The total production of citrus in the United States in the 2018 growing season was 6.1 million tons on a total of 679 thousand acres. California produced 58 percent of the United States total, producing 3.6 million tons of citrus on 278 thousand acres. Approximately 75 percent of California’s citrus production is sold to the fresh market opposed to being processed into other commodities, such as juice. Oranges accounted for 64 percent of the total citrus produced in the United States and were valued at $1.8 billion, according to the United States Department of Agriculture1 . Citrus trees are rarely grown from seed and virtually all commercial citrus is propagated by grafting. This reduces the juvenile phase, allowing for the trees to produce fruit many years earlier than those grown from seed . The most significant impacts are on growth and vigor, tree nutrition, stress resistance, and fruit quality. In citrus, phenotypic differences in fruit quality have been well documented. However, understanding of the molecular mechanisms underlying these differences is lacking, especially regulatory mechanisms. Previous studies in apple, grape, sweet cherry, and other fruit crops have examined transcriptome changes in various rootstock-scion combinations. In citrus, growing blueberries gene expression profiling has been used to understand rootstock effects on growth of trees and responses to biotic and abiotic factors.

Many transcriptomic studies have also been performed in citrus to elucidate fruit ripening and development in commonly grown citrus cultivars. To date, none of these reports have linked the genetic effects of citrus rootstocks to fruit quality. Fruit growth and development and the mechanisms underlying fruit quality are complex. Signal transduction systems regulate many aspects of fruit ripening. During citrus development, the ABA-signal pathway may act as a central regulator of ripening, combined with other hormones, including auxin and ethylene. A recent study showed that ABA is a positive regulator of citrus ripening and exogenously applied ABA regulates citrus fruit maturation, suggesting that ABA metabolism plays a crucial role in citrus fruit development and ripening. Previous studies identified Protein phosphatase 2C family proteins as negative regulators of ABA signaling. PP2C dephosphorylates and inactivates a SNF1-related kinases family 2 protein, which is a positive regulator of ABA-response pathways. Plants with an inactive form of PP2C were hypersensitive to ABA, causing increased activation of ABA-responsivegenes. ABA-signaling response has also been linked with drought-stress tolerance. This study suggested that ABA accumulation is associated with a decrease in relative water content and Romero et al. suggest that ABA increases caused by dehydration upregulate levels of PP2C34 . Auxin, another phytohormone important for fleshy fruit development, regulates many growth and development processes. The auxin-signaling pathway regulates transcription of hundreds of auxin-inducible genes.

Promoters of these auxin-responsive genes contain auxin-responsive elements , which bind the auxin-response factor family of transcription factors41. ARF activity is regulated in part by Aux/IAA genes, which are transcriptional repressors of the auxin response. In the absence of auxin, Aux/IAA proteins dimerize with ARFs and recruit corepressors of the TOPLESS family, which in turn recruit chromatin-remodeling proteins that stabilize the repressed state. When auxin is present, it acts as a “glue” between Aux/IAAs and F-box proteins that are part of a ubiquitin protein ligase complex. This causes polyubiquitination and subsequent degradation of Aux/IAAs, which releases its repression, leading to the activation of auxin-regulated genes. Together with ABA and other hormones, auxin regulates several aspects of fruit development, including fruit set, fruit size, and ripening related events . Additionally, prior studies have indicated that small RNAs may play a regulatory role in fruit development and ripening. Small RNAs are a type of single-stranded, non-coding RNA that is typically 20-24 nucleotides in length, of which microRNAs are the most extensively researched class and are known to post-transcriptionally downregulate the expression of target mRNAs through mRNA cleavage or translational inhibition .In strawberry, miR159 was shown to act as a ripening regulator by targeting a MYB transcription factor, which plays a crucial role in the ripening process. Several examples of miRNA involvement in fruit development and maturation have been described in a variety of crop species, including apple, grape, peach, blueberry, date palm, and tomato. miRNAs that suppress specific transcription factors that are thought to be regulators of citrus fruit development and ripening have also been identified.

However, the expression profiles of miRNAs in various scion-rootstock combinations and their subsequent impact on fruit quality have not yet been evaluated. In this study, trees grafted on four rootstocks were chosen from a rootstock trial at the University of California, Riverside to assess for various fruit quality traits; Argentina sweet orange, Schaub rough lemon, Carrizo citrange, and Rich 16-6 trifoliate orange. In general, rough lemon rootstocks produce the highest yield and fruit size, but fruit is of lower quality, containing lower acidity and lower levels of total soluble solids, also known as the “dilution effect”. On the other hand, trees on trifoliate orange produce smaller, high quality fruit with high yield on often smaller trees. Carrizo citrange rootstocks produce intermediate yield with good fruit quality. Sweet orange rootstocks produce good quality fruit, but trees are very susceptible to various citrus diseases. An RNA-seq approach was implemented to investigate differences in gene expression in fruit due to genetically varying rootstocks, with the aim of identifying genes that could potentially play a role in improvement of fruit quality. Furthermore, miRNA expression profiles were obtained for each of the rootstocks to identify potential regulatory mechanisms associated with their target genes.Gene Ontology and pathway enrichment analyses were conducted to explore the functions of genes that were DE in trees on different rootstocks. GO categorization showed that the molecular function GO terms ‘DNA-binding transcription factor activity’ and ‘transferase activity’ were significantly enriched . Genes associated with photosynthesis and located in the photosynthetic membrane were also enriched . KEGG pathway analysis revealed that genes for plant-hormone signal transduction, carotenoid biosynthesis, and fructose and mannose metabolism were significantly enriched when comparing fruit grown on trifoliate to rough lemon rootstocks . The hormone-signaling-related pathway included DEGs involved in auxin, gibberellin , abscisic acid , ethylene , and jasmonic acid signaling . Visualization of fold changes using MapMan software revealed that several genes in the ABA and GA pathways were down-regulated in fruit grown on rough lemon compared to trifoliate rootstocks. Genes in the ethylene and auxin pathways were both up and down regulated . Many genes involved in other cellular responses, as well as transporters were also DE .Among the identified DEGs, there were many genes that have potential roles in fruit quality. Over 130 citrus genes were DE in this study that have been previously linked to fruit development and ripening. Of these DEGs, 8, 6, and 22 belonged to starch-related, fructose-related, and hormone-signaling proteins, respectively. Additionally, 25 of these DEGs were annotated as transcription factors . Most of the significant differences in transcriptional changes occurred at the second and third time points, and many of the expression levels changed from time two to time three . Figure 1.11, square plant pots which shows all DEGs between genotypes, displays this trend.To further understand the genetic influence of rootstocks on fruit quality, we focused on the expression changes of miRNAs and their target genes. miRNAs are posttranscriptional regulators that cause downregulation of target genes. Therefore, if a target gene is down-regulated by a miRNA, a negative correlation between miRNA expression and the target mRNA expression is expected. No statistically significant differentially expressed miRNAs were observed in our fruit small RNA seq data, so we instead predicted miRNAs that are potentially targeting DEGs found in fruit tissues. This approach was taken due to the complex regulatory networks that are known to exist in plants and other higher organisms. One miRNA may regulate many genes as its targets, while one gene may be targeted by many miRNAs. Both of these scenarios were observed in citrus roots in response to dehydration and salt stress.

To evaluate these potential relationships in this study, the psRNATarget program was utilized, which accepts a list of known plant miRNAs in citrus and the coding sequence of the DEGs reported here to predict miRNAs according to the criteria described by Meyers et al. . Over 15,000 miRNA-mRNA interactions were predicted using psRNATarget. The RNAseq data was then utilized with an in-house R-script in order to select potential interacting pairs with an expected negative correlation in gene expression. After removal of genes that did not have any functional annotation, there were 366 combinations of miRNA-mRNA pairs that showed reciprocal expression patterns. Comparing these genes with the enriched GO terms and KEGG pathways led to several candidate miRNA-mRNA interactions that could be causing changes in fruit traits when differentially expressed between rootstocks. These genes included transcriptional regulators, hormone signal transduction genes, transporters, and sugar metabolism genes.Based on the interacting pairs predicted and their relevance to fruit quality, 10 pairs of miRNAs and target mRNAs were selected for validation via qRT-PCR analysis . Samples collected at timepoints two and three were chosen for validation due to the larger differences in expression levels of genes in fruit grown on trifoliate orange compared to rough lemon rootstocks at those times. For qRT-PCR, two biological replicates and three technical replicates were analyzed to quantify expression of each gene. Three miRNAs were up-regulated at both timepoints, while their target genes were down-regulated, two miRNAs were down-regulated at both timepoints, while their target genes were up-regulated, and the remaining five miRNAs validated were down-regulated at one time point and up-regulated at the other . The correlation between the relative expression level detected by qRT-PCR and by RNA-sequencing was calculated. Pearson correlation values were highly significant with r = 0.94, which strongly supported the sequencing data . However, certain miRNA-mRNA pairs did not have the expected fold changes from one time point to the next. For example, Csi-miR171a shows an increased fold change from September to November. This should correlate with a decreased fold change from September to November in the target gene , but instead, we see an increase in the target mRNA expression from September to November. Only this pair and Csi-miR1863 – ATEXP1 show this inconsistency. The results for the remaining eight pairs were consistent with their expected expression levels. Figure 1.12 shows that seven of the miRNAs had increased expression levels in November compared with September, while three miRNAs decreased in expression from the during fruit development.The objective of this study was to correlate changes in gene expression of grafted citrus trees to effects in fruit quality due to varying rootstocks. In this study, four rootstocks were chosen from a rootstock trial with Washington navel orange scion in Riverside, CA to assess for various fruit quality traits; Argentina sweet orange, Schaub rough lemon, Carrizo citrange, and Rich 16-6 trifoliate orange. Fruit quality data was collected from fruit grown on each of the four rootstocks at the end of the growing season when fruit were ripe. In the present study, weight, height, width, rind color, rind texture, peelthickness, internal texture, juice weight, percent juice, total soluble solid and titratable acid levels were measured. The total yield and average fruit weights were markedly higher in navel orange fruit from trees grafted onto rough lemon compared to sweet orange, Carrizo citrange, or trifoliate orange rootstocks. The rind thickness was also greatest on rough lemon rootstocks compared with the other rootstock-scion combinations. The most substantial differences could be seen in total soluble solids and acid levels. The highest levels of sugars and acids were found in fruit grown on Carrizo citrange. Trifoliate orange and sweet orange rootstocks produced fruit with only slightly lower sugar and acid levels, while rough lemon produced fruit with significantly lower levels . This is consistent with the previously mentioned reports of rootstock effects on fruit quality. Presently, there is very little understanding of how rootstocks influence citrus fruit quality, especially at the level of gene regulation. In this study, an integrated mRNA and miRNA high throughput sequencing analysis in fruit grafted onto genetically diverse rootstocks was performed to help resolve potential mechanisms of rootstock-scion effects on fruit quality.In the present study, RNA-seq was used to investigate transcriptome differences in the fruit of ‘Washington’ navel sweet orange grafted onto different rootstocks and explore genes that may influence fruit quality traits. Juice vesicles from fruit grafted onto four genetically diverse rootstocks at four different fruit development periods were sequenced. The RNA-seq approach detected a similar number of genes in all samples.