Network Studies On Postharvest Senescence Revealed By Transcriptome And Metabolomics Profiling In Citrus Fruits

Fruit ripening and senescence are inevitable and irreversible processes in plant life history. Citrus, a typical non-climacteric fruits, ranks the most dominant position among global fruit industry and shows world-wide economic importance. The study of senescence processing in citrus has important scientific value and socioeconomic benefits. However, biological behaviors underlying citrus fruit ripening and postharvest senescence remain uncertain. Citrus could be regarded as particularly advantageous models for woody perennials tree fruits studies for senescence. In present study, citrus fruit postharvest senescence processes were comprehensively analyzed based on ?Omics‘ information. Transcriptome, metabolomics and physiological quality datasets were collection from four main citrus fresh fruit varieties: Satsuma mandarin(Citrus unshiu Marc)(S), Ponkan(Citrus reticulata Blanco)(K), Newhall sweet orange(Citrus sinensis L. Osbeck)(O) and Shatian pummelo(Citrus grandis Osbeck)(P). We analysis all datasets characteristic at three independent levels: differences between four varieties, differences between rind and flesh tissues, and differences of time trend. We described relationships among datasets by the correlation of network analysis. Furthermore, we also compared senescence process gene expression of two model fruits tomato and grape to citrus. All data sets were processed in order to gain novel insights into citrus senescence: further to discovery the communication between rind and flesh tissues of hesperidium, to find out specialties among citrus varieties with different storage behaviors, to reveal important changes during storing process, and to uncover citrus special non-climacteric characterizes.In addtionly, we employed ortholog-based and domain-based methods to predict orange protein-protein interaction(PPI) network. K-nearest neighbors(KNN) classification method is used to verify and filter the network. The final predicted PPInetwork named Citrus Net. The availability of Citrus Net provides valuable resource for the study of biological functions in sweet orange. The main results were showed as follow:(1) Integrated analysis of metabolites and transcription by Citrus CycTo have a global view of postharvest senescence, here we constructed a genome-scale Citrus Metabolic Pathway Network(Citrus Cyc) based on Ara Cyc by sequence similarity. The Citrus Cyc contains 3,490 reactions, 2,614 metabolites and all major metabolic pathways including sugars/acids metabolic reactions, lipid and amino acid reactions, most phytohormones related reactions, etc. Co-expression networks of reaction genes expression(RE-values)(Networks-RR) were constructed and clustered by Markov Cluster(MCL) Algorithm in the rind and flesh respectively. Major clusters showed modularity character in Networks-RR. Additionly, correlation network between metabolites and RE-values(Network-MR) was calculated by Jaccard distance. It integrated analysis of metabolites and transcription in Citrus Cyc and focused on some important metabolites have significant effect on fruit quality like sugars and acids, and identify the active reactions of those important metabolites. Results indicated that: the sugars reactions more activity in flesh; the organic acids reactions more activity in rind and had great influence in metabolic network.(2) Analysis of the co-expression networks with DEGs at varieties variation levelCo-expression network of differentially expressed genes(DEGs) at varieties variation level(Network-DV) demonstrated modularity feature. The abiotic and biotic stress related TFs were respectively clustered in sub-networks of DEGs in loose-skin and tight-skin fruit, which act as critical distinction factors between tight-skin and loose-skin citrus and indicated loose-skin citrus storing may have activity gene regulate specific signaling pathways in response to abiotic stress factors.Analysis of correlation network of metabolites between the rind and the flesh(Network-Mrf) suggested energy source transportation from flesh to rind during citrus storage life after harvested from tree. We also focused on transporters to analysis energy source transportation process. All the results support the mode of connection between hesperidium tissues during storage is a transporter-mediated transportation process thattransfers water, nutrients like glucose and fructose(probably sucrose), etc. from flesh to rind.Co-expression networks of DEGs at time variation level(Networks-DT) combined with time-trend analysis of metabolic and physiological data could comprehensive represent citrus fruit senescence process. Generally, fruit ripening and senescence are associated with series changes in color, texture, aroma, and nutritional components caused by change of affecting genes. There were several senescence-associated genes(SAGs) have expression during citrus storing. Furthermore, Senescence regulation was illustrated by distinct distribution of TFs in Co-expression networks of DEGs at time variation level. It indicated the rind were more important than flesh during storing.(3) Comparison of phytohormone related gene expression with model fruit reveals citrus have specific non-climacteric characteristicWe also comparied the phytohormones expression patterns among citrus, tomato and grape to revealing the special characteristics of citrus fruit. Our results proved that citrus senescence is co-regulated by both ethylene and abscisic acid, which seems exist a system-2-like pathway of ethylene production similar climacteric fruit tomato and have weaker abscisic acid synthesis genes activity than that in non-climacteric fruit grape.(4) Model of flesh-rind communication and citrus senescence processingWe offered and verified a hypothesis that speed distinction in flesh-rind substances transport channel formed by anatomic structural difference between citrus varieties might be the important factor influence tight/loose-skin fruit senescence behavior. In addtionlly, tight/loose-skin citrus senescence processing was modeled in detail: At early stage of senescence, fruit rind is direct exposured to environment as a focus of energy expenditure for biotic and abiotic stresses resistance, the citrus fruit exports nutrient from flesh to rind through different channels for maintain bioactivity of whole fruit. The substances depletion in rind with different flow velocity resulted in abiotic stress(mainly water deficit stress) of varying degrees, induced phytohormones signals, transcriptional factors regulation and series of physiological and biochemical reactions.(5) Prediction and functional analysis of sweet orange protein-protein interaction networkCitrus Net, a predicted protein-protein interaction network of orange, contains 8,195 proteins with 124,491 interactions. The quality of Citrus Net is evaluated by gene ontology(GO) and Mapman annotation, which confirms the reliability of the network. In addition, we calculated the expression difference of interacting genes(EDI) in Citrus Net with RNA-seq data from four sweet orange tissues, and analyzied the EDI distribution and variation in different kinds of sub-networks. We also demonstrated that gene expression in Citrus Net has significant modular feature.