The Reproductive Phenology of Citrus. II: Citrus Floral Ontogeny

The reproductive phenology of Citrus. II: Citrus floral ontogeny

Jakkie (OPJ) Stander Citrus Research International, Department of Horticultural Science, University of Stellenbosch, Private Bag X1, 7602, Matieland. E-mail: [email protected]

OPSOMMING: • Hoё vlakke van interne ouksiene (IAA) three articles on the physiology of citrus • Meerderheid sitrus-spesies blom gedurende afkomstig van vrugte, inhibeer die vorming flowering, and will focus on the important die lente. Die verskynsel word voorafgegaan van nuwe blom dra-posisies gedurende die chronological events pertaining to the devel- deur ŉ blom-ontwikkelingsproses gedurende somer. opment of a citrus flower, as influenced by die herfs en winter (Mei-Julie). • Vroeё verwydering van vrugte (uitdunning endogenous factors, environmental condi- • Blom-induksie (BI) is die eerste en bepalen- in “aan”-jaar) verminder die inhiberende ef- tions and cultural practices. de stap in die blom-ontwikkelingsproses. fek van ouksiene op ontwikkeling van nuwe • Water stres inisieёr BI in warm, somer- lote (apikale dominansie) en stimuleer die Inleiding reёnval streke en genoegsame lae tempera- vorming van nuwe dra-posisies. Sitrus is ŉ immergroen boom wat ŉ kom- ture (15-20°C) in koeler, winter-reёnval • Lae koolhidraat-vlakke gedurende periode plekse boomstruktuur onderhou deur een streke. van blom-ontwikkeling beperk die potensi- tot drie jaarlikse vegetatiewe groei-fases. Na • Die belangrike FT-geen, is onlangs in aal vir blom-knop ontwikkeling en kan lei tot induksie van nuutgevormde vegetatiewe lote blom-plante geidentifiseer en sy uitdrukking swak blom-ontwikkeling. deur genoegsame geakkumuleerde lae tem- is direk gekoppel aan tyd van blom, sowel as • Verwydering van vrugte gedurende ŉ aan- perature en/of water stress gedurende die blom-intensiteit. jaar deur chemiese- of handuitdunning is ŉ herfs en winter, ontwikkel blomme in die • Water stres en/of lae temperature lei tot die suskesvolle manier om ŉ balans tussen bel- lente vanuit knoppe van lote wat gedurende uitdrukking van die sitrus FT-geen (CiFT) in angrike interne hormone soos GA en IAA te die vorige 12-maand seisoen gevorm het. blare, lote en knoppe tydens BI. Die CiFT- handhaaf, asook om die opbou van koolhi- Hierdie artikel is die tweede in ŉ reeks mRNA word transleer na CiFT-proteien, drate tot ŉ genoegsame vlak te bewerkstellig. van drie artikels wat handel oor die fisiolo- wat daarna in die phloeёm van die loot na • Bemesting en besproeingspraktyke ge- gie van blom-ontwikkeling in sitrus en die ontvanklike knoppe vervoer word. Genoeg- durende die somer en herfs moet so aangepas belangrike stappe van blom-ontwikkeling, same akkumulasie van die proteien in die word om nuwe vegetatiewe groei te stimuleer soos beinvloed deur interne plant-kondisies, knop, onder omgewingskondisies gunstig vir en sodoende nuwe blom dra-posisies te be- omgewingsfaktore en produksie-praktyke. groei, lei tot blom-ontwikkeling gedurende werkstellig. die lente. Flower induction • Natuurlike interne gibbereliene afkomstig Introduction In subtropical climates, citrus flowering (with van sade en die skil van vrugte onderdruk Citrus trees are perennial evergreens, which the exception of lemon and lime) occurs dur- die uitdrukking van die CiFT-geen en die sustain a complex tree structure with one ing spring and is preceded by an intricate, produksie van die CiFT-proteien. Dieselfde to three distinct annual vegetative growth synchronized flower development process respons word verkry van blaartoedienings flushes. After a sufficient induction period, during autumn and winter (April-July) (Kra- van sintetiese gibbereliene (GA) gedurende with the onset of growth-promoting condi- jewski and Rabe, 1995). Flower induction is herfs en winter. tions during spring (increase in temperatures the first and essential step in this process and • In kultivars wat laat in die seisoen ge-oes and sufficient water and mineral nutrient precedes the first microscopic signs of flower word, kan die aanwesigheid van ŉ uiter- supply), flowers develop from buds on veg- initiation by 3 to 4 weeks (Monselise and matige hoё aantal GA-ryke vrugte (“aan”- etative shoots that originated from vegetative Halevy, 1964). Citrus is day-length neutral jaar) die ekspressie van die CiFT-geen on- shoots in the different growth flushes of the (Davenport, 1990; Moss, 1969), with the pre- derdruk. Dit kan lei tot die ontwikkeling van previous 12-month season. dominant promoting stimulus for flower in- min of geen blomme. This article is the second in a series of duction in warm, summer-rainfall areas be-

TECHNOLOGY | CRI 65 AUGUST | SEPTEMBER 2015 currently provides a viable model of flower induction in both annual/biennial and perennial flowering plants. The accepted model states that FT-gene expression in plant leaves in response to an environmental cue results in FT-protein synthesis and transport via the phloem to a susceptible meristem. Bud tran- Figure 1. Citrus trees develop new potential flowering shoots during vegetative growth flushes sition from a vegetative to reproductive state in the spring, summer and autumn (1). Sufficient hours of cool temperature (15-20°C) and/or only initiates under growth-promoting envi- a prolonged period of water stress, lead to the expression of the citrus FT-gene (CiFT) in leaves ronmental conditions such as warm temper- and buds on the newly developed shoots (2A). The CiFT-mRNA is translated to the CiFT-protein atures and increased water supply, once the and transported via the phloem to susceptible buds (2B). Under growth-promoting conditions, accumulation of FT-proteins in the bud have a threshold level of CiFT-protein will result in development of flower parts within the bud and reached a certain species-specific threshold sprouting of flowering shoots. level (Nishikawa et al., 2007). In citrus, a sufficient period of cool temperature (15-20°C) and/or a prolonged period of water stress, are the main inductive stimuli that lead to expression of the citrus FT-gene (CiFT) in leaves, stems and buds (Nishikawa et al., 2007) (Fig. 1). After sufficient induction, the CiFT-mRNA is translated to the CiFT-protein, which is transported via the phloem to susceptible buds, and results in reproductive development under growth-promoting conditions during spring (Fig. 1). Loss of function in the CiFT-gene delays flowering, whereas over-expression Figure 2. Fruit rinds and developing seeds are major sources of endogenous hormones. Auxins leads to very early and/or intense flowering (IAA) transmitted within the shoot during summer, reduce the sprouting of lateral buds and (Nishikawa et al., 2010). development of new potential flowering shoots (1). During flower induction in the winter, gibberellins (GAs) reduce the expression of the key flower promoting gene, the citrus flowering Flower initiation and locus T (CiFT) in the leaves and buds (2A+B) and results in a lack of flowers during spring (3). differentiation Flower initiation in citrus involves the tran- ing that of water stress (Moss, 1969) and low understanding flower induction has recently sition of bud meristematic tissue from a veg- temperatures in cooler, winter-rainfall areas come about with the successful identifica- etative to reproductive state in response to (Valiente and Albrigo, 2004). Both flower- tion of a key flowering gene, the FLOWER- levels of sufficiently accumulated CiFT-pro- promoting stimuli result in the cessation ING LOCUS T (“FT-gene”) in the model teins in the bud (Davenport, 1990). Flower of root growth, which strongly resembles a plant, Arabidopsis, and with the expression initiation is the process within the bud, dur- common mechanism of flower induction in- of its homologues and the direct relationship ing which the flower parts start to develop itiation. For many years, the hypothesis sur- to flowering also being conﬁrmed in peren- at a molecular level into a state significantly rounding this common mechanism was that nial fruit tree species such as citrus (Nishi- distinguishable from vegetative, non-flower- of an obligatory/required tree “rest” period, kawa et al., 2007). The intensity of FT-gene ing buds (Lord and Eckard, 1985). The dif- similar to that of winter dormancy in pome expression in reaction to species-specific en- ferentiation of a bud occurs with the onset of and stone fruit species. vironmental stimuli strongly correlates with growth-promoting conditions during spring However, a significant breakthrough in time of flowering and flower intensity, and and results in bud transformation into either

TEGNOLOGIE | CRI 66 AUGUSTUS | SEPTEMBER 2015 a vegetative or reproductive state. Non-dif- cultivar differences in response to climate In addition, summer fruit removal (hand ferentiating buds remain dormant as a result and water supply (Davenport, 1990), and thinning) results in a significant increase in of sprouting inhibition incurred by presence time of harvest (early vs. late) (Verreynne and the size of remaining fruit (Stander, 2014) of fruit (Verreynne and Lovatt, 2009), insuf- Lovatt, 2009) all point to the presence of an and results in the decrease in an important ficient growth-promoting conditions (Moss, endogenous, self-regulating system that per- group of endogenous hormones, gibberellins 1969) or age of the bud/shoot (Schneider, tains to the intensity of flowering response. (GA), which directly influence flower-bud 1968). Concomitantly, the majority of research development at the molecular level. Flowering buds differentiate first - ap findings strongly support a self-regulatory proximately 3 to 4 weeks earlier than those model that involves an endogenous signal Gibberellins (GA): The inhibiting effect of on vegetative shoots, with the terminal transmitted within the plant (hormones), gibberellins (GA’s) on flowering of peren- flower (generally leafy) developing before with the intensity of transmittance resulting nial fruit trees is well documented (Bang- the flowers in the lateral positions (generally in inhibition or up-regulating of flowering erth, 2009). The effect on flower induction leafless) (Abbott, 1935). Flowers differentiat- response in the presence of growth-promot- in citrus is furthermore supported by the ing from the terminal position of a shoot are ing conditions. In addition, the molecular lack of reproductive inflorescence obtained more advanced (Lord and Eckard, 1985), and response to the hormonal signal is to a large after exogenous applications of gibberel- with the exception of lemon, lime, grapefruit degree dependent on the measure of avail- lic acid (GA3) under prevailing flower in- and satsuma mandarin, fruit set is higher and able plant metabolic energy (carbohydrates) ductive conditions (low temperature and/ eventual fruit size larger compared to lateral and the prevalence and intensity of factors or water stress) (Goldschmidt et al., 1985; positions (Krajewski and Rabe, 1995). influencing the availability thereof. Guardiola et al., 1982; Monselise and Halevy, Except for an increased photo-assimilate 1964). Only recently has light been shed on supply to ovaries obtained from the addi- The role of endogenous hormones the metabolic method of inhibition of exog- tional leaves on leafy inflorescence, these Auxins (IAA): Citrus trees sustain a com- enously applied GA during flower induction, flowers are also a stronger sink than those plex tree structure with the development of with Muñoz-Fambuena et al., (2012) demon- of leafless inflorescences due to better devel- new shoots during vegetative growth flushes strating a reduction in the expression of the oped xylem and a higher number of vascular in spring, summer and autumn. After a suf- key flower-promoting gene, the citrus flow- bundles (Erner and Shomer, 1996). Delayed ficient induction period and with an increase ering locus T (CiFT) in the leaves and buds bud-break and sprouting usually produces in temperature and water and mineral-nu- of sweet orange (Citrus sinensis), in response leafy inflorescence (Lenz, 1966), which could trient supply, flowers develop from buds on to GA3. A similar response was obtained in be explained by increased temperatures at these vegetative shoots. The development of mandarin (Citrus reticulata) by Goldberg- the time of bud-break, because warm tem- new flowering sites/bearing units during the Moeller et al. (2013). peratures prior to and during flower differ- various vegetative growth flushes is therefore Fruit rinds and developing seeds are entiation are associated with the sprouting of of critical importance to ensure a sufficient a major source of endogenous GA (Ben- predominantly leafy inflorescence, and low return bloom. Fruit is a major source of nat- Cheikh et al., 1997) and the suggestion of the temperatures with the development of leaf- ural, endogenous auxins (most importantly inhibiting effect of endogenous GA on flow- less inflorescence (Moss, 1969). indole acetic acid (IAA)), and their presence ering is substantiated by Koshita et al. (1999) during the summer is directly correlated to who related the lack of flower bud formation The regulation of flower the inhibition of new flower-bearing units to a three-fold increase of GA-content in development events (shoots) (Fig. 2) and resultant poor return the leaves of fruit-bearing shoots of satsuma The lack of flowering response on receptive bloom - a mechanism proposed to be similar mandarin (Citrus unshiu) during the time shoots after exposure to favorable environ- to apical dominance (Verreynne, 2005). of flower induction (Fig. 2). The regulatory mental conditions is proof that additional Early removal of fruit high in IAA, al- role of GA conforms to the evolutionary factors influence bud-fate and flowering po- lows for the sprouting of new potential flow- adaptation of perennial plants that restricts tential. The presence of fruit at different phe- er-bearing units, increase in available flow- saturation of buds by flowers and allows for nological stages (Garcia-Luis et al., 1995; Ko- ering sites and intensified flowering during sprouting of vegetative shoots to maintain shita et al., 1999; Valiente and Albrigo, 2004), return bloom (Verreynne and Lovatt, 2009). plant longevity beyond two growing seasons.

TECHNOLOGY | CRI 67 AUGUST | SEPTEMBER 2015 However, from a commercial perspective, reynne, 2005; Van der Merwe, 2012) (Fig. 3). transportation of other, more specific signals even a slight imbalance of a particular factor Concomitantly, an increase in flower num- (such as hormones) that may control or have such as GA, can lead to detrimental conse- bers is obtained from de-fruiting treatments a significant influence on flower formation. quences to successful citrus production over of trees during an “on”-year (Goldschmidt Taking all collective evidence into considera- the short- and long-term. In years of heavy and Golomb, 1982; Stander, 2014). By early tion, the fact that a minimum, threshold level fruit loads, especially in late-maturing culti- removal of a large number of fruit, a strong of carbohydrates is a prerequisite for flower vars that are harvested after winter, such as sink for photo-assimilates is removed and development in citrus, is evident. ‘Nadorcott’ and ‘Or’ mandarins (Citrus re- therefore results in the restoration of carbo- ticulata) as well as ‘Valencia’ sweet orange hydrate levels in leaves and shoots to a level Conclusions (Citrus sinensis), the inhibition of CiFT- sufficient for development and maintenance Taking all of the above-mentioned aspects expression by excessive number of fruit (the of an adequate number of flowers during the into consideration, flowering response in major source of endogenous GA), is one of return bloom. However, the sole effect of citrus is a factor of three important variables the main factors resulting in the manifesta- carbohydrates on flowering following fruit that can only be optimized if fully under- tion of alternate bearing. removal is inconclusive, as removal of fruit stood and carefully considered by a citrus also involves the elimination of a major natu- grower, as follows: The role of carbohydrates ral source of endogenous hormones such as 1. Flowering is a factor of the amount of Although suggested by Koch (1996), the di- GA and IAA that may have influenced bud- available flowering sites/bearing units. rect involvement of total carbohydrates or fate (Moss, 1971). Are the production practices embracing nu- its different components, such as the non- Girdling during flower induction pro- tritional, irrigation and crop management reducing sugar, sucrose, in the transcrip- vides a useful experimental tool to manipu- practices that will ensure the development tional control of flower-gene expression in late tree carbohydrate levels without hav- of new vegetative flush during August/Sep- perennial fruit trees has thus far not been ing to remove fruit. The increase in flower tember, November/December and February/ conclusively proven. Nevertheless, the indi- intensity resulting from autumn girdling March? rect implementation of carbohydrates’ effect (Fig. 4) is strongly correlated to a spike in on flowering is very well documented and carbohydrate availability for energy utiliza- 2. Flowering is a factor of the level of self- adheres to the “nutritional hypothesis” as re- tion during flower induction (Goldschmidt, regulation within a bearing unit. Is the par- lated to citrus flower development. 1999). However, once again the important ticular cultivar under consideration an early Strong and convincing evidence suggests interplay of hormone/carbohydrate balance or late cultivar (fruit harvested after flower that the lack of flowers developing after a se- comes into play. The breaking of vascular induction) and in the case of the latter, how vere “on”-year in alternate bearing cultivars connectivity between potential flowering can fruit load be managed to a commercially as a consequence of extreme depletion of buds and other organs resulting from the acceptable level, without fruit inhibiting carbohydrates in tree storage organs (Ver- girdling treatment could possibly restrict the CiFT-gene expression? 3. Flowering is a factor of the intensity level of flower induction conditions. Will enough low temperatures accumulate during May-July in the particular production area, and if not, can sufficient water stress be applied during this period to stimulate neces- sary CiFT-gene expression, without neglect- ing fruit quality and tree health?

Figure 3. A sufficient level of carbohydrates is needed for development and maintenance of an ad- The worst-case scenario for citrus flower de- equate number of flowers during the return bloom. On-trees [high fruit load, (A)] enter flower in- velopment will be where inadequate new duction with lower available leaf carbohydrates compared to off-trees [low fruit load (B)] and lower bearing shoots are formed in the summer or their potential to obtain a sufficient return bloom (graph adapted from Van der Merwe, 2012). autumn immediately prior to winter induc-

TEGNOLOGIE | CRI 68 AUGUSTUS | SEPTEMBER 2015 Lenz, F. 1966. Flower and fruit development in ‘Va- lencia Late’ oranges, as affected by type of inflorescence and nutritional status. Hort. Res. 6:65–78. Lord, E.M. and KJ. Eckard. 1985. Shoot development in Citrus sinensis L. (Washington navel orange). 1. Floral and inflorescence ontogeny. Botanical Ga- zette. 146: 320-326. Monselise, S.P. and A.H. Halevy. 1964. Chemical inhibition and promotion of citrus flower bud induction. Proc. Am. Soc. Hortic. Sci. 84: 141-146. Moss, G.I. 1969. Influence of temperature and photo- period on flower induction and in florescence development in sweet orange (Citrus sinensis L. Osbeck). Journal of Horticultural Science. 44:311–320. Moss, G.I. 1971. Effect of fruit on flowering in rela- tion to biennial bearing in sweet orange (Citrus sinensis). Journal of Horticultural Science. 46:177–184. Figure 4. Autumn girdling (left) is strongly correlated with a spike in carbohydrate availability Muñoz-Fambuena, N., C. Messejo, M.C. González- for energy utilization during flower induction and results in increased flower intensity during Mas, D. J. Iglesias, E. Primo-Millo, M. Agustí. return bloom, compared to non-girdled trees, with lower carbohydrate reserves (right). 2012. Gibberellic acid reduces flowering intensity in sweet orange [Citrus sinensis (L.) Osbeck] by repress- tion conditions and crop load cannot be regu- 1995a. Leaf carbohydrates and flower formation in ing CiFT gene expression. Journal of plant growth regulators. 31:529–536. lated to i) support the development of new Citrus. Journal of the American Society of Horticul- tural Science. 120:222–227. Nishikawa, F., T. Endo, T. Shimada, H. Fujii, T. bearing shoots and ii) avoid the depletion of Goldberg-Moeller, R., L. Shalom, L. Shlizerman, Shimizu, M. Omura and Y. Ikoma. 2007. Increased carbohydrate reserves. Cultivar choice, such S. Samuels, N. Zur, R. Ophir, E. Blumwald, and CiFT abundance in the stem correlates with floral induction by low temperature in Satsuma mandarin as the planting of a seedy (seed = high in GA), A. Sadka. 2013. Effects of gibberellin treatment during flowering induction period on global gene expres- (Citrus unshiu Marc.). Journal of experimental Bota- late-maturing cultivar (harvested after winter) sion and the transcription of flowering-control genes ny. 58: 3915-3927. in an area with historically moderate average in Citrus buds. Plant Science 198:46-57.Goldschmidt, Nishikawa,F.,T.Endo,T.Shimada,H.Fujii,T. autumn and winter temperatures (>15°C) (lit- E.E. and A. Golomb.1982. The carbohydrate balance of Shimizu,Y. Kobayashi, T. Araki, and M. Omura. alternate bearing citrus trees and the significance of re- 2010. Transcriptional changes in CiFT –introduced tle or no CiFT expression) will place an addi- serves for flowering and fruiting. Journal of the Ameri- transgenic trifoliate orange (Poncirus trifoliate L. Raf.). tional burden on flowering potential. can Society of Horticultural Science 107:206–208. Tree Physiol. 30:431–439 Goldschmidt, E.E., N. Aschkenazi, Y. Herzano, Schneider, H. 1968. Anatomy of Citrus. In: Reuther, References cited A.A. Schaffer, and S.P. Monselise. 1985. A role for W. (Ed.), The citrus industry. Univ. Calif Div Agr Sci, carbohydrate levels in the control of flowering in cit- Berkeley, Vol. 2, p. 1–85. Abbott, C. E. 1935. Blossom-bud differentiation in rus. Scientia Horticulturae. 26:159–166. Stander, O.P.J. 2014. The benefits of hand thinning citrus trees. American Journal of Botany 22(4):476– Goldschmidt, E.E. 1999. Carbohydrate supply as a Nadorcott mandarin (Citrus reticulata Blanco). South 485. critical factor for citrus fruit development and produc- African Fruit Journal, 14(3) Dec/Jan: 59-61. Bangerth, K.F. 2009. Floral induction in mature, tivity. HortScience. 34:1020-1024. Van der Merwe, I.S. 2012. Studies on the phenology perennial angiosperm fruit trees: Similarities and Guardiola, J.L. 1981. Flower initiation and develop- and carbohydrate status of alternate bearing 'Nador- discrepancies with annual/biennial plants and the in- ment in Citrus. Proc. Intl. Soc. Citricult. 2:242–246. cott' mandarin trees. MSc Thesis, Univ. Stellenbosch, volvement of plant hormones. Scientia Horticulturae. South Africa. 122:153–163. Guardiola, J.L., C. Moneri, and M. Agusti. 1982. The inhibitory effect of gibberellic acid on flowering in Valiente, J.I. and L.G. Albrigo. 2004. Flower bud Ben-Cheikh, W., J. Perez-Botella, F.R. Tadeo, M. Citrus. Physiologia Plantarum. 55:136–142. induction of sweet orange trees [Citrus sinensis (L.) Talon, and E. Primo-Millo. 1997. Pollination in- Koch, K.E. 1996. Carbohydrate modulated gene ex- Osbeck]: Effect of low temperatures, crop load, and creases gibberellin levels in developing ovaries of seed- pression in plants. Annual reviews of plant physiology bud age. Journal of the American Society of Horticul- ed varieties of citrus. Plant physiology. 114:557-564. and plant molecular biology. 47:509–540. tural Science. 129:158–164. Davenport T.L. 1990. Citrus flowering. In: Horti- Koshita, Y., T. Takahara, T. Ogata and A. Goto. Verreynne, J.S. 2005. The mechanism and underly- cultural Reviews. Vol. 12 (ed. J. Janick), pp. 349–408. 1999. Involvement of endogenous plant hormones ing physiology perpetuating alternate bearing in ‘Pixie’ Timber Press, Portland. (IAA, ABA, Gas) in leaves and flower bud formation mandarin (Citrus reticulata Blanco). PhD thesis, Uni- Erner, Y. and L. Shomer. 1996. Morphology and of Satsuma mandarin (Citrus unshiu Marc.). Scientia versity of California, Riverside. anatomy of stems and pedicels of spring flush shoots Horticulturae. 79:185-194. Verreynne, J.S., Lovatt, C.J. 2009. The effect of crop associated with citrus fruit set. Annals of Botany (Lon- Krajewski, A. and E. Rabe. 1995. Citrus flowering: load on budbreak influences return bloom in alternate don). 77:537–545. a critical review. Journal of Horticultural Science. bearing 'Pixie' mandarin. Journal of the American So- García-Luis, A., F. Fornes, and J.L. Guardiola. 70:357–374. ciety of Horticultural Science. 134 (3):299–307.

TECHNOLOGY | CRI 69 AUGUST | SEPTEMBER 2015