Huanglongbing (HLB) is a bacterial disease associated with Candidatus Liberibacter asiaticus (CLas), Candidatus Liberibacter americanus, and Candidatus Liberibacter africanus. These bacterial species are transmitted by hemipteran psyllids. For many years, this disease has caused substantial damage to citrus around the world and in recent years, it has significantly impacted citrus production in the U.S.A. The lack of effective control and curative measures against psyllids and the pathogen are the major factors contributing to HLB-related damages. In addition to these factors, limitations such as longer generation time, limited genomics resources and regulatory restrictions hinder rapid research on HLB. Herein, I report a novel model plant-pathosystem of Arabidopsis thaliana and Candidatus Liberibacter psyllaurous (CLps) that has potential to facilitate rapid development of control strategies for HLB disease. Based on 98% identity in 16S rDNA sequence and other genome sequence similarities, CLps is considered as a close relative to CLas. The CLps pathogen can infect solanaceous plants such as tomato, potato and pepper. To our knowledge, this is the first study demonstrating Liberibacter infection in Arabidopsis. The CLps pathogen can be transmitted by psyllid Bactericera cockerelli (BC) into Arabidopsis. Over time an increase in bacterial titer was evident in 19 genetically diverse ecotypes and defense-compromised lines of Arabidopsis. Similar to CLas, CLps systemically infects Arabidopsis. I identified several similarities between the phenotypic and transcriptional changes in CLps infected Col-0 and NahG (a salicylic acid deficient Col-0) lines and the responses in CLas infected citrus. Some unique changes such as development-related transcriptional changes associated with floral transition and a novel phenotype of excessive growth of small size axillary/cauline leaves were observed only in CLps infected Arabidopsis. A foliar spray technique for testing candidate defense-inducing chemicals with Arabidopsis/BC/CLps system and detached leaf assay with tomato/BC/ CLps system were evaluated for their effects on Liberibacter infection. The timeline of experimental assays with Arabidopsis-CLps is considerably shorter than that for HLB-citrus related assays reported previously. Our model plant-pathosystem can be used as a resource in HLB-related preliminary studies to more rapidly understand the effect of (or manipulate) potential genes and defense inducing chemicals on Liberibacter infection.

High throughput genotyping technologies, such as SNP genotyping arrays that have emerged in recent years, have revolutionized the way we can analyze the origin and genetic control of traits in crops and have proven to be high-throughput and cost-effective tools. I developed Affymetrix™ Axiom® Citrus HD Genotyping Array and Affymetrix™ Axiom® Citrus Genotyping Array with 1.4 M and 58 K autosomal SNPs. DNA from 924 citrus accessions was hybridized to one or both of these arrays and yielded an extensive genomic inventory of the Citrus Variety Collection (CVC). High concordance was generally observed between DNA sequence and array data, and a simulation study suggested that errors in the range observed in more distant taxa have little to no effect on analyses such as principal component analysis (PCA), phylogeny and admixture analysis. This data was also exploited in a systematic study of genetic diversity, including phylogeny, admixture, and patterns of introgression at the chromosomal level. The study identified non-admixed accessions belonging to citrus progenitor species, and revealed the scale of diversity found in the CVC. Putative ancestors were identified for the identified admixed accessions and many species named by Tanaka were found to be hybrids. The study also suggested papeda’s polyphyly, in addition to Citrus hongheensis being a possible progenitor species to Citrus maxima. Lastly, we analyzed nucellar embryony (apomixis), an important trait in citrus because it significantly influences strategies for both scion and rootstock breeding. Some of the seedlings arising from a seed of a genotype with nucellar embryony are genetically identical to the seed parent. To better understand the genetic basis of nucellar embryony, a genome-wide association study (GWAS) was conducted separately in Citrus and Poncirus for two distinct but related traits, nucellar embryony and polyembryony. In this study, an average of 46 germinated seedlings from each of 209 accessions were genotyped with KASP SNP markers to determine whether they originated from a nucellar or a zygotic embryo. Using a high density SNP array for citrus to genotype the seed source trees, the GWAS analysis identified a major dominant locus in Citrus and a separate locus in Poncirus.

Citrus, one of the most economically important fruit crops in the world, is commercially propagated through grafting. Varying scion-rootstock combinations cause substantial effects on trees that often influence yield and fruit quality traits. Presently, the explanation for these differences has not been extensively studied at the molecular level. One potential reason for rootstock effects on fruit quality is the presence of small RNAs, molecules known to affect gene expression and plant development.

It was hypothesized that grafting diverse rootstocks influences small RNA populations in citrus, which can greatly impact fruit quality. By using mRNA-seq and small RNA-seq, an integrative analysis of microRNA and mRNA expression profiles was performed on grafted citrus. In this study, fruit and root samples were collected from sweet orange scions grafted onto four genetically differing rootstocks (trifoliate orange, Carrizo citrange, rough lemon, and sweet orange). Differentially expressed microRNA (DEM) and mRNA profiles were identified according to fold change analysis and the relationships between these molecules were identified.

A total of 1,633 unique genes in the fruit tissues and 11,368 in the root tissues were identified as differentially expressed between genotypes in at least one collection date during fruit development. The majority of the differences came from the comparisons between trifoliate and rough lemon rootstocks. GO and KEGG functional annotation analyses revealed that genes differentially expressed between trifoliate and rough lemon rootstocks were related to defense response, cell wall organization and biosynthesis, and carbohydrate metabolism. DEGs in fruit tissues grown on these rootstocks were related to transcription regulation and plant hormone signal transduction.

The small RNA sequencing data identified 603 known, conserved miRNAs. Of these, 188 were differentially expressed between trifoliate and rough lemon rootstocks. Reads that were not homologous to any know plant miRNAs were assessed as potentially novel small RNAs. 17 potential novel microRNAs were identified in citrus, ten of which were differentially expressed between genotypes. Quantitative reverse transcriptase-polymerase chain reaction (qRT-PCR) analysis confirmed the results from the sequencing data and revealed a negative correlation between several DEMs and their respective target genes, whose function may play a crucial role in citrus fruit development.

ABSTRACT OF THE DISSERTATION

A Molecular Genetic Approach to Evaluate a Novel Seedless Phenotype Found in Tango, a New Variety of Mandarin Developed From Gamma-Irradiated W. Murcott

by

Jennifer Robyn Crowley

Doctor of Philosophy, Graduate Program in Genetics, Genomics, Bioinformatics University of California, Riverside, December 2011

Dr. Mikeal Roose, Chairperson

Since the beginning of its first known cultivation as described in ancient European literature (circa 300 BCE) citrus fruit has been one of the most important consumer products in the world. Currently, there is a great deal of research in developing new citrus varieties and in improving existing citrus varieties. Creating and improving citrus varieties can be achieved in one of several different ways: cross-breeding, genetic transformation, artificial selection, and irradiation. Molecular alterations within plant tissue exposed to radiation, such as genetic mutations, can and do occur, making the method of plant irradiation a desirable choice when attempting to create unique citrus varieties.

The intent of this study is to evaluate and compare the molecular genetics of one particular variety of citrus, a seedless mandarin-producing Tango, with its predecessor variety, a seedy mandarin-producing W. Murcott. The Tango variety was developed and propagated from gamma-irradiated W. Murcott budwood. In the case of Tango, fruit produced by a mature tree retained the same characteristics of size, color, easiness to peel, and sweetness as that produced by W. Murcott; Tango fruit differed from W. Murcott fruit only in its seedlessness.

To determine the underlying molecular differences between W. Murcott and Tango, molecular regions prone to alteration due to radiation exposure were targeted for analysis. These included chromosomes, LTR (long-terminal repeat) retrotransposon regions, and methylated/unmethylated CpG regions. For chromosome analysis, stained images of chromosome spreads taken from W. Murcott and Tango cells frozen in their meiotic metaphase state were captured via a Leica SP2 Confocal microscope. Images were compared and contrasted between the two citrus varieties to determine any significant differences in chromosome alignment. CIRE1, a sequenced, intact LTR retrotransposon in the Citrus sinensis genome was targeted for quantitative analysis in the W. Murcott and Tango genomes. Primer sets were designed based on the CIRE1 sequence found within the NCBI database archives, targeting three regions of CIRE1: a long-terminal repeat, an integrase gene, and a reverse transcriptase gene. CpG region methylation patterns were compared and contrasted between W. Murcott and Tango DNA extractions using LUMA analysis followed by pyrosequencing. HpaII (methylation insensitive) and MseI (methylation sensitive) restriction enzymes were used to digest W. Murcott and Tango DNA subsequent to pyrosequencing; the digestions were compared with their respective sequenced untreated controls as well as to each other to determine any significant deviations in CpG island methylation levels. Methylation-sensitive Amplified Fragment Length Polymorphism (ms-AFLP) analysis, which targeted methylated cytosine sites in the Citrus genome, was also performed by Lisa Mu to determine whether or not polymorphic bands between W. Murcott and Tango could be visualized on a polyacrylamide gel.

The results from experimentation showed that W. Murcott and Tango are nearly genetically identical. Both W. Murcott and Tango were found to have 14 copies of the CIRE1 retrotransposon in their genomes, and there were no significant differences in the percentage of methylated cytosines within their genomes (ranging from 53.1% to 58.6%) nor were there deviations in methylated cytosine patterns. However, there are some notable molecular genetic differences between the two varieties; in particular, chromosome misalignment in the meiotic metaphase state was detected in several images of Tango chromosome spreads, while only normal chromosome alignment in the meiotic metaphase state was ever detected in the images taken of W. Murcott chromosome spreads. This suggests that chromosomal abnormalities resulting from gamma radiation exposure could be the main underlying genetic cause of seedless fruit produced by Tango citrus plants.

The history of recorded citrus can be traced back to prior to 800 B.C. and citrus species and related genera are considered to be from the tropical and subtropical region from south-east Asia to north-eastern India, and southern China, Indo-Chinese peninsula and Malay Archipelago, and later spread to the Middle East by Arab traders and then the rest of the world. As one of the valuable consumer products, citrus has attracted great attention from researchers. Citrus breeding can create and/or improve citrus varieties using a variety of approaches, including cross-breeding, genetic transformation, mutation breeding, etc. Mutation breeding accelerates the process of mutation and provides a great opportunity to develop new traits of citrus.

The purpose of this study is to identify the genomic differences between a very low-seeded citrus cultivar, Tango mandarin, and its seedy predecessor cultivar, W. Murcott mandarin. Tango mandarin was developed by gamma-irradiation mutation breeding of W. Murcott budwood. Both varieties share the same characteristics of size, color, sweetness and easiness of peeling, and they differ mainly in Tango having much lower pollen viability and in fruit seed content. W. Murcott fruit will be seedy unless grown in an isolated condition without cross-pollination while Tango is very low-seeded in all conditions. The intents of this study are to determine whether Tango is genetically uniform or has different genetic compositions within different cell layers, and to better characterize the putative cause of seedlessness of Tango fruit.

To determine the underlying genomic differences between W. Murcott and Tango, Illumina next-generation sequence (NGS) data of both W. Murcott and Tango leaf DNA was primarily used in analysis. Sequence data was aligned to the haploid Clementine reference genome and SNPs and Indels were predicted with the Pindel and Dindel programs. Allele-specific PCR, high-resolution melting real-time PCR and TA-cloning followed by sequencing were used for verification. For chromosome rearrangement prediction, Illumina NGS data was analyzed by novoBreak program. Copy number variation was determined by analyzing the depth of coverage of Illumina NGS reads and by loss of heterozygosity analysis with Axiom® Citrus 15AX SNP array data. PCR with primers flanking the putative breakpoints followed by Sanger sequencing was used to verify the chromosome rearrangements and determine the position of chromosome rearrangement breakpoints. For chimerism analysis of Tango, Tango pollen and fruit albedo were used to represent cell layer II and juice vesicles were used to represent cell layer I. Previously verified short deletions and chromosome rearrangements were tested with DNA from these tissues to determine their genetic composition.

These experiments identified and/or confirmed three SNPs and seven Indels which were heterozygous in Tango while they were homozygous in W. Murcott. These SNPs and Indels were developed into molecular markers that differentiate Tango and W. Murcott. The copy number variation results showed no large insertions or deletions present while one heterozygous translocation of a 6 Mb segment on chromosome 2 and one heterozygous inversion of 4 Mb on chromosome 4 in Tango were verified, and their breakpoints were determined. A potential translocation between chromosome 3 and chromosome 1 were also found. These two chromosome rearrangements could be responsible for misalignment and abnormal segregation of chromosomes during meiosis as previously observed, and the resulting inviable gametes could reduce seed content of fruits. Tango pollen and fruit tissue analysis proved that two short deletions and all chromosome rearrangements were present in cell layer II but absent in cell layer I of Tango meristem, while the other five short deletions may be present in both cell layer I and cell layer II. This suggested that cell layer II, which gives rise to gametes, was affected by the initial gamma irradiation, and inviable female gametes give rise to seedless Tango fruit and inviable male gametes greatly reduce male fertility.

NA