SpCas9p is one version of codon-optimized SpCas9 (Ma et al., 2015b), which was successfully employed to edit citrus genome (Jia et al., 2021). From 35S-SpCas9p:DunLOBP (Jia et al., 2021), codon-optimized SpCas9p was amplified using primer Cas9p-5-AflII (5′-AGGTCTTAAGGACAAGAAGTACTCGATCGGCCTCGCCATCG GCACCAACAGCGTCGGCTGGGCGGTGATCAC-3′) and Cas9p-3-MluI (5′-AGTCACGCGT CTTCTTTTTCTTAGCCTGTCCGGCCTT-3′). After digestion with AflII and MluI, part of SpCas9p was cloned into nCas9-PBE vector from Addgene (Addgene plasmid #98164) to form nCas9-mPBE vector. nCas9-PBE (plant base editor) was used to perform base editing in rice, wheat and maize (Zong et al., 2017). To construct GFP-p1380N-CmYLCV-nCas9-mPBE, the BamHI- EcoRI-cut nCas9-mPBE fragment was ligated with BamHI- EcoRI-cut GFP-p1380N-CmYLCV-nCas9-PBE (Huang et al., 2023).

From 35S-SpCas9p:DunLOBP (Jia et al., 2021), the AtU6-26 promoter was amplified using AtU6-26-5-XhoI (5′-AGGTCTCGAGTCGTTGAACAACGGAAACTCGACTTGCCTT-3′) and AtU6-26-3-phos (5′-phosphorylated-aatcactacttcgactctagctgt-3′), and the sgRNA-ALSBE-NosT fragment was PCR-amplified using sgRNA-ALSBE1-P1 (5′-phosphorylated-GcaggtcccTcggaggatgatGTTTTAGAGCTAGAAATAGCAAGT-3′) and NosT-3-SpeI (5′-aggtactagTCCGATCTAGTAACATAGATGACA-3′). Through three-way ligation, XhoI-cut AtU6-26 and SpeI-digested sgRNA-ALSBE1-NosT were inserted into XhoI-XbaI-treated pUC-NosT-MCS to construct pUC-NosT-AtU6-26-sgRNA-ALSBE1. pUC-NosT-MCS was constructed previously, which harbors the XhoI-AscI-XbaI-PmeI multiple enzyme sites (Jia et al., 2019a). From pUC-NosT-AtU6-26-sgRNA-ALSBE (Huang et al., 2023), the AtU6-26-sgRNA-ALSBE2 fragment was amplified using AtU6-26-5-XhoI and NosT-3-BsaI (5′-ATTCGGTCTCCCATGTATGAT AATCATCGCAAGACCGGC-3′), and the AtU6-26-sgRNA-ALSBE1 was PCR-amplified using AtU6-26-5-BsaI (5′-TCGAGGTCTCCCATGTCGTTGAACAACGGAAACTCGACTTGCCTT-3′) and NosT-3-SpeI from pUC-NosT-AtU6-26-sgRNA-ALSBE1. Through three-way ligation, XhoI-BsaI-cut XhoI-AtU6-26-sgRNA-ALSBE2-BsaI and BsaI-SpeI-digested BsaI-AtU6-26-sgRNA-ALSBE1-SpeI were inserted into XhoI-XbaI-treated pUC-NosT-MCS to construct pUC-NosT-AtU6-26-sgRNA-ALSBE2-ALSBE1. Subsequently, the EcoRI-NosT-AtU6-26-sgRNA-ALSBE2-AtU6-26-sgRNA-ALSBE1-PmeI fragment from pUC-NosT-AtU6-26-sgRNA-ALSBE2-ALSBE1 were cloned into EcoRI-PmeI-cut GFP-p1380N-CmYLCV-nCas9-mPBE to generate GFP-p1380N-CmYLCV-nCas9-mPBE:ALS2:ALS1.

Finally, the AscI-PmeI-cut CmYLCV-nCas9-mPBE:ALS2:ALS1 fragment GFP-p1380N-CmYLCV-nCas9-mPBE:ALS2:ALS1 was clone into AscI-PmeI-cut vector GFP-p1380N-ttLbCas12a:LOBP1-AscI-XbaI-PmeI, which was constructed previously (Huang et al., 2023), to form GFP-p1380N-ttLbCas12a:LOBP1-mPBE:ALS2:ALS1 ( Figure 1 , Supplementary Figure 1 ). Notably, GFP-p1380N-ttLbCas12a:LOBP1-mPBE:ALS2:ALS1 contained both GFP and nptII selectable genes that can be used for selection of putative non-transgenic transformants ( Figure 1 , Supplementary Figure 1 ) with GFP being an easier option for visual selection in our experience (Huang et al., 2023).

The binary vector GFP-p1380N-ttLbCas12a:LOBP1-mPBE:ALS2:ALS1 was transformed into A. tumefaciens strain EHA105 via electroporation. Recombinant Agrobacterium cells were employed for Xcc-facilitated agroinfiltration or citrus epicotyl transformation.

Grown in a greenhouse at 28°C, C. sinensis cv. ‘Hamlin’ was pruned to generate uniform shooting before Xcc-facilitated agroinfiltration. It should be pointed out that ttLbCas12a performed better at 28°C than at 22°C in a previous study (Schindele and Puchta, 2020).

Xcc-facilitated agroinfiltration was performed as described previously (Jia and Wang, 2014b). Briefly, the fully-expanded young Hamlin leaves were first treated with Xcc, which was re-suspended in sterile tap water at a concentration of 5 × 10⁸ CFU/mL. Twenty-four hours later, the Xcc-treated leaf areas were inoculated with Agrobacterium cells harboring vectors GFP-p1380N-ttLbCas12a:LOBP1-mPBE:ALS2:ALS1, or p1380-AtHSP70BP-GUSin. p1380-AtHSP70BP-GUSin was constructed before (Jia and Wang, 2014b), and used as a no-GFP-fluorescence control. Four days after agroinfiltration, GFP was observed and photographed, and genomic DNA was extracted from the Hamlin leaves treated by agroinfiltration.

As described previously (Jia et al., 2019b), Hamlin epicotyl transformation was conducted with minor modifications. Briefly, Hamlin epicotyl explants were co-incubated with Agrobacterium cells harboring the binary vector GFP-p1380N-ttLbCas12a:LOBP1-mPBE:ALS2:ALS1. After cocultivation in darkness for 3 days at 25°C, the epicotyl explants were placed on regeneration medium containing 50 mg/L kanamycin for one week at 28°C, and then the epicotyl explants were placed on regeneration medium containing 40 μg/L chlorsulfuron at 28°C (Veillet et al., 2019). All regenerated plants were subjected to genome editing analysis and GFP inspection. It is noteworthy that we initially placed Hamlin epicotyls on medium with 100 mg/L kanamycin for two weeks; three rounds of selection with 40 ng/mL chlorsulfuron, each lasting two weeks as reported previously (Veillet et al., 2019). However, very few shoots were regenerated. Based on our observations, we adapted the protocol by reducing the kanamycin concentration to 50 mg/L for one week, and conducting two rounds of selection with 40 ng/mL chlorsulfuron, each extending to three weeks.The GFP-positive and transgene-free regenerated shoots were selected and micro-grafted on ‘Carrizo’ citrange rootstock plants (C. sinensis (L.) Osbeck x Poncirus trifoliata (L.) Raf.) for further analysis. Six months later, they were used for PCR analysis with the primers Npt-Seq-5 (5′-TGTGCTCGACGTTGTCACTGAAGC-3′) and 35T-3 (5´-TTCGGGGGATCTGGATTTT AGTAC-3′).

Genomic DNA was extracted from the Xcc-facilitated-agroinfiltrated Hamlin leaves or each regenerated Hamlin line. To analyze GFP-p1380N-ttLbCas12a:LOBP1-mPBE:ALS2:ALS1-mediate CsALS and LOBP mutations, PCR was carried out using either primers CsALSP1 (5′- Atctgtatcgccacctcggggcccggc-3′) and CsALSP2 (5′- TGGCCGCCCAGATGTTGCTAAAAGG-3′), or primers LOB21 (5′- ACACCTTGGTAATTTTGACATTAGGTA-3′) and LOB22 (5′- TGAGAGAAGAAAACTGTTGGGTTGTAG-3′). The PCR products were sequenced through cloning and colony sequencing. Primers CsALSP1 and CsALSP2 were designed to amplify the sgRNA-target CsALS region for the mutation analysis via Sanger sequencing, and primers LOB21 and LOB2 were employed to analyze LOBP. For PCR product direct sequencing, the CsALSP1/CsALSP2-amplified and LOB21/LOBP22-amplified PCR products were purified and subjected to sequencing using CsALSP3 (5′- tggtcagcgggctcgccgacgcgct-3′) as to CsALS, and primer LOB4 (5′-CGTCATTCAATTAAAATTAATGAC-3′) as to LOBP. Direct sequencing of PCR products was employed to genotype CRISPR/Cas-mediated indels (Ma et al., 2015a; Jia et al., 2016). Ten random colonies for each transgenic and transgene-free Hamin line were selected for sequencing. Mutation rates were calculated by the mutant colonies/10 colonies. Chromas Lite program was used to analyze the sequencing results.

An Omax camera was installed onto a Zeiss Stemi SV11 dissecting microscope for photographing GFP fluorescence. Under illumination of the Stereo Microscope Fluorescence Adapter (NIGHTSEA), GFP fluorescence of the Hamlin leaves treated by Xcc-facilitated agroinfiltration and GFP-p1380N-ttLbCas12a:LOBP1-mPBE:ALS2:ALS1-transformed Hamlin was observed. Subsequently, the Hamlin leaves were photographed with the Omax Toupview software connected to the Omax camera.

Wild type, transgenic and transgene-free Hamlin plants were grown in a greenhouse at the Citrus Research and Education Center, University of Florida. Prior to Xcc inoculation, all plants were trimmed to generate new shoots. Leaves of similar age were infiltrated with either Xcc or XccΔpthA4:dLOB1.5 (5 × 10⁸ CFU/mL) using needleless syringes. At three, six and nine days post inoculation (DPI), canker symptoms were observed and photographed.

Genomic DNA of transgene-free Hamlin plant #Ham_NoGFP4 was subjected to whole genome sequencing at Novogene (Sacramento, CA, USA). The whole genome data of #Ham_NoGFP4 were released at NCBI (NCBI Bio-project ID: PRJNA1073671). DNA Library construction, sequencing, and data analysis were performed as follows. Following the manufacturer’s protocol of short read DNA sequencing from Illumina (Kozarewa et al., 2009), the library was prepared. After quality control, quantification, and normalization of the DNA libraries, 150-bp paired-end reads were generated using the Illumina NovaSeq 6000 platform according to the manufacturer’s instructions at Novogene. The raw paired-end reads were filtered to remove the low-quality reads using fastp program version 0.22.0 (Chen et al., 2018). To assess the target site mutations of mutated plants, the high quality paired-end short genomic reads were mapped to sweet orange (C. sinensis) (Wang et al., 2021) reference genome using Bowtie2 software version 2.2.6 (Langmead and Salzberg, 2012). The mutations (single nucleotide polymorphisms, deletions and insertions) for the mutated plant genomes were identified using the SAMtools package version 1.2 (Li et al., 2009) and deepvariant program version 1.4.0 (Poplin et al., 2018). The identified mutations were filtered by quality and sequence depth (mapping quality > 10 and mapping depth > 10). The mutations of target sites were visualized using IGV software version 2.15.4 (Robinson et al., 2011). The off-target sites were predicted using CRISPR-P 2.0 (Liu et al., 2017) and the Cas-OFFinder (Bae et al., 2014) program and aligning target sequence with whole genome using blast program. Based on the mapping results, mutations of off-target sites were detected using the SAMtools package version 1.2 and deepvariant program version 1.4.0. To detect the potential foreign DNA contamination, the high quality paired-end short genomic reads were mapped to the plasmid sequences using Bowtie2 software version 2.2.6.

In a previous study, GFP-p1380N-ttLbCas12a:LOBP1-PBE:ALS was constructed to produce transgene-free pummelo, which is a highly homozygous diploid (Wu et al., 2018). One sgRNA was designed to target both alleles of acetolactate synthase (CsALS) and another sgRNA for both alleles of EBE_PthA4-LOBP in pummelo (Huang et al., 2023). C. sinensis cv. Hamlin is a heterozygous hybrid between pummelo (C. maxima) and mandarin (C. reticulata). In this study, the promoter region of CsLOB1 (CsLOBP) from mandarin was designated as Type I LOBP, whereas CsLOBP from pummelo was named as Type II LOBP. As a result, Hamlin has Type I LOBP and Type II LOBP. Notably, Type I LOBP has one more G than Type II LOBP downstream of EBE_PthA4, whose detailed sequence is ATAAACCCCTTTTGCCTT ( Figure 1 ). Similarly, two alleles of CsALS from mandarin and from pummelo were named as Type I CsALS and Type II CsALS, respectively. As a result, Hamlin also has Type I CsALS and Type II CsALS. Notably, Type I CsALS has T at the position of 9, and Type II CsALS has G (the PAM labeled as positions from 21 to 23) ( Figure 1 ). Consequently, two different sgRNAs were designed to target the ALS gene of Hamlin in vector GFP-p1380N-ttLbCas12a:LOBP1-mPBE:ALS2:ALS1 ( Figure 1 ). On the other hand, one crRNA was designed to target the conserved sequence of both alleles of EBE_PthA4-LOBP of Hamlin. Based on the reference genome of sweet orange genome (V3) from HZAU (http://citrus.hzau.edu.cn/index.php) (Liu et al., 2017), the genomic position for target 1 (TCTATATAAACCCCTTTTGCCTT) is from 3970724-3970746 on chromosome 7. The genomic positions for target 2 (caggtcccGcggaggatgat) and targets 3 (caggtcccTcggaggatgat) are from 15794167-15794186 on chromosome 7, since target 2 and targets 3 are two alleles of CsALS ( Figure 1A ). To enhance editing efficiency, base editor nCas9-PBE of GFP-p1380N-ttLbCas12a:LOBP1-PBE:ALS was optimized to form nCas9-mPBE in GFP-p1380N-ttLbCas12a:LOBP1-mPBE:ALS2:ALS1, which harbors the codon-optimized SpCas9p backbone for plants ( Figure 1A ). SpCas9p has been used to produce homozygous/biallelic mutations for multiple plant species in the T0 generation, such as Arabidopsis, rice, and citrus (Ma et al., 2015b; Jia and Wang, 2020).

Xcc-facilitated agroinfiltration of Hamlin leaf was used to test whether GFP-p1380N-ttLbCas12a:LOBP1-mPBE:ALS2:ALS1 could be employed to edit citrus EBE_PthA4-LOBP by ttLbCas12a and edit CsALS by nCas9-mPBE. It must be pointed out that cestrum yellow leaf curling virus (CmYLCV) promoter was used to drive ttLbCas12a and nCas9-mPBE expression in GFP-p1380N-ttLbCas12a:LOBP1-mPBE:ALS2:ALS1 ( Figure 1 , Supplementary Figure 1 ), since CmYLCV outperformed CaMV 35S for citrus genome editing (Huang et al., 2022a). In addition, in order to promote editing, the coding sequence of hammerhead ribozyme (HH) at the 5’ end of crRNA, and the coding sequence of hepatitis delta virus ribozyme (HDV) were placed at the 3’ end of crRNA ( Figure 1 , Supplementary Figure 1 ) (Tang et al., 2017). Genomic DNA was extracted from GFP-expressing Hamlin leaf and subjected to PCR, ligation and E. coli transformation. Sanger sequencing results showed that one colony contained ttLbCas12a-directed indels in EBE_PthA4-LOBP among one batch of 100 colonies sequenced, and one colony contained the nCas9-PBE-mediated cytosine-to-thymine base conversion in CsALS among another batch of 100 colonies sequenced ( Figure 1B ). Therefore, GFP-p1380N-ttLbCas12a:LOBP1-mPBE:ALS2:ALS1 was functional to edit both EBE_PthA4-LOBP and CsALS in Hamlin. Though the same genomic DNA was used as PCR template, two pairs of primers, CsALSP1/CsALSP2 and LOB21/LOBP22, were separately used to analyze mutant CsALS and LOBP. It is likely that ttLbCas12a-directed indels in LOBP and the nCas9-PBE-mediated cytosine-to-thymine base conversion in CsALS took place in the same cells even though we could not totally exclude other possibilities.

Hamlin epicotyls were transformed with recombinant Agrobacterium cells harboring GFP-p1380N-ttLbCas12a:LOBP1-mPBE:ALS2:ALS1 ( Figure 1 ) (Jia et al., 2019b). Three-day after co-cultivation, the epicotyls were first selected on 50 mg/L kanamycin for one week, then transferred to selective medium containing 40 μg/L chlorsulfuron for six weeks, during which new chlorsulfuron-containing medium was used after three-week cultivation. After six week herbicide selection, Hamlin epicotyls were cultivated on chlorsulfuron-free medium. In addition, the shoot generation was done at 28°C to facilitate LbCas12aD156-mediated EBE_PthA4-LOBP editing (Schindele and Puchta, 2020; Jia et al., 2022).

In the presence of chlorsulfuron, 77 shoots were established. Among them, 8 shoots were GFP-positive, named as #Ham_GFP1 to #Ham_GFP8. Unexpectedly, five lines (#Ham_GFP4, #Ham_GFP5, #Ham_GFP6, #Ham_GFP7 and #Ham_GFP8) died after grafting, however, three GFP-positive shoots (#Ham_GFP1, #Ham_GFP2 and #Ham_GFP3) survived ( Figure 2A ). On the other hand, 69 shoots had no GFP expression, designated as #Ham_NoGFP1 to #Ham_NoGFP69 ( Figure 2A ).

Based on the results of direct sequencing of PCR products from transgenic Hamlin, ttLbCas12a-directed indels and mPBE-mediated C-to-T conversions were observed in #Ham_GFP1, #Ham_GFP2 and #Ham_GFP3 ( Figure 3 ). Among the 69 no-GFP-expressing shoots, four shoots, #Ham_NoGFP1 to #Ham_NoGFP4, had mutations in CsALS and EBE_PthA4-LOBP ( Figure 4 ), whereas the rest, from #Ham_NoGFP5 to #Ham_NoGFP69, had wild type CsALS and EBE_PthA4-LOBP ( Figure 4 ). Therefore, the four no-GFP-expressing lines (#Ham_NoGFP1, #Ham_NoGFP2, #Ham_NoGFP3, #Ham_NoGFP4) are likely transgene-free EBE_PthA4-LOBP-edited lines.

To further verify whether the four no-GFP-expressing lines are transgene-free, we analyzed nptII gene in GFP-positive shoots (#Ham_GFP1 to #Ham_GFP3) and no-GFP-expressing shoots (#Ham_NoGFP1 to #Ham_NoGFP4). Using a pair of primers Npt-Seq-5 and 35T-3, nptII gene was subjected to PCR analysis. Three GFP-positive Hamlin had the expected nptII PCR products ( Figure 2B ), which verified that #Ham_GFP1 to #Ham_GFP3 were transgenic. No nptII PCR products were observed in no-GFP-expressing Hamlin lines from #Ham_NoGFP1 to #Ham_NoGFP4. The results confirmed that #Ham_NoGFP1, #Ham_NoGFP2, #Ham_NoGFP3, and #Ham_NoGFP4 were transgene-free plants ( Figure 2B ).

Sanger sequencing was employed to analyze mutation genotypes of the edited lines through cloning. As for EBE_PthA4-LOBP, the results demonstrated that #Ham_GFP1, #Ham_GFP2 and #Ham_GFP3 contained biallelic, chimeric and chimeric mutations, respectively ( Supplementary Figures 2 , 3 ). As for ALS, #Ham_GFP1, #Ham_GFP2 and #Ham_GFP3 contained chimeric, biallelic and biallelic mutations, respectively ( Supplementary Figures 2 , 3 ). Intriguingly, #Ham_GFP3 harbored C to T substitutions at C₁₀ of CsALS sgRNA ( Supplementary Figure 3A ), which is somehow out of nCas9-PBE-mediated C-to-T conversion ranges from position 3 to 9 within the protospacer, counting the PAM as positions 21-23 (Zong et al., 2017). Previous study already showed that CBE could convert C to T outside the editing window (Lv et al., 2020). In addition, 7 colonies had the C to T substitution at C10 among 10 colonies sequenced. Thus, the C to T substitutions at C₁₀ of CsALS in #Ham_GFP3 could attribute to nCas9-PBE activity rather than sequencing errors. #Ham_GFP2 contained 24 bps deletions from EBE_PthA4-LOBP, including part of PAM ( Supplementary Figure 2D ).

For the transgene-free edited lines, Sanger sequencing results showed that #Ham_NoGFP1, #Ham_NoGFP2, #Ham_NoGFP3 and #Ham_NoGFP4 contained biallelic, biallelic, biallelic and homozygous mutations, respectively, in EBE_PthA4-LOBP ( Figures 5B, D , 6, B ). As for ALS, Sanger sequencing results indicated that #Ham_NoGFP1, #Ham_NoGFP2, #Ham_NoGFP3 and #Ham_NoGFP4 harbored biallelic, chimeric, biallelic and homozygous mutations, respectively ( Figures 5A, C , 6A, C ). Notably, #Ham_NoGFP4 had homozygous mutations in both EBE_PthA4-LOBP and CsALS ( Figures 6C, D ).

Sanger sequencing results of transgenic and transgene-free Hamlin demonstrated that EBE_PthA4-LOBP mutations were independent of those of ALS, which is consistent with the previous study in pummelo (Huang et al., 2023). ttLbCas12a deleted several bps from the target site, which occurred ≥10th bp distal to the PAM site except that of #Ham_GFP2 ( Figure 5 , 6 , Supplementary Figures 2 , 3 ), which is consistent with previous work in citrus (Jia et al., 2022; Huang et al., 2023). The #Ham_GFP2 harbored 24 bps deletions overlapping with PAM in EBE_PthA4-LOBP ( Supplementary Figure 2D ). nCas9-mPBE catalyzed the targeted conversion of cytosine to thymine from position 6 to 8 within the protospacer except that of #Ham_GFP3, which is consistent with previous work in rice, wheat and maize (Zong et al., 2017). Notably, the EBE_PthA4-LOBP mutation rates were 100% in #Ham_GFP1, #Ham_GFP2, #Ham_NoGFP1, #Ham_NoGFP2, #Ham_NoGFP3 and #Ham_NoGFP4 ( Supplementary Figure 2 , Figures 5 , 6 ).

Since Hamlin plants from #Ham_NoGFP5 to #Ham_NoGFP69 had no mutations in EBE_PthA4-LOBP ( Figure 4 ), they were not tested with Xcc inoculation. Three transgenic plants (#Ham_GFP1-3) and four transgene-free plants (#Ham_NoGFP1-4) were evaluated for canker resistance. Wild type Hamlin was used as a control. All plants were inoculated with Xcc and XccΔpthA4:dLOB1.5 at a concentration of 5 × 10⁸ CFU/mL. XccΔpthA4:dLOB1.5 was included as a positive control because dLOB1.5 is a designed TALE, which binds to the sequence 5´- TAAAGCAGCTCCTCCTCATCCCTT- 3′ in the promoter region of LOB1, away from the EBE region, to activate LOB1 expression ( Supplementary Figure 4A ) (Jia and Wang, 2020). Therefore, dLOB1.5 can mimic PthA4 function, and XccΔpthA4:dLOB1.5 can cause citrus canker, giving that its recognizing sequence is intact (Jia and Wang, 2020) whereasXccΔpthA4 cannot cause canker (Hu et al., 2014). Sanger sequencing results indicated that the dLOB1.5 binding sites were intact among all Hamlin plants ( Supplementary Figure 4B ).

At three, six and nine days post inoculation (DPI) with Xcc, canker symptoms were observed on wild type Hamlin plants, whereas no canker symptoms were observed on transgene-free #Ham_NoGFP1, #Ham_NoGFP2; #Ham_NoGFP3, #Ham_NoGFP4 and three transgenic Hamlin plants ( Figure 7 , Supplementary Figure 5 ). The results indicated that all plants, transgenic or not, harboring 100% mutations in EBE_PthA4-LOBP ( Figures 5 , 6 , Supplementary Figure 2 ), resisted against Xcc infection as expected. Intriguingly, #Ham_GFP3 containing 90% mutation in EBE_PthA4-LOBP also resisted against Xcc ( Figure 7 , Supplementary Figure 3 ). Consistently, Cas9/sgRNA-transformed D_LOB9 harboring 89.36% mutation rate in CsLOB1 also showed Xcc resistance in a previous study (Jia et al., 2017a). At three, six and nine DPI with XccΔpthA4:dLOB1.5, canker symptoms developed on all treated plants ( Figure 7 , Supplementary Figure 5 ), since there was no editing in the dLOB1.5 binding sites ( Supplementary Figure 4B ). Therefore, it was the EBE_PthA4-LOBP disruption that conferred Hamlin resistance against Xcc.

To further confirm whether the transgene-free genome-edited lines are indeed free of plasmid sequences, we conducted a whole genome sequencing analysis of #Ham_NoGFP4. #Ham_NoGFP4 was selected because it contains homozygous EBE_PthA4-LOBP and CsALS mutations ( Figures 6C, D ). We obtained high quality paired-end short genomic reads for #Ham_NoGFP4 which were mapped to the T-DNA sequence corresponding to GFP-p1380N-ttLbCas12a:LOBP1-mPBE:ALS2:ALS1 using Bowtie2 software version 2.2.6. No T-DNA sequences were mapped to the genomic DNA of #Ham_NoGFP4, confirming it was transgene-free. Furthermore, whole genome sequencing analysis indicated that #Ham_NoGFP4 harbored homozygous mutations in both EBE_PthA4-LOBP and ALS ( Supplementary Figure 6 ), which was consistent with Sanger sequencing results ( Figure 6 ).

To analyze off-target mutations, CRISPR-P 2.0 (Liu et al., 2017) and the Cas-OFFinder (Bae et al., 2014) program were used to find the potential off-targets of crRNAs and sgRNAs of GFP-p1380N-ttLbCas12a:LOBP1-mPBE:ALS2:ALS1. No potential off-target was identified for EBE_PthA4-LOBP and CsALS (mismatch number ≤3). Since no potential off-target was identified for EBEPthA4-LOBP and CsALS (mismatch number ≤3), we changed the mismatch number ≤4. Consequently, 1 potential off-target was identified for EBEPthA4-LOBP, whereas 1 and 4 potential off-targets were identified for type I and type II CsALS alleles, respectively. However, analyses of the whole genome sequencing for the potential off-targets (mismatch number ≤4) did not identify any off-target mutations for both EBEPthA4-LOBP and CsALS.