DETERMINING THE ROLES OF CANDIDATE GENES IN CITRUS-HLB INTERACTIONS AND CREATING HLB-RESISTANT CITRUS CULTIVARS

<br>What was accomplished under these goals? Objective 1. Validate candidate gene expression in inoculated citrus through RNAseq: To understand the underlying mechanism of pathogenesis of HLB by CLas bacteria, transcriptomic analysis was designed to identify differentially expressed genes that could be involved in the process. Leaf samples were collected from citrus varieties that had been previously inoculated with CLas bacteria. Among the population, Sugar Belle and Murcott were selected for the first batch of RNAseq analysis. RNAs were extracted from CLas-infected and mock-inoculated triplicate samples collected at three time points of the seedlings. Reads from Illumina sequencing were mapped to a published C. clementina genome and a list of significant differentially expressed genes (DEGs) were obtained with Blast2Go and DESeq2. The group of DEGs were compared and crossed-referenced with genes from previous studies in our lab and from other published literature. Some of the genes of interest will be used for experiments in objective 3. RNAseq for other citrus varieties in the population that were inoculated in the same time period are in progress. Their transcriptome will be sequenced and analyzed for validation of candidate gene expression. Objective 2. Identify sequence polymorphisms in candidate genes from citrus accessions with different responses to HLB and dissect the gene structure and genomic organization of candidate genes: Knocking out candidate susceptibility genes: SWEET (sugars will eventually be exported transporters) genes have been identified as some of the susceptibility genes for bacterial pathogens in rice, cotton and Arabidopsis. Gene expression analyses seems to indicate that this group of genes may also be involved in citrus susceptibility to bacterial pathogens, including CLas that has been strongly associated with HLB. Two SWEET genes were cloned and sequenced, and a CRISPR/Cas9 vector was developed to knock out each of the SWEET genes. Objective 3. Understand the roles of candidate genes by over-expressing them in HLB-susceptible citrus cultivars: Poncirus trifoliata, a close relative of sweet oranges and grapefruit, is highly tolerant to resistant to Huanglongbing (HLB) or citrus greening disease. A number of Poncirus trifoliata genes have been identified as candidate genes for its citrus greening tolerance. Four of these genes were cloned and placed under the 35S promoter for over-expressing in sweet orange. A large number of sweet orange stem segment explants were co-cultivated with Agrobacterium tumefaciens carrying the cloned Poncirus genes to transfer these genes into sweet orange to produce transgenic lines. More than four dozen transgenic sweet orange lines have been produced, including 27 for CDR2, 23 for CDR8, and several for two other candidate genes. These transgenic shoots have been transferred onto rooting media, and ten of them have rooted. Meanwhile, the cloned CDR2 and CDR8 have been introduced into Arabidopsis to expedite the confirmation of the potential roles of the cloned genes in plant disease resistance. Results show that both genes restored the hypersensitive responses of Arabidopsis mutants to a bacterial species when the transgenic plants were challenged with the bacterium. Transgenic shoots were successfully generated from transformation of Valencia sweet orange in another round of experiments for over-expression studies, using other candidate genes. We cloned candidate genes from gDNA and cDNA and are testing them in transgenic citrus plants; these genes include beta-glucanase, cellulase, callose synthase, receptor-like kinases, and NBS-LRR proteins, among others. The cloned genes were fused with green fluorescent protein in expression vectors for easy visual identification of transgenic lines. Batches of transgenic plants are being generated for further testing and characterization. Roots were also regenerated from these seedlings. Transgenic seedlings are being maintained in the growth chambers and will be transferred to greenhouse and field tests in the near future. Objective 4. Development of CRISPR-mediated technologies suitable for breeding of nontransgenic HLB resistant citrus: Development of technologies for producing gene-edited citrus trees that are transgene free: Using a model plant, we have developed an Agrobacterium-mediated transient Cas9 and sgRNA expression method to produce non-transgenic mutant plants (Chen et al., 2018). Using the same method, we produced non-transgenic gene-edited mutant plants for which no chemicals for transgenic plant selection were used. However, most of the plants are chimeric (a single plant contains both gene-edited and wild-type cells). We have also determined the efficiencies of a DNA recombinase to delete all transgenes from gene-edited citrus tissues. We have observed that the deletion was not uniform with transgenes left in some cells. Previously it has been demonstrated that knockout mutations of a particular gene in Arabidopsis and rice can lead to plant's resistance to a chemical, but no alterations in plant growth and development have been observed. To overcome the chimeric plant problem, we have identified an orthologue of the Arabidopsis and rice genes from citrus. We are currently testing whether knockout mutations of that gene in citrus can also lead to resistance to that particular chemical. If so, that gene may be conveniently used to create uniformly gene-edited citrus plants that are transgene-free. Furthermore, we have identified two candidate genes whose knockout mutations may lead to HLB tolerant rootstocks of citrus. We are currently testing efficiencies of creation of the chemical resistance described above along with production of knockout mutations of these two genes using Cas9 in the same cell. Gene editing of mature citrus tissues: Genetic transformation of mature citrus tissues has been very difficult, with very low efficiencies observed. To accelerate applications of HLB tolerant/resistant citrus trees that are bred using genome editing technologies, one approach is to use mature citrus tissues as starting materials for gene editing. In addition to improving efficiencies of traditional transformation methods for mature citrus tissues, we are working to develop an "in planta" method. We have successfully used an "in-planta" transformation method to produce transgenic and gene-edited plants using juvenile tissues of citrus. Gene editing of specific targets: For the SWEET1 gene target described above, nearly two dozen transgenic lines have been produced and analyzed for sequence changes. Observed sequence changes include addition of one nucleotide and/or deletion of up to 23 nucleotides at the gRNA-targeted sites. Several gene-edited lines showed >80% mutation. For the SWEET10 gene target, CRISPR-induced mutations include addition of one nucleotide and/or deletion of up to 60 nucleotides. Of the total 10 lines analyzed, several lines had 99% to 100% mutation. Two of the gene-edited lines seem to express increased resistance to the citrus canker pathogen. These lines will be propagated and inoculated with CLas to test their resistance to HLB. Exploring other CRISPR systems: This year, we started using a new set of CRISPR constructs. YAO promoter, which is highly expressed in tissues with active cell proliferation, could improve the DNA editing efficiency in citrus. We used YAO promoter to drive the expression of Cas9 gene in our constructs. In our experiments, selected candidate genes were targeted using the new constructs containing YAO promoter. Positive Valencia transgenic shoots resistant to Kanamycin were produced and are being evaluated and checked for their mutation in the target gene area in their genome. <br><br><b>Publications</b><br>

<br>What was accomplished under these goals? Objective 1. Gene expression profiling of HLB-affected citrus using RNAseq and microarray supported the active regulation of candidate gene expression upon infection by CLas bacteria. The grafted citrus accessions were inoculated with CLas and kept in the greenhouse. More leaf samples were collected for the remaining time points designed for the experiment. Real-time PCR experiments with replicates were performed to confirm the successful infection by CLas in the citrus population used for leaf sample collection. RNA samples have been extracted for some of the citrus genotypes and selected time points. Verified RNA samples will soon be sent for sequencing and subsequent gene expression analysis will allow us to re-examine the candidate genes for their roles during the pathogenesis of CLas in citrus. Objective 2. The miraculin genes (mir) from citrus accessions with different responses to HLB were characterized. Several dozens of potential mir genes were identified in the C. clementina and C. sinensis genome databases and re-sequenced Poncirus and Poncirus hybrid accessions. The identified mir genes encode proteins ranging from 74 to 1,077 amino acids (aa) in length with an average of 197 amino acids. Most of the mir genes contain one intron. A number of mir genes are clustered on chromosome 3. The open reading frame (ORF) length ranges from 222 bp to 3,231 bp. The phylogenetic analysis clustered mir genes into two major clades. Several hundred SNPs were detected in Poncirus and Poncirus hybrids. Approximately 1/3 of the polymorphisms occur in the coding regions. Approximately one quarter of the sequence variants were predicted to have moderate effects. In total, we now have collected whole genome resequencing data for 180 citrus accessions and relatives, including HLB-sensitive sweet orange, grapefruit; mandarins, pummelos, citrons, lemons, limes, sour oranges; HLB-tolerant Poncirus, Microcitrus and Citrus latipes. Following the best practice pipeline of GATK, different variants were identified for all these samples, including SNP, indel and structural variations. Compared with the Clementine reference genome, between 2 to 5 million SNPs were identified for each sample, depending on their heterozygosity level. Hybrids between original species usually have many more SNPs than other accessions. SNP number also reflects the taxonomic distance from the reference. The effects of these variants were studied by SNPeff. The distribution of high effect variants across gene families was biased towards gene families involved in biotic stresses, such as NB-LRR and receptor-like protein kinase. Objective 3. To continue the work of last year on citrus transformation, the transgenic shoots from previous experiments were kept on medium for regeneration. There were issues of fungi and bacterial contamination for some of the transgenic shoots, which are being replaced in repeated experiments to get new transgenic shoots and seedlings. For this year, Duncan grapefruit and Madam Vinous sweet orange were also used in producing transgenic citrus plants. New over-expression constructs were made for genes in key basal defense factors and anti-microbial peptides, and they are being used in transformation of citrus. New CRISPR constructs with the sequence of a more efficient Cpf1 and Cas9 enzymes aiming to mutate the process of callose formation and gibberellic acid signaling were made. These constructs have been introduced into Agrobacterium EHA 105 to transform citrus and generate transgenic shoots on medium. Two candidate genes were cloned from Poncirus into two expression vectors. The vectors were introduced in Agrobacterium tumefaciens. More than 3,000 stem segment explants were co-cultivated with the Agrobacterium strains. A number of putative transgenic shoots were obtained and rooted. In the meantime, the GUS gene was cloned into an expression vector not containing the GFP gene, and this vector was used as a control. When the GUS-containing vector was used, about 4% of the regenerated citrus shoots were transgenic. To achieve efficient screening of regenerated citrus shoots and increase transformation efficiency, efforts were made to develop a new expression vector. The new vector has enabled early detection of transgenic shoots based on the GFP expression. Another two candidate genes have been cloned from Poncirus. To better regulate the expression of the Poncirus genes in sweet orange, a translation regulation system is under development. As multiple candidate genes need to be tested for their function in HLB resistance/tolerance, mutant Arabidopsis lines have been adopted to facilitate the analysis of Poncirus candidate genes' roles in plant-pathogen interactions. More than two dozens of transgenic Arabidopsis lines have been generated. The disease resistance of these transgenic Arabidopsis lines will be assessed soon. Objective 4. We have developed and published a rapid, cost-effective, and high-throughput mutant screening protocol based on Illumina sequencing followed by high-resolution melting (HRM) analysis. Using tetraploid tobacco as a model plant and the phytoene desaturase (PDS) gene as a target, we successfully created and expediently identified non-transgenic mutant plants, which were verified as tetra-allelic mutants. We produced pds mutant shoots at a rate of 47.5%, without the use of antibiotic selection. Among these pds plants, 17.2% were confirmed to be non-transgenic, for an overall non-transgenic mutation rate of 8.2%. Our method is reliable and effective in creating non-transgenic mutant plants without the need to segregate out transgenes through sexual reproduction (manuscript submitted). This method should be applicable in citrus and we are currently testing the efficiency of this method in citrus. We are also comparing Agrobacterium- and gene gun-mediated transformation methods to produce transgenic and non-transgenic mutants of citrus. Our results have demonstrated that the efficiency of gene gun mediated method is much lower than that of the Agrobacterium method. Further, we have started testing and develop non-tissue culture methods to for using CRISPR to produce non-transgenic mutants of citrus. <br><br><b>Publications</b><br>

<br>What was accomplished under these goals? Objective 1. Previous studies from our lab indicate that Poncirus, a close citrus relative, shows resistance to HLB, and that the constitutive disease resistance 1 genes (CDR1) might be involved in HLB resistance/tolerance in Poncirus. This gene was first identified and cloned in Arabidopsis. Its product is implicated in disease resistance signaling. Overexpression of a rice CDR1 gene leads to constitutive activation of defense response and enhanced resistance in rice and Arabidopsis against bacterial and fungal pathogens. The expression of the entire CDR gene family in HLB-susceptible (Valencia sweet orange, Duncan grapefruit, and Ruby Red grapefruit) and HLB-tolerant genotypes ('US-812', 'US-897', and 'US-942') after CLas infection were analyzed using RNA-Seq and real-time quantitative PCR with 16 sets of gene-specific primers to understand the response of each copy of CDR genes to HLB. Twenty million sequence reads were obtained for each of the four leaf transcriptomes [two Poncirus accessions (DPI 50-7 and Flying Dragon) grown under two different conditions]. The FPKM values for each CDR gene were calculated. Transcripts of four Poncirus CDR genes were detected in Poncirus leaf transcriptomes. Two Poncirus CDR genes were in lower abundance in CLas-exposed, field-grown DPI 50-7 than in non-inoculated, greenhouse-grown DPI 50-7. Another two Poncirus CDR genes were in higher abundance in CLas-exposed, field-grown plants than in non-inoculated, greenhouse-grown plants. Real-time PCR results indicated that the expression ofmost of the CDR genes was higher in CLas infected 'US-812' and 'US-897' when compared with their respective non-infected samples. CDR transcripts amplified by five primer pairs had higher expression levels in the HLB-tolerant genotypes. Overall, Poncirus CDR genes in one of the clades of their phylogenetic tree (clade V) were highly expressed in HLB-tolerant genotypes compared with HLB-susceptible genotypes. Two of the CDR genes seem to be very interesting for further functional characterization. Our lab also has data from rough lemon demonstrating the tolerance of HLB by rough lemon. Transcriptome analyses of Poncirus and rough lemon revealed that certain groups of genes functioning in defense, cell wall metabolism, and signal transduction cascades were deregulated. A collection of citrus accessions were propagated in the greenhouse, including HLB-susceptible sweet orange and mandarin, HLB-tolerant new breeding lines plus rough lemon, and Poncirus and its hybrids. The grafted population was inoculated with budwood containing CLas bacteria. We have three mock controls and more than five CLas-infected plants for each accession. Different time points were being used when collecting leaf tissue samples. The collected samples were saved in -80 freezers for future RNA extraction and sequencing analysis. Objective 2. CDR genes were first identified in published citrus genome sequences, and their structures were characterized. The predicted protein sequences were then used to identify CDR genes in resequenced Poncirus and Poncirus hybrid genomes. CDR gene sequences from Citrus and Poncirus accessions were compared to reveal sequence polymorphisms, gene structure, and genomic organization among CDR genes. A total of 17 CDR genes were identified in C. × clementina and C. sinensis genomes. Eight copies of CDR genes were identified in the re-sequenced Poncirus samples. Compared to Citrus CDR genes, Poncirus CDR genes carry a multitude of SNPs, including 70 SNPs in four expressed Poncirus CDR genes. Among the SNPs, 20% were nsSNPs, and most of these nsSNPs are present in regulatory regions. Most of the identified SNPs were of the transition type, and the ratio between transition and transversion was greater than one for four Poncirus CDR genes. Most of the sequence variants found in Poncirus hybrids ('US-897' and 'US-812') were heterozygous in nature. The CDRgenes identified in Citrus and Poncirus genomes fell into five clades. Most of the CDR genes contain no or only one intron. CDR genes in the same clades share similar motifs. A few CDR genes seem to have arisen from a recent tandem duplication and may have a redundant function in citrus. The expansion of the CDR genes seems to result from segmental and tandem duplication, which could be the result of chromosomal rearrangement and fusions. The predicted CDR proteins vary in length from 185 to 480 amino acids and share an overall amino acid identity of 53.4%. All 17 predicted citrus CDRproteins contain at least one ASP domain. Two of the Poncirus CDR genes carry a large number of amino acid changes, and their predicted protein structure is altered. We have collected whole genome resequencing data for 101 citrus accessions and relatives, including several HLB-tolerant Poncirus and Poncirus hybrids with sweet oranges/mandarins. Their genetic variants were called using a GATK pipeline. On average, 3.5 million SNPs were identified for each sample. Whole genome wide association analysis was performed based on these variants and their HLB tolerance (such as bacterial titer and disease severity rating). Many genome locations overlapping with our candidate genes were identified. The SNP effects in overlapping genes were examined by SNPeff. Objective 3. The postdoc research associate working on over-expressing Poncirus candidate genes had to leave due to a visa issue. The two Agrobacterium strains carrying two CDR candidate genes were sent to a citrus transformation facility to produce transgenic citrus. A number of Agrobacterium-mediated transformation experiments were conducted at the transformation facility to deliver the Poncirus candidate genes into sweet orange; several shoots emerged from sweet orange stem segment explants that were co-cultivated with the Agrobacterium and selected with the antibiotic kanamycin. None of these shoots rooted in rooting media. Micro-grafting was attempted to rescue the roots but was unsuccessful. Efforts are being made to overcome the low transformation efficiency with these gene constructs. Over-expression constructs containing genes on callose decomposition, basal defense response, and hormone signaling were made successfully. Transient expression of the constructs were visually inspected in Nicotiana benthamiana leaves by checking the fused green fluorescent protein under microscope. The verified constructs were then transformed into sweet orange genotype Valencia and explants were screen with appropriate antibiotics. Young transgenic shoots were obtained and transferred to rooting medium for regeneration of transgenic citrus. Gene constructs containing CRISPR Cas9 and gRNA were also tested in transforming citrus but these experiments were suspended due to the problems found within the constructs later on. Objective 4. We have constructed the proposed genes and produced transgenic citrus plants using Agrobacterium-mediated transformation methods. However, the efficiency of the CRISPR in citrus is very low based on the rates of mutations created in the PDS gene. We have been testing what are the most important factors that contribute to the low efficiency. We have been working on development of a method to use Agrobacterium for transient expression of the T-DNA genes so that we can express the Cas9 gene and sgRNA genes to create non-transgenic citrus with desirable mutations. Our results indicate that the transient expression levels of the T-DNA genes are very high in both tobacco (a model plant for a proof of concept study) and citrus. Our preliminary results have shown that the method can be used to efficiently express Cas9 and sgRNA genes to produce non-transgenic mutants. Objective 6. We have presented findings and results from this project at the 5th International Research Conference on Huanglongbing and the International Citrus Congress in Brazil. <br><br><b>Publications</b><br>

<br>What was accomplished under these goals? Objective 1. Validate candidate gene expression in inoculated citrus through RNAseq: The leaf transcriptomes of Poncirus genotypes under two conditions (CLas inoculated or mock-inoculated) have been sequenced and analyzed to reveal or validate the gene expression profiles of candidate genes that have shown intimate involvement in plant defense or immunity. More than a dozen of NB-LRR genes were found to be up-regulated in response to CLas infection in both Poncirus genotypes, while more than two dozens of NB-LRR were down-regulated in response to CLas infection. Data from real-time PCR confirmed that nearly 10 of the NB-LRR genes were also up-regulated in Poncirus hybrid genotypes. One candidate gene was up-regulated several fold after CLas inoculation in Poncirus genotypes. Real-time PCR data confirmed the up-regulation of this gene in Poncirus genotypes. Another 11 candidate genes were expressed in the Poncirus genotypes. To achieve greater statistical significance and discriminatory power in future RNA-seq and real-time PCR analysis, 15 genotypes known to have different levels of HLB tolerance/resistance were grafted onto sour orange rootstock. The grafted plants (12 per genotype) have been grown in the greenhouse for 6 months, and they will be inoculated with CLas-infected rough lemon budwood in the near future, to yield at least 3 biological replicates for each genotype x condition. The genotypes selected for CLas inoculation include sweet orange (HLB-susceptible), irradiated sweet orange (with apparent tolerance to HLB), rough lemon (HLB tolerant), two breeding lines (HLB tolerant), one Citrus species (HLB resistant/tolerant), Poncirus trifoliata and hybrids (HLB resistant/tolerant), and Microcitrus. Leaf tissues will be collected from these genotypes at various time points after CLas inoculation, and the transcriptomes of these leaf tissue samples will be sequenced to validate the expression levels of candidate genes. Objective 2. Identify sequence polymorphisms in candidate genes from citrus accessions with different responses to HLB and dissect the gene structure and genomic organization of candidate genes: Prior to a major focus on candidate gene polymorphisms, we chose to look more broadly at cross-genome polymorphisms, based on 37 publicly available and in-house generated sequencing data sets. Different variants were identified for all these samples, including SNP, indel and structural variations. On average, about 3.5 million SNPs were identified for each sample. The SNP number identified is not related to the read depth, while it reflects the taxonomic distance from the reference. The distribution of high effect variants across gene families was biased towards gene families involved in biotic stresses, such as NB-LRR and receptor-like protein kinase. High quality SNPs were identified in three groups of candidate genes. Assemblies from short sequence reads (HiSeq 2000) did not show SNPs in the exons of some candidate genes; however, direct sequencing of PCR products amplified from citrus genomic DNA indicated the possibility of multiple copies of candidate genes with SNPs in the exons. More specific PCR primers have been designed to understand the sequence polymorphism, structural variation, or copy number variation in these candidate genes. Objective 3. Understand the roles of candidate genes by over-expressing them in HLB-susceptible citrus cultivars: Emphasis has been on developing expression constructs for genes involved in disease resistance, callose deposition, cell wall synthesis, etc. We have cloned six candidate genes by PCR or RT-PCR. The cloned genes are being shuffled into transformation vectors containing the green fluorescent protein (gfp) for production and identification of transgenic plants. Objective 4. Develop CRISPR-mediated technologies for breeding of non-transgenic HLB-resistant citrus: We performed the following experiments and produced following results: 1) Construction of CRISPR constructs with sgRNAs for the PDS genes in tobacco and citrus. 2) Transformation of tobacco and citrus plants with the CRISPR gene cassettes. 3) Characterization of transgenic plants produced. 4) Successful generation of white color (PDS mutation) mutants of both citrus and tobacco. 5) Cloning and sequencing the T-DNA border sequences for deletion of the CRIPR cassettes from the host plant genome. 5) Starting development of additional technologies to create CRISPR-mediated mutant plants that are non-transgenic. Objective 6. Outreach and disseminate project results to stakeholders and the public: We have presented findings and results from this project at the USDA-NIFA-SCRI Citrus Disease Subcommittee Meeting, the 4th International Research Conference on Huanglongbing, the XXIV International Conference on Plant and Animal Genomes, and the 2nd international Horticulture Research Conference. <br><br><b>Publications</b><br>