FRUIT INHIBITS FLOWERING IN ALTERNATE BEARING CITRUS VARIETIES. HORMONAL, GENETIC AND EPIGENETIC REGULATION

Transcripción

1 DIPARTIMENTO DI SCIENZE AGRARIE E FORESTALI Dottorato di Ricerca in Frutticoltura Mediterranea FRUIT INHIBITS FLOWERING IN ALTERNATE BEARING CITRUS VARIETIES. HORMONAL, GENETIC AND EPIGENETIC REGULATION SSD AGR/03 Arboricoltura generale e coltivazioni arboree Ph.D. CANDIDATE Natalia Muñoz Fambuena CO-SUPERVISOR Dr. Domingo J. Igleia Fuente SUPERVISOR Dr. Carlo Meejo Conejo CO-SUPERVISOR and Ph.D. COORDINATOR Prof. Maria Antonietta Germanà CYCLE XXVI - ACADEMIC YEAR 2015

2

3 DIPARTIMENTO DI SCIENZE AGRARIE E FORESTALI Dottorato di Ricerca in Frutticoltura Mediterranea FRUIT INHIBITS FLOWERING IN ALTERNATE BEARING CITRUS VARIETIES. HORMONAL, GENETIC AND EPIGENETIC REGULATION SSD AGR/03 Arboricoltura generale e coltivazioni arboree Ph.D. CANDIDATE Natalia Muñoz Fambuena CO-SUPERVISOR Dr. Domingo J. Igleia Fuente SUPERVISOR Dr. Carlo Meejo Conejo CO-SUPERVISOR and Ph.D. COORDINATOR Prof. Maria Antonietta Germanà CYCLE XXVI - ACADEMIC YEAR 2015

4

5 AGRADECIMIENTOS/ RINGRAZIAMENTI Eta tei ha upueto un largo viaje que empezó hace cai 7 año, mucha gente ha influido en ella y por eo me gutaría agradecérelo. Una tei implica mucho trabajo, que a vece no parece vere premiado, pero el reultado final compena TODO. En primer lugar me gutaría agradecer a Manuel Agutí el haberme abierto la puerta en u equipo en el IAM y haberme permitido trabajar y crecer cada día má, con u conejo y u apoyo incondicional, abiendo iempre como motivarme a eguir adelante. Cómo dice él, No hay que rendire nunca. Mucha gracia. A mi director tanto de mi trabajo final de carrera, de teina de máter y ahora finalmente de mi Tei Doctoral, Carlo Meejo. El cuál ha etado iempre a mi lado apoyándome y dirigiéndome en la dirección correcta. Ademá de no permitirme nunca deanimarme, levantándome en lo momento duro y iempre con una viión alegre y poitiva de la vida. A ti gracia por todo. A Carmina por tranmitirme u buena actitud y conejo, apoyándome iempre en la conecución de mi objetivo. A Amparo, gracia, por er mi amiga y por ayudarme iempre diciéndome y haciendo aquello que neceitaba en ee momento. A mi do grande compañero, el ya doctor Roberto y a la futura doctora Ana, por u apoyo, alegría y amitad, eta tei no habría ido igual in vootro. A Vicent por u buena diponibilidad iempre. A todo lo que han paado por el IAM en eto año: Bartolomé, Giovanni, Caterina, Salvatore, Andrea, Nicola, Víctor, Fran, Alfredo, María, Diego, Alba, Marta, Alejandro, Helena, Tono. Quiero agradecer en epecial a MªJoé y a Sara u ayuda y u buena actitud iempre. A todo ello gracia por la ria y lo bueno momento en la Sala de Becario. A mi compañero doctore y futuro doctore, por eo almuerzo y comida en la que hemo compartido tanta confeione entre ria y iempre intentando ver el lado poitivo de la coa: Alex, Critina, Altea, Juan Antonio, Vicky, David y Tono. A Mercede, la cual iempre ha etado dipueta a ayudarme con u alegría, bueno conejo y con la que he compartido grande momento. Quiero agradecer del IVIA a Eduardo Primo, por animarme y eneñarme a invetigar, permitiéndome aprender de u experiencia orientándome en la dirección correcta. También quiero agradecer a mi co-director Domingo J. Igleia, que me ha apoyado iempre y del que he aprendido mucho durante eta tei, demotrándome que la palabra confianza exite en ete mundo en el que trabajamo. Gracia a lo do por lo bueno conejo. A Ana e Ia, mi mami durante mi periodo en el IVIA. Gracia por cuidarme, ayudarme y hacerme ver que todo tiene u lado bueno. Del IBMCP quiero agradecer a Miguel Blázquez, por abrirme la puerta de u laboratorio y hacerme todo tan fácil. A Paco, por u gran ayuda durante toda eta última fae,

6 gracia por tu buena actitud y diponibilidad iempre. A Aurelio Gómez y a Julián Cueva, gracia por reviar y mejorar eta tei. A mi amigo, lo cuále han tenido que aguantar mi ida y venida durante todo eto año.gracia Queta tei è tato un lungo viaggio che ha iniziato quai ette anni fa, molte perone hanno influenzato e quindi vorrei ringraziare. Una tei comporta un acco di lavoro, che a volte non embra eere premiato, ma il riultato finale vale TUTTO. Mi piacerebbe ringraziare alla mia co-direttora Maria Antonietta Germanà per avermi aperto il uo laboratorio permettendomi di imparare a viluppare una linea di invetigazione ed appoggiandomi durante tutta la mia tei dottorale. Grazie mille. A Benedetta Chiancone, grazie, non olo per l'aiuto ricevuto durante il mio periodo italiano nel laboratorio, nel quale, i problemi non erano problemi. Ma oprattutto voglio ringraziarti che mi accogliei nel tuo mondo, facendomi entire come nella mia propria caa, eendo come una orella maggiore per me. A i profeori Franca, Caruo, Marra, Paolo Inglee, Ricardo e Vittorio e a Laura, Roberto, Valeria, Germana, Fabrizio, Ahmed, Valeria G., Letizia, Marine, a tutti voi grazie per l aiuto e le riate condivie. A Pina, per condividere con me la ua poitività ed amicizia. Alle mie ragazzine, Martina, Serena, Giulia e Roalia per eere tato lì la mia famiglia. A mio fratello, Aleio, grazie per prenderti cura di me, aiutarmi empre in tutto ed includermi nella tua vita. Grazie per includermi nel tuo gruppo di amici e che i traformarono anche in miei, grazie a: Giueppe, Franceco, Critina, Aleia, Mariella, Giovanni, Aleia F., Luca, Emilio, Andrea. Ai miei compagni e futuri dottori, per quelli momenti in cui abbiamo condivio tante confeioni, ridendo e cercando empre di vedere il lato poitivo delle coe: Selene, Pino, Antonio, Fabrizia e Martina. In particolare, mi piacerebbe ringraziare a queto dottorato, l'avermi permeo di conocere quella perona, la quale i è traformata nella mia forza ed appoggio, e per la quale mi alzo ogni giorno, Walt Ademá, quiero agradecer a mi hermano (Ia, Nuri, Mari, Nel-lo y Cri) a mi cuñado (Omar, Santi, Jaime y Lorena) y por upueto a mi obrino (Jaume, Ale, Núria, Pau, Ia, Sami, Elena y Pep), el tranmitirme iempre u energía, alegría y buena actitud iempre hacia la vida. Finalmente, a mi padre Manuel e Iabel, gracia por apoyarme, orientarme, aguantarme y quererme cada día. O quiero, nada de eto ería poible in vootro. Haz que tú onria cambie el mundo, no que el mundo cambie tú onria

7 Table of content

8

9 Table of content 5 Abbreviation 11 Abtract 17 Introduction Flowering Flower induction, initiation and differentiation When do plant flower? Exogenou and endogenou factor inducing 24 flowering 1.3 Floral induction in annual/biennial plant compared to mature fruit 26 tree 2. Molecular bai of flowering Floral integrator gene and flowering pathway FT-FD in the meritem and floral initiation gene Epigenetic control of flowering Flowering in fruit tree pecie Qualitative molecular-genetic regulation of flower induction in tree Quantitative long ditance ignal regulation of flower induction Hypothei and objective Hypothei Objective 50 Material and method Plant material and experimental deign Previou experiment Characterization of Flowering Locu C Determination of Flowering Locu T reponible for floral induction Section Experiment I Experiment II Experiment III Experiment IV Section Section Method Yield, prouting and flowering evaluation Fruit characteritic Culture Biological material Seed preparation and terilization of Arabidopi thaliana Tranformation of Arabidopi thaliana with Agrobacterium 60 tumefacien by FLC characterization Microhoot terilization and culture of Mandarino Tardivo di 60 Ciaculli Bacterial train 61 7

10 Tranformation of Echerichia coli Tranformation of Agrobacterium tumefacien Molecular biology technique Extraction Protein RNA and genomic DNA Extraction of plamid DNA Protein analyi and meaurement Fluorecent labelling D electrophorei Gel imaging and data analyi Protein identification by ma pectrometry (MALDI, MS/MS) 65 analyi Starch analyi Catalae activity Gene ontology analyi Biulphite treatment Gene expreion analyi by qrt-pcr Sequence analyi and phylogenetic tree Generation of contruction in plamid vector Polymerae chain reaction (PCR) Ligation in pgem-t Eay I vector Blue-white creening Digetion Cloning FLC into a plant cloning vector Hormone iolation, purification and quantification Evaluation of leaf number Statitical analyi Supplementary figure. Phylogenetic tree Supplementary data. FLC equence 76 Reult 79 Section 1. Hormonal inhibition of flower induction: Gibberellin Fruit development and flower induction gene expreion Time-coure of gibberellin metabolim in the exocarp, node and leave 86 from ON and OFF tree 1.3 Exogenou control of flowering: Treatment with Paclobutrazol (PBZ) 92 and GA3 1.4 Branch independence: the hort ditance effect of the fruit. Fruit-hoot 98 phloem tranport interruption by peduncle girdling 1.5 In vitro culture of excied bud from ON and OFF branche 104 Section 2. Proteomic analyi of Moncada mandarin leave and bud with 108 contrating fruit load 2.1 Comparative proteome analyi 108 8

11 2.2 Identification of differentially expreed protein Claification of identified protein Specie from which the identified protein proceed Gene ontology analyi Starch, ABA and JA content and catalae activity in ON and OFF tree 128 Section 3. Epigenetic control of flowering Alternate bearing: biennial reproductive-vegetative development cycle DNA methylation profile During the flower induction period Flowering promoter Flowering inhibitor Before the flower induction period Flowering Locu C Time-coure of methyltranferae gene expreion Promoter of FLC expreion Inhibitor of FLC expreion Effect of defruiting on flowering, methyltranferae and flowering gene 142 expreion and FLC-DNA methylation Effect of defruiting on flowering Effect of defruiting on flowering gene and methyltranferae relative 143 expreion Flowering gene Methylation gene DNA methylation profile 145 Dicuion 147 Concluion 165 Reference 169 9

12

13 Abbreviation

14

15 12ATX. TRITHORAX 1,2 5ATX. TRITHORAX 5 5-AzaC. 5-Azacytidine 7ATX. TRITHORAX 7 a/b. annual/biennial ABA. Abciic acid AGL24. AGAMOUS LIKE 24 AGP. ADP-glucoe pyrophophorylae AP1. APETALLA1 bzip. Baic Leucine Zipper C/N. Carbon/Nitrogen CC. Companion cell CcFLC. Citru clementina FLOWERING LOCUS C CDF1. CYCLING DOF FACTOR 1 CEN. CENTRORADIALIS CETS. CENTRORADIALIS/ TERMINAL FLOWER 1/SELF-PRUNING CG.Citoine/guanine CHG.Citoine/ any nucleotide except G/guanine CHH. Citoine/any nucleotide except G/any nucleotide except G CiFT. Citru unhiu FLOWERING LOCUS T CK. Cytokinin CO. CONSTANS CLFY. Citru ineni LEAFY CT. Control DAT. Day after treatment DEF. Defruiting DNA. Deoxyribonucleic acid ELF8. EARLY FLOWER 8 Exp. Expected FD. FLOWERING LOCUS D FKF1. FLAVIN-BINDING, KELCH REPEAT, F-BOX 1 (FKF1) FLC. FLOWERING LOCUS C FM. Floral meritem FT. FLOWERING LOCUS T GA. Gibberellin GA20ox. Gibberellin 20-oxidae GA3. Gibberellic acid GA3ox. Gibberellin 3-oxidae GBSS. Granule-bound tarch ynthae GFP. Green fluorecent protein Abbreviation 13

16 Abbreviation GI. GIGANTEA GO. Gene Ontology H3. Hitona 3 H3K9. Hitone H3 lyine 9 HP1. Heterochromatin Protein 1 HPLC. High-performance liquid chromatography IAA. Indol acetic acid IM. Inflorecence meritem IP. Iopentyl adenine JA. Jamonic acid LB. Lyogeny broth LC/MS/MS. Liquid chromatography ma pectrometry LD. Long-Day LFY. LEAFY MADS. MCM1/AGAMOUS/DEFICIENS/SRF MALDI, MS/MS. Ma pectrometry MD. Malate dehydrogenae MFT. MOTHER OF FT MS. Murahige Skoog MYB. MYELOBLASTOSIS N. Nitrogen NCED3. 9-ci-epoxycarotenoid dioxygenae OFF. Without fruit ON. With fruit PBZ. Paclobutrazol PCR. Polymerae chain reaction PEBP. Phophatidylethanolamine binding protein pi. Ioelectric point PMF. Peptide ma fingerprinting PPM. Plant Preervative Mixture QTL. Quantitative trait locu RNA. Ribonucleic acid RT-PCR/qRT-PCR. Real time polymerae chain reaction SAM. Shoot apical meritem SD. Short-Day SE. Sieve element SE. Standard error SIM. Selected Ion Monitoring SKB1. PROTEIN ARGININE METHYLTRANFERASE 5 SOC. Super Optimal broth with Catabolite repreion 14

17 SOC1. SUPPRESSSOR OF OVEREXPRESSION OF CONSTANS 1 SPL3. SQUAMOSA PROMOTER BINDING PROTEIN-LIKE 3 SVP. SHORT VEGETATIVE PHASE t-aba. Tran-ABA TEM1. TEMPRANILLO 1 TFL1. TERMINAL FLOWER 1 Theo. Theoric TSF. TWIN SISTER OF FT TSS. Total oluble olid Tz. Tran-zeatina UPLC-MS/MS. Ultra high preure liquid chromatography, ma pectrometry VRN. VERNALISATION X-Gal. 5-bromo-4-chloro-3-indolyl-β-D-galactopyranoide Abbreviation 15

18

19 Abtract

20

21 Abtract In mature citru tree, flowering i regulated by exogenou ignal (i.e., low/high temperature, water hortage/water upply) that control flowering time, and by endogenou ignal that control flower intenity. Actually, the ratio of endogenou promoter to inhibitor i conidered a reponible for flowering; however, there i evidence uggeting that the inhibitor and their metabolim are the ole factor controlling flowering in Citru. All the meritem have the information and ability to flower unle a negative factor hamper the proce. The main endogenou factor controlling flower intenity i the fruit. Attending to the aforementioned, in thi PhD. thei, the following hypothei wa teted: Fruit inhibit flowering when ripening begin by exporting of hormone that induce epigenetic upregulation of flowering inhibitor gene in the leave, and interfering in flower bud differentiation. The main finding are: 1. The increae in FLC gene expreion in ON-tree leave coincide with fruit color change and low temperature in the flower induction period. CiFT2 i expreed in OFF-tree but not in ON-tree leave. 2. The fruit produce and export GA and ABA to the leave overlapping flower induction inhibition wherea thee hormone decreae in OFF-tree leave. GA3 treatment reduce CiFT2 gene expreion but not FLC gene expreion. 3. After bud prouting, the inhibitory ignal (FLC, TEM1, SVP) are not expreed in the new leave. Every bud beide a fruit need to retart vegetative growth to gain the flowering ability. 4. The fruit activate protein from the tre repone and the oxidoreductae activity while in OFF-tree the primary metabolim and ynthei of tarch i upregulated. 19

22 Abtract 5. During the flower induction period, the fruit modifie the DNA-methylation profile of the FLC gene and increae the expreion of methyltranferae that increae FLC gene expreion. In defruited tree the proce i reverted to the OFF-tree tate. 20

23 Introduction

24

25 1. Flowering Introduction Flowering comprie all of the development neceary for the irreverible commitment by the meritem to produce a flower or an inflorecence (Kinet, 1993). Thee tage include flower induction, flower initiation, flower differentiation and anthei. The overall flowering proce involve many tep, uually tarting with perception of an environmental factor and terminating with the differentiation of threedimenional tructure, the floral primordia (Zeevaart, 1976). During development, plant mut undergo a juvenile phae before tranition to adulthood. A period of juvenility i characteritic to all higher plant. Juvenility wa defined a the period during which a plant cannot be induced to flower (Wareing, 1959; Hackett, 1985; Poethig, 1990; Goldchmidt and Samach, 2001). Thi tage finihe with the firt flower. During the juvenile phae of plant development, meritem acquire reproductive competence becoming able to ene and repond to ignal that induce flowering. In annual/biennial plant thi period i very hort but in angioperm tree it lat for everal year. 1.1 Flower induction, initiation and differentiation In the adult tage, the plant i enitive to inductive factor. Flower induction i the tranition of the meritem from the vegetative to the reproductive phae. During thi period, the leaf and the meritem receive flowering ignal, and the gene required for flower development are turned on. A a reult, nutrient, hormone and protein metabolim change inide the bud. Flower initiation i the period when a erie of hitological change are underway, but no viible morphological difference are oberved. Flower differentiation i characterized by the development of the primordia of floral organ. Chailakhyan (1937) propoed that a flower-inducing hormone called florigen caue flowering. Movement of florigen inide the plant wa demontrated by mean of grafting experiment. In annual/biannual plant thi floral promoting factor i trong enough to even induce other plant a well if, for example, a ingle leaf i grafted onto a non-induced plant (Zeevaart, 1976), while in woody plant even though the inducing agent may be aturating a high number of the meritem alway tay vegetative (Davenport et al., 2006). 23

26 Introduction Searle (1965) tated that flower induction of photoperiod-enitive plant i controlled by florigen that i produced in leave and tranported to bud. The export of the florigen in the hort day plant Perilla p. from the leaf to the hoot apical meritem wa related with the idea of the long ditance tranport through the phloem (King and Zeevart, 1973). But the poibility of a low tranfer of the timulu from cell to cell wa alo uggeted from experiment with graft-induced plant where no vacular connection were formed (Welleniek, 1970). Florigen wa thought to play a poitive role activating gene or a negative one, blocking gene-repreor. After induction, the hoot apex change from the vegetative to the flowering tage a a reult of an increae in it mitotic activity (Bernier, 1971; Gifford and Coron, 1971). Thi proce wa tudied in the annual plant Sinapi alba, in which the rate of cell diviion increaed eight-fold in the central zone of the meritem and ix-fold in the peripheral zone. Following the cell diviion and ynchronization of the cell in the apex, a peak in DNA and protein ynthei wa oberved, thi being aociated with initiation of the firt flower bud (Jacqmard et al., 1972). 1.2 When do plant flower? Exogenou and endogenou factor inducing flowering Plant have developed the ability to detect eaonal change in order to decide when to flower, according to their poibilitie of development. Thu, the timing of flowering i regulated by autonomou and environmental factor, and four major pathway have been decribed to induce flowering: photoperiod, vernalization, gibberellin (GA) and autonomou pathway (Blázquez et al., 2006; Wilkie et al., 2008). Many flowering plant ue photoreceptor protein, uch a phytochrome or cryptochrome, to ene eaonal change in night length, or photoperiod. Long-Day (LD) and Short-Day (SD) plant flower in repone to a change in the length of the dark period. LD plant flower when the night length fall below their critical photoperiod. Thee plant typically flower during late pring or early ummer, a day are becoming longer. Example of thee plant are pea, barley, Arabidopi thaliana or wheat. On the other hand, SD plant flower when the night length exceed their critical photoperiod. Thee plant typically flower during ummer and fall. Example of thee plant are cotton or rice. Facultative LD plant can flower under SD condition wherea obligate LD plant cannot. 24

27 Introduction Neutral-day plant do not initiate flowering baed on change in photoperiod. Intead, they initiate flowering in repone to alternative environmental timuli depending on climate type. In temperate climate, a period of low temperature (vernaliation) induce flowering in many plant (e.g. tomato, roe, cucumber). When leave are expoed to low temperature, ugar accumulate and cold-regulated gene are induced to contribute to the cold acclimation proce (Goruch et al., 2010). Plant are able to detect the end of the cold period (winter) and, thu, flower in pring. In tropical climate, where no ignificant variation in photoperiod or temperature occur, ignal uch a water hortage have the ability to induce flowering (Southwick and Davenport, 1986). The hormonal (GA) and the autonomou pathway alo induce flowering, and are largely independent from environmental influence (Parcy, 2005). In the autonomou pathway flowering i induced by internal cue at particular tage of plant development, wherea GA act directly on the meritem at the flower initiation tage interacting with the ucroe content (Blázquez and Weigel, 2000). Therefore, the nutritional tatu alo play a role, at leat a a ource of energy, to induce flowering. In fruit tree a large number of bud remain vegetative in pring. Thi phenomenon wa explained by a deficit of nutrient. Baed on practical experience in fruit growing, the C/N-ratio hypothei wa ued to explain difference in flower bud differentiation in fruit tree. A high C/N ratio favor flower formation and exceive N fertilization inhibit it. However, it eem that nutrient are not the limiting factor for flower formation when a threhold level i reached (Lang, 1965). Finally, ome of thee factor interact to induce flowering. Studie of the photoperiodic repone of plant have demontrated coniderable diverity in the critical length of the dark period, the age at which eedling are ripe-to-flower and the effect of temperature on photoinduction (Zeevart, 1976). In general, all wild population of rice exhibit a trong SD repone while cultivated varietie how a lower enitivity (Katayama, 1971). Other plant, a ugarcane, require intermediate photoperiod, around 12 ½ hour, for floral initiation (Julien, 1973). The C3 and C4 pathway of photoynthei alo have been correlated with the photoperiodic requirement for flowering (Purohit and Tregunna, 1974). There i a group of plant originally conidered to be trictly photoperiodic, which can alo be induced to produce flower bud by factor other than daylength. One of thee factor i the temperature. Thu, the 25

28 Introduction photoperiod requirement, beide the high or low temperature, regulate flower formation (Blondon and Harada, 1972; Deronne and Blondon, 1973). In the pat, it ha been very difficult to detect difference in metabolim or in a chemical compoition between induced and non-induced plant, and therefore, reult were difficult to interpret given change in photoperiod or temperature. However, the development of genomic and trancriptomic tool ha contributed to a better undertanding of the metabolic and molecular procee involved in floral biology. Mot of our knowledge about flower induction ha come from tudying flowering regulatory gene in Arabidopi thaliana. 1.3 Floral induction in annual/biennial plant compared to mature fruit tree Floral induction in mature fruit tree i ditinct from that of annual/biennial (a/b) plant. In tree, it i a quantitative proce with a ignificant proportion of the above ground meritem remaining vegetative, while in a/b-plant all the meritem are induced at once, which terminate the life of the plant (Bangerth, 2009). In a/b-plant, quantitative difference in flowering between pecie refer to the time of flowering during a given eaon (Zeevart, 1976). In mature tree, reproductive competence varie between meritem, both apical and lateral, o that when they are expoed to favourable environmental condition only competent meritem perceive flower inductive ignal and differentiate into inflorecence and flower (Walton et al., 1997; Battey and Tooke, 2002; Martin-Trillo and Martinez-Zapater, 2002; Bangerth, 2009). Although there are difference between a/b plant and perennial plant, the genetic of flower induction and floral organ formation eem to be imilar among them (Tan and Swain, 2007). Bangerth (2009) preented two comparative model involved in floral induction of a/b and perennial plant: the qualitative molecular-genetic model and the quantitative long-ditance ignal model (Table 1). Both level of regulation are at work in adult tree, and the expreion of a number of Arabidopi floral integrator, flowering time and flowering identity gene, and ortholog of them, alo occur in apple, citru, grape and other perennial tree and perennial herbaceou plant (Kotoda et al., 2000; Pilliteri et al., 2004b; Sreekantan et al., 2004; Böhleniu et al., 2006; Lifchitz et al., 2006; Carmona et al., 2007; Nihikawa et al., 2007; Muñoz-Fambuena et al., 2011). 26

29 Introduction However, thi doe not necearily imply that thee gene have the ame or imilar function in tree and in Arabidopi. Table 1. Comparative cheme of two model involved in floral induction of annual/biennial- (left) and perennial (right) plant. Lited are deciive factor involved in the floral induction proce. Obviou i the number of factor involved in FI in p-plant compared to a/b-plant (from Bangerth, 2009). Qualitative molecular-genetic model Quantitative long-ditance ignal model Floral inducing cue Floral promoting cue Photoperiod: low temperature (vernalization): Exogenou factor: low temperature (ubtropical tree) gibberellin; unknown autonomou factor timulate flower induction (+); high light intenity (temperature fruit) (+); water hortage (ome tropical tree) (+); high photoynthei (ufficient carbohydrate) (+); production and tranport of long-ditance ignal to acceible/non-acceible meritem Florigen (FT protein): production and tranport from leave or other tiue to the Shoot Apical Meritem (SAM) Endogenou factor: high number of fruit/eed (biennial bearing) (-); trong vegetative growth (high GA concentration) (-); type of roottock (trong/weak) (-/+); localization of bud along the hoot (acceible/non-acceible to floral promoter) (+/-) Expreion of flowering time/identity gene Horticultural factor: optimal nutrition, particularly (morphogenetic tranition) with nitrogen (+); partial removal of young fruit (thinning) (+); girdling, coring (+); hoot bending (+); hoot pruning (winter/ummer) (-/+); root pruning (+) Floral morphogenei Hormone a long-ditance ignal: gibberellin (-); indoleacetic acid (-); cytokinin (+); ethylene (+?) (+) promote floral induction; (-) inhibit floral induction. 2. Molecular bai of flowering Arabidopi thaliana ha proven to be an ideal organim to tudy plant development although it ha no commercial value a it i conidered a weed. During the pat two decade, many flowering-related gene have been identified in Arabidopi thaliana. Stimulator and repreor of flowering antagonize in metabolic pathway activating or inhibiting, repectively, the expreion of gene that caue floral tranition (Figure 1) (for more information ee review by Blázquez et al., 2006; Wilkie et al., 2008; Amaino and Michael, 2010; André and Coupland, 2012; Pin and Nilon, 2012; Blümel et al., 2015). Thi complex interaction of multiple pathway enure the tranition of a plant from the vegetative into the generative phae during favourable 27

30 Introduction environmental condition. Genetic analyi of Arabidopi flowering time mutant reulted in four major pathway controlling the time of floral tranition: photoperiod, vernalization, autonomou and GA pathway. Wherea the photoperiod and the vernalization pathway mediate the repone to environmental factor (light and temperature, repectively), the autonomou and GA pathway are endogenou (Martinez-Zapater et al., 1994; Parcy, 2005; Wilkie et al., 2008). Of thoe, the two main pathway promoting the expreion of floral integrator gene in Arabidopi are photoperiod and GA. Figure 1. Flowering time gene network with known genetic and epigenetic regulator in Arabidopi thaliana. Arrow indicate a promoting. T-end indicate an inhibiting genetic interaction. Round dot at both end mark an interaction without a known direction. Dahed line denote an indirect interaction. Gene attributed a major regulator in the different flowering time pathway are written in bold. Red font indicate the functional characterization of a gene a a flowering time regulator in cultivated pecie although not necearily with the ame function a in Arabidopi by mutant analyi, equencing and complementation analyi or heterologou expreion, RNA interference, or clear linkage with a major QTL. (From Blümel et al., 2015). 28

31 2.1 Floral integrator gene and flowering pathway Introduction Flower development occur in the meritem, but floral induction occur in the leave. The patial eparation between thee organ implie the need of a mechanim by which the floral ignal i tranferred from the leaf to the meritem. Among the network of gene implied in thi mechanim CONSTANS (CO), FLOWERING LOCUS C (FLC), FLOWERING LOCUS T (FT), SUPPRESSSOR OF OVEREXPRESSION OF CONSTANS 1 (SOC1) and FLOWERING LOCUS D (FD) play a main role in identifying and tranmitting (or obtructing, in the cae of FLC) the flowering ignal, wherea LEAFY (LFY), APETALLA1 (AP1) and TERMINAL FLOWER 1 (TFL1) regulate the tranition of the meritem. Work in Arabidopi ha provided evidence that the mall FT protein i (at leat a component of) the floral ignal (Florigen) tranferred from the leaf to the meritem, and work in other pecie ha trengthened thi concluion. FT gene expreion i activated in the leaf and the derived protein i tranported to the meritem (Kobayahi et al., 1999; Michael et al., 2005; Teper-Bamnolker and Samach, 2005; Corbeier at al., 2007; Giakounti and Coupland, 2008; Jang et al., 2009). The FT protein i a member of the CETS (CEN1; TFL1; FT) family (Pnueli et al., 2001) and i related to phophatidylethanolamine binding protein (PEBP) family (Kardailky et al., 1999; Kobayahi et al., 1999). In angioperm, PEBP homologue are divided into three group: MOTHER OF FT (MFT), FT, and TFL1 (Kobayahi et al., 1999). In Arabidopi, MFT and FT gene familie function promoting flowering, wherea the TFL1 family delay flowering. Activation of the trancription of FT-like gene in leave ha been oberved in other pecie, and appear to be a highly conerved apect of floral induction. Expreion of uch gene ha been hown in rice (Komiya et al., 2008), barley (Faure et al., 2007), poplar (Böhleniu et al., 2006; Hu et al., 2011) Ipomea nil (Japanee morning glory) (Hayama et al., 2007), tomato (Lifchitz et al., 2006), apple (Kotoda et al., 2010) and citru (Nihikawa et al., 2007; Muñoz-Fambuena et al., 2011). But FT protein i alo reported to have nonflowering-related function in ome other pecie, namely tuberization in potato, and bud et in poplar and conifer (Böhleniu et al., 2006; Rodriguez-Falcon et al., 2006; Gyllentrand et al., 2007). The tranport of the FT protein from the leaf to the hoot apex i thought to be through the phloem (Corbeier et al., 2007) (Figure 2). The evidence for FT protein movement from the vacular tiue to the meritem come from everal experiment 29

32 Introduction achieved in different pecie. Experiment of fuion protein uch a FT::GFP expreed in phloem companion cell of Arabidopi or rice were detected at the meritem, demontrating their capacity for long-ditance movement (Corbeier et al., 2007; Tamaki et al., 2007). Two mechanim are propoed to elucidate a FT protein enter the phloem ieve element from the companion cell: diffuion of the FT protein or active tranport. Giakounti and Coupland (2008) propoed that the ieve element of the phloem are made up of enucleate cell that are connected to each other at their end wall. When maturity i acquire, the ieve element connect the photoynthetically active leave to the growing part of the plant (Oparka and Santa-Cruz, 2000; Zambryki, 2008). The ieve element are cloely connected to companion cell, which are nucleate. FT-mRNA i induced in thee cell (Mathieu et al., 2007). The companion cell are connected to the ieve element by pecialized branched plamodemata (Oparka and Santa-Cruz, 2000). Plamodemata act to facilitate the entry of macromolecule uch a ugar, RNA, and protein into the ieve element. FT i maller than the ize excluion limit of thee plamodemata, and therefore could move into the ieve element by diffuion. However, Giakounti and Coupland (2008) thought that ince the protein i expreed at extremely low level in wild-type plant, a pecific mechanim that permit the movement of FT into the ieve element may be reponible for the movement of FT. Moreover, other author oberved that in C.mochata the movement of FT-like protein into the ieve element eem to be regulated by photoperiod, uggeting the involvement of a pecific mechanim that can be influenced by photoperiod rather than diffuion (Lin et al., 2007). 30

33 Introduction Figure 2. FLOWERING LOCUS T (FT) a a ytemic ignal. (a) CONSTANS (CO) protein allow for the trancription of FT and TWIN SISTER OF FT (TSF) in the phloem companion cell. FT protein i uploaded into the ieve element either by diffuion through plamodemata or by an unidentified active tranport mechanim (white circle). The imilarity between FT and TSF protein ugget they behave imilarly. (b) Long-ditance tranport of FT toward ink tiue occur in the phloem tranlocation tream. FT may aociate with other a yet unknown factor (X) during thi tep. (c) FT unloading from the phloem and tranport within the apex probably involve cell-to-cell tranport through plamodemata. The yellow area indicate a poible gradient of FT and TSF protein ditribution in the hoot apical meritem (SAM). Induction of SUPPRESSOR OF OVEREXPRESSION OF CONSTANS (SOC1) i the firt detectable event in the inflorecence meritem (IM) and it depend on the preence of FT and FLOWERING LOCUS D (FD). FT and FD interact phyically and the complex i directly involved in APETALA 1 (AP1) trancriptional activation, which occur during the formation of the firt floral bud. AP1 directly repree SOC1 in the floral meritem (FM). CC, companion cell; SE, ieve element; boxe, mrna; circle, protein; olid black arrow, experimentally confirmed interconnection; broken arrow, inferred interconnection. (From Turck et al., 2008). The photoperiod flowering pathway i mainly directed via the expreion of the gene CO. Expoure of plant to long day condition reult in a higher accumulation of the protein CO. Thu, the leave hould previouly ynthetize the RNAm of CO but it i only ynthetized during the night and late afternoon. In fact, more RNA mean more 31

34 Introduction protein, but thi protein i table only with light. For thi reaon the flowering in thi type of plant i produced in pring-ummer, when the high level of RNAm i formed, thi moment coinciding with the maximum tabilization of the protein (Blázquez et al., 2011). At when CO reache thi level, CO work a a trancription factor of the FT gene, and the related gene TWIN SISTER OF FT (TSF), which are activated in the leaf by CO (Kobayahi et al., 1999; Samach et al., 2000; Michael et al., 2005; Teper- Bamnolker and Samach, 2005; Jang et al., 2009). Fuion of the promoter of CO and FT to marker gene are expreed in the phloem, wherea the FLAVIN-BINDING, KELCH REPEAT, F-BOX 1 (FKF1) and CYCLING DOF FACTOR 1 (CDF1) gene, which encode regulator of CO, are alo expreed mainly in thi vacular tiue. Beide CO, FHA and GIGANTEA (GI) are activated in the roette leave of Arabidopi in a circadian rhythm (Solti et al., 2002) under long day condition. The vernalization pathway of FT trancription i negatively modulated by other trancription factor. The MADS box trancription factor FLC and SHORT VEGETATIVE PHASE (SVP) directly bind to FT and repre it trancription (Searle et al., 2006; Lee et al., 2007). The FLC and SVP mrna are both expreed in the SAM a direct repreor of SOC1 trancription (Lee et al., 2000; Jang et al., 2009; Liu et al., 2009). FLC block FT trancription until the plant i expoed to low temperature that repre FLC trancription, allowing for the induction of FT the following pring a the photoperiod lengthen. Cold i perceived in the hoot apical meritem by activation of the cold-repone gene VERNALISATION 1 (VRN1), VRN2 and VRN3 and by change in DNA methylation (Finnegan et al., 1998). Thee factor uppre FLC expreion; thu, vernalization promote flowering by uppreing FLC. The reduction in FLC expreion in low temperature involve expreion of an antiene RNA (Swiezewki et al., 2009) and machinery that modifie the tail of hitone 3 at the FLC gene, particularly trimethylation of lyine 27 (Finnegan and Denni, 2007). Thee procee reult in the diminihed FLC mrna expreion in the cold and table repreion of FLC when plant are returned to normal growth temperature. SVP expreion i reduced during the floral tranition a SOC1 mrna tart to rie. So that oon after SOC1 i trongly expreed in the meritem, SVP mrna i not detectable (Albani and Coupland, 2010). The mechanim by which SVP i repreed in the meritem i unclear. The GA pathway of flowering induction in Arabidopi ha little influence under LD condition. But in SD, in the abence of the photoperiod flowering pathway, the GA pathway aume a major role and become obligatory (Mutaa-Göttgen and 32

35 Introduction Hedden, 2009). GA promote flowering in Arabidopi through the activation of gene encoding the floral integrator SOC1, LFY, and FT in the inflorecence and floral meritem, and in leave, repectively. In other LD pecie (i.e. Lolium temelentum) there i trong evidence that GA act a a mobile ignal to tranmit the photoperiodic flowering timulu (FT). While the role of GA in flowering ha become etablihed for a limited number of pecie, GA i clearly not a univeral flowering timulu. In fact, it inhibit flowering in woody pecie. Thi paradoxical obervation (GA act poitively in the witch to reproductive development in annual plant but negatively in woody pecie) wa recently tudied by Yamaguchi et al (2014). Thee author found dual oppoite role of GA in Arabidopi flowering promotion: GA promote termination of vegetative development but it inhibit flower formation. To overcome thi effect, the trancription factor LFY induce expreion of a GA catabolim gene, and conequently, increae LFY activity cauing reduced GA level and flower formation. 2.2 FT-FD in the meritem and floral initiation gene In the meritem, FT i believed to activate trancription of pecific target gene by interacting with the bzip trancription factor FD (Abe et al., 2005; Wigge et al., 2005). Both FT and FD are required for the upregulation of SOC1, but it i till not known whether thi i a direct or indirect effect (Albani and Coupland, 2010). SOC1 i an activator of floral initiation in the hoot apical meritem (SAM) and i upregulated oon after the hift from SD to LD (Giakounti and Coupland, 2008) or by GA (Moon et al., 2003). Searle et al. (2006) howed that SOC1 i expreed in leave and hoot meritem, but it promote flowering more powerfully when it i ectopically expreed in the meritem compared to the leaf. SOC1 trancription i repreed by FLC (Hepworth et al., 2002; Searle et al., 2006; Sheldon et al., 2006). Therefore, SOC1 act a an integrator of the photoperiod, GA, and vernalization pathway (Albani and Coupland, 2010). See Figure 3. 33

36 Introduction Figure 3. Spatial pattern of expreion and molecular cacade aociated with reprogramming of the floral meritem upon floral induction. (a) Vegetative meritem (no floral induction). FD mrna (purple) i preent in low amount throughout the meritem while TFL1 mrna (blue) i expreed at low abundance in the center of the meritem. (b) Tranition meritem (floral induction ha occurred but no floral primordia are viible). Upon arrival from the phloem to the apex, the FT protein (green) interact with the FD protein (red). Thi reult in the direct upregulation of SOC1 mrna (yellow), one of the earliet known molecular marker of floral induction in the meritem. The FT FD protein complex alo upregulate, with a mall delay compared to SOC1, AP1 mrna expreion in the flank of the meritem, in a region which will develop into a floral primordium (dahed red circle). (c) Floral committed meritem, early tage (floral commitment ha occurred but no floral primordia are viible). SOC1, together with a gene called AGL24 encoding another MAD box protein, participate in a poitive feedback loop which eventually upregulate LFY expreion in the flank of the IM (Lee et al., 2008) with a mall delay compared to AP1. (d) Floral committed meritem (floral primordia are viible) TFL1 mrna (blue) i now trongly expreed in the center of the IM while the protein (green) tranport intercellularly (blue arrow) throughout the whole IM and repree LFY and AP1 trancription in the IM. At the ame time both AP1 and LFY protein (orange) enure TFL1 doe not accumulate in the floral primordia, eparating thee two feature of the apical meritem. LFY and AP1 maintain their expreion in the developing floral primordium through reciprocal upregulation. IM: inflorecence meritem. (From Giakounti and Coupland, 2008). Among the floral identity gene LFY and AP1 have the mot ignificant influence. Both gene are expreed trongly in young floral primordia. Overexpreion 34

37 Introduction of LFY or AP1 are ufficient to confer floral identity to the SAM (Mandel and Yanofky, 1995; Weigel and Nilon, 1995; Jack, 2004). Strong LFY expreion occur rather late in the meritem during floral induction and i activated by everal clae of trancription factor. Both SOC1 and SPL3 have been hown to bind directly to LFY in vivo uing chromatin immunoprecipitation (Lee et al., 2008; Yamaguchi et al., 2009). Alo, at leat in vitro, the AtGAMYB AtMYB33 trancription factor from the GA pathway bind directly to LFY. Activation of LFY by thee trancription factor can explain why LFY act downtream from everal flowering pathway (Blázquez and Weigel, 2000). LFY directly activate AP1 expreion (Wagner et al., 1999), and AP1 can alo provide feedback to activate LFY (Ferrandiz et al., 2000). In thi way LFY and AP1 enhance each other expreion reinforcing floral identity and initiating floral organ pecification and development. 2.3 Epigenetic control of flowering Trancriptional control of a gene involved in flower induction i due, to a coniderable extent, to hitone modification that are able to keep particular gene completely ilent even in the preence of promoting ignal. Floral induction in Arabidopi ha been tudied in ome detail with regard to hitone modification (Sung and Amaino, 2004). Nelien et al. (2007) how that ATP-dependent modification of core hitone in the nuclei of cell have a great influence on the acceibility of the DNA of particular gene to trancription factor or the whole trancriptional complex. In brief, hitone conit of four type of contitutively preent protein onto which the DNA i tightly-packed (heterochromatin). In thi configuration, DNA i trancriptionally inactive and, therefore, non-acceible and not trancribed. However, ignal reaching the cell urface can rapidly activate phyiological mechanim in the cell that covalently modify core hitone tail, making DNA acceible (euchromatin) for the binding of trancriptional factor. Thee covalent modification of core hitone tail affect almot all apect of plant development, including flower induction (Nelien et al., 2007). The phyiological mechanim needed to accomplih thee modification are methylation, acetylation, phophorylation, and ubiquitination, together with their repective enzyme, which modify particular amino acid in the tail of the hitone, thu cauing conformational change reulting in a looening or tightening of DNA-hitone binding. Thee methylation or acetylation pattern among 35

38 Introduction the hitone protein are partially inherited and, due to change in the former terminology, conidered a epigenetic and a an additional order of regulation not involving the DNA code. While methylation of a particular hitone tail often repree gene trancription, acetylation uually ha the oppoite effect and promote DNA trancription. The acetylation and de-acetylation of hitone are readily reverible enzymatic procee and enable increaed trancription in addition to complete ilencing of particular gene (Nelien et al., 2007). Among the example that have been invetigated more intenively for floral induction i the cool treatment induced of Arabidopi and other plant (vernalization) caued by an epigenetic ilencing of the gene FLC. Thi ilencing i due to the hitone methylation of FLC, which reduce it repreing effect on downtream gene, leading to earlier flowering. However, more detailed tudie have hown that thi proce i more complicated inofar a it i accompanied by the deacetylation of the hitone H3, whoe table epigenetic repreion i executed by the methylation of the hitone HP1 (Batow et al., 2004). Thi hitone modifie another hitone, H3K9, which i involved in heterochromatin formation and, thu, in the downregulation of FLC activity (Mylne et al., 2006). Vernalization coniderably increae the activity of a number of gene (Bäurle and Dean, 2006), ome of which are known to code for trancription factor involved in hitone modification and ultimately in floral induction. Aide from vernalization, light-dependent floral induction ha been examined with regard to hitone modification and chromatin modeling/re-modeling. Studie into light duration effect in Arabidopi demontrated that light (imilar to vernalization), gibberellin, and the autonomou pathway ue the FLC gene to regulate floral induction (Takada and Goto, 2003; Bo et al., 2004; Deal et al., 2007). In thee light procee hitone modification and cromatin re-modeling were again involved. However, in contrat to vernalization, thee hitone modification were different for all four of the previouly mentioned floral induction pathway. An alternative regulatory poibility for the ilencing of gene relevant to floral induction i the pot-trancriptional action of microrna (mirna), which alo eem to interfere with the action of gene like FLC (Schmid et al., 2003). 36

39 3. Flowering in fruit tree pecie Introduction Fruit tree pecie how two main characteritic in relation to flowering: 1) a long juvenile phae (juvenility) of tree propagated by eed, during which all the meritem in the tree cannot be induced to flower and remain vegetative; and 2) flowering i a quantitative proce during the mature tage, with a ignificant proportion of the aboveground meritem remaining vegetative or dormant. Thi mechanim guarantee a long life pan for thee plant. Very little i known about how a tree achieve thi trait. The long juvenile phae i a major problem in breeding program but not in fruit production, becaue fruit tree are propagated by grafting and thu are in their mature tage from planting. During the mature tage, fruit tree flowering eem to be rarely aociated with photoperiod and vernalization (Wilkie et al., 2008). In many ubtropical and tropical tree pecie, uch a mango, lychee, macadamia, avocado, and orange, flowering i induced by cool temperature. The temperature required for flowering in uch tree i between 10ºC to 15ºC and are higher than thoe required for vernalization in herbaceou pecie. When cultivated under tropical condition, water hortage alo induce flowering. Unlike ubtropical and tropical tree, flowering i regulated autonomouly in many temperate deciduou tree pecie, uch a apple. Kobel (1954) reported that pecific factor inducing flower bud development are concentrated in the appropriate leave attached to the bud. According to Bünning (1952), the developmental tage of the leave i eential for the formation of flower organ. The role of leave in the formation of flower can be explained in three way: 1) the leaf, a an organ of aimilation, provide carbohydrate needed in flower induction, 2) the leaf i a major ite of hormone ynthei and 3) the leaf i the receptor of environmental ignal. In the pot-genomic era when gene have been identified a controling floral development, it became evident that important floral-inducing gene are active in leave. It hould alo be mentioned that defoliating leave before flower induction can greatly inhibit flower formation in apple (Li et al., 1995) and citru (Muñoz-Fambuena et al., 2012b). The tage of flower induction wa tudied by removing leave at different time before and after full bloom in everal pecie (Fatta del Boco, 1961; Li et al., 1995). When defoliation ha no inhibitory effect on flowering, the implication i that meritem have paed the flower induction period. 37

40 Introduction But the mot outtanding point, and common to mot fruit tree pecie, i the inhibitory effect of fruit on flower induction, even with the preence of exogenou flowering ignal. In mot pecie, fruit development coincide with the time of flower induction. Chan and Cain (1967) demontrated that the preence of eed i crucial in fruit inhibition of apple flower induction. But in parthenocarpic citru pecie (Martínez-Fuente et al., 2010; Muñoz-Fambuena et al., 2011), the fruit alo inhibit flower induction, o the mechanim mut not be excluively attributed to eed. The fruit inhibit flowering from the time it complete it growth, not before, and tart ripening. The proce ha been oberved in orange (Martínez-Fuente et al., 2010), mandarin (Muñoz-Fambuena et al., 2011), loquat (Reig et al., 2014), olive (Dag et al., 2010) and avocado (Ziv et al., 2014). The inhibition of flowering induce alternate (biennial) bearing, a critical horticultural problem characterized by large yield of mall-ized fruit in the on-year, and low yield of overized fruit in the off-year (Handchack and Schmidt, 1990). Thi i caued by the advere relationhip between fruit development and flower bud induction or differentiation (Monelie and Goldchmidt, 1982). Genotype i probably the major caue for alternative bearing. Within a given pecie, there are regular bearing cultivar and cultivar whoe trend i to be very biennial bearing (Fig. 4). Figure 4. Schematic repreentation of the vegetative and reproductive development of biennial bearing woody pecie 38

41 Introduction Beide, gravimorphim play an eential role in fruit tree. Vegetative growth i affected negatively by gravity (Smith and Wareing, 1964). Shoot growth and flower bud formation i obviouly affected by hoot orientation, namely placing hoot in a horizontal poition increae flower-bud formation and reduce growth (Longman et al., 1965). Tromp (1982) found that hoot growth wa reduced and flowering wa increaed by hoot bending and by application of a chemical growth inhibitor. In all cae a promotion of hoot growth increae apical dominance and vice vera, an inhibition of hoot growth upport lateral flower-bud formation. The poition of an axillary meritem on the plant, it age/ize, and time of outgrowth might alo determine it developmental fate. Vegetative development in apple and Arabi alpine i maintained by axillary meritem cloe to the SAM (Foter et al., 2003; Wang et al., 2009). Additionally, in kiwifruit the ize of the meritem prior to bud break in the pring can determine the fate of the econd-order axillary meritem (Walton et al., 1997). Flowering probability alo depend on both hoot and tree age. In ome pome fruit pecie, newly developed long hoot have a lower flowering ability than the older hoot. There i a decline correponding to a growth reduction and an increae in the probability of flowering from the center of the tree toward the periphery. Thi centrifugal gradient ha been found in certain tree pecie (Cote et al., 1992; Sabatier and Barthelemy, 2001). However, other pecie uch a citru, the younger hoot prout and flower with a higher intenity than the older one (Agutí, 2003). According to Bangerth (2009), two level of flowering regulation act at once during flower induction in fruit tree: the qualitative molecular-genetic regulation and the quantitative long ditance ignal regulation. 3.1 Qualitative molecular-genetic regulation of flower induction in tree In fruit tree pecie, the expreion of flowering gene orthologou of Arabidopi have been tudied in relation to the pecific ignal which inhibit or promote flowering. A number of thee have been identified in non-related pecie, which ugget a certain degree of evolutionary conervation. For intance, the expreion of the FT gene ha been identified in apple (Kotada et al., 2010), mango (Nakagawa et al., 2012), avocado (Ziv et al., 2014), orange (Muñoz-Fambuena et al., 2012a), mandarin (Nihikawa et al., 2007), poplar (Böhleniu et al., 2006), etc. In thi 39

42 Introduction Ph.D thei, pecie from the genu Citru were the model tree elected to tudy the inhibitory effect of fruit on flowering. Therefore, thi ection of the reviewed literature focue on tudie conducted with Citru p. In Citru, the flowering-related gene characterized in Arabidopi eem to be conerved (Dornela et al., 2007). For thi reaon, on the bai of data from Arabidopi, citru homologue of FT, LFY, AP1, TFL1 and SOC1 have been identified and characterized (Kobayahi et al., 1999; Pilliteri et al., 2004a, 2004b; Endo et al., 2005; Nihikawa et al., 2007; Tan and Swain, 2007). In Citru clementina, there are three loci encoding FT-like protein (Samach, 2013). Overexpreion of CiFT, CLFY, CAP1, or CSL induce an early flowering phenotype in Arabidopi (Kobayahi et al., 1999; Pilliteri et al., 2004b; Tan and Swain, 2007), wherea Arabidopi plant ectopically expreing CTFL how late-flowering phenotype (Pilliteri et al., 2004a). In addition, reearch by tranformation with flowering-related gene ha highlighted the long juvenile period in citru (Peña et al., 2001; Endo et al., 2005). In thee experiment, contitutive expreion of Arabidopi AP1 or LFY caued early flowering and fruiting in many of the trangenic citrange [Citru ineni (L.) Ob. x Ponciru trifoliate (L.) Raf.], which initiated flowering in pring, month after their tranfer to the greenhoue (Peña et al., 2001). The contitutive expreion of CiFT tarted to flower a early a 12 week after tranfer to a greenhoue, wherea wild-type plant uually have a long juvenile period of everal year (Endo et al., 2005). The fact that the flowering phenotype reulting from ectopic expreion of CiFT i an early flowering, reearcher have ued thi method for the rapid evaluation of trangenic citru flower and fruit (Endo et al., 2009). Nihikawa (2013) revied the flowering in Citru. The genu Citru i different from it cloe relative, Ponciru and Fortunella. Both Citru and Fortunella are evergreen, wherea Ponciru i deciduou. In atuma mandarin, floral induction occur during fall and winter, floral organ are firt obervable from January through a microcope, and the three different type of hoot in May (Iwaaki, 1959; Inoue, 1990). In repect to the Kumquat (genu Fortunella), the bloom i produced during ummer and fall, and the floral organ development i oberved jut before flowering (Abbott, 1935). In trifoliate orange (genu Ponciru), evocation happen during early ummer (Reuther et al., 1967). The development of flower bud top during fall and winter, then begin again and the tree different type of hoot are formed in pring. Change in CiFT expreion are aociated with the eaonal periodicity of flowering (Nihikawa et 40

43 Introduction al., 2007, 2009, 2011; Muñoz-Fambuena et al., 2011, 2012b). In atuma mandarin (Nihikawa et al., 2009) and Moncada mandarin (Muñoz-Fambuena et al., 2011), CiFT expreion increae during fall and winter, in which floral induction occur. Thi increae reult from the expreion of CiFT2 which i alo induced during fall and winter in murcott mandarin (Citru reticulate Blanco) (Shalom et al., 2012). In kumquat and trifoliate orange, CiFT expreion increae during early ummer, in which period differentiation of flower organ ha been initiated, or jut before evocation. The mrna level of CLFY remain low during floral induction (winter and fall) in Moncada mandarin (Muñoz-Fambuena et al., 2011) and atuma mandarin (Nihikawa, 2013); ubequently, level increae in February and April, repectively, and during thi period floral organ develop rapidly. In kumquat and trifoliate orange, mrna level of CLFY increae during ummer when floral induction and evocation i produced (Nihikawa, 2013). Expreion level of CAP1 are moderately high during pring and ummer and low during winter (Nihikawa, 2013). In all of thee genera, CTFL relative expreion wa detected at high level during late pring or early ummer, then decreaed to undetectable level. Only CiFT howed a cloe correlation with floral induction in all genera (Nihikawa, 2013). In Citru, the cool temperature i a deciive factor in floral induction. In ome experiment related in atuma mandarin tree, the plant were expoed to cool temperature (15 C) and CiFT expreion increaed imultaneouly with floral induction (Nihikawa et al., 2007). But when the tree were grown at high temperature (25 C), vegetative growth wa maintained and flowering wa not induced (Nihikawa, 2013). At the ame time, CiFT expreion remain at a low level. With thee experiment Nihikawa (2013) howed that CiFT expreion i regulated by temperature. Moreover the ame author howed that in other flowering-related gene, uch a CLFY, CAP1, and CTFL, the change in mrna level howed no aociation with the period of floral induction in tree placed in a growth room at 15 C (Nihikawa, 2013). In fact, CLFY and CAP1 relative expreion level increaed during the period of evocation of flower bud in atuma mandarin and weet orange tree under prolonged 15 C treatment or a ubequent change to high temperature (Nihikawa et al., 2007; Pillitteri et al., 2004b). Only CiFT expreion howed an increae in repone to cool temperature in adult tree and not in juvenile plant (Nihikawa et al., 2007; Nihikawa, 2013). On the other hand, CTFL relative expreion wa detected at higher level in young rather than adult plant in tree of atuma mandarin and weet orange 41

44 Introduction expoed to cool temperature (Nihikawa et al., 2007; Pillitteri et al., 2004a). Nihikawa (2013) hypotheized that the uppreion of flowering in young plant might correlate with low CiFT expreion and high CTFL expreion under cool-temperature condition. In the lat year, change in gene expreion in repone to fruit bearing have been reported in Murcott (Shalom et al., 2012), Moncada (Muñoz-Fambuena et al., 2011; 2012b), and atuma mandarin (Nihikawa et al., 2012). In thee genotype, the level of CiFT expreion appear to be high in light-loaded or without fruit OFF tree and low in fully-loaded ON tree during the floral inductive period. In atuma mandarin, it wa demontrated that the total CiFT expreion during fall how a clear, trong correlation with fruit weight per leaf area (Nihikawa et al., 2012). In addition, the CiFT gene expreion wa aociated with the timing of fruit harvet (Nihikawa, 2013). Another experiment done by Nihikawa (2013) conited in analyzing the CiFT relative expreion in different branch of the ame tree, and CiFT expreion wa high in the branch from which flower were removed in May and low in the branch from which fruit were harveted in November. Thi indicated that CiFT expreion i uppreed by a long fruit-bearing period, and the uppreion of CiFT expreion by fruit bearing i limited to the vicinity to the fruit-bearing portion of the tree. Moreover, the flower number in the following pring howed a poitive and high correlation with CiFT expreion level during fall and winter in atuma mandarin (Nihikawa et al., 2012). Taken together, it i conidered that exceive fruit amount, a longer period of fruit bearing, and the vicinity to the fruit bearing portion of the tree reduce CiFT expreion in the tem of vegetative hoot during fall and winter, and correpond with flower number in the following pring (Muñoz-Fambuena et al., 2011, 2012b; Nihikawa, 2013). Given that the flower number and CiFT expreion how a cloe correlation (Muñoz-Fambuena et al., 2011; 2012a; 2012b; Nihikawa et al., 2012) and CiFT ha a function in the promotion of flowering (Endo et al., 2005; Kobayahi et al., 1999), it i believed that fruit bearing uppree flower number in the following pring via uppreion of CiFT expreion (Muñoz-Fambuena et al., 2011, 2012a, 2012b). Defoliation of the tree inhibit the flower formation in the following pring (Muñoz-Fambuena et al., 2012b). Moreover, experiment for atuma mandarin that conited in removing or decreaing the number of leave uppreed CiFT expreion in tree grown under a floral-inductive condition (15 C) (Nihikawa et al., 2013) and at the ame time, floral induction wa alo uppreed in tree from which all leave had 42

45 Introduction been previouly removed (Muñoz-Fambuena et al., 2012b; Nihiwaka, 2013) or which carried fewer leave (Nikikawa, 2013). Thu, defoliation uppree both CiFT expreion and floral induction. Gibberellin, an inhibitor of flowering in Citru, reduce CiFT expreion in bud of Orri mandarin (C. reticulata Blanco C. temple Hort. ex Y. Tanaka) (Goldberg-Moeller et al., 2013) and leave of weet orange (Muñoz-Fambuena et al., 2012a). In weet orange, paclobutrazol (PBZ), a GA bioynthei inhibitor, increae flowering by improving CiFT relative expreion (Muñoz-Fambuena et al., 2012a), which indicate that endogenou GA or application with GA3 can inhibit flowering via uppreion of CiFT expreion. With repect to FLC in citru, a direct inhibitor of flowering, there i little information. FLC-like analyzed by Muñoz-Fambuena et al., (2011;2012a; 2012b) did not eem to be relationhip with the inhibition of flowering induction in leave and bud due to the expreion initiated after of thee period in leave (Muñoz-Fambuena et al., 2011). On the other hand, other author analyzed other FLC-like, and thi howed no difference of expreion between treated and untreated bud with GA3 on OFF tree in the floral induction period (Goldberg-Moeller et al., 2013). 3.2 Quantitative long ditance ignal regulation of flower induction The main function of the econd level of regulation in perennial tree need to be a phyiological/molecular mechanim that protect/ilence particular gene/meritem to be acceible to a floral promoter (Bangerth, 2009). Bangerth (2009) hypotheized that the regulation of flower induction in tree could be regulated far uptream to floral integrator gene, and that they are more imilar to general repreor gene like the FLC gene or a even epigenetic in nature, which could control the expreion of gene and would therefore be particularly appropriate to ilence floral gene. Becaue the epigenetic mechanim i able to act over longer ditance of time, it action could alo explain phenomena like the flower induction inhibition by the fruit, although the fruit had been harveted long ago (Lavee, 1988). Otherwie, the degree of expreion of gene of the firt level may directly or indirectly be altered by the econd level of regulation in a quantitative way, uch a DNA methylation (Zemach and Grafi, 2007) or pot-trancriptional modification by particular non-coding RNA pecie, uch a microrna (Schmid et al., 2003), which, in ome cae, alo have epigenetic characteritic. 43

46 Introduction Although there may be more than jut one mechanim that could prevent part of the meritem of the tree from being induced to flower, the non-acceibility of meritem to floral-promoting ignal, due to epigenetic control, eem to be an attractive hyphotei to at leat partially explain perenniality (Bangerth, 2009). Example with Arabidopi demontrate that thi kind of regulation may in fact be one explanation for difference between annual/biennial and perennial plant a well a for effect like hoot bending, girdling, or fruit inhibition on floral induction in tree. Plant hormone are mot prominent amongt long-ditance-ignal and almot all of them are already extenively ued by the plant to interfere with floral induction and other phyiological procee uch a vegetative growth, dominance phenomena, aimilate partitioning, fruit et and growth, or tre ituation (Hoad et al., 1993; Bangerth, 2000). Endogenou a well a applied gibberellin have been hown to promote floral induction in ome annual/biennial plant and, in addition, to indicate ditinct relationhip with the level one molecular genetic pathway (Zeevaart, 1976; Blázquez and Weigel, 2000; King and Evan, 2005). Comparion of endogenou GA level between roe plant exhibiting eaonal and perpetual flowering habit indicate that GA might play a role in the duration of flowering eaon by regulating the return to vegetative development. In eaonal flowering plant, GA1 and GA4 level increae in the hoot apice jut after flower initiation, wherea in perpetual flowering plant GA level are low throughout the year (Robert et al., 1999). King and Evan (2005) demontrated in detailed experiment with their model plant Lolium temulentum that GA5 i a florigenic ignal, poibly intermediate between long-day perception and FT expreion (King and Evan, 2005). In ome pecie of gymnoperm tree, applied low polar GA are able to induce precociou and profilic floral induction (Phari and King, 1985). However, in mot angioperm tree, GA have the oppoite effect, that i, they coniderably inhibit floral induction applied exogenouly (Guardiola et al., 1982; Tromp, 1982; Oliveira and Browning, 1993; Prang et al., 1997; Mutaa-Göttgen and Hedden, 2009), a evidenced in the aay appliying GA3 (Ayalon and Monelie, 1960). When dealing with endogenou GA, however, the ituation become more complex (Oliveira and Browing, 1993; Bernier and Périlleux, 2005). The problem arie when conidering GA a directed long-ditance ignal to inhibit floral induction in tree. A better undertanding of GA a floral inhibiting long-ditance-ignal began with the finding of Chan and Cain (1967) who demontrated that eeded but not eedle apple fruit inhibit floral induction for the next eaon. Thi finding wa 44

47 Introduction undertandable taking into account the high concentration of GA in the eed of apple and other fruit (Hedden et al., 1993). The ame reult wa obtained in pear (Grigg et al., 1970) and citru (Monelie and Goldchmidt, 1982). In experiment with apple, inhibition of floral induction by fruit/eed wa generally mediated mainly to nearby meritem. The effect of GA on floral induction i confued by a number of tree-pecific interfering factor, like vegetative growth in apple (Neilen and Denni, 1999) and citru (Krawjewki and Rabe, 1995), the number of leave in pear (Huet, 1972). Exogenou treatment uually reduced floral induction in a number of tree pecie (Goldchmidt and Monelie, 1970). However, the reult were often variable and the concentration needed to obtain ignificant effect were generally high and not of normal phyiological concentration. In addition, there are coniderable tructural difference in GA which greatly affect their inhibiting action (Tromp, 1982; Oliveira and Browing, 1993). The mot ignificant effect in GAefficiency in thi repect i the way GA are applied to the plant. Goldchmidt and Monelie (1970) treated ingle bud of orange hoot with GA3 and obtained a 75% reduction in the average number of flower/hoot, with higher concentration cauing a complete inhibition of floral induction. Bertling and Bangerth (1995) injected either GA3 or GA4/7 into the trunk of numerou fruit tree pecie and cv. and obtained a coniderable reduction of the floral induction. Thi method confirmed that when praying GA an obviouly high proportion of the hormone i conjugated and/or metabolized or doe not reach the intended target, which could explain the low or variable efficiency of praying. When application are done in a more appropriate phyiological way, GA are able to keep many above-ground bud in a vegetative nonacceible tate even if applied at phyiological concentration. The way in which thi non-acceibility i maintained i preently not known but may involve chromatin a well a rejuvenation effect. A directed tranport of GA from a defined ource uch a fruit/eed or hoot tip to a pecific target ha not yet been clearly demontrated. In fact, it would be difficult to aume tranport from a fruit, a trong ink, to the apical bud of an apple boure hoot, which i a weak ink. Nonethele, Prang et al. (1997) reported an export of GA out of apple fruit a well a hoot tip with a peak occurring at the time preumed floral induction doe occur, but there wa little correlation between the regular and the alternate-bearing cv. Due to the technique applied, the detination of thee GA-ignal could not be determined. Uing the ame cv. in the ame year, 45

48 Introduction Stephan et al. (2001) ued LC-ESI-MS and internal tandard to invetigate the export and metabolim of GA out of thee fruit. In fruit exudate of the regular bearing cv.,they found ix different GA with GA4 prevailing over GA3. In the biennial bearing cv. Eltar, however, GA3 wa the predominant ignal. When radioactive labelled GA were injected into the core of the fruit and their export and metabolim in the pedicel and pur followed, the ame author found a high proportion of GA-metabolite or glycoylated conjugated, both of which are biologically inactive (Stephan et al., 2001). No intact GA-molecule could be found in the apical bud of the hoot which i uually induced to flower. Seed are not the only ource of GA relevant for floral induction in tree. Shoot tip are alo rich in bioactive GA, particularly GA1, and thee may be involved in the floral inhibition of lateral bud of long hoot, for example, during trong vegetative growth (Forhey and Elving, 1989). Bo et al., (2004) howed that a mutant of Viti vinifera, deficient in it repone to GA, only produced inflorecence and no tendril, wherea the wild type, which howed full GA-enitivity produced only tendril. In citru, there are two key moment in which the exogenou application of GA ha the maximum efficiency on the inhibition: during tranlocation of the flowering ignal of leave to bud (the end of November to early December) and the beginning of the morphologic differentiation of flower (Monelie and Halevy, 1964; Guardiola et al., 1982). A the effect of the exogenou application of GA i imilar to that produced by the fruit, Goldchmidt and Monelie (1970) uggeted that the mechanim by the fruit inhibit the flowering i through the ynthei of GA. In fact, in the Satuma mandarin, the GA concentration (GA1 / 3) i higher in the leave of ON hoot in October (Kohita et al., 1999). But it concentration decreaed to a imilar level in both type of hoot in December, coinciding with the flowering induction in thi genotype (December-January). Therefore, their relationhip with the inhibition of floral induction cannot be determined. Bangerth (2009) uggeted that polar IAA tranport, a long-ditance-ignal, could act a a econdary meenger to GA. IAA i the only plant hormone with a trictly polar, highly-regulated tranport pathway (Muday and DeLong, 2001), and IAA ignal are therefore independent of ink or tranpiration driven tranport. Further, by ome kind of auxin tranport auto-inhibition, thi hormone tranport i able to affect a particular organ, for example, a bud/meritem, without entering it (Bangerth, 2000). The mot prominent example of thi i apical dominance. There are indication which eem 46

49 Introduction to upport a role for IAA tranport in floral induction and it function a econd meenger. The application of the gibberellin GA3, GA4 and GA7 to apple tree coniderably timulated IAA-export of fruit a well a of hoot tip (Calleja and Bangerth, 1997), which would be in line with a econd meenger role for IAA. A particular cla of IAA-tranport inhibitor (morphactin and TIBA) timulate floral induction in a number of fruit tree (Luckwill, 1969; Ito et al., 2001; Bangerth, 2009) and a number of horticultural procedure are ued to increae floral induction, like bending, girdling, hoot tip removal or damage, thee coniderably reducing polar IAA tranport (Blaikie et al., 2004). Latly, application of an auxin to a decapitated hoot tip i alo reported to inhibit floral induction (Tamim, 1996). In contrat to GA and to polar IAA tranport, it ha been hown that the application of cytokinin (CK) promote floral induction in monocarpic and polycarpic plant (Ramirez and Hoad, 1981; Bernier and Périlleux, 2005). Application of low concentration of CK had no effect on floral induction (Bernier and Périlleux, 2005) while trangenic CK-deficient Arabidopi plant never flowered (Werner et al., 2003), hence the eential role of CK in floral induction. At medium CK concentration, floral induction occur, but high concentration promote only vegetative development; thee poitive effect are een in Sinapi alba (Bernier and Périlleux, 2005), Viti vinifera (Srinivaan and Mullin, 1981) and Malu dometica (Bangerth, 2009). Other experiment, (Stern et al., 2003) focuing on the effect of water tre to increae the concentration of xylem CK, found a concomitant quantitative increae in floral induction. Application of other, non-ck compound, like maleic-hydrazide and triidobenzoic acid (TIBA), have alo ignificantly increaed floral induction of apple, pear, olive and mango tree (Luckwill, 1969; Ben-Tal and Lavee, 1985; Ito et al., 2001; Blaikie et al., 2004). Ito et al. (2001) reported coniderable increae in the endogenou CK concentration of lateral and terminal bud on tree treated with thee ubtance. Both maleic-hydrazide and TIBA are potential inhibitor of the IAA polar tranport and/or metabolim, the CK increae may be the reult of a reduced IAA concentration either in the bud themelve or in the root (Bangerth, 2000) o that CK may be the caue for the increaed floral induction rather than the IAA tranport. Another important hormone i ABA, which i tranported acropetally a baipetally (Davie, 2010). In aay with citru Shalom et al. (2014) identified a tranport of ABA from fruit to bud. The content of ABA wa higher in ON. However, the NCED3 gene, which determine ynthei of ABA, had preented a higher 47

50 Introduction expreion in OFF bud. Therefore reearcher ugget that the content of ABA in ON bud i produced in other part of the tree, particularly, the fruit (Shalom et al., 2014). In early tudie relating ABA with the inhibition of flowering, Goldchmidt and Golomb (1982) oberved a higher ABA concentration and it iomer t-aba in leave, tem and bud of ON tree in December, uggeting their relationhip with dormancy and the inhibition of flowering. On the other hand, exogenou ABA treatment to bud of Citru unhiu reduced prouting and flowering of thee bud (García-Lui et al., 1986). Neverthele, in tre condition flowering i promoted (Kohita and Takahara, 2004), and thi i coincident with an increae in ABA content in leave (Gómez-Cadena et al., 2000; Kohita and Takahara, 2004) and an increae in CiFT expreion (Chica and Albrigo, 2013). The relationhip of ABA with flowering in citru ha yet to be clarified. Floral induction in annual/biennial plant ha been linked to the induction and the releae of dormancy (Horvath et al., 2005). It i generally agreed that dormancy of both bud and cambial meritem i largely controlled by plant hormone, motly ABA, cytokinin and gibberellin (Powell, 1987; Lombard et al., 2006; Nieminen et al., 2008) and their interaction with DNA methylation/demethylation and hitone/chromatin modification. 48

51 4. Hypothei and objective Introduction 4.1 Hypothei In mature citru tree, flowering i regulated by exogenou ignal (i.e., low/high temperature, water hortage/water upply) that control flowering time, and by endogenou ignal that control flower intenity. According to Martínez-Fuente et al. (2004), three level of flowering can be etablihed depending in the promoter/inhibitor (P/I) ratio. The firt i one of high flower intenity, defined by a high P/I ratio in which P reache a high level and exogenou inhibitor cannot counteract the promoter. In contrat, a low flowering level i due to a low P/I ratio in which the inhibitor prevail over the promoter and the application of the latter cannot counteract the inhibitor. Only at a nearly balanced P/I ratio can flowering be exogenouly inhibited by applying inhibitor (i.e. GA3), or promoted, by applying promoter (PBZ) (Fig. 5). The ratio of endogenou promoter to inhibitor i conidered a reponible for flowering; however, there i evidence uggeting that the inhibitor and their metabolim are the ole factor controlling flowering in Citru. According to Agutí (1980), all the meritem have the information and ability to flower unle a negative factor hamper the proce. The main endogenou factor controlling flower intenity i the fruit. Bangerth (2009) propoed the exitence of long-ditance ignal (hormonal in nature) needed to ilent the trancription of gene involved in the floral proce through epigenetic mechanim. Figure 5. Schematic diagram of the ratio of endogenou flowering inhibitor and promoter and exogenou flowering inhibitor and promoter and their effect on controlling flowering in Citru (From Martínez-Fuente et al., 2004). 49

52 Introduction In thi PhD. thei, the following hypothei wa teted: Fruit inhibit flowering when ripening begin by exporting hormone that induce epigenetic upregulation of flowering inhibitor gene in the leave, and interfering in flower bud differentiation. 4.2 Objective Therefore, the following objective were etablihed: 1. To correlate the endogenou content of GA and ABA in the fruit, leave and bud, and their bioynthei, with the expreion of flower induction and flower differentiation gene. 2. To identify the effect of bud iolation from the fruit (girdling and in vitro culture) on prouting and flowering. 3. To determine the influence of the fruit in the proteome of bud and leave. 4. To characterize the effect of the fruit in DNA methylation of gene promoter and inhibitor of flower induction. 50

53 Material and method

54

55 Material and method 1. Plant material and experimental deign 1.1 Previou experiment Characterization of Flowering Locu C The material ued a experimental model wa Arabidopi thaliana, concretely the eed of the ecotype Columbia 0 (Col-0) and Landberg erecta (Ler). The culture wa done in the greenhoue, in culture chamber and in fitotron under light condition (16 hour of light and 8 hour of darkne), relative humidity (50-80%) and temperature (21-23ºC) controlled. A ubtrate, we ued a mix of peat mo:perlite:vermiculite (2:1:1) arranged in each container of polyethylene; each container wa irrigated with a mix of nutritive olution (Hoagland, 1920) Determination of Flowering Locu T reponible for floral induction Thi tudy wa conducted comparing Cleopatra mandarin (Citru rehni Hort. Ex Tan.) juvenile plant and adult Moncada mandarin (Clementine Oroval [Citru clementine Hort. ex Tan.] x Kara mandarin [C. unhiu Marc. x C. nobili Lou.]) tree, grafted onto Carrizo citrange (C. ineni L. Obeck x Ponciru trifoliata L.Raf.) roottock, planted 5 m x 5 m apart in a loamy-clay oil, with drip irrigation. The experimental field wa located in the IVIA Reearch Station (Moncada, Spain). Three tree were elected for homogeneity in diameter, canopy height, ize and hape and a randomized complete-block deign wa employed. Three plant of Cleopatra mandarin were elected for each cultivar in July (2012). 30 leave per cultivar were ampled in September, November and February for RNA extraction and RT-PCR analyi for the tudy of Flowering Locu T (FT) expreion. Sample were immediately ground and tored at -80ºC for further analyze. 53

56 Material and method 1.2 Section Experiment I In thi experiment the Moncada mandarin from IVIA wa alo ued. From early September (2013) to the end of February (2014), 30 ON hoot and OFF hoot were collected. In the laboratory, thee hoot were eparated into developed mature adult leave, bud and exocarp from ON (fully loaded) and OFF tree (without fruit). Previouly, the fruit were meaured for their diameter, color of exocarp, chlorophyll and carotenoid content. At the ame time in September a ample of fifteen OFF hoot wa taken and the leave of different prouting were eparated (pring+ummer or fall). From thee, half of the leave and bud were collected for RNA extraction and RT-PCR analyi for the tudy of Flowering Locu T 2 (CiFT2), Flowering Locu C (CcFLC), Gibberellin 20 oxidae 1 (GA20ox1) and Gibberellin 3 oxidae 1 (GA3ox1) expreion and the other half were ued to ample the exocarp of fruit for GA and IAA concentration analyi by UPLC-MS/MS. All ample were immediately ground and tored at -80ºC until the time for analyi Experiment II Experiment were carried out during 4 conecutive year ( ) uing adult tree (10 15 year old) of Salutiana and Navelina weet orange (Citru ineni L. Ob), Hernandina Clementine (Citru clementina Hort. ex Tan.) and the two hybrid Afourer (Citru reticulata C. ineni L. Ob) and Moncada mandarin. Orange Salutiana and Navelina were planted 5 m 6 m apart and grown in a loamy clay oil and andy loamy oil, in orchard located in Valencia (39.28ºN 0.22ºW, 30 m altitude) and Huelva (37.22ºN 6.58ºW, 26 m altitude), Spain, repectively. The Hernandina Clementine and Moncada mandarin tree were planted 5 m 5 m apart in a loamy clay oil in Valencia. The Afourer mandarin tree were planted 4 m 6.5 m apart in a andy loamy oil in Huelva. All the orchard had drip irrigation, and fertilization, pet management and pruning were in accordance with normal commercial practice. Paclobutrazol (PBZ) wa applied during either the floral bud inductive period (November 20 25) or the floral bud differentiation (February 20 24). PBZ wa applied either to the oil, directly to the drip-line zone (1 10 g tree 1 ), or prayed on the canopy 54

57 Material and method by handgun (6 l tree 1, 2500 mg l 1, i.e. 15 g tree 1 ). Early in the ret period (October 10 15) treatment were alo carried out for comparion. Gibberellic acid (GA3), at a concentration of 50 mg l 1, wa prayed on the canopy in the floral bud inductive period for comparion. Six to ten ON- and OFF-tree were randomly elected for each cultivar, orchard and treatment. Untreated ON- and OFF-tree were ued a control. Four homogeneou branche per tree containing pring, ummer and fall fluhe, with ome 400 node per branch, and randomly ditributed around the canopy, were elected on the treatment date for flowering evaluation in the following pring. In a eparate experiment, the repone of bud to PBZ wa tudied by placing a 10 µl drop of a 2500 mg l 1 aqueou olution directly on bud of ON- and OFF-tree of Hernandina mandarin. Thi treatment wa carried out during the floral bud inductive period. One hundred bud were randomly elected from 10 ON-tree and 10 OFF-tree for treatment, and another hundred erved a control. In addition, another experiment wa done to analyze the inhibitory effect of fruit on flowering. Only tree bearing moderate fruit load (*80 kg tree -1 ) were elected of Salutiana weet orange. In early December, 40 mg L -1 of GA3 and 2,000 mg L -1 of PBZ were prayed onto the entire tree with a hand-gun prayer. Untreated tree erved a control. From the treatment date (11 December) to the onet of bud prouting (late February), 30 fully developed, autumn fluh (that i, nonbearing hoot), mature adult leave per tree from control and another 30 from GA3- and 30 from PBZ-treated tree were randomly collected for RNA extraction. Sample were ground and tored at 80ºC for RNA extraction and RT-PCR analyi for the tudy of CiFT, FLC-like, GA20ox1 and LEAFY (CLFY). Six tree were ued for the extraction. All of thee experiment were evaluated the yield, the prouting and the flowering Experiment III Thi experiment wa done in Afourer mandarin tree ituated in Pedralba (Valencia) with normal drip irrigation, fertilization and culture. 100 leafy ingle flowered hoot (ON) were elected from 5-6 node long and 70 of vegetative hoot (OFF) from 10 tree. In half of the ON hoot girdling wa done for the peduncle at the end of Augut (25 Augut, 2014). Girdling wa performed by removing a 2-mm ring of bark from the peduncle 1.0 cm from the calyx. Fruit from ON hoot that were not girdled erved a 55

58 Material and method control. At the end of ummer and pring the prouting and the flowering in 30 hoot per treatment were evaluated. In thee period the characteritic of the fruit were evaluated: weight (g), diameter (mm) and the color the peel and the pulp, the total oluble olid (TSS) and the length of the hoot formed. For the endogenou hormonal analyi, 20 hoot were collected for each treatment before the onet of prouting (11 day after girdling, September 5).In the laboratory OFF hoot were eparated in different node (node 1 nearer of apex than node 5) and the ame ON hoot. In ON hoot the peel of the pulp wa alo eparated. In thee ample GA1, GA4, IAA, Tz, ABA and JA concentration by UPLC-MS/MS were analyzed. All ample were immediately ground and tored at -80ºC until analyzed. Previouly to analyze the ample were lyophilized Experiment IV Thi tudy wa conducted with 10-year-old field-grown tree of Mandarino tardivo di Ciaculli mandarin (Citru reticulata Blanco), grafted onto Citru aurantium L. Thi experiment wa done in an experimental field located at the Univerità degli Studi di Palermo (Palermo, Italy) with normal drip irrigation, fertilization and culture. In thi experiment 20 vegetative hoot (OFF) and 20 leafy ingle flowered hoot were collected in the induction period (November, 2013) and in February. The microhoot from thee hoot were elected to cultivate in vitro. The microhoot were 1.0 cm long and each microhoot had a bud. The MS medium (Murahige and Skoog, 1962) wa ued with and without zeatin. The percentage of prouting wa evaluated every 7 day for 60 day. In the induction period (November 13) and in February the characteritic of the fruit were evaluated: weight (g), diameter (mm), the color the exocarp and number of eed. Moreover, in pring, prouting and flowering were evaluated in 30 hoot per treatment. 1.3 Section 2 Moncada mandarin tree located in the IVIA orchard were ued in thee experiment. Spring fluh leave and bud from 8 month old tree were ued a biological material and were randomly collected from 12 tree (6 on-crop and 6 off-crop) in autumn, which i when the fruit affect floral induction (November 2010). All leave and bud 56

59 Material and method were harveted the ame day at 11:00 a.m. and were immediately frozen and tored at - 80ºC. Thee ample were collected by protein extraction and ubequent analyi. In another experiment, 30 ON hoot and 30 OFF hoot were collected from early September to the end of January. In the laboratory, thee hoot were eparated into developed mature adult leave, bud and exocarp from ON (fully loaded) and OFF tree (without fruit). Previouly, the fruit were meaured for the color of exocarp and the carotenoid content. The ample were collected for ABA and JA concentration analyi by UPLC-MS/MS. All ample were immediately ground and tored at -80ºC until analyi. 1.4 Section 3 In thee experiment Moncada mandarin tree located in the IVIA orchard and Afourer mandarin tree located in Pedralba were ued. In all experiment flowering in pring wa evaluated. In the firt experiment, from early January (Year 1, 2014) to the end of February of the next year (Year 2, 2015), ON and OFF bud (January Year 1) and 10 ON leave and 10 OFF leave were collected for each ample (Year 2). In the econd experiment a election wa made from 3 branche per tree, OFF, ON and defruiting (defruited in Augut 25, 2014) in 3 different tree. 10 ON, 10 OFF and 10 defruiting leave were collected at the induction period (November 25, 2014). In the third experiment 100 ON hoot taken from 3 tree were elected. Half of the hoot were ued a control. The other half of thee leafy ingle flowered hoot were prayed with a hand-gun prayer containing 5-Azacytidine, 350 µm of olution. After the treatment, the ample were taken at 8, 24 and 48 hour. The leave and bud of thee ample were collected for RNA extraction and RT-PCR analyi for the tudy of CiFT2, FLC, SHORT VEGETATIVE PHASE (SVP), TEMPRANILLO 1 (TEM1), EARLY FLOWER 8 (ELF8), TRITHORAX 1,2 (12ATX), TRITHORAX 7 (7ATX), PROTEIN ARGININE METHYLTRANFERASE 5 (SKB1) and TRITHORAX 5 (5ATX) expreion. Moreover, ample from OFF and ON leave in May and November were ued for DNA extraction and DNA methylation analyi of FLC in May and Flowering Locu T 1 (CiFT1), CiFT2, FLC, SVP and TEM1 in November. All ample were immediately ground and tored at -80ºC until analyi. 57

60 Material and method 2. Method 2.1 Yield, prouting and flowering evaluation Sprouted bud, developing hoot and flower a well a leave per hoot were counted and evaluated following to Guardiola et al. (1982). Unprouted node were alo recorded. The total number of flower per branch wa calculated uing the number of flower per hoot and the number of hoot developed per branch. The reult are expreed in flower per 100 node to compenate for the difference in the ize of the branche elected for counting. Fruit were harveted at the appropriate commercial ize and color tandard, and yield (kg tree 1 ) wa determined by weighing and counting all fruit harveted per tree. 2.2 Fruit characteritic Fruit were previouly evaluated for weight (g), diameter (mm) with a caliper (Mitutoyo, Japan), the color of the exocarp and the endocarp by a colorimeter and by Minolta Chromameter CR-300 (Minolta, Japan) and the total oluble olid with a digital refractometer (Atago, Japan) and number of eed. Three meaurement per fruit in the equatorial zone of the fruit were taken from October until January or March in Moncada (Valencia), Palermo (Sicily) and Pedralba (Valencia). The reult are given a a, b and L Hunter coordinate. a Hunter i indicative of peel tranition from green to orange in Citru and becaue zero coincide with the onet of fruit color break. Color reading of a denote green or red when negative or poitive, repectively. b Hunter denote blue or yellow when negative or poitive, repectively, and L Hunter denote black or white when i 0 or 100, repectively. In addition, pigment analyi chlorophyll and carotenoid were extracted from frozen exocarp a decribed in Rodrigo et al. (2003). Chlorophyll a, b and total (a + b) were determined by meauring aborbance at 644 and 662 nm. Calculation were performed following Smith and Benitez (1955) equation. After chlorophyll meaurement, the pigment ethereal olution wa dried and aponified with 10% methanolic KOH olution. Carotenoid were ubequently re-extracted with diethyl ether until the hypophae wa colorle. Total carotenoid content in the ethereal extract wa calculated meauring aborbance at 450 nm and uing the extinction coefficient of β- 58

61 Material and method carotene (ϵ = 2500; Davie, 1976). Three independent extraction were performed per ample. 2.3 Culture Biological material Seed preparation and terilization of Arabidopi thaliana For the terilization of the eed, two protocol were followed: a) Sterilization by chlorine atmophere. Seed were placed in an open microcentrifuge tube and after thee eed were placed in a deiccator which contained a beaker with 100 ml bleach. Next, 3 ml of 37% hydrochloric acid wa added to the bleach and the deiccator wa immediately cloed. It wa left tanding between 1-4 hour. Thi terilization wa performed in a fume cupboard. b) Sterilization with ethanol. In a microcentrifuge tube containing the eed 200 µl of 70% ethanol % Triton X-100 wa added. The tube wa gently haken for 3 minute to wah the eed. The liquid wa removed, and after that 200 µl of 96% ethanol wa added and the tube wa haken for 1 minute. Next, the eed were placed on terile filter paper moitened with 96% ethanol and allowed to dry. For the owing of the eed two protocol were followed: a) Direct owing in ubtrate. Pre-terilization i not neceary. Seed were immered in culture tube with 0.05% agaroe olution. The tube wa haken to ditribute homogenouly. The eed were tratified after 3 day in darkne at 4ºC. Finally, the upenion wa ditributed in polyethylene container with moitened ubtrate. b) Sowing in plate of MS. After terilizing the eed, they were ditributed homogeneouly in Petri plate with MS medium. The plate were cloed with porou tape and they were tratified by treatment over 3 day in darkne at 4ºC. After 3 day, the plate were taken out and placed in a culture chamber. To tranfer the eedling that were grown in vitro to oil, we ued curved forcep. Thee eedling were ditributed in polyethylene container with moitened ubtrate. 59

62 Material and method Tranformation of Arabidopi thaliana with Agrobacterium tumefacien by FLC characterization The method ued for the tranformation wa immerion floral dip. The modified protocol of Clough and Bent wa followed (Clough and Bent, 1998). Firt of all, the plant were cultured until flowering (4-5 week approximately) in individual container with ubtrate. Secondly, the Agrobacterium tumefacien C58 upenion wa prepared with the correpondent contruction. Next, 1 liter of the aturate culture wa ued for the tranformation and thi wa centrifuged 10 minute at 4000 rpm. After thi, the pellet wa reupended in infiltration medium: 5% accharoe % Silwett detergent. Finally, plant were immered for 1 minute in Agrobacterium tumefacien C58 upenion. Then, the plant were placed on paper to remove exce of upenion. They were left in darkne for 24 hour. Subequently, the plant were cultured under normal condition of relative humidity, light and temperature until the eed were collected Microhoot terilization and culture of Mandarino Tardivo di Ciaculli Sterilization and culture: Firt of all, the leave and fruit were removed. Then, the hoot were cut in microhoot of 1 cm with one bud. After that, the material wa wahed with water and 10 ml detergent at the ame time that the each hoot wa cleaned with a toothbruh. Next, the material wa placed in a vacuum pump for 20 minute. Then microhoot were umerged in fungicide 2g l -1 for 10 minute (alo in vacuum pump). Subequently, the material wa wahed with water and they were placed in the laminar flow hood.. Under the laminar flow hood, the microhoot were immered firtly for 5 min in 70% (v/v) ethyl alcohol (vacuum), then in 35% commercial bleach olution with few drop of Tween 20, for 15 min. (vacuum). Finally, they were rined three time with terile ditilled water and immerged in left tanding for 30 minute with water and antibiotic (Cefotaxime 0.2 mg ml -1 ). Ten microhoot were put per each Petri dih containing 25 ml of medium before the incubation at 27±1 C, under cool white fluorecent lamp (Philip TLM 30W/84) with a Photoynthetic Photon Flux Denity of 35 µmol m -2-1 and a 16 h light photoperiod. The culture were oberved for 2 month, every week. 60

63 Material and method Medium compoition: For the culture, MS baal alt medium wa prepared (Murahige and Skoog, 1962). The vitamin upplement conited of 1 mg l 1 thiamine HCl, 1 mg l 1 pyridoxine HCl, 1 mg l 1 nicotinic acid, 2 mg l 1 glycine, 5 mg l 1 calcium pentathenate and 100 mg l 1 myo-inoitol. The medium alo contained 30 g l 1 ucroe and 8 g l 1 agar (agar-agar/gum agar) (Sigma Chemical, St. Loui, MO). The medium wa balanced to ph 5.85 with 1N KOH, and autoclaved at 121 C and 1x10 5 Pa (1.1 kg/cm 2 ) for 21 min. Two media were prepared. The firt MS medium without growth regulator and the econd MS medium with zeatin (1 mg ml -1 ). In addition, 1.5 ml PPM l -1 wa added to each medium Bacterial train Two type of bacterial train were ued: 1. DH5α train of Echerichia coli by the production and maintenance of plamid. 2. C58 train of Agrobacterium tumefacien by tranformation of Arabidopi thaliana. Bacterial cell were grown in LB medium (yeat extract 0.5%, triptone 1% and NaCl 1%) liquid and olid (with agaroe 1.5%) upplemented with the correponding antibiotic to the reitance gene that the plamid had. E.coli wa cultivated for 1 day at 37ºC while A.tumefacien wa cultivated for 2 day at 28ºC, both under agitation. When the plamid contained gene of election by color, 40 µl of X-Gal (40 mg ml -1 ) wa added to the medium. The contruction produced were kept in a freezer at -80ºC in glycerin form. Next, 700 µl of a aturate liquid culture of the bacterial train wa mixed with 300 µl of 50% glycerol. Finally, it wa frozen in liquid nitrogen and tored until ue Tranformation of Echerichia coli Two method of tranformation were ued: a) Tranformation by heat hock method. DNA ample wa dialyzed (50 µg when i a ligation) before that it wa introduced into the bacteria. The drop wa ituated on a nitrocelluloe filter of µm in form of dic of Millipore. It wa placed in Petri dih with miliq water for 10 minute. After that, DNA wa collected and 61

64 Material and method mixed with chemically competent cell that were in ice at 4ºC. Thi mixture wa left to tand for 30 minute in ice. After thi incubation in ice, a mixture of chemically competent bacteria and DNA wa placed at 42 C for 1 minute (heat hock) and then placed back in ice. 500 µl of SOC media wa added and the tranformed cell were incubated at 37 C for 1 hour with agitation. Finally, the mix wa poured into the Petri dih with the LB and the antibiotic. b) Tranformation by electroporation. DNA ample wa dialyzed and mixed with electro competent cell that were in ice at 4ºC. The mixture wa introduced in electroporation cuvette, dry and cool. Thee cell were ubjected to an electric pule of 1500V for 5 m. 250 µl of SOC media wa added in the cuvette and thi mix wa tranferred to a microcentrifuge tube. After that, the tube wa incubated at 37 C for 1 hour with agitation. Finally, the mixture wa poured into the Petri dih with the LB and the antibiotic Tranformation of Agrobacterium tumefacien For thi, tranformation by electroporation wa ued. The DNA ample wa dialyzed and mixed with electro competent cell that were in ice at 4ºC. The mixture wa introduced in electroporation cuvette, dry and cool. Thee cell were ubjected to an electric pule of 1440V for 5 m. 1 ml of SOC media wa added in the cuvette and thi mixture wa tranferred to a microcentrifuge tube. After that, the tube wa incubated at 28 C for 2-3 hour with agitation. Finally, the mixture wa poured into the Petri dih with the LB and the antibiotic. 2.4 Molecular biology technique Extraction Protein Frozen Moncada leave and bud from each tree were eparately pulverized in liquid nitrogen with 0.05% PVP (twelve ample). Then, ample were homogenized in an extraction buffer (50 mm Tri-HCl, ph 7.5, 1 mm PMSF, 0.2% β-mercaptoethanol) uing a petle and mortar. Homogenate were centrifuged for 20 min at 20,000 g. The 62

65 Material and method upernatant were mixed with an equal amount of cold 20% TCA. The mixture were incubated for 1 h at 4ºC and centrifuged at 20,000 g for 15 min at 4ºC. The protein pellet were wahed three time with acetone. After the lat centrifugation at 20,000 g for 15 min at 4ºC, pellet were re-upended in a lyi buffer (7 M urea, 2 M thiourea, 4% CHAPS). Protein aliquot analyzed by 2D electrophorei were tripped of non-protein contaminant uing a 2D Clean-Up Kit following the manufacturer intruction (GE Healthcare). The cleaned protein wa re-olubilied in a lyi buffer the conventional 2D analyi or in a Tri-buffered olution (7 M urea, 2 M thiourea, 4% CHAPS, 20 mm Tri- HCl ph 8.5) for 2D DIGE analyi. Protein concentration wa determined with the Bio- Rad protein aay uing bovine erum albumin (BSA) a tandard RNA and genomic DNA Total RNA wa extracted from frozen tiue and ubequently treated with DNae I (RNae Free DNae Set, Qiagen, USA). The amount of RNA wa meaured by pectrophotometric analyi (NanoDrop NDB1000 pectrophotometer, Thermo Fiher, USA). The abence of DNA contamination wa checked by performing a no revere trancription aay which conited of a PCR with each RNA ample uing the Citru actin primer (Supp. Table 1). No amplified product were detected, which confirmed the purity of the RNA extract. RNA extracted wa tored in a freezer at -80ºC until ue. Genomic DNA wa extracted from frozen tiue. DNA extracted wa frozen at -20ºC until ue Extraction of plamid DNA For the plamid DNA extraction of Echerichia coli 5 ml of the liquid culture wa ued and the ytem ued for thi wa Plamid Mini Kit I (E.Z.N.A) Protein analyi and meaurement Fluorecent labelling Protein ample were labelled uing the CyDye DIGE fluor (Cy2, Cy3 and Cy5) according to the manufacturer intruction (GE, Healthcare). Three different protein 63

66 Material and method ample (internal tandard, on and off ample) were labelled individually with the three dye. The internal tandard wa created by pooling an aliquot of all biological ample analyzed in the experiment and it wa alway labelled with Cy2. Six biological replicate were analyzed in thi experiment; thu, twelve biological ample were ued to make the internal tandard. The ample of each on-crop tree never wa mixed with ample of other ON-crop tree (the ame for OFF-crop ample). The ON ample and the OFF ample were labelled with Cy3 or Cy5, alternatively, depending on the biological replicate, thu avoiding the label effect. Equal amount (50 mg) of ON (Cy3, for example), OFF (Cy5) and internal tandard (Cy2) ample of the ame biological replicate were pooled. A lyi buffer wa added to make the volume up to 40 ml. Then, the ample wa mixed with 40 ml ioelectrofocuing (IEF) rehydration buffer (8 M urea, 4% CHAPS, 0.005% bromophenol blue) containing 65 mm DTT and 1% IPG buffer ph 3-11 and loaded into the gel (one gel for each biological replicate) D electrophorei For 2D analyi, trip of 24 cm in length with immobilized ph gradient of 3-11 were hydrated overnight at room temperature with 450 ml IEF rehydration buffer, containing the reagent Detreak and Pharmalyte ph 3-10, according to the manufacturer intruction (GE, Healthcare). CyDye labelled ample (150 mg of protein) were loaded in hydrated trip. IEF wa performed on an IPGphor unit (GE Healthcare) at 20ºC and a maximum current etting of 50 ma per trip, uing the following etting: 300 V for 1 h, an increaing voltage gradient to 1000 V over 6 h, an increaing voltage gradient to 8000 V over 3 h, before finally holding at 8000 V for a total of 32,000 V h. After IEF, each trip wa equilibrated eparately for 15 min in 10 ml equilibration olution I (0.05MTri-HCl buffer, ph 8.8 containing 6 M urea, 30% glycerol, 2% SDS, 200 mg DTT per 10 ml buffer) followed by equilibration olution II (ubtituting DTT for 250 mg iodoacetamide per 10 ml buffer and adding 0.01% bromophenol blue) before being applied directly to the econd dimenion 12.5% SDS- PAGE gel. Six gel were run imultaneouly at 20ºC, applying 2 W/gel for 30 min and 20 W/gel for the remaining 5-6 h, uing an Ettan DALTix unit (GE, Healthcare). A running buffer of 25 mm Tri ph 8.3, 192 mm glycine and 0.2% SDS wa ued. Each gel howed the differential protein expreion between an ON ample (from a ingle ONcrop tree) and an OFF ample (from a ingle OFF-crop tree). 64

67 Material and method Gel imaging and data analyi After SDS-PAGE, CyDye-labelled protein were viualized by canning uing a Typhoon Trio canner (GE Healthcare) with the relevant wavelength for each CyDye. Cy2 image were canned uing a blue laer (488 nm) and an emiion filter of 520 nm band pa (BP) 40. Cy3 image were canned uing a green laer (532 nm) and a 580 nm BP 30 emiion filter. Cy5 image were canned uing a red laer (633 nm) and a 670 nm BP 30 emiion filter. All gel were canned at 200 mm (pixel ize) reolution. The photomultiplier tube (PMT) wa et between 500 and 600 V uing normal enitivity. The canned gel were then directly tranferred to the ImageQuant V5.2 oftware package (GE, Healthcare). Image gel analyi wa carried out uing the DeCyder 2D Software V6.5 following the manufacturer intruction (GE Healthcare). The image were exported to the DeCyder Batch Proceor module and DIA (Differential in-gel analyi) and BVA (Biological Variation analyi) module were made automatically. The DIA module wa ued for pot detection, pot volume quantification, and volume ratio normalization of different ample in the ame gel. The BVA module wa ued to match protein pot among different gel and to identify protein pot that exhibit ignificant difference. Manual editing wa performed in the biological variation analyi module to enure that pot were correctly matched between different gel and were not contaminated with artifact, uch a treak or dut. The paired t-tet wa ued for tatitical analyi of the data. A fale dicovery rate (FDR) correction wa applied to eliminate fale poitive. Protein pot that howed a tatitically ignificant change in abundance between ON and OFF ample uing a Student t-tet (p < 0.05) were conidered a being differentially expreed Protein identification by ma pectrometry (MALDI, MS/MS) analyi To elect pot of interet, gel were firt tained with Silver Staining Kit, Protein (GE, Healthcare). Protein of interet were manually excied from analytical gel and were ditained with two 5-min wahe with acetonitrile (ACN)/water (1:1, v/v), followed by rehydration with 50 mm ammonium bicarbonate for 5 min and 25 mm ammonium bicarbonate in 50% (v/v) ACN for 15 min. Gel piece were then digeted with equencing grade trypin (Promega) a decribed elewhere (Shevchenko et al., 1996), and ubject to 65

68 Material and method PMF and/or LC/MS/MS analye. The digetion mixture wa dried in a vacuum centrifuge, reupended in 7 ml of 0.1% TFA (trifluoroacetic acid, Sigma), and 1 ml wa potted onto the MALDI target plate. After air-drying the droplet at room temperature, 0.5 ml of matrix (5 mg/ml CHCA) (acyano-4-hydroxycinnamic acid, Sigma) in 0.1% TFA-ACN/H2O (1:1, v/v) wa added and allowed to air-dry at room temperature. The reulting mixture were analyzed in a 4700 Proteomic Analyzer (Applied Bioytem, Foter City, USA) in poitive reflectron mode (2000 hot each poition). Five of the mot intene precuror (according to the threhold criteria: minimum ignal-to-noie: 10, minimum cluter area: 500, maximum precuror gap: 200 ppm, maximum fraction gap: 4) were elected for every poition for the MS/MS analyi. And, MS/MS data were acquired uing the default 1 kv MS/MS method. The MS and MS/MS information wa ent to MASCOT via the Protein Pilot oftware (Applied Bioytem). Databae earche on NCBI, Swi-Prot, and HarvEST: Citru databae were performed uing MASCOT earch engine (Matrix-Science). HarvEST: Citru contain bet BLASTX hit from UniProt, the Arabidopi genome and Phytozome verion Poptr1.1 ( Searche were performed with tryptic pecificity allowing one mied cleavage and a tolerance on the ma meaurement of 100 ppm in MS mode and 0.6 Da for MS/MS ion. Carbamidomethylation of Cy wa ued a a fixed modification factor, while oxidation of Met and deamidation of An and Gln a variable modification. The ample without a poitive identification were analyzed by LC/MS/MS. Peptide eparation by LC-MS/MS wa performed uing an Ultimate nano-lc ytem (LC Packing) and a QSTAR XL Q-TOF hybrid ma pectrometer (AB Sciex). Sample (5 ml) were delivered to the ytem uing a FAMOS autoampler (LC Packing) at 30 ml/min, and the peptide were trapped onto a PepMap C18 pre-column (5 mm & 300 mm i.d.; LC Packing). Peptide were then eluted onto the PepMap C18 analytical column (15 cm x 75 mm i.d.; LC Packing) at 300 nl/min and eparated uing a 30 min gradient of 5-45% ACN. The QSTAR XL wa operated uing an information dependent acquiition mode, in which a 1- TOF MS can from 400 to 2000 m/z wa performed, followed by 3- product ion can from 65 to 2000 m/z on the three mot intene doubly or triply charged ion. The MS/MS information wa ent to MASCOT via the MASCOT DAEMON oftware (MATRIX SCIENCE). The earch parameter were defined a for MS-MS/MS analyi. 66

69 Material and method Starch analyi Leave (2 g) were dried in the oven (60ºC) and then were treated with 80% ethanol. The remaining pellet were gelatinized by autoclaving and then odium-acetate buffer and amyloglucoidae were added to the gelatinized extract, according to Schaffer et al. (1987). Enzymatic digetion were performed for 2 h at 55ºC. After filtration, releaed glucoe wa quantified with a Water HPLC ytem equipped with a carbohydrate column (4.6 x 250 mm, 5 mm, Tracer Carbohydrat Tecknokroma, Barcelona, Spain) and a 2410 differential refractometer. A binary iocratic phae coniting in ACN:water 75:25 (v/v) wa ued and the retention time for glucoe wa 11.5 min. Quantification wa performed by external tandard calibration. Starch content wa expreed in mg g -1 dry weight Catalae activity Leave (2 g) were homogenized in a Polytron 3100 (Kinematica, Lucerne, Switzerland) uing 10 ml of 50 mm phophate buffer, ph 7.0, containing 2mM EDTA and 2% polyvinylpolypyrrolidone (PVPP, Sigma, Barcelona, Spain). The crude extract wa centrifuged at 12,000 rpm at 4ºC for 30 min, and the upernatant wa filtered (0.45 mm; Nylon) and ued for the catalae aay within 1 h. The protein concentration in the upernatant wa determined in the TCA precipitate uing bovine erum albumin a tandard (Lowry et al., 1951). The reaction medium (2 ml) contained 100 mm phophate buffer ph 7.0, and 100 ml of the upernatant. The reaction wa tarted by adding 100 ml of 10 mm H2O2. Catalae activity wa pectrophotometrically determined by the decreae in hydrogen peroxide (Tewari et al., 2005). The reaction wa monitored at 240 nm in a pectrophotometer UV-1610 (Shimadzu Corp., Kyoto Japan), at room temperature. The molar extinction coefficient ued wa 43.6 M -1 cm -1. Catalae activity wa expreed a mmol H2O2 conumed g -1 protein min -1 after 3 min reaction Gene ontology analyi Total amount of iolated protein were analyzed eparately in two group, which were etablihed according to the ratio expreion uing FatiGO ( (Dimmer et al., 2008). The databae at 67

70 Material and method wa ued to earch for the Arabidopi protein that were homologou to the protein identified in thi tudy. Etablihing homologie enabled u to undertand the main biological procee in which protein are involved Biulphite treatment ng genomic DNA wa ubjected to one treatment of odium biulphite converion uing the EpiTect Biulphite kit (Qiagen) according to the manufacturer intruction. The reaction wa then purified once more uing the PCR purification kit (Qiagen, USA) Gene expreion analyi by qrt-pcr The trancript preent in 1 μg of total RNA were revere-trancribed uing the QuantiTect Revere Trancription Kit (Qiagen, USA) in a total volume of 20 μl. A 2.5 μl aliquot of a 4-time diluted firt-trand cdna wa ued for each amplification reaction. Quantitative real-time PCR wa carried out on a Rotor Gene Q 5-Plex (Qiagen, USA) uing the QuantiTect SYBR Green PCR Kit (Qiagen, USA). The reaction mix and condition followed the manufacturer intruction with certain modification. The PCR mix contained 2.5 μl of diluted cdna, 12.5 μl of QuantiTect SYBR Green PCR Mater Mix (Qiagen, USA), 1.5 μl of 0.3 μm primer F, and 1.5 μl of 0.3 μm primer R, the final volume being 25 μl. The cycling protocol for the amplification conited of 15 min at 95ºC for pre-incubation, then 40 cycle of 15 at 94ºC for denaturation, 30 at 60ºC for annealing and 30 at 72ºC for extenion. RT-PCR reaction were repeated three time for each gene and monitored in real time with the Rotor Gene Detector. After amplification, the melting-curve analyi excluded artefactual amplification. The relative expreion of RNA trancript wa quantified with the threhold cycle value (Ct) obtained from each ample uing the 2 -DDCt method (Livak and Schmittgen, 2001). Expreion level were calculated relative to the contitutively expreed ACTIN gene (Table S1). The relative gene expreion level i given by 2 -DDCt. Normalization wa performed to the lowet value between the ample for each experiment. Two or three independent biological ample under each experimental condition were evaluated in technical triplicate. 68

71 Material and method Table S1. Primer equence ued in RT-PCR amplification reaction. Annotation CDS 5 -Direct primer- 3 Predicted product (bp) 5 -Revere primer- 3 ACTIN Ciclev m TTAACCCCAAGGCCAACAGA 141 TCCCTCATAGATTGGTACAGTATGAGAC CiFT1 Ciclev m GGGAGGCAGACTGTTTATGC 150 TCATCGTCTGACAGGCCTTC CiFT2 Ciclev m TCTAGCAGGGACAGAGATCCT 53 AGAACATCACCAACAACGCG CiFT3 Ciclev m GGCTGAGGGAGTACTTGCAT 139 TGCCGGAACAACACGAAAAC FLC Ciclev m GGCAACTTGAAGGTCCAAAC 124 GCCCAATGAGCATAGGAATG CLFY Ciclev m TCTTGATCCAGGTCCAGAACATC 63 TAGTCACCTTGGTTGGGCATT GA20ox1 Ciclev m ACCAAGTGGGTGGTCTTCAG 96 TGAAGGTGTCGCCAATGTTA GA3ox1 Ciclev m CAACGCAAGATGTCAAATGG 85 CAGGCCGGGTAGTAATTCAA SVP Ciclev m AGTGGCGGAGGTATCAAATG 119 TGAGGGAGGTGTCTGAGCTT TEM1 Ciclev m GCAAATGTCTTGTGCTGGAA 104 TGTGCTTCCTCAGCATATCG ELF8 Ciclev m TCTCGATCTCTGGCTCATCA 91 AGGACTAGACCCTTCCTCCAA SKB1 Ciclev m GGGTTCGCTGGTTATTTTGA 67 CCGTTGATGGCTCAATACCT 12ATX Ciclev m TGTGTTGGCTGCCAGTTTAG 52 AATATCCCCAGGCTCGAGTT 7ATX Ciclev m CGAACACATTTATGCCAACG 70 GCACCCATTAGTGGAGCAAT 5ATX Ciclev m TAGGGTGAAAGGTTCCATGC 121 CTGCTTCCGCTCCTTCATAG Sequence analyi and phylogenetic tree CiFT1-3 equence (Nihiwaka et al., 2007; Samach, 2012) and CLFY equence (Pilliteri et al., 2004) were ued for thee tudie. FLC (Supplementary data S1), GA20- oxidae, GA3-oxidae, SVP, TEM, ELF8, SKB1 and ATX amino acid equence of A.thaliana and other annual plant pecie and woody evergreen and deciduou tree pecie were obtained from the NCBI databae ( Sequence were aligned againt the Citru clementina genome uing the TBLASTN tool of Phytozome v10.3 databae ( Baed on thi equence imilarity, ome putative homolog of thee gene were identified in the Citru clementina genome. The GA20ox tudy wa baed on the imilarity to the characterized amino acid equence of CcGA20ox1 and CcGA20ox2 from the citru hybrid citrange Carrizo [(Ponciru trifoliata Raff. x Citru ineni (L.) Ob.)]. The GA3ox, FLC, SVP, TEM, ELF8, SKB1 and ATX from A.thaliana. Phylogenetic tree are given in Supplementary figure at the end of 69

72 Material and method thi chapter. Primer ued for qrt-pcr analyi ( are lited in Supplementary Table 1 and primer by PCR analyi ( are lited in Supplementary Table Generation of contruction in plamid vector Polymerae chain reaction (PCR) The biulphite treated DNA wa amplified uing Hot tart Platinum Taq DNA Polymerae (Invitrogen). The thermal cycling program wa et at 95 C for 1 min followed by 40 cycle of 95 C for 30, annealing 50º for 30, and extenion at C for 30, ending with a 3 min extenion at C. PCR product were analyzed in agaroe gel of 1%. Forward (F) and revere (R) primer for biulphite equencing PCR were deigned uing MethPrimer: deigning primer for methylation PCR, given in Supplementary Table 2. Table S2. Primer equence ued in PCR amplification reaction. Annotation CDS 5 -Direct primer- 3 Predicted product (bp) 5 -Revere primer- 3 CiFT1 Ciclev m TTATTTTTTTTATTTATGTTTATTTTTGGTTTTTTTG 112 TTTCTAATTAATCCAAAAAAAAAAAATAATCACTTAC TTATATATGTATGTAGGTTAAGAGATTGTG 211 CCCTCTAACAATTAAAATAAACAAC TTAATTGGTTGTGGTTAGTGATTTTT 355 TAAACAATCTACCTCCCAAATTACC CiFT2 Ciclev m GTAGGAGGAGGAGGTTTTAGTATTG 278 AAAACAACCTCTCACCTCTAAACAT FLC Ciclev m TTGTTAGTATTGTTGTGTTTTATTTTATAT 387 AAAAAACCTAAATAATTATATTCTATTCAT TTGGGATTATTTTTAAAATTTTGTA 386 AAATTCTCTACTTCTAACAAATCCA GTTTTGATGAAATATTTTAAAAGTAA 340 ATAAAACTAATCACCTACAATACACC SVP Ciclev m TGAGAGGAATAAAGTTTGAAGTAAT 363 AAAAAACACTACAACTCTCCTTAAC TEM1 Ciclev m TGGGAAAAATTTAAGTTAAAATGTA 297 TTTAATAATTTTATTTTACCACCAC 70

73 Material and method Ligation in pgem-t Eay I vector PCR product have adenine over-hang that form by Taq DNA polymerae activity and can bind with thymine ticky end of the pgem-t vector. Hence, directly from PCR reaction tube 2 µl of amplified DNA fragment with 5 µl 2x Rapid ligation buffer for T4 DNA Ligae, 1 µl of linear pgem-t vector and 1 µl T4 DNA ligae (3U/µl) (Promega) were mixed and then incubated at 37ºC for 1 hour (alternatively at 4ºC for 24 hour) Blue-white creening E.coli cell, which are tranformed with pgem-t vector, have an ampicillin reitance gene and a LacZ operon which encode a β-galactoidae. LacZ expreion i induced by IPTG, a lactoe analogue that cannot be metabolized by E.coli. LacZ releae indole from the ubtrate X-gal (a glycoide compoed of galactoide and indole). Indole molecule form blue dimer. When pgem-t vector contain an inert in the cloning ite, the LacZ gene i interrupted and no color will be produced. So blue colonie indicate recloure of plamid DNA reulting in LacZ gene function, wherea white colonie indicate that the DNA of interet wa inerted into the plamid and caued interruption of the LacZ gene. For blue-white creening 1 ml of grown cell were poured on LB agar plate. Plate were urface dried in the laminar air flow and incubated at 37ºC over night. Then, the white colonie were collected and grown in 5 ml LB + Ampicillin over night at 37ºC Digetion DNA of interet with the plamid wa digeted with EcoRI enzyme for 2-3 hour at 37ºC. Then 6-12 individual clone were equenced Cloning FLC into a plant cloning vector FLC codification region wa amplified by PCR. The cloned wa obtained from IVIA1 library, IC0AAA56AF11. Subequently, it wa cloned in pcr8 and after that, all of thi wa introduced in pearlygate 201 (kanamycin and glufoinate reitant). Next, thi contruction wa introduced in Agrobacterium (gentamicine and rifampicin reitant). 71

74 Material and method 2.5 Hormone iolation, purification and quantification Material frozen in liquid nitrogen wa ground into a fine powder. Aliquot (about 50 mg freh and dry weight) of material were extracted with 80% methanol containing 1% acetic acid. Internal tandard were added and mixed with the aliquot at 4 C for 1 hour. The internal tandard for quantification of each of the different plant hormone were the deuterium-labelled hormone. The extraction protocol ued i that decribed in Seo et al. (2011) with certain modification. In brief, for dealination, the extract were paed through revere phae column HLB (Water). The plant hormone were eluted with 80% methanol containing 1% acetic acid and conecutively applied to cation exchange MCX column (Water). The fraction containing the acidic ABA, GA, IAA, IP, Tz and JA wa applied through ion exchange WAX column (Water). The final reidue wa diolved in 5% acetonitrile-1% acetic acid, and the hormone were eparated uing an auto ampler and revere phae UPHL chromatography (2.6 μm Accucore RP- MS column, 50 mm length x 2.1 mm i.d.; ThermoFiher Scientific) with a 5 to 50% acetonitrile gradient containing 0.05% acetic acid, at 400 μl/min for 14 min. The hormone were analyzed with a Q-Exactive ma pectrometer (Orbitrap detector; ThermoFiher Scientific) by targeted Selected Ion Monitoring (SIM). The concentration of hormone in the extract were determined uing embedded calibration curve and the Xcalibur 2.2 SP1 build 48 and TraceFinder program. 2.6 Evaluation of leaf number To evaluate the effect of citru FLC. Three line of Arabidopi thaliana cv. Columbia were tranformed. The firt wa control (col) and the econd line wa introduced the FLC of citru (col-ccflc). When all of thee plant flowered the number of roette and caulinar leave wa evaluated. 2.7 Statitical analyi Parameter were tatitically teted by analye of variance (ANOVA), uing the leat ignificant difference (LSD) tet for mean eparation. The experimental data were analyzed with Statgraphic Plu 5.1 oftware (Statitical Graphic, Englewood Cliff, NJ). 72

75 Material and method 2.8 Supplementary figure. Phylogenetic tree Figure S1. Phylogenetic analyi of GA20 oxidae in Citru clementina. Only the CDS relationhip with GA20 oxidae 1, Ciclev m, wa elected according to it imilarity with citrange Carrizo. Figure S2. Phylogenetic analyi of GA3 oxidae in Citru clementina. Only the CDS directly related with GA3 oxidae of Solanum lycopericum and with the GA3 oxidae 1 of Arabidopi thaliana, Ciclev m wa elected. 73

76 Material and method Figure S3. Phylogenetic analyi of FLC in Citru clementina. Only the CDS inide the FLC/MAF branch wa elected, it wa directly related with FLC of Ponciru trifoliata and Viti vinifera, Ciclev m. 74

77 Material and method Figure S4. Phylogenetic analyi of SVP and TEM in Citru clementina. Only the CDS in the middle of the SVP and TEM, repectively, of each branch, Ciclev m (SVP) and Ciclev m (TEM) were elected. Figure S5. Phylogenetic analyi of ELF8 and SKB1 in Citru clementina. Only the CDS inide the branch of the ELF8 and SKB1, repectively, Ciclev m (ELF8) and Ciclev m (SKB1) were elected. 75

78 Material and method Figure S6. Phylogenetic analyi of 1, 2, 5 and 7ATX in Citru clementina. Only the CDS inide each branch of the 1, 2, 5 and 7ATX, repectively, Ciclev m (1,2ATX), Ciclev m (5ATX) and Ciclev m (7ATX) were elected. 2.9 Supplementary data. FLC equence Data S1. Genomic equence of FLC, obtained of IVIA1 library, IC0AAA56AF11 clone. In Phytozome v10.3, it correpond with Ciclev m.g. TTTATAATTTTATAATATATATGTAGATATATGAAATTTTATTAAATATAAAAATAATTATCACTATAAGATGTGATAATTTTATAAAAAGTAATTTATAAAAAATTTATCAAATATCAAAATTAACTTTTAACTTTTAAACATAATAATTGAGCTT CTAAAAAATTAATTCCAAACGGCCTTAATGTATGAATAATTTTTTTTTTAATTACAAAATATATATTTTTATTAAATAATTCAATCATCAAAAAAGTTTTGCTGTCTTAGGAGAATTAATTTGTAGGACCCACAGGCGCGTACACGAAAAAAATG ATAAAATGAGACTCAGGTCTATCTGCAGTAATAAACTAAGCTAACATGTTGACTTGCAAAAAGTTAATTTATTTGGTGTCGGCACTTAAGTGATTCCAGTAGAATAATTATACATCAAATTTGGAGGTTTGACGTGAGAAGACATTGGCTGGC CTTGGCTGGTTAATCTGGTAACTTACTTTAATCATACAGTTTATCCACTAGCCACTGGGTCCTATCGCAACAAAATGTTGTTTTTATTTTTAAAAAAATAAACAAAAGGGTGTAAACAAGTTCACTCTTTCTAATTAGAGTAATTTATTTGTTTCT AAAAAGAGTAATCTAATCTTTCTTCGAAATAAAATAATTAATAAATAAATAAATAAATAAATAAAACTAGGCGTGCCGAAAGTCCCCCAAAAAAAAAAAAACCTTAAAAATACGACAACAAATAACACGCACGGCACACGTAGCATTAAAAA ACAAATAAAAAAGCAGCCGCAAGCCCGCAACAAAAGCACGAATGATGCCCGCGGAGACACGTGTCGAATTGTCAGGCGTCCTCGCCTCCGACTTTCGATTGATCGATTTATTCACACACTGTCAGCACTGCTGTGCTTTACTCTACATAAGT ACCACCAAGCGACCCTAATTTCCGCCTTATCTGCTGTGCTCCTTTTCTCAATCTAGGGCGGCATTGGGTCAGAAAATGGGCCGGAAAAAGCTTCAACTGCAGCGAATCGAAAACAAAAGCCGATGTCAAGTAACGTTCTCAAAGCGGCGTAG 76

79 Material and method 77 CGGATTGATCAAGAAGGCTCGTGAGCTCTCCGTCCTCTGCGACGTCGACTTGGCGCTCGTCATCTTCTCCAGCCGCGGTAGGCTTTACGAGTTCTGTAGCGCCGACAGGTGCGCCTTCCCCTCCTCCTTTTCTTTTCTTTTTTTCCCCCTTTCCA AAAATGTTAATGTTGTATGAATAGAATACAATTATTTAGGCTCTCTACAGTAAACTGGGAAGGCAATTTAGTCGGTTTCTTGACTGCGTCATGGTCACGCGGTGTTCACTGATCAGTTTAAAATAAGTGTCATCTCAAAAAAGGAGAACTGTT TGGTGATGTGATTGCTATCATCCGTGCGACAAAATTGCTCTCCACGTGTAATGTGGCAGCAGTAGCGCTAACCGACTTGATTTTGGGATAATTTCTGCTCACCTGCGATGGAAGTGATATTTCTTCATCCACAGTGCCCGCTCCTCTATCTAGT TTCTTTTTTAATGTTGATTAATTAGGGAGCTCAGGACTGACCCTCTTTAGCTTCTTTATTAATTTTAAAGTTGCCGTAAGATTTAAGAAAGTAAGAATGATCATTGAAACGTGTCTAAAATAAATCTGGAATTTGAAAATTAATTTTTTGAAGCC TTCTTATTAAGACAATTCTTCTTACAACCAGTTACTTTTATATGTTGGATTGAAATTTAATGGTTGGGAAATAATTTTTGGGAAGTGAGAATTAGTGGGCCACGAGCTGAATTAGTTTGGTATTATGGGCTGTACGTAAATGTTACTTACTAAA ACGCTCTTATTAAAAGTATTTTATGAGATTTTGTTGCATTTCACGCTTTAAAAGGAAAAAATCTATTTTCAGATTGGGAACATGATCTAATGAGCACCTGAACATATGTTATGATTCTATGACAGCTTCCGAAGAGCACTGAAGTTGTGGATTA GTCAGATAATGTCACGGGACGTTTGTTTTCACGTTCACTAGATCATTACTATATATGGCTAATGATCTAGGTGGTGACGGTGCTCACTGTCACATGTTGCTGTTTAGTTAATCCATTTCTTAAACATTATTGACACGAAGACATAAAATTCTGCC ATGTGTCCCATGCCCACCATATGGTGGGTCTTGGTTCTCACTAGATCATCACCCGACATACATATTTCGAAATGTTTTTATAAGTTAGTTCAATGAGCTTGGAGGCCACAAATCATTTTATTACGAAAGATTAAATTTATTCTTTAATTTTTCTGT CATATTCATTATTTTTATGAGTTCTTAAGTTTTCAAGGGATGTTTTCGAAGAAATCTAAGAAAATAAATTTTAGAAGGATAATGTGACTTCTAAGTGAGTTCTTTTGATTTAGATGATTCAATTCAATTGGTTTTTTTCAACACTGCTAGCTAGC GAAGTATGAGCTTTGGATTTTCTATGTAAAAAAAATATTTTTAAACTAAGTAAATATGATTCTTAAATAATGATTTTGATAAAAATAGTAAAGATATTATAAATTTTTCATTATAGACGTAGTGAATCGAATTAATGTAGATTCTAATTTACAGT AGTTAATAGTTATCAAACGCTGTAGTTTTTAAACTACAATCGTCCAACCTTAATATCAAATAGGGTCTAAATAGGGCTCGAAGAGCAGTCAGGACCACTCGTATCAATTTATATGACTTATTCCAAGTATTGATTATGCATCAAATAATTTGTA GTACTCTAAGCAAAAAGTATTCGTATTTAGTAAAAAAGCAATAGTAATAGATATTAGATTAAGAATAAATTGGTTCTAAGTGTAATCGATACTTGTGGTTAGATTAATTTATATTTTAAAATAAACATGCTTATGTTTAGGAGAACTTGTATTC CAAGGAAGAGTCTTCAATATACTTGTTTGTGTGAGAGGATTTAGGGCTAAATTAGAGCTGATTTTTATTGTTTGAATCAAAAGAAAGAACAATCCTGACCTACAATTCAAAAGATAAGTTAATCTTTTTTCAAAACTCATAATTACTCCTAATA CTTGAGAAAAATGAGCATAAACGCTAAATGACTACTTAATCTCTTATTTATACCTGAACCCTAATTAGTGAATAAGTAAAATTACTAAATTGTCCCTCCCCTTTTATATTTGCACTTAAAATAAGTTTTCATTGGTCTAACATTACTCCCCTCCCA AAAAGACACCTTGTCCTCTAGGTGAAATCCAAGAAATTACTGCCTCTTCTTCCTTGGCTTCCTCTTCGCCTTCCTCCTCATCACACACAGTCAGGACTTATAGCGATTTGTCCTTACATCGATGCCCCGATGAAAATTTCTCATCACATCAAAAAC ACAACTCCTTTGCCCTTTTGTCCTGAATCTCTATCTCAGTGAGGCATTTGAAAGTCACCCTTGATCGTGCCTCACTAGCCCGATCGCTTTGGGAATCTCTTTGTAACTCCATCATTTACGTTTTCTGGCCTTCCGATGGTGTGCTTTGTGAACTCT CACTGAAACCCCAAGGTCTGGGTTTTCTATCCTACCAATCGTGTGTTGTTTCGGTATGAAAAACCGACTGAATTGATCTTATTAACCTATTCTGCTGTTTTCTTATCCTCAATCATCTGTGCCAACTCCATTGTCTCCCGTAACCCTTTAGGCCTT AGCAAGCGTAACGAGGCTCGGATATTTGGTTTCAGCCCTTCATAAAATCCCCTTCCAATACCGCCTCTGAGATTCCCCGTAAATGCCCCGATAGCAACTCAAACTTCTCCCTATACTCTATGACTGTTCTCTCTTAGGTCAGGGTAAAAAATTGT TCATGCAAATCTCCCACTTGAGTGGCCCGAAACCTCTCTAACAGTCGGTCCTTAAATTCACCCTACGACCTCAATGGTTGTCGTTGTTCTCGCCATTGAAACCATGCAAGGGCTTTTCCCTCAAACACAACAGTGCCGCCATTAGTTTTTCCTTT TCTGAGAGTCCATTGATGGCGAAATAACGCTCTACCCTATACACAACCATAGGAAATGGGCATTTTCAGTTTCTTTACCCTAGAATCCGCCCCAGATTGAGTTTACCTTTCGCGCCATTGGAGATCTCCAACCTCTTAAATCTACTCTAATTTCT ACAAATCCCGAAGTCTCAGCCTAAAATTCTCCTTGCCCAGCTGCATTCAGGTTATAATTGCTTCATCTCTCACCGGAAACCCCTCTTCCCAACCATCCGCCGAGTTGCCAAGCCTAACCATCCCCAACTTCTATGCACCCATTTCAGTGCCGAAA CGCCAATTCCTAACCTTAAGCTGCCTCCTCCAGCTGTGTAACCCTCACATCCACCACCCCCAGCCGCGACCCAGCCTTCTCCGCCACACCCGACGTCAGGAAAGGCTCCCCTAGGCCGCGAGTCCCCAAAACCGACCTCCCCGCCACCAAATCC ACCATGACTCGTTGCCTTTGCCTGTGTGAGATTGGTGGTTTCGACCCTCTGCGCTGCTGCCGACCACGGGTTCTATACCGCCGGATCCTTCTCCTGCCATGTTGCCGTGGCCCGTCGATATTGCCGGTAGTGGGTTAAAGTGGAGTTTCGTCA ATTTTAGCAACATGGTTTGATGGGATGAAAATTGTTGTTGAACAGTGTTAAATTGCTCGTTGATGGGTGGCCTTGAGCGATCCTTCGCTTCCAATTTCTCTTCAAGTGCGTCCACCTTCTTCGCTGCCATTGTCCCATAATTTTTGCTCTGATGC CACTTTGATAGGATATAGGGCTAAATTAAGGCTGATTTTTATTGATTGAATAAAAGAAAGAACAATGCTGACATACAATTCGAGAGGTAGATCAATCTCTCCCAGAACTCCTAATTACTCATAATACTAGACAAAACTCAGTATAAACTCTAA ATGATTCCTTAATCTCTTATTTTTACTTGAGCCCTAATTAGTGAATAAGTAAAATTACTAAACTGCCCCTACCTATTTCTATTTATGCCTAACATAAGTTTTCATTGGTCTAACATTGTGTACTTAAGAAACTCTACATTTTTATCCTAACCACACTT AACAAGGTAGTACTTTATGCTTATGCATAAGATTTCTTTATCTTCTCTAACTTTACAAATTTTGATAGATAGTATAAGAATCACACACCTATGCTTTCACACATATTCTCTCCTTGAATTAAGGTGAGAATTTGTTTAATTGGAGAATGGATCTTG GGATCCATTTGCTTTGATGTTTCACTAACATTTTGGTTGTTAGCTATTGCTGTTGTAATATTTTTTTTATTCTACTAAGTTAAATTTGATTTTTCTAAAACGATGTCCGTCTCTTGACTCCTTAAGTGACAAGAAAAACTTAAACTAAGGCTTGAAT GACCATTTCTTTTAATATGTAGATGTGATACAAGACTTGGAACAAGCAACTAGAGAGTAAGGGAGTATTTTTCCTATTTCTTTTTATTTTTCTTTTTTTGTATCTCTGTTAGTAGGCTTATATAGGTCAACTATTGATCCCTAATGTATAACCACC CAAGTATAGAACTAGGTGAAACCAATATTTATCTCCACAAAACTCTTTTACAAATTCTACTTCAAACTTGTTTGGAGGTCTCAAACAAACCTTAAAACCATTGGATGTACTTATACTTATGTCAAAAATTAAGAGGAAACTAGTTATATTAATAT TGTAAACTATACTATTTATATTAATGCATGAAAACGGGATGAAGCTAAAGTACCTATTACACCAACCTTTGAATGTACTCCACAATCATCATCCTTCACTCATTATTTCTCGAGCTAGCCATCGTTTATCTACTTGGATTCATGGTCAAGTGTCCT CTTAAGCCGGCCTTAGTAATTACTTTTATATTGTTGAAAGTCAGAAAGGACATTTAACTAAATGATTAGCCAAGCCTACAAAACCCTAGGACACCTTTACTTTGCGAAATGCATCCGCATCTAACTAATTGTAAATATTCTTATCAATAACAAA GAAAATATATTCACTAAAGGGTTAACTCTTAATTGCCCCCTTCGTGAAATGATAATCTTCAAAATATCCCCTAAATTTTTGTGAACCTCAATTTGCCCCTTACTTTTGTTAAGAAGACAAAAAAAATTCATTAACCGCCCCATTATTCCTTATCTTT TGAATAAAAATAAGGGTAAGTTCAAGTTTACAGTAGTTTAGGGGTATCTTGAAGAGTGTCATTTCATTAGGCGGTTATCTATAATTTACTCTTCAACACAATTAAATGTTTATGACACGAATATTAATTTGCCAATCTATCACATTTCATATTTT GGATGAATTATTATTTATTAATTATTGTGTTGTAATAAATTTACATGATTAAATTTTTATCTTTGATTAGTTTAAGTTTGTATGTTGTTTATTATTTCTTAATCATATATTCTATTTTATTTTATCTTTAAAAGTTGAATTATGAAAGTATTTCAGTGT TAATCAAGTTGTTAATCGTATTATCGTGTCTTAATCGTATAATTGTGTCTTAATTGTGTTATCATAATCATTTCATAATCGTGTAATTGTGTCAAATAGTAACATGTCAAATGTGTTGCGTCGTGACGTGTCACCTATTTATTAAATGGGTCGTAT CTGTGTTTGAATTTTTTGACATGATTATAAATGTGTCGGATCCATATTTAGTAAAATTGGACATGACACGTTTATGACATGCCACAAATATAACATGATTACTTTTACACAAATTGCCACCTCTAATATTTAACATGATTTTTATGTTGTATATAA CATTCATGTGATTAGTGAGATGTTATATCATGTTTGGATTTCTCTTAGCTTATTGGTTAATGCTAAATAATTAGTTAGTTAAATTTCTCAATGCGTTTAACTAATTTTCCTTTTACTTTACTCAATTGAATATTAATTCTCTATACCTTTAGTCGTCC TCTTTTTTTCTCTTTCCCTCTTTTTTTTCCCCCAATGGCTTGTTCTGTGTAAGCTATAGATGGTTAGGCCATCTCCAATTTGACACCAAAACTAACACCAAACCACTTTTTACACTATTTTTGGTATAAAATGCTTCTCCAATGTAACACCAAAACTA ACACCAAAATGGTGTTATGAAATAGTATTGCACCAAATATGGTGCAATACTATTCATTTGGTGTTAGGGAAAAAAAGTCACAAATATCCTTATTACTTTTTTTTATTTACATATATGTCCCATTTTATATAATATTTTAAGTTATTATTATTATTTT TATTCATTTTTTAATTTTTGAATATTATACTAATATTTAAATAAAAATATTCAACACAAAAATTAATAATAAAAATAAAATGCCCAATAACATTAAATTACTACATTACATATTACAAAATTAAAAATACAAAATTAAATTAACACAAATTAAAAA AGCTACTACTACTAGCATAAATTAGAACAATTATTTGATAATCTGGGAGATTATTATTAGATCTTCCAATATCATTAAAATATTGGCCAATATTATTGGAAGTATTTTGAGATGCTTGGCCTTGTTGAGATCTTCTCTGTAAAATCTTAATTTGTT CATTTTGAAAATATTGATTTAGTTAATTAAATGCAAATTTAATAATTTTCTATATTTTTTTTGTTAATCGATATAAAATTACTTATTACTTAATTATCTATATTAATAGTAATATTTATTGTGTTAATTGAATTAATTCTTGAGAATTAAAAGAATAA AAAGAAGAATATTCTTTTTTGTGTAAAATTTGGTGTAAATGGGTTGGAGATGAAATAGTAATTTTGGTGTGAAAAAAGTACATTTTTGGTGTTGGTGTTACACCAAAATGATATTAACCATCGGAGATGCCCTTAGTTATGGAAAGTTATAAA TTAAAGAGTCCAAAATGAGAACAACTAAAGGGGACAAAAGCATTCTTAAAAAAGAGCAAGAAATTTATCAGAGTGGGATTGTAAGCTTGAGGTTATCGTTACACTTTACATATATAGAACTGCAATTAACAACATTTATGGATTTATAAAGC AGAAATATATTGAAACAATTTATATATGTTTACATTTGTGTCGAAATAATTAATAAACAGTGACATCCGGTTTGAATTTTTTGTATATGAAAAAAATATCAAAATAAGATTAATTTAAAGCAATTCGAAGTTCATTAATTTATATTTGTTATCAC ATCATAGAAAGTATAAGTCTAAGGTAATTATTGATGAAGACTTGGATAAATCTTTTAATTAGTAATATTTATAAATGATGTTGTAGTACTTTGAGAAGCATCTGTAGCTGTCATAGTTAAAGAGCTCCACATGAGAAATTGCACATGTTAATCA ATTGATTACCAAATATCTATCTTATTCTAGCATTATATGATACGAACTCATTAAAACTTCTTTATATATGTGCATGTTCTATGTTTTGTTATATTTGTATGCATATAAATCTAATCATGGTTAGCTTGAACCTAGACTTCTACCTTGAAACATTTTTT TCATTCTTTTGTGAAAAATGAGAAAATCTAGATTTGAGTCTTTTTCAAGCCTTTAGATCAAGTAATGAGTAAACTAAAAGCAAATGAGATGTTAAGAACTACAAATTTCTTTAATATTTTACCTAATTGCAGCTGAGTTGTTCTTTGTTTTTTTCT TTTTGCTCCTCATGATTTATGGAGAAGATGATCTCCAAGCCTAAAAATAAATCATTTGATTATTCAATCATTTGCCTTTTTAAAAAAAATTATTCTTCAATCATAAAAGCAAATGAGATTTGAGGAACTCCAAATTTCTTTTAATATTTTACTTAA TTACATCCATTTCTTTCTTGTCTTTATTTTTTCCTTCCTCGTGATCTTCTTCCCTTCCAACTAAAGAAATGGAGAAAATTAATCTTCAAAGGCTGAGCACCATAGTCTCCTCCACTATATGTAAATCTNNNNNNNNNNNNNNNNNNNNNNNNN NNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNCCTGCATTACTCATAAAGCGTTCTATCCGTCCTCTCAAACAAGTCGAAATGCAAATACTCCGACTA TTTTATCTCTCTTCATTCTCTTTCCGTAGCCTACCCACAAGGATATGATATTGGCCACTTTAAGCTAGCATAGGTTTGTTATTGGGACCATTTCTAAAACCTTGTATTATCTGTGAGACAACATGTAGATACGTGCCCTGGTTATATATGCTAACC AAATATTGGCAGACTTCATTACATCGCTCACATCTCATCCCACTTTTGCCTTCTAACTTGCTTTATAGATGCAAGCGTGTTATTCATAAGTTTACTCATCTTAAAAACTATATCAAATATCCAACGGACTATACAATATCGGTCATCTTAGTCTAA CTTCTAAAGTTGCTCTTATTATCCTCCATTCTAAAGGAACCACGCTTGATGACATTAGCCAAGCACAAATGATTATTTTACATTTACATGTAGGATTGATTTTTATTTTTGAAACATCATCTAGGTTTATTGGATTTGCTAGAAGCAGAGAACCT ATTCGACTAAAATATATATATATACATTGGCAATAGACATCCCTAAATATAATAAGCAAGCTTTGATATAATCTTCCCAGAACCTCATTTGAAATAGACATAATGTTCGCTCTTTTCTTTTTTACCTCCTATTTAGTCGATGCTTTTCTTTTACATC TTACTTTATTTTTCTCTTTTATTTTACTTTGTAGACTTTTACCAATAATACAAAGTCTCTGTCGGACTCATAATATTTTTATCTTCTTGGATCTTTACATTTGTACTTAACGCTCCTTTCTTCTCTATATTTATTCCTCATTCACTTAAATTATATTCCGT TCGATGTATTGAAATGTTGCTCGTCTCCATCGATTCCCTCCTATTCAAACTCTTACACTCCTCGTATAACCTTCTCATGGTACAAGAATATATTTTTCTCAAATTTTCTTCTTTAAAATGCTTTGATGAAATATTCCAAAAGCAATACTGAATGGAA TAAAATGTTCACGCTAATAGTTGTCCTAATATACAGCTTTTATAATTAGTCAAAATAGAACATACTCCTCAAAAAAATACACCATATCCTGACTTACGTAAACACAAAAGTAAATCAATTTTACAACATAGAGATCTAAGGAACAAGAAAAACT ATTCAATCATCAAAGAGCTCATATAAACATCCTATAAACCCGAAAGAGCAGGGTGCACCCTAATTATAAATATATTTGTGCTTTGTGGGAGTAGTTGGTGTATTGGGAAATATTACTTCGGAGGTGTACTGCAGGTGATTAGTTTCACTTTGG AAGTTTATCATTGATGCAAGTGTTGCATTATTAGTTGGTTGAAAATTTGGGAATGCCCATACTTTAGGGGAAACTCTTCCCAAATTTTTTGCAAGGACAAGGACATTGTTTTGAGTGATTGTTCCATTACAACCCAGGTTTAAGAATGAATTTT TTGCCTCATGTTAGTCTCATTTTTACTTGGAAATTCTTTTTAGGTTTTGAATAAATTTCACTATTGCTGAATTCCCGCTAGAGATTCTTATGGTTGAATTTCTGTTTTGAAGTAATAGGAAGTGCAAAGCTAACTTTGAGTAGCCTACTATAGTAC TATTGCATCTTCACGGACGATAGAAATGACCTTTTTGTACACTTGTGATAGAGCATTAGAACTGACTTTCTCAACAAATTCTCAAGACTGATATGGCTACATGACATATAGAATTGCTGAAGAAGAAGACGGCATTCTTTAATACCTTTTCAAG GTCTGCGCCTCTGTTTTGGCATTTTAACTACCAAAAAGGGATGCTCATCCCTACTTTCTGCCGGCCTTCATAATTACCACTTTTCATTGTATTTTCCTCTTTCTAATATTTTGAAGAACCACCATATCTCCACAATCTAGAAAATTATTACATGATC AAGAAATCTGAATAATTAATACTGATAATGCAGTTTGGCCAGCATCCTTGAGCGCTACCAGAGTCGCATTGCAGAGGAAGCTGTAGCTTCAGGAGCACCTAAGGCAGAGGTAAACTGTTCTCTCTCTATTTAGAACACATTAGCAATGGCTA CGCTAGGTCTAAAGTCTAAACCTATCCTGTCTTCAGCATGCTCTCCATGAAGTGGATCAACTCCTAGGGATGAATTTATGCATCAAATTGTGAATAATCAATCTAAGTAAGACTAAGTAGTAGCTTTGACTTAATTTCCTATTATAGAGTGTTT GGACGAGAACTGTAAGGGTAACTGATATTTTCAATATTATACTCTGCCCATACAATTTCGCATATAACTTTAATATCTAAAACCCTCTAACTGGAAGTATTCTAGGTTCTCCTGAATATGTCAACCAACTTGGTCTAACCTGGTTTATACACTCA ATCAATCTATATGCTGACGAAGTTCTTTCTAAATGCCTATGGGATATGAAACCCAGTGAAAATGGCCAAAATTGGACTGTCATTGTGAATATGAGGACATGCTTGAGAGTGATGACCATTAAGTCACAGCCTCCTGCAGTATAATCCCTTAAT TTATGTCTGAGTACTATCTCCATACATGTGCCAATACAACATATCTGGGAATTTTGTAGGTAAACATAATGTTGTGACTACTCCTAGCATGCCTGATAGAAAAACTTCATACAACCGCATCGATTTGAATCAATTTAGTACTATTGCTTGGCTCT TACCTACCAACTTAAGTTATCAAGTCAGGCTTTAACTTAATATACTTTTGGAGCTTTTTTGATCAATAGGTCTAGTGTCCAAATCCTCCTGCGAGATTTTCCTCTGTAGTTGTTTCTATCTTCAATCTTCCTTTAGGCTCATCTCTATGTTTGGATTT GTTTATAGCTTTCATCTTTGGTTGAGCTTCTTATGCAGGACAAGACCATTTGGTATGGCCTTTGCATGTGAAGAGCTCCGTTAGGGCTTACTATGATGTTCTAGGTTACGACCTAATAGCTTTAGGGTCAATTGGTAAAGTAACACAGTACCA GAAACCTAGATATCCAGTGCTATTATACCCTGACATGGTTGCCTTTTGGGTTTGCCTTCTGTCTTCCATCTGTGGTTGGGCTTGGTCCACATAAAAGCTCATTTATACATCAGTCTTGGTCCACTTGACAAGCTGATTTATACATCGGTCTCGCT GATATAAGGGGTTGTCAAAGGTTAATAGTATTGTTCAAGTCAGGCAGGCCAGAATTAACGTCATTTATCTATAATGCATTTGCGTACAAAACTTTTCTTTTCCTTTTTTTTTTTTGCCTGGGTGGTTGGGGGGCCTCCAGGGGTTGTTTCATGG CTAACTTTCGGGATCTCTACACTGACAGGACCCTCAGGCTGGATATTCGAGCCTGCAACTGCTGAAAATAGTTCAAAGGCAATTTCTCCAAGTATTCTATACATTTTCCCAGAATGGCACATTATGTGTTTGTGTGTGTGTTTGTGTGTATTAC TTTATTGATCTGTATTTTGCATCCACAGGCAACTTGAAGGTCCAAACTTTGAGCAGTTCACTGCAGCTGATCTTCTGCAGTTAGAGAATCTACTCAACGATACACTGACAGAAATTAGAGCTGTGAAGGTCAGGATTATAGTACTTTTGCAATT GAAATCTCGAGTGTCTAAAAATAATAGGAATATGGTTATGAGATTACATATTGACTATCAAGAAGATAATGTTCATGTATTGTTGATGCCTTTGCCGCTTGCTCCTATATCTTGGTTCTCCTTTATTCAGACCTTTCCAAGAGATTATGATACTA CAAATTCTTAGTCCCATGTTTGATCATCTTAAAATTCAACCAAGCCAACAAATACACAAACAATCTGCAACTTCCATCAATACTGGTTGCTGATTTTCACACGTTTACCTAACTGAGAATGCTTTAACCACTGATTGCTATATCACATTAATCTTT TCTTTCCTGTCTATCCAAGCACTTCTTTTTGGTTTAGCAGCTTTTCACTAGGTTCAATTATCTGCAGCTGCTTACTGCTGAAACTATGGAGAAGACCTATGGTTGTGCGCGTGAACTATGAAGCGCACTTGAGTTTGCTTTCTGTAGCCCCTGTT GTTGAAATACTCGTTTTGCCACCAAAAATACCTTACCTTAATAATAGATTCTGGCCACCTTAAAGTAAAACAGAGGATGCACCCCTGCTCTAGTTTGACCTGTTGTCTGTCACTTGATTATTTGAACTTGGAACTACTATTGCATTATCAATTAG ATCCTGTATGCGCCTTTTTTTTGCGCCCTCCTTTGTGCCTGCTTGAGGCTTGTTATACTATCCATGAAGGTTCATTCCTATGCTCATTGGGCATCTTTGGTGCAGACACAACTGATGATGGGATCCATTAGAACACTCCAAGAGAAGGTACACTT AATTAAATACTTTCCATACCTTACTCTGTGGAATTGTTTATTCCCGGTAAATCTTGAAGAAAAACCTAAGCTCCTAGCTGGATGCAAATTGCAAGTGTCAACTTTTATGTGGTTAAAATGACTGCAAATATTTCTGTTTCACTACGACAGGGTCG TTTAACATTAGATATCATGGTTCCAATTTAGTTGGTGGAAAATTTTCTTCCTATTCATTAAAAAGCAGGATCTTATTGCAAGTGAAACTGTACAGCTGGCTGTGTCAACAGTAGGCTTTCAGAGTCTCCCGGGCTCATTTGTTTGCCATTTGCAT GAAGTTTTATCCAATAGATCCTTTTTATTTTCCATGCCTTTAATTCCTGCCTTCTGAGAAACCTACAGCTTTCTATTTCACTACTTTCTAGTTGTGAGTGTCCTTTGTCCTTGAGAACCTACCAGCTTTAAGAATTTCCAAGCACCTCATAATTGAT CAAAATCAGATTGGAAAAGGAAGAATGGAGATCTAATGCTGGTGAAACCATGTGATGATGCTGTCTTTTTGTGGCATAAAACCTTTGTACCTCCCTATATTAGGTAGCTAAGGTGGTAAAATGAAGGACCATATAGCACCTTAGTGCCGCTT TGGTTAATTTGAACACTACTATTCGAGCTTCTAATATAGTACTCTAATCTTGGATTATCATGTAATCAAGTCTAGTCAACCATATTCATGCATTCCATTGTATTAGAAATACTTTCTCAAAGATTGATATATCTGATGAAGAATCTAAAGCACCG GTGACATCTGGATAGCTAGACTTGTTTTCTAAGAGGGGAAAAAAAGGGTTGTTTTAGTTATTAGATTCATAATGAATCAGGGTTTGGCAGACGTTTAGTTTGTTGTCATGGTTACCTGAAAAACTCAATTTCTGCATTATCGCGACCACATCA CCTAATTTTTTTCTCGGTAATTTGTACAGCATGTGTGTGTACAGACATGCATGTATCTAATGAACAAGCTGGATCCCATCTGTGGATGTTTCTGGTTTATCATTTCATATTTCTTCTATTTTAAACAGGAAAAGCTGCTGAGAGAAGAGAAAAA GCTTCTAGAGGGAAAGGTAAAAGTATTACGTACTAAATGCATATCTTGTTCTTCAAAAGAGTCTAATAATTGTCAGTTGTCACTTACAGATTGCAGCAGGGGAGAAGAGTGGTAATGATACAACAGCAGCGTCTATGGCAACCCTCCCTTTG CTTAGATAAGCTGAGGAGAGGCGCTTTTCCAAGTGTACTTTTCAGCCGCCCATCTGTTGTACTTGGTGAATATGTAAGTCTGTTATCAGTGGCTAATTCTACGTAGCGAATAAATTGTAATGCTCTGTACTACTATAAGGCGGCTAGATAATC AGGGAGCGGCTTCTGTAGAAGTTCCCACTGACTACGTTTTTATCATCTGCCATCATCAAACTTGCAGAATGTTTTATTATTTTGTGATTTGTGTAACCATGCTTGAGCAGAAATGCAACAACATATGAGACAACAATTTATAAATAATGGCCTT AAAAAACTGCACAAACTGGCAGGAGGTTAGATGAAATATTATATGAAGAGTAATGATACAGTTATAAATTTTTATATAAATTTATTTTGTACAAACTGATGTGGTATTAATTCATTGATTGAATTAAATATCTTTTGGTTTATATGATTTATTTT TATTATTTTATATTTTTATTTAATTAATAAATTAATGTCACTTCAATTTGTACAAAATAAATTTATACAAGAATTTGTATTTGTATCATCACTCGTTATGAGACATAAAATAATAGCCTTGTTGACCAGGTGCCATGCCAATGAAGGTTCAGCTTG GGGGCTATTTGCACCTGTCAAGCCGTGGGCCGTGCTTGGCAAGCAGCCTGTTGCTTTTTGCTTTTTCGGAATTGGGCTTCTTGTTCTAAAATAGAAAAAATATACATACATACATATATATATATATATATATATATATATTATTAGGGAAATTT ACAAAAATAACCATTAATATTTTATTATTTTCGTAATTAACCCTCTCAACTTTTTTCTATCAACATTAGCCAAACACAAGCCTTTCTTCCCAAATTACCCTTCTTTACAGAACAAAACCAACATTGTTTCTCCCATCTCGCCAAACCAAAAGCAGCT TTTCATTCTTGAATCATAAGCCAACGGCGACCGCAAAAGACAACTGTAAAAGTGATATCCGATCATAATCTATCCAAAT

80

81 Reult

82

83 Reult Section 1. Hormonal inhibition of flower induction: Gibberellin In citru, when it reache it final ize fruit inhibit flower induction by repreing the expreion of flowering gene. But it i not known which long (or hort) ditance ignal fruit end produce to hamper the proce. GA have been thought to be thi ignal. But a directed tranport of GA from the fruit to the bud or the leaf ha not yet been clearly demontrated. Beide, the relationhip between endogenou GA and flower induction it i not clearly known. The objective of thi ection i to dicu the effect of fruit in endogenou GA ynthei and tranport and flowering. The Moncada mandarin tree preent an abolutely natural 100% alternation between flowering and fruiting. Thu, it wa elected a a model ytem to tudy the proce. The tree produce up to kg year -1 (ON-tree). The fruit ripen in January and harvet-time lat until February. Thereafter, in pring, the tree i not able to flower and only develop vegetative hoot (Fig. 1.1). Therefore, the tree doe not produce a fruit during a complete year (OFF-tree). Flowering i induced in autumn-winter, and the following pring the tree prout and flower profuely, developing motly generative hoot (inflorecence), but alo vegetative hoot and mixed hoot with flower and leave (Fig. 1.1). A large number of thee flower et a fruit, the tree becoming ON again. A B Figure 1.1 Flowering of Moncada mandarin ON (with yield) and OFF (without yield) tree. A: flowering intenity. B: type of prouted hoot. V: vegetative hoot; M: mixed-type hoot (flower and leave); G: generative hoot (inflorecence or ingle flower). Data are mean ± ES of 6 tree. *: indicate ignificant difference (p 0.05). 81

84 Reult 1.1 Fruit development and flower induction gene expreion To tudy flower induction, FT gene expreion (main promoter) and FLC gene expreion (repreor) were determined. FT gene ha been decribed a the florigen in everal pecie, Citru included (Nihikawa et al., 2007). But it ha alo been reported to regulate both vegetative and reproductive development in tree due to the preence of at leat 2 paralog (FT1 and FT2) (Hu et al., 2011). In Citru, there are 3 paralog encoding thi type of protein, CiFT1, CiFT2, CiFT3. Their role regulating vegetative or reproductive development ha not yet been demontrated. To determine which of thee paralog are related to flower induction, a preliminary experiment wa performed comparing CiFT1, CiFT2 and CiFT3 expreion in juvenile tree (without the ability to flower) and adult OFF-tree. The relative expreion wa tudied in September, November and February, coinciding with the fall prouting, the floral induction tage and pring prouting, repectively (Fig. 1.2). CiFT1 did not differ ignificantly in the juvenile and adult tree, howing increaed expreion in fall and pring prouting tage, and reduced expreion in the floral induction tage. On the other hand, CiFT2 wa not expreed in juvenile plant in the period tudied, while in adult tree the relative expreion of CiFT2 wa increaed 12-fold in the floral induction tage and 35-fold during flowering. The relative expreion of the CiFT3 gene howed no variation in adult tree and a ignificant decreae over time in juvenile tree. In ummary, only CiFT2 howed ignificant difference in the floral induction period between adult and juvenile plant, and, accordingly to thi, it may regulate flower induction in Citru. Additionally, CiFT1 correlated with prouting in juvenile and adult tree, uggeting a role in the control of vegetative growth, a occur in poplar (Hu et al., 2011). The citru FLC gene ha been equenced (Supplementary data S1) and characterized. Thu, the FLC gene wa introduced in Arabidopi thaliana to tudy the delaying effect on flowering time. Thi wa evaluated by counting the leave at the moment of flowering (Fig. 1.3). Arabidopi plant tranformed with CcFLC had more leave than the Arabidopi control plant, which indicate that CcFLC-Arabidopi plant preented delayed flowering. 82

85 Reult * * * * * Figure 1.2 CiFT, CiFT2 and CiFT3 expreion in leave of Moncada mandarin adult tree and leave of a juvenile plant of Cleopatra mandarin. Data are mean ± ES of 3 qrt-pcr replicate. *: indicate ignificant difference (p 0.05). * Figure 1.3 Number of leave before flowering of Arabidopi thaliana cv. Columbia. Influence of the FLC gene expreion. Data are mean ± ES of 40 plant by column. *: indicate ignificant difference (p 0.05). 83

86 Reult From the beginning of July, Moncada mandarin fruit underwent a phae of rapid development to reach it final ize around mid-november. From the end of September until mid-november, the fruit grew to about 20% of it final ize (Fig. 1.4A). Thi period till coincided with high temperature (Fig. 1.4B) and in September and October the color of the peel remained dark green (Fig. 1.4A). Thereafter, the end of fruit growth coincided with the beginning of fruit degreening, and chlorophyll content ignificantly declined (from 350 mg/g PF in October to 110 mg/g PF in January) (Fig. 1.4C). The color of the peel changed progreively from pale green to orange a/b (Hunter) around the end of November (Fig. 1.4A) coinciding with a ignificant reduction in temperature (Fig. 1.4B). The ditinctive orange color of the fruit wa gradually developed until late January (Fig. 1.4A). Thi effect wa produced by the increae in carotenoid content (from OD 440 at the end of November to OD 440 at the end of January) (Fig. 1.4D). The end of fruit growth and the beginning of fruit ripening, due to the decreae in temperature, coincided with the upregulation of FLC and the inhibition of CiFT2. On the other hand, CiFT2 wa ignificantly upregulated and FLC downregulated in the leave of the OFF tree (Fig. 1.4E and F). 84

87 Reult * * * * * * * * Figure 1.4 Evolution of fruit diameter and the color of the peel (a/b Hunter) (A), air temperature (B), chlorophyll content in the peel (C), carotenoid content (D) and CiFT2 and FLC expreion related to chilling hour (E-F) in ON and OFF tree of Moncada mandarin. The data of fruit diameter and color are the mean ± SD of 10 replicate ample. The data for chlorophyll content are the mean ± SD of two meaurement. The data of gene expreion are mean ± ES of 3 qrt-pcr replicate. *: indicate ignificant difference (p 0.05). 85

88 Reult 1.2 Time-coure of gibberellin metabolim in the exocarp, node and leave from ON and OFF tree Higher concentration were formed for gibberellin from the 13-hydroxilation pathway than gibberellin from the non-hydroxilation pathway in both ON and OFF tree (Fig. 1.5 and 1.6). In ON tree, the beginning of chlorophyll breakdown in the flavedo coincided with a ignificant reduction in GA12 GA19, GA20 and GA29. Concentration of GA12, GA20 and GA29 rapidly reduced from 0.82 ng g -1, 0.88 ng g -1 and 0.81 ng g -1 in September to 0 in October. GA19 concentration dropped progreively from 2.47 ng g -1 in September to 0 in January (Fig. 1.5). Neither GA1 or GA8 wa found in the flavedo. Gambetta et al. (2012) uggeted that citru fruit might export GA to change color, which i in agreement with thee reult. Concomitantly, GA20 and GA1 concentration ignificantly increaed (4-fold) in the leave during November, a well a the GA20 catabolite, GA29. The GA20 precuror, GA44, alo followed thi trend although reaching a ignificantly lower concentration (Fig. 1.5). Thi increae in GA concentration wa not upported by a ignificant increae in GA ynthei, a deduced from the analyi of GA20ox1 (which catalyze the GA52-GA44-GA19-GA20) or GA3ox1 gene expreion in the ON-tree leave during October and November (Fig. 1.7). Although thee experiment do not demontrate the tranport of GA to the leave, it might be uggeted, according to the reult. On the other hand, GA concentration in the leave from the OFF tree wa high in September and low in November. Specifically, in the younget leave (coming from the fall fluh), GA12, GA19 and GA1 concentration fell from 19 ng g -1, 2.1 ng g -1 and 2.9 ng g -1 in September to 0 in October, being almot nil until the end of the tudy (Fig. 1.6). Accordingly, GA20ox1 and GA3ox1 gene expreion alo peaked in September and progreively diminihed until the end of the tudy (Fig. 1.7). Leave from the pring fluh howed a imilar trend for thee GA but at lower concentration. The highet GA ynthei hown in September may be due to the development of the leave of the fall fluh. A detailed repreentation i hown in Figure 1.8. The growing leave from the fall fluh had higher GA1 and IAA concentration than the adult leave from the pring fluh. Moreover, thee adult leave upported the growth of the young leave, i.e the adult leave of a node with a growing hoot had higher GA1 and IAA concentration than the adult leave of a node without a growing hoot (Fig. 1.8). In the node, GA1 concentration followed the ame trend in ON and OFF tree being lightly higher in ON tree (1.71 ng g -1 ) compared to OFF tree (1.10 ng g -1 ) (Fig 86

89 Reult 1.5 and 1.6). No GA ynthei wa found during the dormant period of the bud, which ugget that GA1 found in the node i the reult of GA-ynthei and tranport from other organ (i.e. leaf or fruit). On the other hand, GA20ox1 and GA3ox1 gene expreion wa triggered at the bud prouting tage (Fig. 1.7) Although the non-hydroxilation pathway i quantitatively le important in Citru, it i intereting to note the time-coure of GA4 concentration in the fruit from ON tree. GA4 increaed from September to October and remained contant at a concentration of 2.3 ng g -1 until the fruit color wa complete. When carotenoid ynthei wa triggered, GA4 wa depleted (Fig.1.5). In brief, in thi experiment in fruiting hoot from ON tree, a peak in GA wa found in the leave when FT gene expreion wa hampered. Thi alo coincided with the moment the fruit reached it maximum ize and began the ripening tage. GA in the fruit were alo depleted; GA1 wa continuouly found in the node and GA ynthei in the leaf did not increae. Therefore, reult might ugget a hort-ditance tranport from the fruit to the proximal node and leave. Three major quetion arie: 1) I thi GA increae reponible for FT inhibition in ON-tree? 2) If GA ynthei i inhibited, can the bud from ON-tree flower? 3) If tranport between the fruit and bud i interrupted, can bud flower? 87

90 Reult Figure 1.5 Endogenou gibberellin content of the non-13-hydroxylation pathway, (GA12, GA15, GA24, GA9, GA4, and catabolite GA51), and 13-hydroxylation pathway, (GA53, GA44, GA19, GA20, GA1, and catabolite GA29, GA8), in ON tree of Moncada mandarin. GA were meaured in the pring leave, bud and exocarp. Date are mean ± ES of two et of 10 ingle flowered leafy hoot. 88

91 Reult Figure 1.6 Endogenou gibberellin content of the non-13-hydroxylation pathway, (GA12, GA15, GA24, GA9, GA4, and catabolite GA51), and 13-hydroxylation pathway, (GA53, GA44, GA19, GA20, GA1, and catabolite GA29, GA8), in OFF tree of Moncada mandarin. GA were meaured in the pring leave, bud and exocarp. Date are mean ± ES of two et of 10 ingle flowered leafy hoot. 89

92 Reult * * * * * Figure 1.7 GA20ox1 and GA3ox1 expreion in Moncada mandarin leave and bud from September to February. Data are mean ± ES of 3 qrt-pcr replicate. *: indicate ignificant difference (p 0.05). 90

93 Reult * * * * * * * Figure 1.8 Endogenou gibberellin content of non-hydroxylation pathway, GA 12, GA 15, GA 24, GA 9, GA 4, and catabolite GA 51, and 13-hydroxylation pathway, GA 53, GA 44, GA 19, GA 20, GA 1, and catabolite GA 29, GA 8, in OFF tree of Moncada mandarin. GA were meaured in the pring (SP) adult leave from node with and without new fall hoot, and in fall leave. Date are mean ± ES of two et of 10 vegetative hoot. *: indicate ignificant difference (p 0.05). 91

94 Reult 1.3 Exogenou control of flowering: Treatment with Paclobutrazol (PBZ) and GA3 Medium-yield tree PBZ (1 g tree 1 ) applied at the flower induction period ignificantly increaed flowering intenity (70%) in Hernandina Clementine mandarin (70 flower per 100 node) a compared to untreated tree (40 flower per 100 node) (Fig. 1.9). Doubling the amount of PBZ applied (2 g tree 1 ) did not improve the repone. Gibberellic acid had the oppoite effect a the number of flower per 100 node dropped by 37%, from 40 (untreated tree) to 25 flower per 100 node (50 mg l 1 GA3 treated-tree) (Fig. 1.9); the effect of GA3 prevailed over that of PBZ when the two regulator were applied together (Fig. 1.9). Figure 1.9 Effect of paclobutrazol and gibberellic acid applied to Hernandina Clementine mandarin. Value are the mean of 10 tree per treatment. Standard error are given a vertical bar. Thi effect doe not depend on the pecie. Thu, the application of 40 mg L -1 of GA3 at the flower bud inductive period alo reduced the number of flower per 100 node of Salutiana weet orange by 72% (p 0.01) in comparion to control tree (Table 1.1). Thi treatment alo reduced bud prouting by 40% (p 0.05) compared to the control. Leafle ingle-flowered hoot and leafle inflorecence were reduced on average from 3.6 to 1.3 and from 5.1 to 1.9 per 100 node, repectively, due to treatment, with difference being tatitically ignificant (p 0.05). Among flowered leafy hoot, only inflorecence were ignificantly reduced by application GA3 from 5.6 to 1.2 per 100 node (p 0.05). Converely, GA3 ignificantly increaed vegetative hoot from 3.8 to 9.0 (p 0.05). PBZ applied at a concentration of 2,000 mg L -1 produced an oppoite 92

95 Reult trend. The number of flower per 100 node and percentage of prouted bud were increaed by 123% and 74%, repectively, compared to the control (p 0.05). For leafy hoot, neither ingle-flowered nor inflorecence were ignificantly altered by thi treatment. However, for the leafle hoot, both ingle-flowered and inflorecence were affected riing from 3.6 to 8.8 (p 0.05) and from 5.1 to 16.3 hoot per 100 node (p 0.01), repectively. PBZ ignificantly reduced the number of vegetative hoot per 100 node (0.8) compared to the control (3.8; p 0.05). Interetingly, the number of flower per hoot of both leafy and leafle inflorecence wa not ignificantly altered by GA3 in comparion to the control, with 3.9 and 3.4 flower per leafy inflorecence and 3.7 and 3.4 flower per leafle inflorecence, repectively, wherea PBZ increaed flower number ignificantly, up to 4.9 and 4.2 flower per hoot for leafy and leafle inflorecence, repectively (p 0.05). Neither GA3 nor PBZ changed the number of leave per hoot in any cae, even that of vegetative hoot (data not hown). The time coure of the relative expreion of CiFT in leave throughout the tudy wa ignificantly affected by GA3 (Fig. 1.10A). Significant difference in mrna trancript between GA3 treated tree and control tree were detected from 8 day after treatment (DAT) onward. The expreion in control tree leave increaed progreively up to 32 DAT (mid-january), decreaing thereafter to almot the initial value (Fig. 1.10A). Gene expreion in leave of GA3 treated tree paralleled that of control tree but wa reduced by 16% on average, except for 80 DAT (late February) when no ignificant difference were found between control and treated tree (Fig. 1.10A). On the other hand, PBZ treatment tree ignificantly booted the relative expreion of CiFT in leave (by 30% on average) from 8 DAT up to the end of February, which i the onet of bud prouting. In thi cae, leaf gene expreion alo paralleled that of control tree, but with ignificantly higher value throughout the entire period tudied. Figure 1.10B how the time coure of the relative expreion of the FLC-like gene in leave from control and GA3 and PBZ treated tree. From early December to the onet of bud break, no difference in gene expreion were found. Activity in leave remained almot tationary between 0.70 and 1.03, regardle of the treatment. 93

96 Reult Table 1.1 Effect of GA 3 (40 mg L -1 ) and PBZ (2.000 mg L -1 ) applied to entire tree during the floral bud inductive period (10 December) on bud prouting and flowering of Salutiana weet orange tree. Each value i the mean of ix tree ± SE. Different letter in the ame line indicate ignificant difference (p 0.05). a Sprouted bud expreed a percent of total bud. b Number of hoot and flower expreed per 100 node. A B Figure 1.10 Effect of gibberellic acid (GA 3, 40 mg L -1 ) and PBZ (2.000 mg L -1 ) applied on December 10th on the time coure of CiFT and FLC-like expreion in the leave of the Salutiana weet orange. Different letter for the ame ampling date indicate ignificant difference (p 0.05). 94

97 Reult Control and GA3-treated tree did not differ ignificantly in the relative expreion of CLFY in leave, remaining almot contant between 0.66 and 1.21 throughout the tudy (Fig. 1.11). However, in PBZ-treated tree there wa no treatment effect until late in January when mrna trancript in the leaf ignificantly increaed 1.8-fold (1.94) compared to control (1.12). In pite of the ubequent decline, a ignificantly higher relative expreion of CLFY in comparion to control and GA3-treated tree wa recorded in PBZ-treated leave at bud break. Thi effect coincided with the ignificant inhibition of GA20ox1 activity, which wa oberved (Fig. 1.11). Figure 1.11 Effect of gibberellic acid (GA 3, 40 mg L -1 ) and PBZ (2.000 mg L -1 ) applied on December 10th on the time coure of CLFY and GA20ox1 expreion in the leave of the Salutiana weet orange. Different letter for the ame ampling date indicate ignificant difference (p 0.05). Reult ugget that, under medium-yield condition, treatment that modify GA metabolim can quantitatively modify flowering, i.e. the number of flowering hoot produced, by interfering in either flower induction gene (FT) or flower differentiation gene (LFY). The inhibitory flowering effect of GA ha been previouly related with the leaf and flower induction (Guardiola et al., 1982) but alo with the bud and flower differentiation (García-Lui et al., 1986; Goldberg-Moller et al., 2013). 95

98 Reult ON and OFF tree Figure 1.12 how the influence of fruit load on flowering in the following pring for untreated and PBZ-treated tree. A ignificant invere relationhip wa found between fruit load and flowering intenity in control and PBZ-treated tree of Salutiana weet orange (A) and Afourer mandarin (B). Reult alo ugget that PBZ increaed flowering only in low- and medium-yield tree and not in thoe with a heavy fruit yield, the threhold value being about kg tree 1 for Salutiana and Afourer. For larger crop load, no ignificant crop load flowering intenity relationhip wa detected, wherea for lower crop load a cloe relationhip wa oberved. Thu, for intance, when PBZ wa applied to both Salutiana weet orange and Afourer mandarin tree bearing 50 kg tree 1, the number of flower per 100 node doubled from 33 to 64 and from 59 to 110, repectively; thi effect wa not achieved when PBZ wa applied to tree with a crop load 125 kg tree 1 (Fig. 1.12A and 1.12B). A B Figure 1.12 (A-B) Fruit load and flowering intenity relationhip in Salutiana weet orange (A) and Afourer mandarin (B) control and PBZ-treated tree (n = 23). PBZ (1 g tree -1 ) wa applied to the oil wherea GA 3 (50 mg l -1 ) wa applied a a foliar pray on November 25. Regreion [y = (a + b* ln x) 2 ] are ignificant at p < Furthermore, under ON-tree condition, PBZ failed to promote flowering regardle of the doe applied, date or method of treatment or cultivar (Fig. 1.13). There wa a ignificant increae in flowering intenity the following pring in OFF-tree of Salutiana weet orange and Afourer mandarin receiving 1 g tree 1 or 10 g tree 1 to 96

99 Reult the oil during the floral bud inductive period. However, ON-tree receiving the ame doe flowered in proportion imilar to untreated ON-tree, i.e. next to nothing (Fig. 1.13A). A imilar repone wa obtained with 15 g tree 1 PBZ prayed onto the canopie of ON- and OFF-tree of Moncada mandarin (Fig. 1.13A). Advancing the treatment date of PBZ (1 g tree 1 ) to earlier in the ret period (October) or delaying it until floral bud differentiation (February) did not counteract the effect of fruit load, nor did it produce the effect obtained when the ame doe of PBZ wa applied in the floral bud inductive period (November), regardle of the cultivar or method of treatment (Fig. 1.13B). A B Figure 1.13 (A) Effect of PBZ doe (0, 1, 10, 15 g tree -1 ) applied at floral bud inductive period (November 20-25) and (B) treatment date (October 10-15, November 20-25, February 20-24) (1 g tree -1 ) (on flowering) in ON- and OFF-tree of Salutiana weet orange, Hernandina clementine mandarin a well a Afourer and Moncada hybrid mandarin. PBZ wa applied to the oil or a a foliar pray, and GA 3 wa applied a a foliar pray. Data are mean of 6-10 tree per cultivar. Standard error are given a vertical bar. Different letter for the ame cultivar indicate ignificant difference (p 0.05). CT: control tree. Treatment were alo applied directly to the bud to further demontrate it inability to flower in the preence of fruit. Thu, 2.5 µg PBZ applied locally to the bud of OFFtree at the floral bud inductive period promoted flowering (113.5 flower per 100 node) when compared to untreated bud (59.5 flower per 100 node), but when PBZ wa applied to the bud of ON-tree, it did not affect the number of flower (7.6) compared to untreated ON-bud (7.8 flower per 100 node). 97

100 Reult 1.4 Branch independence: the hort-ditance effect of the fruit. Fruit-hoot phloem tranport interruption by peduncle girdling Tree of Salutiana weet orange that bear 50 kg tree -1 on average are conidered medium-yield tree (Fig. 1.12). Under thee condition, the tree i able to flower in pring. However, the inhibitory effect of the fruit i alo evident if ON and OFF branche in the ame tree are tudied eparately. For thi reearch prouting and flowering were evaluated in pring attending to the poition of the preceding yield in branche of about node, i.e, at leat 2 year old. Reult indicate a clear invere relationhip between fruit poition and flowering, uggeting a hort-ditance effect of the fruit in flower induction inhibition (Fig. 1.14). A B Figure 1.14 Schematic repreentation of prouting and flowering in pring of two branche ( node, 2 year old) from the ame tree of Salutiana weet orange. Branche had 13 (A) and 4 (B) fruit and flowered very low and very high, repectively. An invere relationhip between fruit poition and flowering i oberved. 98

101 Reult Therefore, in order to induce flowering in fruit bearing branche, an experiment wa deigned to interfere in thi hort ditance dominance. Spring branche with 8 node and a fruit in terminal poition were girdled in the fruit peduncle at the end of tage two of fruit development, i.e., end of ummer (Augut 25, 2014) (Fig. 1.15A). Ungirdled branche and OFF branche were elected for comparion. Girdling did not caue the abciion of the fruit, which remained on the branch until harvet. However, girdling ignificantly modified fruit development. At harvet, 7 month after girdling (March), girdled fruit were ignificantly maller and their exocarp were greener than control fruit. The pulp changed color a did control fruit but the girdled fruit accumulated fewer TSS (total oluble olid) (Table 1.2). Table 1.2 Characteritic of fruit from ON and Girdled hoot (7 month after girdling) in Afourer hybrid mandarin. Each value i the mean of 10 fruit ± SE. Control Girdling Significance Weight (g) 92.8± ±8.7 * Diameter (mm) 57.5±4,8 43.2±2.9 * Exocarp color a 28.8± ±0.89 * b 33.9± ±1.2 * a/b 0.9± ±0.1 * Pulp color a 10.8± ±0.4 N b 19.6± ±0.3 N a/b 0.5± ±0.02 N TSS (ºBrix) 16.4± ±0.2 * Eleven day after girdling, the proximal bud prouted in 15% of the girdled branche, and 22 day after girdling bud prouting wa achieved in 45% of the branche. In contrat, un-girdled branche (ON) did not prout. Non-bearing branche (OFF) alo prouted naturally (fall fluh) (Fig. 1.15). Bud prouting occurred in the 4 node cloet to the girdle. From node #5 to node #8, bud did not prout (Fig. 1.16). All the prouted bud produced only vegetative hoot, which were longer in the girdled branche than in the OFF branche (Fig. 1.17). 99

102 Reult SAM A Girdling B Figure 1.15 A. Repreentative figure of the type of hoot ued for the experiment, vegetative hoot (OFF), ingle flowered leafy hoot (ON and Girdling). B. Evolution of the percentage of prouting at the end of ummer in ingle flowered leafy hoot (ON) and vegetative hoot (OFF) of Afourer mandarin. Effect of girdling the peduncle of the fruit (Girdling). Each value i the mean of 50 hoot. Vertical bar repreent the tandard error. The blue bar hown on the left of the figure correpond to LSD interval. Figure 1.16 Ditribution of prouting at the end of ummer along the node in ingle flowered leafy hoot (ON) and vegetative hoot of Afourer mandarin. Effect of girdling the peduncle of the fruit (Girdling). Each value i the mean of 50 hoot. Node #1 correpond to the node cloet to the fruit/apex, while node #5 i the farthet. Vertical bar repreent the tandard error. The blue bar hown on the left of the figure correpond to LSD interval. 100

103 Reult Figure 1.17 Length of the hoot at the end of ummer, along the node in ingle flowered leafy hoot (ON) vegetative hoot (OFF) of Afourer mandarin. Effect of girdling the peduncle of the fruit (Girdling). Each value i the mean of 50 hoot. Node #1 correpond to node cloet to the fruit/apex, while node #5 i the farthet. Vertical bar repreent the tandard error. The blue bar hown on the left of the figure correpond to LSD interval. Bud prouting wa related to change in the hormonal balance in the node. In the un-girdled ON branche, the IAA content in the fruit wa 7 ng g -1 in both the exocarp (ex) and the meocarp+pulp (me+p) (Fig. 1.18). Further, the IAA content wa ignificantly higher in the node than in the fruit. The third node howed the maximum IAA level (200 ng g 1 ). No GA1 wa found in the fruit while in the node GA1 concentration wa around 1 ng g -1 (Fig. 1.18). Girdling modified the hormonal balance; thu, it ignificantly reduced IAA concentration in the pulp (2 ng g -1 ) and in the node (100 ng g -1 in the third node), and alo t-zeatin in the node compared to the ungirdled branch (Fig. 1.18). On the other hand, girdling ignificantly increaed GA1 in the node (3.4 ng g -1 ). Other hormone (GA4, ABA and JA) were not modified by the treatment (Fig. 1.18). Nonethele, the mot outtanding reult i that girdling of fruit peduncle triggered flowering the next pring while the ON branche did not flower. The girdled branche produced 45 flower per branch, on average (Fig. 1.19). But flower were only produced on the girdled branche that prouted in September, that i, on bud from the fall hoot (Fig. 1.19). Girdled branche that did not prout in September did not flower in March (Fig. 1.19). Thi reult ugget that flowering in the fruiting branche wa not directly due to girdling but to prouting. On the other hand, all the meritem from OFF branche had the ability to flower (even without fall hoot). In thi cae, fall prouting alo increaed flowering (Fig. 1.19). 101

104 Reult Figure 1.18 Effect of girdling on the endogenou GA 1, GA 4, IAA, TZ, ABA and JA content along the OFF, ON and Girdled hoot 11 day after girdling (Augut 25) in Afourer hybrid mandarin. N1-N5: node next to the fruit; Girdling: of fruit peduncle. Data are mean ± ES of 2 et of 10 vegetative hoot/ leafy ingle flowered hoot/girdled leafy ingleflowered hoot. The blue bar hown on the left of the figure correpond to LSD interval. 102

105 Reult * * * * * * Figure 1.19 Girdling (Augut 2014) influence on prouting and flowering in pring (March 2015). Experiment performed in the Afourer hybrid mandarin. Number of pring hoot (A and C) and flower (B and D) per node, in ON hoot (A and B), control and girdled, and OFF hoot (C and D); with fall hoot: thi hoot prouted in autumn and the pring prouting in 2015 wa analyzed on pring and autumn node of 2014; without fall hoot: pring prouting in 2015 wa analyzed on pring node of Data are mean of 10 hoot. Standard error are given a vertical bar. *: indicate ignificant difference (p 0.05). Thi important reult ugget that meritem beide a fruit mut be retarted to activate have the flowering ability. If thi renewed ability i due to the development of new leave, new bud or both then it i worthy of further examination. The lack of flowering in un-prouted bud could be due to a continuou inhibitory effect even after the fruit reache it final ize and ripen, and not only from the fruit located in the terminal poition but alo from other fruit on the branch. To tet thi hypothei, bud (node) from ON and OFF branche, imilar to thoe from the girdling experiment, were excied and cultured in vitro (without the effect of the tree) before and after the flower induction period. Sprouting and flowering were evaluated. 103

106 Reult 1.5 In vitro culture of excied bud from ON and OFF branche A derived in Material and method, the experiment wa conducted with the eedy mandarin Tardivo di Ciaculli, which i imilar to Moncada mandarin. Both are late ripening cultivar related to the Clementine mandarin pecie. In pring, the OFF branche prouted 22.5 hoot 100 node -1 while ON branche prouted ignificantly fewer, 2 hoot 100 node -1. OFF branche had the ability to flower (producing 10 flower 100 node -1 ) while ON branche did not (Fig. 1.20). * * Figure 1.20 Number of hoot and flower per 100 node of vegetative hoot (OFF) and leafy ingle flowered hoot (ON) of 'Mandarino Tardivo di Ciaculli mandarin. Data are mean ± ES of 30 hoot per treatment. *: indicate ignificant difference (p 0.05). Node from OFF and ON branche were cultured in vitro during (November) and after (February) the flower induction tage. At thee date, the fruit preented the characteritic decribed in the Table

107 Reult Table 1.5 Characteritic of fruit from leafy ingle flowered hoot in 'Mandarino Tardivo di Ciaculli' mandarin. Each value i the mean of 10 fruit ± SE. November February Diameter (mm) 25.1 ± ± 0.8 Exocarp color a ± ± 0.7 b 22.8 ± ± 0.7 a/b ± ± 0.05 Number of eed 7.2 ± 0.5 n.d In November, bud prouting wa triggered in vitro during the firt 2 week of the experiment (Fig. 1.21A and B). OFF-node prouted up to 30% wherea only 5% of ONnode prouted. To timulate bud prouting, zeatin wa added to the growing medium, which wa baed on carbohydrate, macronutrient and micronutrient. Zeatin increaed bud prouting ignificantly in both OFF-node (52%) and ON-node (12%) (Fig. 1.21B). No flowering wa achieved in vitro. Sprouted bud only produced vegetative hoot in both OFF and ON node (Fig. 1.22). In order to enure that the flowering ignal (FT) had reached the bud due to winter low temperature, at leat in OFF branche, a econd et of node were harveted and cultured in vitro in February. At thi time, bud prouting wa alo triggered during the firt 2 week, but a higher percentage of prouted bud wa found compared to November, particularly for that of ON-node (Fig. 1.21). Thu, 15 day after culture, ON-bud prouted up to 33% and 55% in medium without and with zeatin, repectively, (6% and 12% in November) (Fig. 1.21B and C). Sixty day after culture the ON-bud percentage of prouting reached 51% and 71% in medium without and with zeatin, repectively (Fig. 1.21B and C). In February, difference in bud prouting between ON and OFF bud were ignificantly narrowed compared to November. Reult ugget that OFF-bud are able to prout when exogenou factor are favorable wherea ON-bud cannot prout until the fruit inhibitory ignal diappear (a it eemed to occur in February). It i worth nothing that flowering wa inhibited in both ON and OFF bud cultured in vitro. That i, ON bud could not flower a they did on the tree, and OFF bud alo lot 105

108 Reult their ability to differentiate flower when eparated from the tree, even after the flower induction period. The fact that a meritem beide a fruit need to be retarted to gain the flowering ability, and that the application of PBZ doe not induce flowering under heavy ON-tree condition, ugget that flowering i the reult of complex interaction at the metabolic and molecular level, involving multiple endogenou mechanim of regulation different from hormone concentration and tranport. It i till unclear which main metabolic route and biological procee are up-regulated and down-regulated in tree with contrating ability to flower due to the preence of fruit. Therefore, to attempt to clarify thee quetion, a proteomic tudy of leave and bud from ON and OFF tree wa conducted during the flower induction period. Figure 1.21 Effect of the fruit and the medium compoition (-GR/+GR) on the prouting in vitro of microhoot with one bud in Mandarino Tardivo di Ciaculli. Each value i the mean of 10 dihe with 5 microhoot. GR: growth regulator (zeatin (1 mg/ml)). *: indicate ignificant difference (p 0.05). 106

109 Reult Figure 1.22 Growth of microhoot in the culture medium. Photograph taken after 7 (A), 14 (B) 21 (C), 28 (D), 35 (E), 42 (F), 56 (G), 60 (H) day of culture. OFF microhoot. 107

110 Reult Section 2. Proteomic analyi of Moncada mandarin leave and bud with contrating fruit load 2.1 Comparative proteome analyi The protein pot were identified that were up- and down-regulated in OFF tree comparing with ON tree. Therefore, leaf and bud ample from OFF and ON tree were analyzed by 2D DIGE gel, two-dimenional difference gel electrophorei. Gel were of high quality with reproducible protein pattern among replicate of the ame ample (Fig. 2.1). Approximately 1436 and 1162 pot in gel image from ample were reolved in leave and bud, repectively. To ae global difference in the expreion level between OFF and ON ample, gel were compared and quantified uing the DeCyder Differential Analyi Software. Among the total protein, 176 pot howed a ignificant quantitative differential accumulation (t-tet < 0.05) between ON and OFF ample in leave, while there were 350 pot in bud. 110 pot in leave and 192 in bud were confirmed with a good match and a ufficient volume for ubequent identification by ma pectrometry. To reliably determine quantitative change in protein expreion and therefore overcome error impoed by technical and biological variation, protein were identified a up-regulated in OFF or ON ample if they were found to have an average expreion level at leat 1.10 higher than the other ON or OFF ample, repectively. Among the 110 protein in leave, 43 had increaed expreion in the OFF ample compared to ON ample (Av ratio +), while 67 howed a decreaed expreion in the OFF ample (Av ratio -). In bud, 97 diplayed an increaed expreion in the OFF ample a compared to the ON ample (Av ratio +), while 95 exhibited a decreaed expreion in the OFF ample (Av ratio -). 108

111 Reult Figure 2.1 Repreentative 2D DIGE gel of protein extracted from Moncada mandarin leave (A-B) and bud (C-D). Equal amount (50µg) of ON ample (Cy5, red), OFF ample (Cy3, green) and internal tandard (Cy2, blue) were loaded in the ame gel. (A-C) Protein up-expreed in the OFF appear in green, thoe down-expreed in OFF appear in red and protein unaffected appear in yellow. (B-D) Protein elected for the analyi by ma pectrometry. Spot number correpond to thoe indicated in Table 2.1 and Identification of differentially expreed protein The 110 protein in leave and 192 in bud were manually excied with a good match from a preparative 2DE gel to further identify 90 leave and 88 bud by MALDI- MS analyi, and the other 20 leave and 104 bud protein by LC-MS/MS analyi. Table 2.1 and Table 2.2 provide the pot number, the function of each protein together with the putative protein name, the acceion code, the organim baed on the protein identified, the homologue in C. clementina etablihed by the databae in 109

112 Reult (Phytozome v9.1), the homologue in Arabidopi thaliana etablihed by the databae in the value for theoretical and experimental pi and molecular ma, the expreion ratio and p-value, and the MASCOT core together with the equence coverage and peptide matched. 110

113 Reult 111

114 Reult 112

115 Reult 113

116 Reult 114

117 Reult 115

118 Reult 116

119 Reult 117

120 Reult 118

121 Reult 119

122 Reult 120

123 Reult 121

124 Reult 122

125 2.3 Claification of identified protein Reult The protein identified in leave can be claified into even group according to their biological function: (i) primary metabolim (33 pot: 26 pot being aociated with photoynthei and carbohydrate metabolim, 4 pot related to Kreb cycle, 1 pot related to pentoe phophate pathway, and 2 pot related to nutrient reervoir activity); (ii) oxidoreductae activity (8 pot: one of them, pot 79, up-regulated in OFF ample, preented the highet ratio among all pot, and 5 pot were catalae, all of them upregulated in ON-crop ample, with pot 381, 384, 386 and 394 matching the ame EST equence); (iii) tre repone (3 pot, all up-regulated in ON ample); (iv) ignal tranduction (1 pot); (v) protein ynthei and degradation (6 pot); (vi) expanin (1 pot); (vii) other protein (58 pot: thi i the larget group, mot of thee protein with unknown function). The relative percentage of protein both in ON leave and in OFF leave are given in Fig On the one hand, ome of thee pot were identified a the ame protein uch a catalae, for pot 381, 384, 386 and 394 (oxidoreductae group); NADP-iocitrate dehydrogenae, for pot 515 and 516 (Kreb cycle ubgroup, up-regulated in ON ample); RuBiCO large ubunit-binding protein ubunit beta chloroplat, for pot 300, 301 and 307; granule-bound tarch ynthae Ib precuror, for pot 327 and 328 (Fig. 2.3); putative cinnamoyl-coa reductae, for pot 743, 744 and 748. The lat three group of protein are related to primary metabolim and all of them are up-regulated in OFF ample. On the other hand, ome of the pot were identified a the ame protein, but diplayed different pi and molecular ma value and might account for ioform or pot-tranlationally modified form of thee protein. Example of the thee pot are miraculin-like protein 1 (pot 1016 and 1049) or NADP-dependent glyceraldehyde-3- phophate dehydrogenae (pot 420 and 437), both related to carbohydrate metabolim, or protein diulphide iomerae (pot 104 and 126, belonging to the protein ynthei and degradation group), all of them up-regulated in OFF ample. The protein identified in bud were claified into 10 group according to biological function involving each protein (Table 2.2): (i) primary metabolim (91 pot: 70 pot aociated with photoynthei and carbohydrate metabolim, 7 pot related to Kreb cycle, and 14 pot related to repiration); (ii) oxidoreductae activity (17 pot); (iii) tre/defene repone (12 pot); (iv) ignal tranduction (1 pot); (v) amino acid metabolim (25 pot); (vi) protein metabolim (18 pot); (vii) the 123

126 Reult mevalonate pathway (3 pot); (viii) flavonoid bioynthei (1 pot); (ix) cell wall metabolim (2 pot); (x) other protein (22 pot; thee protein are involved in unknown biological procee). The relative percentage of protein in both the ON and the OFF bud are given in Fig Figure 2.2 Functional claification of the protein identified and found to be upregulated in ON-crop or OFF-crop leave and bud. The relative percentage of protein in each category are hown. 124

127 Reult Figure 2.3 Repreentative protein analyzed uing DeCyder Software (pot 327, 328, 381, 462, 784 and 1018). Differential expreion analyi in ON and OFF leave of repreentative pot from Fig. 2.1A and 2.1B. 2.4 Specie from which the identified protein proceed Fifty-nine pot were identified in leave when matched againt Citru equence, the number of matched peptide being between 3 and 24, with 14-70% equence coverage. Reult of a imilar order were found for the other 51 protein, identified by homology to equence from other pecie uch a Arabidopi thaliana (15 pot), Prunu perica (5 pot), Medicago truncatula (3 pot), Viti vinifera (2 pot), Phaeolu vulgari (2 pot) or Solanum lycopericum (2 pot). Ninety-nine pot were identified in bud by matching the Citru equence, and the number of matched peptide wa between 1 and 21, with % equence coverage (mean 32.44). Reult of a imilar order were found for the other 93 protein, 125

128 Reult identified by homology with equence from other pecie uch a Pinu trobe (14 pot), Arabidopi thaliana (13 pot), Oryza ativa (13 pot) and Glycine max (8 pot). 2.5 Gene ontology analyi Total amount of protein iolated in leave were analyzed eparately in two group etablihed according to ratio expreion, uing the Web tool FatiGO ( The databae in wa ued to earch for Arabidopi protein homologou to protein identified in thi tudy (Table 2.1 and Table 2.2). The etablihment of thee homologie allowed to know the main biological procee in which the identified protein are involved. The larget group of protein with AV ratio + were compoed of protein involved in carbohydrate and tarch bioynthei, carbohydrate metabolim, protein folding and repone to metal ion in the OFF-crop ample. The larget group of protein with AV ratio - wa compoed of protein involved in hydrogen peroxide catabolic proce and in repone to tre, but many other biological procee are related to the up-expreed protein in the ONcrop ample (Fig. 2.4 leave). The ontology tudy for thee protein in bud revealed that the larget group of protein that were up-expreed in the OFF bud ample in November (AV ratio +), a compared to the ON bud ample (AV ratio -), were compoed of the protein not only involved in carbohydrate and amino acid metabolim, but alo thoe expreed in repone to timuli, uch a reactive oxygen pecie (Fig. 2.4 bud). The larget group of the protein with an AV ratio - wa made up of the protein involved in carbohydrate metabolim and alo thoe expreed mainly in repone to oxidative tre. 126

129 Reult Figure 2.4 Biological procee related to protein up-regulated in ON and OFF leave and bud according to FatiGO tudy. Term at higher level of the hierarchy decribe more general procee while term at lower level are more pecific. Color intenity in the graph quare reflect the level of contribution to the total biological procee (the more intene the color, the higher the level of contribution). 127

130 Reult 2.6 Starch, ABA and JA content and catalae activity in ON and OFF tree According to the main biological function found to be different between ON and OFF tree (primary metabolim, oxidoreductae activity and tre repone), tarch concentration, catalae activity and ABA and JA concentration were tudied. The larget group of protein up-expreed in the OFF-tree leave compared to the ON-tree leave, i involved in carbohydrate and tarch bioynthei, both of them related to primary metabolim. Among them, the following protein may be highlighted: Granule-bound tarch ynthae Ib precuror (GBSS, pot 327 and 328), malate dehydrogenae glyoxyomal precuror (MD, pot 764) and ADP-glucoe pyrophophorylae mall ubunit (AGP, pot 462) (Fig. 2.3). The activation tate of thee protein i correlated with the accumulation of tarch and oluble ugar in everal tiue from rice, wheat or tomato (Kawaaki et al., 1996; Smidanky et al., 2007; Centeno et al., 2011). In fact, the tarch level wa found to be ignificantly higher in OFF leave than in ON leave, being thee reult conitent with the up-expreion of GBSS in OFF-tree leave (Fig. 2.5A). For protein up-expreed in the ON leave, the larget group are thoe involving a hydrogen peroxide catabolic proce, uch a catalae (pot 381, 382, 384, 386 and 394) or monodehydroacorbate reductae (pot 474). Thi i conitent with tudie in apple that howed how catalae activity remained high during tage of fruit growth (Abai et al., 1998). To validate thi in Citru, the catalae activity wa meaured in ON and OFF leave; the reult of thi meaurement confirmed that catalae activity in ON ample wa twice a high that in OFF ample (Fig. 2.5B). Although to a lower extent, the tre repone were alo differentially expreed in ON and OFF tree. ABA and JA are related with tre. In fact, an Abciic acid tre ripening-like protein (pot 802) wa found to be up-regulated in ON bud. The quantification of ABA in OFF and ON tiue howed ignificant difference (Fig. 2.6). The highet ABA content in the fruit exocarp wa found at the end of November, coinciding with fruit color change, and then the ABA content ignificantly decreaed (Fig. 1.3, Section 1). At the ame time that the ABA content increaed in the exocarp it alo increaed ignificantly in bud and leave (Fig 2.6). On the other hand, the ABA content ignificantly decreaed only in OFF leave and wa maintained almot contant in bud (Fig. 2.6). 128

131 Reult * * Figure 2.5 A. Starch concentration in ON and OFF leave expreed in mg g -1 dry weight. B. Catalae activity in ON and OFF leave expreed in µmol H 2O 2 conumed g -1 protein min -1. Data are mean of ix independent replicate (n = 6). There are ignificant difference between ON and OFF leave (p 0.05). *: indicate ignificant difference. Figure 2.6 Endogenou abciic acid content in the ON and OFF tree of C. clementine cv. Moncada. ABA wa meaured in the pring+ummer (SP+S) leave (ON and OFF), fall leave (OFF), bud (ON and OFF) and exocarp (ON). Data are mean ± ES of 2 et of 10 hoot. 129

132 Reult Miraculin-like protein have been related with defene repone and JA (Maerti et al., 2011), and were up-regulated in ON bud (Table 2.2). The JA content howed ignificant difference between ON and OFF tiue (Fig. 2.7). ON hoot howed higher JA content in bud in October and in leave at the end of November compared to the other tiue at the ame time. In the cae of the OFF hoot, the JA content howed ignificant difference between bud and leave during the entire period of the tudy. But the highet difference were oberved from September to November (2500 ng gdw -1 in bud and ng gdw -1 in leave). After November JA content decreaed in both tiue. The JA content wa higher in OFF than in ON hoot. Figure 2.7 Endogenou jamonic acid content in ON and OFF Moncada tree. JA wa meaured in the leave, bud, and fruit exocarp. Data are mean ± ES of 2 et of 10 hoot. 130

133 Reult Section 3: Epigenetic control of flowering 3.1 Alternate bearing: biennial reproductive-vegetative development cycle Citru alternate bearing tree repeat biennial cycle of reproductive-vegetative development (for more detail, ee Figure 4 in the introduction). Alternation of the reproductive development in the Moncada mandarin i hown in Fig Twelve tree were elected in ummer according to their fruit yield in the year 1: tree 1-6 were ON (red, Fig. 3.1 A) wherea tree 7-12 were OFF (blue, Fig. 3.1 B). The preence of fruit in the year 1 (red tree) dratically inhibited the number of flower produced in pring (0.7 flower/100 node) in year 2 compared to OFF tree (blue) (Fig. 3.1 C and D). Blue tree flowered profuely (142 flower/100 node) and et a high number of fruit in year 2 (Fig. 3.1 B) which inhibited flowering in year 3 (Fig. 3.1 D). On the other hand, red tree were OFF in year 2, et a very low yield (Fig. 3.1 A), and developed 5300 vegetative hoot tree -1 (Fig. 3.2), thu producing a high number of new bud receptive to exogenou flowering ignal. Accordingly, red tree flowered profuely in year 3 (145 flower/100 node) becoming ON tree again (Fig. 3.1 A and B). The relative expreion of the flowering promoter CiFT2, LFY and AP1, and the flowering inhibitor FLC, SVP and TEM1 alo howed alternate bearing in ON and OFF tree, ince they were trongly affected by the fruit (Fig. 3.3). In Year 1 jut before bud prouting (January), CiFT2 (Fig. 3.3 A), LFY and AP1 (Fig. 3.3 A, bar chart) relative expreion wa ignificantly higher in OFF bud than in ON bud. A a reult, OFF tree prouted and flowered profuely in pring (Year 2) (becoming ON) wherea ON tree prouted le and did not flower (becoming OFF). Later, CiFT2 howed ignificant difference in leave between ON and OFF tree from November onward. In OFF-tree leave, CiFT2 expreion wa up-regulated, becoming more than 15-fold higher than that of ON-tree leave in January, and remaining almot contant up to late in February. CiFT2 in ON-tree leave did not vary ignificantly during the tudy period (Fig. 3.3 A). On the contrary, the fruit timulated the FLC gene expreion in ON-tree leave from October to the end of February (Fig. 3.3 B). But what i more important, the expreion of FLC in bud in January, and in leave in May did not differ ignificantly between ON and OFF tree. Thi ugget a kind of mechanim that inhibit flower induction in the adult leaf (and CiFT2 protein i not tranferred to the meritem), but allow the meritem to retart and develop new leave without the inhibitory ignal. Afterward, when leave and fruit 131

134 Reult reach the adult tage, FLC i again expreed in ON-tree leave. The SVP gene, which directly inhibit FT together with FLC, howed a imilar trend to that of FLC. In Year 2 the SVP relative expreion wa higher in ON than OFF leave, but difference were only ignificant in December (Fig. 3.3 C). Finally, TEM1 relative expreion wa increaed 80-fold in ON-tree mature leave in January (Fig. 3.3 D). In the Moncada mandarin, alternate bearing i a natural proce. A defruiting treatment in ON tree in year 1 only increaed flowering in year 2. But the third year, tree began again the ON-OFF cycle (Fig. 3.1 E and F). Therefore, an endogenou/genetic mechanim may regulate the proce. Figure 3.1 Alternation of the reproductive development of 24 Moncada mandarin tree in 4 year. Tree 1-6 (red, A: fruiting; C: flowering) and 7-12 (blue, B: fruiting; D: flowering) were ON (with fruit) and OFF (without fruit), repectively, in the firt year (A and B). Tree (green, E) and (orange, F) were ON tree in year 1 defruited by hand (in ummer) up to 33% and 66%, repectively. Year 3 were ON again a it did the red tree. 132

135 Reult * * * * Figure 3.2 Vegetative growth in Moncada mandarin ON (black) and OFF (white) tree. Data are mean ± ES of 6 tree. *: indicate ignificant difference (p 0.05). * * * * * * * * * * * * * * * * * * * * * * Figure 3.3 CiFT2, LFY and AP1 (A), and FLC (B), SVP (C) and TEM1 (D) relative expreion in Moncada mandarin bud (Year 1) and leave (Year 2) from January to February of the next year. Data are mean ± ES of 3 qrt-pcr replicate. *: indicate ignificant difference (p 0.05). 133

136 Reult Reult ugget that flowering promoter might be up-regulated by exogenou cue (i.e. low temperature) wherea flowering inhibitor might be up-regulated by endogenou cue. The objective of thi chapter i to determine the exitence of DNA methylation mark that control expreion of the gene that negatively regulate flowering (FLC, SVP and TEM1). Methylation of cytoine in DNA i a common epigenetic ignaling tool that cell ue to lock gene in the "OFF" poition. Cytoine methylation occur in three equence context, cytoine-guanine (CG), the mot abundant, but alo in CHG and CHH (where H i any nucleotide except G). A a general rule, the more DNA methylation found near the promoter region the more gene ilencing. Depending on the DNA methylation profile gene expreion can be up-regulated or otherwie blocked. Hitone methylation i another epigenetic mechanim that regulate gene expreion. Several tudie have indicated that DNA methylation and hitone methylation at certain poition are connected. 3.2 DNA methylation profile During the flower induction period Flowering promoter DNA-methylation wa tudied in leave ampled on November 30 th (flower induction period), when a clear difference wa found in CiFT2, FLC and SVP gene expreion between ON and OFF tree (Fig. 3.3). DNA methylation wa tudied in both CiFT1 and CiFT2. In the CiFT1 gene equence, three region were found to preent CG ite: promoter-intron (in 2 poition, 1 and 2), intron (in 3 poition, 3, 4 and 5) and exon (in 6 poition, 6-11) (Fig. 3.4 A). However, no cytoine wa found to be metilated in thee CG in ON and OFF leave (Fig. 3.4 B and C). Neither CHG nor CHH were found to be methylated in the CiFT1 gene. Thu, they were not repreented in Fig. 3.4 A. On the other hand, only the promoter region in the CiFT2 gene wa tudied. Thi region preented 7 CG ite (Fig. 3.4 D). Only the firt CG howed cytoine methylation in 4 (out of 10) analyzed equence of the ON leave (Fig. 3.4 E). Thi carce DNA methylation profile ugget no relationhip with CiFT2 ilencing in ON-tree leave. The CiFT2 gene howed no DNA methylation in 134

137 Reult OFF-tree leave (Fig. 3.4 F). Reult ugget that difference in CiFT2 gene expreion between ON and OFF tree (Fig. 3.3 A) are not regulated by DNA methylation. Figure 3.4 DNA methylation profile of CiFT1 and CiFT2 in ON and OFF tree in the flower induction period (November). A and D: Scheme of the analyzed zone. Red number indicate CG ite in the equence of the analyzed gene. B, C, E and F: reult of the biulphite equencing analyi of at leat 7 clone per tree. White bar indicate no cytoine methylation wherea black bar indicate cytoine methylation. DNA wa collected from ON and OFF leave during the floral bud inductive period (November) of Moncada mandarin. 135

138 Reult Flowering inhibitor The FLC gene preented 24 CG ite in the promoter region and 10 CG ite in the intron (Fig. 3.5 A). After the biulfite treatment, 4 CHH ite (poition 20, 30, 32 and 35) howed difference in cytoine methylation between ON and OFF leave and thu were included in Figure 3.5. In the promoter region, CG ite howed no methylation in either ON or OFF tree, and only the poition 20 (CHH) howed partial methylation (4 out of 10 clone) in OFF leave (Fig. 3.5 B and C). On the other hand, ON-tree leave howed more cytoine methylation in the intron than OFF-tree leave. ON-tree leave howed methylation change compared to OFF-tree leave in the poition 27 (100% v 33% CG), 29 (0% v 66% CG), 30 (33% v 0% CHH), 31 (66% v 33% CG), 32 (0% v 33% CHH), 35 (100% v 0% CHH) and 37 (100% v 0% CG). Thee ignificant difference in the methylation profile may explain the ignificant difference found in the FLC gene expreion in November between ON and OFF tree (Fig. 3.3 B). The SVP and TEM1 methylation profile did not how ignificant difference between the ON-tree and OFF-tree leave. While SVP howed no methylation in either ON- or OFF-tree leave (Fig. 3.5 E and F), TEM1 howed all the CG ite methylated (Fig. 3.5 H and I), which were found in the promoter region (Fig. 3.5 G). Epigenetic regulation of the TEM1 gene cannot be dicarded ince it relative gene expreion did not differ between ON and OFF tree in November (Fig. 3.3 D). However, the temporal pattern of TEM1 gene expreion in relation to CiFT2 ugget a econdary role of thi gene negatively regulating flowering in citru. On the other hand, FLC might play a key role regulating the proce. To better relate the methylation profile of the FLC gene with it expreion, the DNA-methylation tudy wa repeated in young leave ampled in May (before the flower induction period), when FLC gene expreion did not differ between ON and OFF tree (Fig. 3.3 B) Before the flower induction period Flowering Locu C Before the flower induction period (May), the FLC methylation profile of ONand OFF-tree leave howed high imilarity. Almot all the ame ite howed DNA methylation. Only the poition 34 howed an abolute difference (0% ON v 100% OFF 136

139 Reult CG), and the poition 29 (100% v 54.5% CG) and 31 (18.2% v 100% CG) howed partial methylation (Fig. 3.6). Figure 3.5 DNA methylation profile of FLC, SVP and TEM1. Scheme of zone analyzed. Biulphite equencing of promoter and intron region wa performed on DNA collected from ON and OFF leave during the floral bud inductive period (November 30) of Moncada mandarin. 137

140 Reult Figure 3.6 DNA methylation profile of FLC. Scheme of zone analyzed, and biulphite equencing of promoter and intron region wa performed on DNA collected from ON and OFF leave (May 13) of Moncada mandarin. It i worth nothing that while in May the FLC gene howed DNA methylation in the ame poition for ON and OFF leave (except poition 34) (Fig. 3.7), and alo the ame relative expreion (Fig. 3.3), in November, both DNA methylation and relative expreion of the FLC gene differed ignificantly between ON-tree and OFF-tree leave (Fig. 3.7 and 3.3). In November, DNA methylation increaed in ON-tree leave compared to May, wherea it decreaed in OFF-tree leave (Fig. 3.7). Thi reult raie the quetion of which methyltranferae modify their activity in ON and OFF leave between May and November coinciding with the up-regulation of FLC gene expreion. 138

141 Reult Figure 3.7 Schematic repreentation of the DNA methylated ite in the FLC gene in May (white) and November (black) in ON and OFF tree of the Moncada mandarin. 3.3 Time-coure of methyltranferae gene expreion Several methyltranferae have the ability to promote or repre FLC gene expreion by directly methylating FLC DNA and indirectly, by inducing change in the hitone Promoter of FLC expreion Among the promoter, the methyltranferae ELF8 directly methylate FLC DNA wherea 12ATX and 7ATX are involved in hitone methylation. The relative expreion of ELF8 differed ignificantly in ON and OFF leave in September and during the flower induction period (December) (1.2 ON v 0.4 OFF and 2.9 ON v 0.6 OFF, repectively). Difference were maintained between ON and OFF 139

142 Reult tree until the end of February (Fig. 3.6 A). The relative expreion of 12ATX wa ignificantly higher in bud from ON tree compared to bud from OFF tree (Year 1). After bud prouting (Year 2), the 12ATX expreion decreaed and it did not differ between ON and OFF leave until December, when it ignificantly increaed in ON-tree leave (14-fold). Significant difference were not found in January or in February between ON and OFF leave (Fig. 3.6 B). Similar reult were found for 7ATX gene expreion (Fig. 3.6 C). Thee reult are in concordance with the higher FLC relative expreion found in ON-tree leave compared to that of OFF-tree leave during the flower induction period (Fig. 3.3 B). Figure 3.6 ELF8 (A) 12ATX (B) and 7ATX (C) relative expreion in Moncada mandarin bud (Year 1) and leave (Year 2) from January to February of the next year. Data are mean ± ES of 3 qrt-pcr replicate. *: indicate ignificant difference (p 0.05). 140

143 Reult Inhibitor of FLC expreion Among the inhibitor, SKB1 directly methylate FLC DNA wherea 5ATX i involved in hitone methylation. SKB1 relative expreion wa ignificantly upregulated in OFF-tree leave at the beginning of the flower induction period (November); it differed ignificantly (2.6-fold) from the SKB1 expreion found in ON-tree leave (Fig. 3.7 A). On the other hand, the 5ATX gene expreion wa ignificantly higher in ON-tree leave in December and January and in ON-tree bud in January compared to the OFF-tree (Fig. 3.7 B). Figure 3.7 SKB1(A) and 5ATX (B) relative expreion in Moncada mandarin bud (Year 1) and leave (Year 2) from January to February of the next year. Data are mean ± ES of 3 qrt-pcr replicate. *: indicate ignificant difference (p 0.05). Reult ugget a relationhip between DNA methylation (and hitone methylation) and FLC expreion in citru. To further characterize thi relationhip, the effect of 5-azazytidine on FLC and CiFT2 gene expreion wa tudied. Azazytidine i a chemical analogue of the cytoine nucleoide that at low doe inhibit DNA methyltranferae cauing hypomethylation of DNA. The treatment ignificantly increaed FLC expreion from 10 to 50 hour after treatment. Converely, the CiFT2 gene expreion wa ignificantly reduced, probably due to the increae in FLC (Fig. 3.8). Reult indicate a relationhip between DNA methylation and FLC expreion. However, the effect oberved wa unexpected: in the ON-OFF ytem the more DNA methylation the more FLC expreion. But 5-azazytidine, which caue hypomethylation, alo increaed FLC expreion, uggeting a complex relationhip between methylation, FLC expreion and flowering. 141

144 Reult * * * * * * Figure 3.8 Effect of 5-Azacytidine (5-AzaC, 350µM) applied at floral bud inductive period (November 25) on the time coure of CiFT2 and FLC relative expreion in the leave of the ingle flowered leafy hoot of Afourer mandarin. Treatment wa applied a a foliar pray. Data are mean ± ES of 3 qrt-pcr replicate. *: indicate ignificant difference (p 0.05). Another important quetion that require confirmation i the relationhip between fruit, FLC methylation and FLC expreion. Significant difference in the methylated poition of the FLC gene were found for ON and OFF tree when FLC expreion wa ignificantly different (November); moreover, DNA methylated poition did not differ when FLC expreion wa the ame in ON and OFF tree. However, thi relationhip doe not imply cauality; therefore, further experiment were carried out to demontrate the direct effect of fruit removal upon FLC methylation and gene expreion. 3.4 Effect of defruiting on flowering, methyltranferae and flowering gene expreion and FLC-DNA methylation Effect of defruiting on flowering The effect of defruiting on flowering intenity i hown in Fig A expected, ON tree had the lowet number of flower in pring (6 flower/100 node) in comparion with OFF tree, which had the highet number of flower (100 flower/100 node). Defruited tree (in Augut) flowered in pring a did OFF tree (90 flower/100 node), without differing ignificantly (Fig. 3.9). 142

145 Reult * Figure 3.9 Effect of defruiting on flowering intenity of Afourer mandarin branche. Fruit were removed in Augut. ON: with fruit; OFF: without fruit. Data are the mean of 3 tree ± tandard error. *: indicate ignificant difference (p 0.05) Effect of defruiting on flowering gene and methyltranferae relative expreion Flowering gene A expected, defruiting ignificantly increaed the CiFT2 relative expreion and reduced the FLC gene expreion up to the OFF-tree level. ON tree howed ignificantly lower CiFT2 and higher FLC gene expreion (Fig A and B). On the other hand, defruiting did not ignificantly modify SVP or TEM1 gene expreion (Fig C and D). * * Figure 3.10 Effect of fruit and defruiting on CiFT2 (A), FLC (B), SVP (C) and TEM1(D) relative expreion in Afourer mandarin leave at floral bud inductive period (November 25). Fruit were removed in Augut. ON (black): with fruit; OFF (white): without fruit. Data are mean ± ES of 3 qrt-pcr replicate. *: indicate ignificant difference (p 0.05). 143

146 Reult Methylation gene Compared to the ON tree defruited tree had a ignificant reduction in (6-fold) the expreion of the FLC promoter 12ATX and 7ATX, and it did not ignificantly differ from that of the OFF tree. No ignificant difference were found in the ELF8 gene expreion (Fig A, B and C). Similarly, defruiting did not modify the relative expreion of the FLC inhibitor SKB1 and 5ATX, which howed the highet value in OFF tree (34x compared to ON tree) (Fig D and E). * * * Figure 3.11 Effect of fruit and defruiting on ELF8 (A), 12ATX (B), 7ATX (C), SKB1 (D) and 5ATX (E) relative expreion in Afourer mandarin leave at floral bud inductive period (November 25). Fruit were removed in Augut. ON (black): with fruit; OFF (white): without fruit. Data are mean ± ES of 3 qrt- PCR replicate. *: indicate ignificant difference (p 0.05). 144

147 Reult DNA methylation profile To analyze the effect of defruiting on FLC methylation, the intron region wa tudied (Fig A). FLC DNA profile wa analyzed in November, i.e during the flower induction period. Defruited and OFF tree howed DNA methylation in the ame ite except for poition 29, which wa partially methylated in OFF tree (45.5% OFF v 0% DEF CG). On the other hand, ON tree methylation profile ignificantly differed from Defruited and OFF tree in the poition 26a (45.5% ON v 0% OFF and DEF CHH), 27a (100% ON v 0% OFF and DEF CHH), 29 (100% ON v 45.5% OFF and 0% DEF CG), and 31 (0% ON v 100% OFF and 45.5% DEF CG) (Fig. 3.12). The difference and imilaritie between thee profile can be oberved in Fig Reult ugget a clear relationhip between fruit, FLC methylation, FLC expreion, CiFT2 expreion and, therefore, flowering. Figure 3.12 DNA methylation profile of FLC. Scheme of zone analyzed. Biulphite equencing of promoter and intron region wa performed on DNA collected from leave at floral bud inductive period (November 25) of Afourer mandarin. 145

148 Reult Figure 3.13 Schematic repreentation of the DNA methylated ite in the FLC gene in ON (white), OFF (black) and Defruiting (green) tree of the Afourer mandarin. 146