Dispatch R651

Photoperiodism: The consistent use of CONSTANS Alon Samach and Ayala Gover

Photoperiodic induction of flowering in the long-day Genetic background, the age of the plant, plant hormones plant is mediated by the circadian and environmental signals all have diverse effects on regulated CONSTANS gene. New evidence suggests transcript levels of both genes in a way that correlates with that CONSTANS-like genes have a similar role in short- flowering [6–9]. day induction of flowering of rice and Pharbitis. The SOC1 and FT genes are directly activated by the Address: Institute of Plant Sciences and Genetics in Agriculture, Faculty of Agriculture, The Hebrew University of Jerusalem P.O. product of the CONSTANS (CO) gene, which mediates the Box 12, Rehovot 76100, Israel. photoperiodic induction of flowering [6]. The CO gene was E-mail: [email protected] originally identified because of the late-flowering pheno- type of co mutant plants grown in long-day conditions Current Biology 2001, 11:R651–R654 (Figure 1) [10]. Transgenic plants in which CO is overex- 0960-9822/01/$ – see front matter pressed contain high levels of SOC1 and FT mRNAs, flower © 2001 Elsevier Science Ltd. All rights reserved. very early both in short and long days, and are no longer responsive to daylength [11]. CO encodes a protein with two In non-equatorial regions, the duration of light hours in a conserved domains [12,13]: an amino-terminal zinc finger, 24 hour day expands and decreases systematically in an B-box domain, predicted to mediate protein–protein inter- annual cycle. Living organisms need to adjust their behav- actions; and a carboxy-terminal CCT domain required for ioural programs to such changes in day length, but they nuclear localization and protein–protein interactions [13,14]. also seem to make use of them to initiate developmental As CO lacks any evident DNA-binding motif, other DNA programs at a certain time of the year. In many plants, the binding proteins may recruit it to promoters. onset of sexual reproduction — flowering — is triggered by a change in photoperiod, a phenomenon known as CO is also involved in the photoperiodic response of the . ‘Short-day’ plants remain vegetative until short day rice plant, Oryza sativa [4]. Different rice culti- they perceive a daylength that is shorter than a certain vars show a large degree of genetic variability, from pho- period, whereas ‘long-day’ plants flower only when the toperiod insensitivity to a strong photoperiod sensitivity daylength exceeds a certain limit. that causes a delay of flowering — ‘heading’ — in long days. The heading date is a complex trait controlled by Since photoperiodism was discovered by Garner and multiple genes known as quantitative trait loci (QTLs). Allard, intense physiological investigations have provided One major QTL, named Hd1, was recently identified as a insight into this process [1]. Most experiments were per- CO homolog [4]. The Hd1 has both conserved domains formed on ‘obligatory’ plants, which remain vegetative and also exhibits sequence similarity with CO and CO-like unless they encounter a certain daylength, such as the Arabidopsis proteins in regions between these domains. short-day plant Pharbitis nil. Other plants have less stringent Comparison of the sequences of hd1 alleles from three requirements. This is true, for example of the model plant different rice cultivars showed that certain alleles that — in this case, plants grown in long- cause decreased photoperiodic sensitivity have insertions day conditions flower earlier and with fewer leaves than or deletions that either truncate the B-box domain or those grown under short days. Molecular genetic studies cause the loss of the CCT domain (Figure 1). In the have identified several genes that participate in the pho- obligatory short-day plant Pharbitis, a CO homolog (PnCO) toperiodic control of flowering of Arabidopsis [2]. Recent was identified by differential analysis of induced versus findings [3–5] suggest that these genes play a major role in non-induced tissues. When overexpressed in transgenic photoperiodic control of flowering in both long-day and Arabidopsis plants, PnCO was capable of promoting flower- short-day plants. ing of co mutants [5].

In Arabidopsis, the transition to reproductive development Short-day plants might have the same downstream flower- is normally initiated once flower-promoting proteins have promoting genes — FT and SOC1 — as long-day plants. In accumulated to certain threshold levels. Two such pro- rice, allelic variation in an FT-like gene was recently teins are the products of the FLOWERING LOCUS T (FT) reported to be the cause of an additional QTL for pho- and SUPPRESSOR OF OVEREXPRESSION OF CO 1 toperiod sensitivity — Hd3A, [15]. Overexpression of a (SOC1; also known as AGL20) genes. Loss-of-function SOC1 homolog, taken from the long-day mustard plant, mutations in either of these genes delay flowering, and was found to cause early flowering in the short-day tobacco overexpression of either gene causes very early flowering. variety Maryland Mammoth [9]. R652 Current Biology Vol 11 No 16

Figure 1 suggest that what actually causes the difference between long-day and short-day plants is the interaction of CO or its partners with light via phytochrome. Arabidopsis The photoperiodic control of flowering time requires a mechanism that measures the duration of a photoperiod. In 1936, Erwin Bünning first proposed that the fulfills such a role. The circadian clock is an endoge-

Time to flower nous pacemaker that generates rhythms, which form an coÐ Wild type hy1Ð approximate 24 hour cycle. These rhythms are daily entrained by (or synchronized with) day–night transitions Long days Reproductive development Short days of light and temperature, and can further persist, for a few

cycles, after transition to continuous conditions of light or dark. Bünning and others showed that the inhibitory effects of ‘night break’ treatments on the flowering of short-day plants varied with the time of treatment, showing a rhythmic, circadian (approximately 24 hour) pattern

Time to flower (Figure 2) [1,18]. This suggested that a free-running inter- hd1Ð Wild type se5Ð nal circadian oscillator is creating an output involved in Rice ‘gating’ the effect of light on flowering. There is evidence Current Biology that light input into the oscillator is ‘gated’ as well, allow- ing reduced sensitivity of the circadian clock to transitory Reproductive development in Arabidopsis is induced (arrow) by long fluctuations in light levels during the night. The light days and repressed (T bar) by short days. In rice, short days induce flowering, while long days repress flowering. Mutations in Arabidopsis input pathway is thus rhythmically regulated by feedback CO [12] or in the CO-like ‘Kasalath’ hd1 allele of rice [4] reduce from the oscillator, creating a ‘time-taker’ that could create photoperiodic sensitivity. The ‘Kasalath’ hd1 allele causes a reduction rhythmic input even under constant conditions. in flowering time in long days. In Arabidopsis, some alleles of CO cause slightly earlier flowering in short days, although others do not [13]. Complete loss of function alleles will determine whether or not Recent studies in Arabidopsis have helped to clarify the role the functional protein is actively repressing flowering in non-inductive of the EARLY FLOWERING 3 (ELF3) gene in both ‘gating’ conditions. Chromophore biosynthesis mutants in rice (se5) [17] and responses [19–22]. ELF3 masks light signaling both to the Arabidopsis (hy1) [12] respond similarly to photoperiod. In both clock and to circadian outputs. The product accumulates in mutants, flowering is slightly faster in long days. the nucleus and can potentially interact with the light receptor phytochrome B [21]. Expression of ELF3 is clock regulated, and by altering the levels of ELF3 so that Plants detect changes in light quality through photorecep- expression is greatest at night, the plant is able to restrict tors, mainly the red/far-red-absorbing phytochromes and the sensitivity of its clock to light signals during the night. the blue/UV-A-absorbing . Physiological and When a plant is artificially subjected to constant light con- genetic evidence has shown that both types of photorecep- ditions, the ELF3 product continues to gate light respon- tor have roles in measuring and reacting to changes in pho- siveness in the late subjective day and early subjective toperiod [16]. For example, in the short-day plant Pharbitis, night, allowing sustained high-amplitude rhythms of circa- one long inductive night is required for flowering. A short dian output. In Arabidopsis, most mutations that perturb five-minute interruption of this night with low fluence red circadian clock function result in abnormal responses to light — ‘night break’ — can keep the plant vegetative. photoperiod [23]. Indeed, most clock-associated genes were originally identified as flowering-time mutants. Muta- The effects of mutations have revealed the involvement tions in ELF3 cause early, photoperiod-insensitive flower- of specific photoreceptors in the regulation of flowering ing, while transgenic plants in which ELF3 is overexpressed time [16]. In Arabidopsis, 2 and PHY- are late flowering in long days [21]. TOCHROME A seem to be specifically required for long- day promotion of flowering, and PHYTOCHROME B Saurez-lopez and colleagues [3] have provided elegant (PHYB) acts as a floral repressor. Interestingly, the reduction molecular and genetic evidence that CO mediates the or absence of functional phytochromes in chromophore flowering phenotype of clock mutants. CO mRNA levels biosynthesis mutants causes the normally short-day rice are under circadian clock regulation, with peak expression (se5) [17] and long-day Arabidopsis (hy1) [12] to respond occurring 16–20 hours from the transition to constant con- similarly to photoperiod (Figure 1). In both mutants, flow- ditions. There is still no evidence that CO protein cycles, ering is slightly faster in long-day conditions. This might although indirect evidence suggests that the protein is Dispatch R653

unstable and should therefore follow a similar pattern [3]. Figure 2 How does CO specifically promote flowering in long days? In Arabidopsis, long days do not cause an overall increase in (a) CO mRNA. At certain time points during the day, the levels of CO mRNA are higher in long-day plants than in short-day plants [3]. In fact, the peak of CO expression is broadened as days become longer, an effect common to other circadian regulated genes in plants and mammals [23,24]. As a result, high levels of CO mRNA during the Flowering ‘light on’ period can only be found in long-day plants [3]. If the same pattern exists for the protein, then the activa- tion of flower-promoting genes by CO might require light 24 48 [3,23]. Probably as a result to loss of gating of light signals, the elf3 mutant shows abnormally high levels of CO mRNA (b) at dawn in both short and long photoperiods, which might cause early flowering [3]. ELF3 mRNA itself seems to be affected in a similar way in the elf3 mutant [20].

Interactions between the circadian clock and photoperiod, expression and the way they combine to regulate gene expression leading to seasonal behavioral changes, have also intrigued PnCO researchers working on mammalian systems. The Siberian hamster responds to short photoperiods by undergoing 24 48 gonadal regression, molting to a winter pelage, and a decrease in body and lipid mass. In mammals, the central (c) circadian pacemaker is located in the suprachiasmatic nucleus (SCN). Mammals show a form of memory for day length at the behavioral level. This is clear from afteref- fects of day length on several aspects of free-running circa-

dian rhythms. For example, the intrinsic rhythmicity of expression the SCN in constant dark is dependant on prior day length CO [25]. Pharbitis plants show a similar effect of day length on the free-running circadian rhythm of night breaks [18].

24 48 Pharbitis PnCO mRNA levels are also under circadian Current Biology regulation. In fact, the free-running circadian peaks of Arabidopsis CO and Pharbitis PnCO genes are quite similar, Circadian rhythms in short-day and long-day plants. (a) When dark- even though the entrainment and constant conditions used grown Pharbitis are exposed to a 4 h photoperiod followed in studies were quite different for the two species — light by a 48 h dark inductive period, ‘night beak’ interruptions are most in Arabidopsis and dark in Pharbitis (Figure 2). Other CO- efficient (reduction of flowering) at ~15 and ~39 from the beginning of the light period, suggesting a circadian rhythm [18]. (b) Circadian like genes in Arabidopsis are circadian regulated, yet their expression of PnCO in Pharbitis under continuous dark shows an peak expression times and overexpression phenotypes maximum at ~16–24 h [5]. (c) Circadian expression of CO in suggest that they differ from CO and PnCO in function [26]. Arabidopsis under continuous light shows a similar pattern [3]. White Unfortunately, there are still no data available on PnCO and black bars represent light and dark, respectively. Hatched bars represent subjective nights. expression in normal photoperiods. PnCO mRNA accumu- lated in the dark, and long periods of light seemed to reduce the CO mRNA level. A night break treatment, although completely delaying flowering, had no clear level in Arabidopsis co mutant plants [5]. It remains possi- immediate or longer term effect on PnCO mRNA levels [5]. ble that the differences lie within the regulatory domains of the down-stream FT and SOC1 genes. Important answers The exact mechanism by which CO mediates the pho- will likely come from comparative data on gene expres- toperiodic response in long-day and short-day plants is still sion, post-transcriptional regulation and in situ localization. not clear. It is not likely that the differences are intrinsic to Further studies on the CO product, its interacting proteins the CO protein, as Pharbitis PnCO and Arabidopsis CO were and its possible modification by light will most likely found to have similar effects when expressed at a high provide additional insights. R654 Current Biology Vol 11 No 16

In plants, changes in photoperiod are thought to be per- 13. Robson F, Costa MMR, Hepworth S, Vizir I, Pineiro M, Putterill J, Coupland G: Functional importance of conserved domains in the ceived by the leaves, which when induced produce an flowering-time gene CONSTANS demonstrated by analysis of unknown flowering stimulus known as ‘florigen’. This sys- mutant alleles and transgenic plants. Plant J 2001, in press. temic signal is thought to move from the induced leaf to 14. Strayer C, Oyama T, Schultz TF, Raman R, Somers DE, Mas P, Panda S, Kreps JA, Kay SA: Cloning of the Arabidopsis clock gene TOC1, the meristem [1,27]. This was nicely demonstrated when a an autoregulatory response regulator homolog. Science 2000, leaf exposed to inductive conditions caused flowering 289:768-771. 15. Kojima S, Monna L, Fuse T, Sasaki T, Yano M: Map based cloning of when grafted to a plant grown in non-inductive conditions. the rice photoperiod sensitivity QTL, Hd3A. Plant & Animal Genome Could this grafted, ‘induced’ leaf introduce a new rhythm IX Conference 2001; P116. into the non-induced plant? Experiments in mammals 16. Whitelam GC, Devlin P F: Roles of different phytochromes in Arabidopsis photomorphogenesis. Plant Cell Environ 1998, show that when explanted from the body, the rhythm 20:752-758. within the mammalian SCN reflects the day length the 17. Izawa T, Oikawa T, Tokutomi S, Okuno K, Shimamoto K: whole animal had been exposed to prior to removal [25]. Phytochromes confer the photoperiodic control of flowering in rice (a short-day plant). Plant J 2000, 22:391-399. Plants do not have a centrally localized pacemaker, rather 18. Lumsden PJ, Millar AJ: Biological Rhythms and Photoperiodism in they contain a spatial array of autonomous circadian oscil- Plants. Oxford: BIOS Scientific Publishers; 1998. 19. McWatters HG, Bastow RM, Hall A, Millar AJ: The ELF3 zeitnehmer lators [28]. Although different organs on a plant can be regulates light signaling to the circadian clock. Nature 2000, entrained independently [28] perhaps at least part of the 408:716-720. elusive floral stimulus is the transmissible change in rhyth- 20. Hicks KA, Albertson TM, Wagner DR: EARLY FLOWERING 3 encodes a novel protein that regulates circadian clock function micity introduced by the ‘induced’ tissue. and flowering in Arabidopsis. Plant Cell 2001, 13:1281-1292. 21. Liu XL, Covington MF, Fankhauser C, Chory J, Wagner DR: ELF3 encodes a circadian clock-regulated nuclear protein that functions Acknowledgements in an Arabidopsis PHYB signal transduction pathway. Plant Cell We would like to thank George Coupland and Claire Périlleux for insightful 2001, 13:1293-1304. comments and the sharing of unpublished data. 22. Covington MF, Panda S, Liu XL, Strayer CA, Wagner DR, Kay SA: ELF3 modulates resetting of the circadian clock in Arabidopsis. References Plant Cell 2001, 13:1305-1315. 23. Samach A, Coupland G: Time measurement and the control of 1. Thomas B, Vince Prue D: Photoperiodism in Plants. San Diego: flowering in plants. Bioessays 2000, 22:38-47. Academic Press 1997. 24. Nuesslein-Hildesheim B, O’Brien JA, Ebling FJP, Maywood ES, 2. Périlleux C, Bernier G: The control of flowering: do genetical and Hastings MH: The circadian cycle of mPER clock gene products in physiological approaches converge? Plant Reproduction Annual In . the suprachiasmatic nucleus of the Siberian hamster encodes Plant Reviews. Vol.6. Edited by O’Neill SD, Roberts JA. Sheffield: both daily and seasonal time. Eur J Neurosci 2000, 12:2856-2864. Sheffield Academic Press; 2001, in press. 25. Ledger S, Strayer C, Ashton F, Kay S, Putterill, J: Analysis of the 3. Suarez-Lopez P, Wheatley K, Robson F, Onouchi H, Valverde F, function of two circadian-regulated CONSTANS-like genes. Plant J CONSTANS mediates between the circadian clock Coupland G: 2001, 26:15-22. and the control of flowering in Arabidopsis. Nature 2001, 26. Jagota A, de la Iglesia HO, Schwartz WJ: Morning and evening 410:1116-1120. circadian oscillations in the suprachiasmatic nucleus in vitro. 4. Yano M, Katayose Y, Ashikari M, Yamanouchi U, Monna L, Fuse T, Nature Neurosci 2000, 3:372-376. Baba T, Yamamoto K, Umehara Y, Nagamura Y, Sasaki T: Hd1, a 27. Ma H: From photoperiodism to Florigen. Curr Biol 1998, major photoperiod sensitivity QTL in rice, is closely related to the 8:R690-R692. Arabidopsis flowering time gene CONSTANS. Plant Cell 2000, 28. Thain SC, Hall A, Millar AJ: Functional independence of circadian 12:2473-2483. clocks that regulate plant gene expression. Curr Biol 2000, 5. Liu J, Yu J, McIntosh L, Kende H, Zeevaart JAD: Isolation of a 10:951-956. CONSTANS ortholog from Pharbitis nil and its role in flowering. Plant Physiol 2001, 125:1821-1830. 6. Samach A, Onouchi H, Gold SE, Ditta GS, Schwarz-Sommer Z, Yanofsky MF, Coupland G: Distinct roles of CONSTANS target genes in reproductive development in Arabidopsis. Science 2000, 288:1613-1616. 7. Kardailsky I, Shukla VK, Ahn JH, Dagenais N, Christensen SK, Nguyen, JT, Chory J, Harrison MJ, Weigel D: Activation tagging of the floral inducer FT. Science 1999, 286:1962-1965. 8. Lee H, Suh SS, Park E, Cho E, Ahn JH, Kim SG, Lee JS, Kwon YM, Lee I: The AGAMOUS-LIKE 20 MADS domain protein integrates floral inductive pathways in Arabidopsis. Genes Dev 2000, 14:2366-2376. 9. Borner R, Kampmann G, Chandler J, Gleissner R, Wisman E, Apel K, Melzer S: A MADS domain gene involved in the transition to flowering in Arabidopsis. Plant J 2000, 24:591-599. 10. Koornneef M, Hanhart CJ, van der Veen JH: A genetic and physiological analysis of late flowering mutants in Arabidopsis thaliana. Mol Gen Genet 1991, 229:57-66. 11. Onouchi H, Igeno MI, Perilleux C, Graves K, Coupland G: Mutagenesis of plants overexpressing CONSTANS demonstrates novel interactions among Arabidopsis flowering-time genes. Plant Cell 2000, 12:885-900. 12. Putterill J, Robson F, Lee K, Simon R, Coupland G: The CONSTANS gene of Arabidopsis promotes flowering and encodes a protein showing similarities to zinc finger transcription factors. Cell 1995, 80:847-857.