP-p1380N-ttLbCas12a:LOBP. LB and RB, the left and right borders of the T-DNA region; CsVMV, the cassava vein mosaic virus promoter; GFP, green fluorescent protein; 35T, the cauliflower mosaic virus 35S terminator; CmYLCV, the cestrum yellow leaf curling virus promoter; NosP and NosT, the nopaline synthase gene promoter and its terminator; ttLbCas12a, temperature-tolerant LbCas12a containing the single mutation D156R; AtU6-26, Arabidopsis U6-26 promoter; target, the 23 nucleotides of Type II LOBP highlighted by blue, was located downstream of protospacer-adjacent motif (PAM); HH, the coding sequence of hammerhead ribozyme; HDV, the coding sequence of hepatitis delta virus ribozyme; NptII, the coding sequence of neomycin phosphotransferase II.

Xcc-Facilitated Agroinfiltration in Pummelo
Pummelo (Citrus maxima) was grown in a greenhouse at around 28 °C and was pruned for uniform shooting before Xcc-facilitated agroinfiltration. It should be kept in mind that ttLbCas12a performed better at 28 °C [41].
Xcc-facilitated agroinfiltration was performed as described previously with minor modification [43]. Briefly, the fully-expanded young Pummelo leaves were pre-treated with XccΔgumC [44], which was re-suspended in sterile tap water at a concentration of 5 × 10 8 CFU/mL. Twenty-four hours later, the pre-treated leaf areas were inoculated with Agrobacterium cells harboring GFP-p1380N-ttLbCas12a:LOBP or p1380-AtHSP70BP-GUSin. GFP was observed and photographed four days after agroinfiltration. p1380-AtHSP70BP-GUSin was used as a control as described elsewhere [43]. Part of the Type II LOBP sequence and its chromatogram were presented, in which EBE PthA4 was highlighted by red rectangles. A crRNA was designed to target EBE PthA4 -LOBP, which was indicated by blue. (b) Schematic diagram of GFP-p1380N-ttLbCas12a:LOBP. LB and RB, the left and right borders of the T-DNA region; CsVMV, the cassava vein mosaic virus promoter; GFP, green fluorescent protein; 35T, the cauliflower mosaic virus 35S terminator; CmYLCV, the cestrum yellow leaf curling virus promoter; NosP and NosT, the nopaline synthase gene promoter and its terminator; ttLbCas12a, temperature-tolerant LbCas12a containing the single mutation D156R; AtU6-26, Arabidopsis U6-26 promoter; target, the 23 nucleotides of Type II LOBP highlighted by blue, was located downstream of protospacer-adjacent motif (PAM); HH, the coding sequence of hammerhead ribozyme; HDV, the coding sequence of hepatitis delta virus ribozyme; NptII, the coding sequence of neomycin phosphotransferase II.

Xcc-Facilitated Agroinfiltration in Pummelo
Pummelo (Citrus maxima) was grown in a greenhouse at around 28 • C and was pruned for uniform shooting before Xcc-facilitated agroinfiltration. It should be kept in mind that ttLbCas12a performed better at 28 • C [41].
Xcc-facilitated agroinfiltration was performed as described previously with minor modification [43]. Briefly, the fully-expanded young Pummelo leaves were pre-treated with Xcc∆gumC [44], which was re-suspended in sterile tap water at a concentration of 5 × 10 8 CFU/mL. Twenty-four hours later, the pre-treated leaf areas were inoculated with Agrobacterium cells harboring GFP-p1380N-ttLbCas12a:LOBP or p1380-AtHSP70BP-GUSin. GFP was observed and photographed four days after agroinfiltration. p1380-AtHSP70BP-GUSin was used as a control as described elsewhere [43].

Agrobacterium-Mediated Pummelo Transformation
Pummelo transformation was conducted as described previously with minor modifications [12]. Briefly, Pummelo epicotyl explants were co-incubated with Agrobacterium cells harboring the binary vector GFP-p1380N-ttLbCas12a:LOBP. After cocultivation in darkness for 2 or 3 days at 25 • C, the epicotyl explants were placed on regeneration medium at 28 • C, at which ttLbCas12a could edit plant genome more efficiently [41].

PCR Amplification of Mutagenized LOBP
Genomic DNA was extracted from the Pummelo leaves treated by agroinfiltration or each transgenic Pummelo line. To analyze ttLbCas12a-mediate LOBP mutations, PCR was carried out using primers LOBP3 (5 -AGGTAAGCTTATTCATATTAACGTTATCAATGATT-3 ) and LOBP2 (5 -ACCTGGATCCTTTTGAGAGAAGAAAACTGTTGGGT-3 ). The PCR products were sequenced either through cloning and colony sequencing or direct sequencing using primer LOB4 (5 -CGTCATTCAATTAAAATTAATGAC-3 ). Ten random colonies for each transgenic Pummelo line were selected for sequencing. Chromas Lite program was used to analyze the sequencing results.

GFP Detection
A Zeiss Stemi SV11 dissecting microscope equipped with an Omax camera was used to detect GFP fluorescence of the Pummelo leaves treated by Xcc-facilitated agroinfiltration and GFP-p1380N-ttLbCas12a:LOBP-transformed Pummelo, under illumination of the Stereo Microscope Fluorescence Adapter (NIGHTSEA). Subsequently, the Pummelo leaves were photographed with the Omax Toupview software.

Canker Symptom Assay in Citrus
Wild type and transgenic Pummelo plants were grown in a greenhouse at the Citrus Research and Education Center, University of Florida. Before Xcc inoculation, all plants were trimmed to generate new shoots. Leaves of similar age were inoculated with either Xcc or Xcc∆pthA4:dLOB1.5 (5 × 10 8 CFU/mL) using needleless syringes. Canker symptoms were observed and photographed at five DPI.

Transient Expression of ttLbCas12a to Edit Citrus Genome via Xcc-Facilitated Agroinfiltration
Binary vector GFP-p1380N-ttLbCas12a:LOBP was constructed to edit Pummelo EBE PthA4 -LOBP ( Figure 1). The vector harbors ttLbCas12a, which has the single mutation D156R (Supplementary Figure S1) [41]. It should be noted that cestrum yellow leaf curling virus (CmYLCV) promoter was used to drive ttLbCas12a expression (Figure 1), since CmYLCV outperformed CaMV 35S and ubiquitin promoter for citrus genome editing [10]. Hammerhead ribozyme (HH) gene was placed at both ends of crRNA to promote editing. In detail, the coding sequence of hammerhead ribozyme and the coding sequence of hepatitis delta virus ribozyme (HDV) were placed at the 5 end and the 3 end of crRNA, respectively (Figure 1b and Figure S2A) [33]. In addition, the Pummelo plants were grown at 28 • C, at which ttLbCas12a could edit plant genome with higher efficiency than that at 22 • C [41].

Transgenic Expression of ttLbCas12a in Pummelo
Pummelo epicotyls were transformed with recombinant Agrobacterium cells harboring GFP-p1380N-ttLbCas12a:LOBP [45]. Notably, the shoots were generated at 28 °C to facilitate ttLbCas12a-mediated editing. Eight GFP-positive shoots were established (Figure 3), which were designated as #Pumtt1 to #Pumtt8. The transgenic plants were verified by PCR analysis (Figure 3). Based on the results of direct sequencing of PCR products, ttLbCas12a-mediated indels took placed in all transgenic Pummelo plants except #Pumtt7 (Figure 4). Remarkably, line #Pumtt2 is homozygous, since its chromatogram of direct PCR product sequencing The targeted sequence within LOBP was underlined by black lines, and the mutant site was indicated with an arrow. EBE pthA4 -TII LOBP was highlighted by red rectangles.

Transgenic Expression of ttLbCas12a in Pummelo
Pummelo epicotyls were transformed with recombinant Agrobacterium cells harboring GFP-p1380N-ttLbCas12a:LOBP [45]. Notably, the shoots were generated at 28 °C to facilitate ttLbCas12a-mediated editing. Eight GFP-positive shoots were established (Figure 3), which were designated as #Pumtt1 to #Pumtt8. The transgenic plants were verified by PCR analysis (Figure 3). Based on the results of direct sequencing of PCR products, ttLbCas12a-mediated indels took placed in all transgenic Pummelo plants except #Pumtt7 (Figure 4). Remarkably, line #Pumtt2 is homozygous, since its chromatogram of direct PCR product sequencing Based on the results of direct sequencing of PCR products, ttLbCas12a-mediated indels took placed in all transgenic Pummelo plants except #Pum tt 7 ( Figure 4). Remarkably, line #Pum tt 2 is homozygous, since its chromatogram of direct PCR product sequencing has single peaks (Figure 4). Further analysis revealed that ten nucleotides (taaacccctt) were deleted from EBE PthA4 -LOBP ( Figure 5). Furthermore, colony sequencing was employed to calculate the mutation rates. The mutation rates were 100% among the seven transgenic plants (#Pum tt 1-6 and 8), but not in #Pum tt 7, whose mutation rate was 0 ( Figures 5 and S3-S6). The results indicated that ttLbCas12a could modify citrus genome with high efficiency. has single peaks ( Figure 4). Further analysis revealed that ten nucleotides (taaacccctt) were deleted from EBEPthA4-LOBP ( Figure 5). Furthermore, colony sequencing was employed to calculate the mutation rates. The mutation rates were 100% among the seven transgenic plants (#Pumtt 1-6 and 8), but not in #Pumtt7, whose mutation rate was 0 ( Figures 5 and  S3-S6). The results indicated that ttLbCas12a could modify citrus genome with high efficiency. Figure 4. Detection of genome editing of GFP-p1380N-ttLbCas12a:LOBP-transformed Pummelo by direct sequencing of PCR products. The chromatograms of direct sequencing of PCR products. Primers LOBP2 and LOBP5 were used to amplify LOBP from wild type and transgenic Pummelo. Direct sequencing primer was LOB4. The mutation site or the beginning sites of double/multiple peaks were shown by arrows. The targeted sequence was underlined by black lines and EBEPthA4-TII LOBP was highlighted by red rectangles.
. Figure 4. Detection of genome editing of GFP-p1380N-ttLbCas12a:LOBP-transformed Pummelo by direct sequencing of PCR products. The chromatograms of direct sequencing of PCR products. Primers LOBP2 and LOBP5 were used to amplify LOBP from wild type and transgenic Pummelo. Direct sequencing primer was LOB4. The mutation site or the beginning sites of double/multiple peaks were shown by arrows. The targeted sequence was underlined by black lines and EBE PthA4 -TII LOBP was highlighted by red rectangles.
has single peaks ( Figure 4). Further analysis revealed that ten nucleotides (taaacccctt) were deleted from EBEPthA4-LOBP ( Figure 5). Furthermore, colony sequencing was employed to calculate the mutation rates. The mutation rates were 100% among the seven transgenic plants (#Pumtt 1-6 and 8), but not in #Pumtt7, whose mutation rate was 0 ( Figures 5 and  S3-S6). The results indicated that ttLbCas12a could modify citrus genome with high efficiency. Figure 4. Detection of genome editing of GFP-p1380N-ttLbCas12a:LOBP-transformed Pummelo by direct sequencing of PCR products. The chromatograms of direct sequencing of PCR products. Primers LOBP2 and LOBP5 were used to amplify LOBP from wild type and transgenic Pummelo. Direct sequencing primer was LOB4. The mutation site or the beginning sites of double/multiple peaks were shown by arrows. The targeted sequence was underlined by black lines and EBEPthA4-TII LOBP was highlighted by red rectangles.
. Figure 5. Homozygous line #Pum tt 2. (a) Direct sequencing of PCR product of #Pum tt 2. Upper: The chromatograms of three different leaves from #Pum tt 2 were shown. The chromatograms are consistent with one another, which verified #Pum tt 2 to be homozygous. EBE PthA4 -LOBP was highlighted by red rectangles. The targeted sequence was underlined by black lines, and the mutation sites were indicated with arrows. The chromatogram of wild type Pummelo plant was included for comparison purpose. Lower: The targeted sequence is shown in blue, and the mutations are shown in purple. (b) Sanger sequencing results of #Pum tt 2. Among 10 colonies sequenced, all of them are taaacccctt deletion. Type II LOBP in #Pum1. A part of LOBP sequences and its chromatogram are shown.

Mutation Genotypes of ttLbCpf1 in Transgenic Pummelo
Sanger sequencing results demonstrated that ttLbCas12a deleted ≥2 base pairs (bps) from the target site ( Figures 5 and S3-S6). The deletion took place ≥10th bp distal to the PAM site ( Figures 5 and S3-S6). Although #Pum tt 3 had both deletions and insertions (Supplementary Figure S4), the other six Pummelo lines had only deletion mutation genotypes ( Figures 5 and S3-S6), which is consistent with LbCas12a-mediated citrus genome editing [17].

Canker Resistance of ttLbCas12a-Transformed Pummelo Plants
Next, we tested whether the eight ttLbCas12a-transformed Pummelo plants were resistant to citrus canker. For this purpose, Xcc was used to inoculate wild type and transgenic Pummelo plants at a concentration of 5 × 10 8 CFU/mL. No canker symptoms were observed on seven transgenic plants (#Pum tt 1-6 and 8). Typical canker symptoms were observed on wild type and #Pum tt 7 Pummelo plants at five days post inoculation (DPI) ( Figure 6). The results indicated that the homozygous #Pum tt 2 and the transgenic Pummelo plants (#Pum tt 1, #Pum tt 3, #Pum tt 4, #Pum tt 5, #Pum tt 6, and #Pum tt 8) containing 100% indels were resistant against Xcc infection, which results from EBE PthA4 -LOBP disruption ( Figures 5 and S3-S6). sistent with one another, which verified #Pumtt2 to be homozygous. EBEPthA4-LOBP was highlighted by red rectangles. The targeted sequence was underlined by black lines, and the mutation sites were indicated with arrows. The chromatogram of wild type Pummelo plant was included for comparison purpose. Lower: The targeted sequence is shown in blue, and the mutations are shown in purple. (b) Sanger sequencing results of #Pumtt2. Among 10 colonies sequenced, all of them are taaacccctt deletion. Type II LOBP in #Pum1. A part of LOBP sequences and its chromatogram are shown.

Mutation Genotypes of ttLbCpf1 in Transgenic Pummelo
Sanger sequencing results demonstrated that ttLbCas12a deleted ≥2 base pairs (bps) from the target site ( Figures 5 and S3-S6). The deletion took place ≥10th bp distal to the PAM site ( Figures 5 and S3-S6). Although #Pumtt3 had both deletions and insertions (Supplementary Figure S4), the other six Pummelo lines had only deletion mutation genotypes ( Figures 5 and S3-S6), which is consistent with LbCas12a-mediated citrus genome editing [17].
Finally, Cas-Offinder software (Available online: http://www.rgenome.net/cas-offinder/ (accessed on 12 May 2021)) was used to search the potential off-targets of GFP-p1380N-SpCas9p:PumLOBP crRNA. When up to 3 bp mismatches with the targeting crRNA were used for searching, no potential off-targets were identified (Supplementary Figure S8). Thus, we did not conduct sequencing-based off-target analyses.

Discussion
In this study, we have successfully generated seven EBE-edited Pummelo plants. ttLbCas12a-mediated mutation genotypes in citrus are distinct from those of SpCas9, which are predominantly short indels (1-2 bp) [9,11,14]. Most of the mutations generated by ttLbCas12a are relatively long deletions, which are similar to those in LbCas12a-transformed citrus and other plants [17,19,23,26,[31][32][33][34][35][36][37][38][39]. Different editing features of ttLbCas12a and SpCas9 might result from different cleavage patterns between ttLbCas12a and SpCas9. LbCas12a cleaves DNA at sites distal to the PAM site, leading to 5 staggered ends, which have 4-5 nucleotide overhangs, whereas SpCas9 cuts DNA 3-4 nucleotides upstream of the PAM site, resulting in blunt ends. In addition, it was reported that the stagger cutting of LbCas12a could lead to the longer deletions [24]. It is worth noting that SpCas9 recognizes NGG PAM, ttLbCas12a recognizes TTTV PAM, and SaCas9 PAM recognizes NNGRRT. All three have been successfully used for genome editing in citrus [8,16,17].
Remarkably, the mutation frequencies of seven ttLbCas12a-transformed Pummelo were 100%, and only one plant was not edited. In a previous study, the highest mutation rate was 55% in LbCas12a-transfomred citrus [17]. Thus, ttLbCas12a undoubtedly outperforms LbCas12a for citrus genome editing, which is consistent with the results in Arabidopsis and tobacco [41,42]. However, the mutation efficiency of ttLbCas12-transformed soybean is comparable to that of LbCas12-transformed soybean [46]. This discrepancy might result from different genetic backgrounds of the plants. Although LbCas12a could edit genome in many dicot and monocot plants, the biallelic editing efficacy of LbCas12a remains low. For example, in LbCas12a-transformed rice, the biallelic editing frequencies were less than 50% [29,34,47]. Therefore, it is worth testing whether ttCas12a could improve biallelic editing frequencies in rice and other plants beyond citrus, Arabidopsis, and tobacco. In addition to ttLbCas12a, as one of the most popular genome editing systems, CRISPR/LbCas12a is constantly subjected to improvements, including improved editing efficiency, altered PAM specificities, and multiplexed genome engineering [25,28,41,[48][49][50].
No off-targets were identified for ttLbCas12a-mediated genome editing of citrus. Intriguingly, similar scenario was also observed for LbCas12a-mediated genome editing of citrus [17]. This concurs with that LbCas12a has lower off-target mutation rates than CRISPR/SpCas9 [50,51]. It should be noted that we only analyzed up to three mismatches. It was reported that increasing mismatches between the on-target and potential off-target sequence significantly decrease the likelihood of off-target effects. The off-target mutation rates decreased from 59% when there is one mismatch between the on-target and off-target sequences to 0% when four or more mismatches are present. Thus, off-target mutations caused by ttLbCas12a are probably low, even though we could not totally rule them out.
In summary, ttLbCas12a was successfully adapted to generate homozygous/biallelic Pummelo mutant lines; thus, it can be used as a valuable tool for functional study of citrus genes and breeding. Intriguingly, ribonucleoproteins (RNPs) consisting of LbCas12a and crRNA were employed for transgene-free genome editing [26,52]. Delivery of CRISPR/Cas RNPs bypasses the need to remove selection markers from genetically modified plants. It remains to be determined whether ttLbCas12a and LbCas12a RNPs can be used to create transgene-free genome modified citrus.