Plant Circadian Clocks Increase Photosynthesis, Growth, Survival, and Competitive Advantage
´vei, 3Antony N. Dodd, 1Neeraj Salathia, 2*Anthony Hall, 2. Eva Ke
´ka To ´th, 3Ferenc Nagy, 3Julian M. Hibberd, 1Andrew J. Millar, 2-Re
Alex A. R. Webb 1`
Circadian clocks are believed to confer an advantage to plants, but the nature of that advantage has been unknown. We show that a substantial photo-synthetic advantage is conferred by correct matching of the circadian clock period with that of the external light-dark cycle. In wild type and in long–and short–circadianperiod mutants of Arabidopsis thaliana , plants with a clock period matched to the environment contain more chlorophyll, fix more carbon, grow faster, and survive better than plants with circadian periods differing from their environment. This explains why plants gain advantage from cir-cadian control.
Circadian clocks produce an internal estimate of time that synchronizes biological events with external day-night cycles (1). Clocks with
Department of Plant Sciences, University of Cambridge, Downing Street, Cambridge, CB23EA, UK. 2Department of Biological Sciences, University of Warwick, Coventry, CV47AL, UK. 3Plant Biology Institute, Biological Research Centre of the Hungarian Academy of Sciences, Post Office Box 521, H-6701Szeged, Hungary.
*Presentaddress:Bauer Center for Genomics Research, Harvard University, 7Divinity Avenue, Cambridge, MA 02138, USA.
. Present address:School of Biological Sciences, Uni-versity of Liverpool, Crown Street, Liverpool L697ZB, UK.
-Present address:Institute of Molecular Plant Sciences, Mayfield Road, University of Edinburgh, Edinburgh EH93JH, UK.
`To whom correspondence should be addressed. E-mail:[email protected]
1
similar properties and regulatory architecture have evolved at least four times, indicating that circadian rhythms confer a selective advantage (2). In plants, circadian rhythms control gene expression, stomatal opening, and the timing component of photoperiodism, which regulates seasonal reproduction, but the basis for their contribution to fitness during vegetative growth remains undetermined (3, 4). Indirect evidence suggests a physiological benefit from circadian rhythms during growth under unnaturally short photoperiods (5). Cyanobacteria and higher plants gain an advantage when the endogenous period is matched to the light-dark cycle (6–8). Rhythmic growth inhibitor secretion might cause the growth advantage in cyanobacteria (7, 8), but this hypothesis may not apply to multicellular eukaryotes. We demonstrate that
VOL 309
SCIENCE
when correctly tuned, the Arabidopsis circadi-an system enhances chlorophyll content, photo-synthetic carbon fixation, and growth. We also show that circadian enhancement of photo-synthesis leads to improved survival and competitive advantage.
Biological clocks have evolved so that clock outputs are in phase with the Earth _s ro-tation. We wished to identify and quantify mechanisms by which the clock confers ad-vantage in light-dark cycles. We hypothesized that matching the endogenous clock period (t ) with the period of exogenous light-dark cycles (T)E so called B circadian resonance [(7) ^pro-vides an advantage by optimizing the phase relation between clock-controlled biology and exogenous day-night cycles. Plants having clocks that are dissonant from the environment, therefore, may be disadvantaged. To test this hypothesis, we compared the performance of wild-type plants with lines having mutations that alter clock period length, in a range of environmental period lengths (B T cycles [) that were either matched or unmatched to the en-dogenous clock period.
We used three experimental approaches to test this hypothesis (9). First, wild-type plants with a circadian period of about 24hours were grown in 10hours light–10hours dark (T20),12hours light–12hours dark (T24)and 14hours light–14hours dark (T28)cycles. Second, we grew the long-and short-period mutants ztl-1E t 027.1hours–32.5hours; (10) ^and toc1-1E t 020.7hours; (11) ^in T cycles that were similar to, or dissimilar from, their endogenous clock periods (T20and T28). In these T-cycle experiments, relative perform-ance was measured within, not between, geno-types, which specifically quantified the benefit
63022JULY 2005www.sciencemag.org
R E P O R T S
Fig. 1. Leaves contain more chlorophyll when their clock period matches the environmental period. (A ) Total chlorophyll in Col-0wild type grown under T20, T24, and T28, and in arrhythmic line CCA1-ox under T24. (B ) Total chlorophyll in ztl-1(long-periodmutant) and toc1-1(short-periodmutant) grown in T20and T28. For all groups, n 05; data are means þSEM. In two-sample t tests comparing chlorophyll concentrations to the line having the period matched to the environment, significance of re-sults:*P G 0.05, **P G 0.01.
of circadian resonance and excluded effects not associated with the clock. Last, the effect of circadian arrhythmia on growth and physi-ology was investigated in well-characterized arrhythmic plants overexpressing the molecu-lar oscillator component CCA1(CCA1-ox), and compared with rhythmic wild types (8, 12). Experiments were conducted during vegetative growth, to assess the contribution of circadian resonance to growth and fitness, and to elim-inate pleiotropic effects on life history due to the flowering time alterations that arise in circadian period mutants (13). We assessed the contribution of circadian resonance to carbon fixation, biomass, and leaf chlorophyll. Carbon fixation rates (14–16) and biomass (17) are traits associated with plant fitness; therefore, our study provides information concerning specific mechanisms by which the clock contributes to fitness.
In wild type and in short-and long-period mutants, leaves contained more chlorophyll when the oscillator period matched that of the environment. Leaves of Columbia-0(Col-0) wild-type plants grown for 30days in T24contained more chlorophyll than Col-0grown in T20or T28(Fig.1A). When ztl-1and toc1-1were grown under T20and T28, the long-period mutant ztl-1contained more chlorophyll
Fig. 2. The circadian clock enhances photosynthetic carbon fixation. (A ) Mean C fixation per hour in ztl-1and toc1-1grown in T20and T28. (B ) Mean C fixation per hour in Col-0wild type and arrhythmic CCA1-ox, in T24. (C ) CCA1overexpression (opencircles) abolishes the circadian rhythms of CO 2fixation and stomatal opening that occur in Col-0wild type (filledcircles). (D ) CO 2assimilation and stomatal conductance in CCA1-ox (opencircles) and Col-0wild type (filledcircles) under light-dark cycles (indicatedby bars on the x axis). For these experiments, n 06; data are means T SEM (Aand B) or largest standard error (Cand D). In two-sample t tests comparing net C fixation per hour to the line with the clock period matched to the environment, significance of results:*P G 0.05, **P G 0.01.
after growth in T28than T20, whereas the short-period mutant toc1-1contained more chlorophyll after growth in T20than T28(Fig.1B). Therefore, correct matching of the circadian period with the external period in-creases chlorophyll accumulation. When the Col-0clock was stopped by CCA1over-expression, less chlorophyll was present com-pared with wild-type Col-0under T24, which confirmed the dependence of chlorophyll ac-cumulation on clock function (Fig.1A).
There are circadian rhythms in transcript abundance of genes associated with chloro-phyll synthesis, heme production, chlorophyll accumulation, and synthesis of chlorophyll-binding proteins (11, 18–20). Virtually all clock-controlled genes associated with chlorophyll synthesis and the light-harvesting apparatus exhibit peak circadian transcript abundance 4hours after subjective dawn (18), which suggests that circadian expression of these genes could be important for enhancing light-harvesting capacity. Although the amount of light-harvesting complex pigments and pro-teins remains uniform over the diel cycle (19), the clock might sustain steady-state levels of proteins that exhibit light-induced degrada-SCIENCE
VOL 309
tion, by way of circadian changes in turn-over. This could explain why chlorosis can occur under very long photoperiods (21), because the duration of light-induced degrada-tion of light-harvesting complex proteins ex-tends beyond their period of clock-enhanced transcription.
Because chlorophyll content was greatest under circumstances of matched endogenous and environmental periods, we examined wheth-er circadian resonance improves photosynthesis. We compared net carbon fixation of long-and short-period mutants and CCA1-ox, under T20, T24, and T28. The long-period mutant ztl1fixed 42%more carbon under T28than T20, whereas the short-period mutant toc1-1fixed 40%more carbon under T20than T28(Fig.2A). Circadian resonance, therefore, increases CO 2fixation. Col-0wild type fixed 67%more carbon than arrhythmic CCA1-ox (Fig.2B). The reduction in carbon fixation associated with circadian arrhythmia was, therefore, great-er than the disadvantage caused by the È8-hour period mismatch with the environment that occurred in ztl1and toc1-1. Under continuous light, the rate of CO 2fixation was lower in CCA1-ox than wild type for the first 48hours
www.sciencemag.org 22JULY 2005631
R E P O R T S
of continuous light, but during prolonged constant light, assimilation was higher in CCA1-ox than wild type (Fig.2C). This was reminiscent of the outcompetition of rhythmic cyanobacterial lines by the arrhythmic line CLAb under continuous light (8). CCA1over-expression abolished circadian rhythms of CO 2fixation and stomatal opening in constant conditions. Fourier analysis (22) estimated period lengths of 24.1T 0.6hours and 23.5T 0.3hours for stomatal conductance and carbon assimilation, respectively, in Col-0wild type, but failed to detect circadian regulation for CCA1-ox. Similarly, the toc1-1and ztl1mu-tations cause respective shortening and exten-sion of the circadian period of CO 2fixation and stomatal opening rhythms that occur in continuous light (23, 24).
Under light-dark cycles, rhythmic stoma-tal opening and closure was restored in CCA1-ox, but anticipation of dawn and dusk was absent, which demonstrated that the clock remained stopped in CCA1-ox, even in light-dark cycles (Fig.2D). The stomata continued to open for the entire photoperiod in CCA1-ox, whereas in Col-0, stomatal open-ing ceased around midday. CCA1-ox, there-fore, had higher total transpiration than Col-0during the light period (Fig.2D). Thus, the clock allows stomata to anticipate dusk and might participate in enhancement of water-use efficiency.
Because photosynthesis increased when exogenous and endogenous periods were simi-lar (Fig.2) and because long-term carbon fixation is correlated to leaf chlorophyll con-tent (25), we reasoned that circadian resonance might increase vegetative growth. Col-0wild type grown under T20, T24, and T28had greatest vegetative biomass in T24(Fig.3, A
Fig. 3. Environmentally matched clock period enhances vegetative growth. (A and B ) Dry aerial biomass (A)and visible leaf area (B)in Col-0wild type after growth under T20, T24, and T28. (C and D ) Dry aerial biomass (C)and vis-ible leaf area (D)after growth of the long-period mutant ztl-1and short-period mutant toc1-1under T20and T28. (E and F ) Dry aerial biomass (E)and visible leaf area (F)after 35days’growth of Col-0wild type and arrhythmic CCA1-ox line in T24. Biomass was mea-sured after 32days (Aand C) or 35days (E).For all groups, n 05; data are means þor T SEM. (A,C, and E) In two-sample t tests comparing aerial biomass to the line with the clock period matched to the environment, significance of results:*P G 0.05, **P G
0.01.
and B). Under T20, aerial biomass was reduced by 47%relative to T24, and growth under T28resulted in 42%less biomass relative to T24. When the long-and short-period mutants ztl1and toc1-1were grown under T20and T28, ztl1had maximum aerial biomass and leaf area under T28. In toc1-1, aerial biomass and leaf area were maximal under T20(Fig.3, C and D). These measures of growth were lower in toc1-1than ztl1-1under all conditions. Sepa-rately, we compared growth of Col-0and CCA1-ox under T24. After 32days, the aerial biomass of CCA1-ox was 53%lower than that of Col-0(Fig.3, E and F). We did not compare growth of CCA1-ox and Col-0in continuous light, because wild-type plants become arrhyth-mic under extended continuous light (26). Therefore, circadian resonance enhances growth in wild-type plants and mutants with altered circadian period, and stopping the clock further reduces growth. Enhancement of biomass and photosynthesis by the circadian clock conse-quently indicates processes by which the clock increases fitness (14–17).
Fitness arising from circadian resonance might be reported by seedset (5). We compared seedset from short-and long-period mutants and wild type, under T24. In a single experi-ment, there were no large or consistent differ-ences in mean seed production (toc1-1, 16200seeds per plant; wild type, 15757seeds per plant; ztl-1, 15005seeds per plant; n 015). Because the clock determines flowering time, which becomes altered in period mutants (13), seedset is likely to be an ambiguous marker for the fitness implications of circadian resonance. However, we have demonstrated that circadian resonance increases the established fitness traits of photosynthesis and biomass (Figs.1to 3) (14–17). We performed reciprocal competitions be-tween long-and short-period plants (7), using two short-period mutants of TOC1E toc1-1and toc1-2, t , 20hours (27) ^and two long-period mutants of ZTL E ztl-1and ztl-27, t , 28hours (9, 10) ^. Mixed populations of toc1and ztl were grown under T20and T28, which generated a crowded lawn of plants. In these conditions, interactions among neighboring plants of dif-fering genotypes affect physiological outcomes, in addition to the interaction between each plant and the light-dark cycle that we tested previously. We grew monocultures of each line, to assess the importance of differing growth rates among neighbors compared with growth rates among lines. Testing for com-petitive advantage derived from circadian resonance, using reciprocal competition, re-quires a mixed population of two period-length genotypes; therefore, it cannot be assessed in just wild type. The period-length mutants were appropriate for this experiment because the growth disadvantage that occurred when T m t in wild type also occurred with the period mutants (Figs.1to 3).
In two separate competition experiments, using different ZTL and TOC1mutants, under T20, TOC1mutants grew more successfully than ZTL mutants, as indicated by multiple parameters including chlorophyll content, leaf number, rosette diameter, and aerial biomass (Fig.4). Conversely, under T28, growth of ZTL mutants was enhanced compared with TOC1. This was similar to results obtained when plants were grown without competition (Figs.1to 3). However, competition caused mortality of some plants, which did not occur in the absence of competition. Mortality was greater in ztl-27under T20, and greater in toc1-2under T28(Fig.4B). Circadian resonance, therefore, en-
63222JULY 2005VOL 309SCIENCE www.sciencemag.org
R E P O R T S
hanced growth and survival, and this was more pronounced under competition than during monoculture (Fig.4A). In two separate mono-culture experiments, there was no consistent pattern of T cycle–dependentmortality (Fig.4C). Therefore, both poor individual growth and outcompetition confer a disadvantage under dissonant T cycles. Our data underestimate the true growth advantage that occurs under competition, because physiological parameters were not measured in dead plants. Neither genotype had an advantage in all conditions, which implicated circadian effects of the mutations rather than secondary phenotypes. Competition between toc1-1and ztl-1, and between toc1-1or toc1-2and ztl-27, gave the same result, which discounts the likelihood of background mutations or allele-specific effects (28). Arabidopsis entrains stably to T cycles far from t (24, 29, 30), so the long-term growth advantage was likely due to correct phasing of rhythmic processes relative to the environment in one genotype, and an incorrect phase in its competitor (7). This suggests that a correctly matched circadian clock confers a competitive advantage, whereas the enhance-ment of two key fitness traits E biomass and photosynthesis; (14–17) ^by circadian reso-nance indicates that enhanced photosynthesis
is one mechanism by which the clock in-creases fitness.
Our experiments demonstrate that the circa-dian clock allows plants to increase photo-synthesis and that the clock underlies a doubling of Arabidopsis productivity. This may derive from correct anticipation of dawn and dusk, and synchronization of the synthesis of light-harvesting complex proteins and chlorophyll, both of which are unstable in their unbound state (18). Incorrect matching of en-dogenous rhythms to environmental rhythms reduced leaf chlorophyll content, reduced as-similation, reduced growth, and increased mor-tality. Optimization of these parameters by circadian resonance could represent one of the mechanisms that has selected for circadian clock function during plant evolution. We suggest that selective plant breeding for enhanced crop performance must be performed carefully, be-cause phase and period changes could arise from the close genetic linkage of phase and period loci (31) to the trait under selection, and cause alterations to clock function that might reduce vegetative yield. Clock manipulation could enhance food production during exploration of space and other planets, where the light-dark cycle may differ from the terrestrial 24-hour period. Circadian resonance is likely to provide
an advantage in all kingdoms, because reso-nance of the internal clock with the external light-dark cycle ensures an optimal phase relation between physiology and the day-night cycle and provides the basis for anticipation of changes in environmental conditions.
References and Notes
1. C. H. Johnson, M. Knight, A. Trewavas, T. Kondo, in Biological Rhythms and Photoperiodism in Plants , P. L. Lumsden, A. J. Millar, Eds. (BiosScientific, Oxford, 1998), pp. 1–34.
2. M. W. Young, S. A. Kay, Nat. Rev. Genet. 2, 702(2001).3. M. J. Yanovsky, S. A. Kay, Curr. Opin. Plant Biol. 4, 429(2001).
4. T. P. Michael et al. , Science 302, 1049(2003).
5. R. M. Green, S. Tingay, Z. Y. Wang, E. M. Tobin, Plant Physiol. 129, 576(2002).
6. H. R. Highkin, J. B. Hanson, Plant Physiol. 29, 301(1954).
7. Y. Ouyang, C. R. Andersson, T. Kondo, S. S. Golden, C. H. Johnson, Proc. Natl. Acad. Sci. U.S.A. 95, 8660(1998).
8. M. A. Woelfle, Y. Ouyang, K. Phanvijhitsiri, C. H. Johnson, Curr. Biol. 14, 1481(2004).
9. Materials and methods are available on Science Online. 10. D. E. Somers, T. F. Schultz, M. Milnamow, S. A. Kay,
Cell 101, 319(2000).
´, C. A. Strayer, N.-H. Chua, S. A. 11. A. J. Millar, I. A. Carre
Kay, Science 267, 1161(1995).
12. Z. Y. Wang et al. , Plant Cell 9, 491(1997).
13. R. Hayama, G. Coupland, Curr. Opin. Plant Biol. 6, 13
(2003).
14. A. M. Arntz, E. H. DeLucia, N. Jordan, Ecology 81,
2567(2000).
15. J. K. Ward, J. K. Kelly, Ecol. Lett. 7, 427(2004).16. P. S. Curtis, X. Wang, Oecologia 113, 299(1998).17. T. Mitchell-Olds, Evol. Int. J. Org. Evol. 50, 140
(1996).
18. S. L. Harmer et al. , Science 290, 2110(suppl.data)
(2000).
19. A. Prombona, J. Argyroudi-Akoyunoglou, Plant Sci.
167, 117(2004).
20. K. Kaasik, C.-C. Lee, Nature 430, 467(2004).
21. A. P. Withrow, R. B. Withrow, Plant Physiol. 24, 657
(1949).
22. J. D. Plautz et al. , J. Biol. Rhythms 12, 204(1997).23. A. N. Dodd, K. Parkinson, A. A. R. Webb, New Phytol.
162, 63(2004).
24. D. E. Somers, A. A. R. Webb, M. Pearson, S. A. Kay,
Development 125, 485(1998).
25. H. Griffiths et al. , in Physiological Plant Ecology , M. C.
Press, J. D. Scholes, M. G. Barker, Eds. (Blackwell,Oxford, 1998), pp. 415–441.
26. T. L. Hennessey, C. B. Field, Plant Physiol. 96, 831
(1991).
´et al. , Science 293, 880(2001).27. D. Alabadı
28. N. Salathia et al ., unpublished observations.
29. M. J. Yanovsky, S. A. Kay, Nature 419, 308(2002).30. L. C. Roden, H. R. Song, S. Jackson, K. Morris, I. A.
´, Proc. Natl. Acad. Sci. U.S.A. 99, 13313(2002).Carre
31. D. E. Somers, W.-Y. Kim, R. Geng, Plant Cell 16, 769
(2004).
32. Authors E.K., R.T., A.H., and A.J.M. identified ztl-27; N.S.
and A.J.M. conceived and performed competition and seedset experiments with A.H.; A.N.D., J.M.H., and A.A.R.W. performed all other experiments. We thank the U.K. Biotechnology and Biological Sciences Re-search Council (A.A.R.W.,J.M.H., and A.J.M.); the Royal Society of London and the Isaac Newton Trust (A.A.R.W.);and the Howard Hughes Medical Institute (F.N.).We are grateful to E. Tobin (UCLA)and S. Kay (Scripps)for collaboration and donation of CCA1-ox and toc1-2, respectively. Supporting Online Material
www.sciencemag.org/cgi/content/full/309/5734/630/DC1
Materials and Methods References and Notes
1April 2005; accepted 7June 2005
10.1126/science.1115581
Fig. 4. Correct circadian period enhances growth and survival. (A ) In two separate competition experiments using different mutant lines and one monoculture experiment, comparative growth and survival of toc1and ztl under T20and T28; n 049, except n 017to 20for dry biomass values; data are means T SEM. (B ) Representative individuals from a competition experiment. Mortality in short-and long-period lines after competition (B)or monoculture (C ) in T20and T28.
www.sciencemag.org SCIENCE VOL 30922JULY 2005633
Plant Circadian Clocks Increase Photosynthesis, Growth, Survival, and Competitive Advantage
´vei, 3Antony N. Dodd, 1Neeraj Salathia, 2*Anthony Hall, 2. Eva Ke
´ka To ´th, 3Ferenc Nagy, 3Julian M. Hibberd, 1Andrew J. Millar, 2-Re
Alex A. R. Webb 1`
Circadian clocks are believed to confer an advantage to plants, but the nature of that advantage has been unknown. We show that a substantial photo-synthetic advantage is conferred by correct matching of the circadian clock period with that of the external light-dark cycle. In wild type and in long–and short–circadianperiod mutants of Arabidopsis thaliana , plants with a clock period matched to the environment contain more chlorophyll, fix more carbon, grow faster, and survive better than plants with circadian periods differing from their environment. This explains why plants gain advantage from cir-cadian control.
Circadian clocks produce an internal estimate of time that synchronizes biological events with external day-night cycles (1). Clocks with
Department of Plant Sciences, University of Cambridge, Downing Street, Cambridge, CB23EA, UK. 2Department of Biological Sciences, University of Warwick, Coventry, CV47AL, UK. 3Plant Biology Institute, Biological Research Centre of the Hungarian Academy of Sciences, Post Office Box 521, H-6701Szeged, Hungary.
*Presentaddress:Bauer Center for Genomics Research, Harvard University, 7Divinity Avenue, Cambridge, MA 02138, USA.
. Present address:School of Biological Sciences, Uni-versity of Liverpool, Crown Street, Liverpool L697ZB, UK.
-Present address:Institute of Molecular Plant Sciences, Mayfield Road, University of Edinburgh, Edinburgh EH93JH, UK.
`To whom correspondence should be addressed. E-mail:[email protected]
1
similar properties and regulatory architecture have evolved at least four times, indicating that circadian rhythms confer a selective advantage (2). In plants, circadian rhythms control gene expression, stomatal opening, and the timing component of photoperiodism, which regulates seasonal reproduction, but the basis for their contribution to fitness during vegetative growth remains undetermined (3, 4). Indirect evidence suggests a physiological benefit from circadian rhythms during growth under unnaturally short photoperiods (5). Cyanobacteria and higher plants gain an advantage when the endogenous period is matched to the light-dark cycle (6–8). Rhythmic growth inhibitor secretion might cause the growth advantage in cyanobacteria (7, 8), but this hypothesis may not apply to multicellular eukaryotes. We demonstrate that
VOL 309
SCIENCE
when correctly tuned, the Arabidopsis circadi-an system enhances chlorophyll content, photo-synthetic carbon fixation, and growth. We also show that circadian enhancement of photo-synthesis leads to improved survival and competitive advantage.
Biological clocks have evolved so that clock outputs are in phase with the Earth _s ro-tation. We wished to identify and quantify mechanisms by which the clock confers ad-vantage in light-dark cycles. We hypothesized that matching the endogenous clock period (t ) with the period of exogenous light-dark cycles (T)E so called B circadian resonance [(7) ^pro-vides an advantage by optimizing the phase relation between clock-controlled biology and exogenous day-night cycles. Plants having clocks that are dissonant from the environment, therefore, may be disadvantaged. To test this hypothesis, we compared the performance of wild-type plants with lines having mutations that alter clock period length, in a range of environmental period lengths (B T cycles [) that were either matched or unmatched to the en-dogenous clock period.
We used three experimental approaches to test this hypothesis (9). First, wild-type plants with a circadian period of about 24hours were grown in 10hours light–10hours dark (T20),12hours light–12hours dark (T24)and 14hours light–14hours dark (T28)cycles. Second, we grew the long-and short-period mutants ztl-1E t 027.1hours–32.5hours; (10) ^and toc1-1E t 020.7hours; (11) ^in T cycles that were similar to, or dissimilar from, their endogenous clock periods (T20and T28). In these T-cycle experiments, relative perform-ance was measured within, not between, geno-types, which specifically quantified the benefit
63022JULY 2005www.sciencemag.org
R E P O R T S
Fig. 1. Leaves contain more chlorophyll when their clock period matches the environmental period. (A ) Total chlorophyll in Col-0wild type grown under T20, T24, and T28, and in arrhythmic line CCA1-ox under T24. (B ) Total chlorophyll in ztl-1(long-periodmutant) and toc1-1(short-periodmutant) grown in T20and T28. For all groups, n 05; data are means þSEM. In two-sample t tests comparing chlorophyll concentrations to the line having the period matched to the environment, significance of re-sults:*P G 0.05, **P G 0.01.
of circadian resonance and excluded effects not associated with the clock. Last, the effect of circadian arrhythmia on growth and physi-ology was investigated in well-characterized arrhythmic plants overexpressing the molecu-lar oscillator component CCA1(CCA1-ox), and compared with rhythmic wild types (8, 12). Experiments were conducted during vegetative growth, to assess the contribution of circadian resonance to growth and fitness, and to elim-inate pleiotropic effects on life history due to the flowering time alterations that arise in circadian period mutants (13). We assessed the contribution of circadian resonance to carbon fixation, biomass, and leaf chlorophyll. Carbon fixation rates (14–16) and biomass (17) are traits associated with plant fitness; therefore, our study provides information concerning specific mechanisms by which the clock contributes to fitness.
In wild type and in short-and long-period mutants, leaves contained more chlorophyll when the oscillator period matched that of the environment. Leaves of Columbia-0(Col-0) wild-type plants grown for 30days in T24contained more chlorophyll than Col-0grown in T20or T28(Fig.1A). When ztl-1and toc1-1were grown under T20and T28, the long-period mutant ztl-1contained more chlorophyll
Fig. 2. The circadian clock enhances photosynthetic carbon fixation. (A ) Mean C fixation per hour in ztl-1and toc1-1grown in T20and T28. (B ) Mean C fixation per hour in Col-0wild type and arrhythmic CCA1-ox, in T24. (C ) CCA1overexpression (opencircles) abolishes the circadian rhythms of CO 2fixation and stomatal opening that occur in Col-0wild type (filledcircles). (D ) CO 2assimilation and stomatal conductance in CCA1-ox (opencircles) and Col-0wild type (filledcircles) under light-dark cycles (indicatedby bars on the x axis). For these experiments, n 06; data are means T SEM (Aand B) or largest standard error (Cand D). In two-sample t tests comparing net C fixation per hour to the line with the clock period matched to the environment, significance of results:*P G 0.05, **P G 0.01.
after growth in T28than T20, whereas the short-period mutant toc1-1contained more chlorophyll after growth in T20than T28(Fig.1B). Therefore, correct matching of the circadian period with the external period in-creases chlorophyll accumulation. When the Col-0clock was stopped by CCA1over-expression, less chlorophyll was present com-pared with wild-type Col-0under T24, which confirmed the dependence of chlorophyll ac-cumulation on clock function (Fig.1A).
There are circadian rhythms in transcript abundance of genes associated with chloro-phyll synthesis, heme production, chlorophyll accumulation, and synthesis of chlorophyll-binding proteins (11, 18–20). Virtually all clock-controlled genes associated with chlorophyll synthesis and the light-harvesting apparatus exhibit peak circadian transcript abundance 4hours after subjective dawn (18), which suggests that circadian expression of these genes could be important for enhancing light-harvesting capacity. Although the amount of light-harvesting complex pigments and pro-teins remains uniform over the diel cycle (19), the clock might sustain steady-state levels of proteins that exhibit light-induced degrada-SCIENCE
VOL 309
tion, by way of circadian changes in turn-over. This could explain why chlorosis can occur under very long photoperiods (21), because the duration of light-induced degrada-tion of light-harvesting complex proteins ex-tends beyond their period of clock-enhanced transcription.
Because chlorophyll content was greatest under circumstances of matched endogenous and environmental periods, we examined wheth-er circadian resonance improves photosynthesis. We compared net carbon fixation of long-and short-period mutants and CCA1-ox, under T20, T24, and T28. The long-period mutant ztl1fixed 42%more carbon under T28than T20, whereas the short-period mutant toc1-1fixed 40%more carbon under T20than T28(Fig.2A). Circadian resonance, therefore, increases CO 2fixation. Col-0wild type fixed 67%more carbon than arrhythmic CCA1-ox (Fig.2B). The reduction in carbon fixation associated with circadian arrhythmia was, therefore, great-er than the disadvantage caused by the È8-hour period mismatch with the environment that occurred in ztl1and toc1-1. Under continuous light, the rate of CO 2fixation was lower in CCA1-ox than wild type for the first 48hours
www.sciencemag.org 22JULY 2005631
R E P O R T S
of continuous light, but during prolonged constant light, assimilation was higher in CCA1-ox than wild type (Fig.2C). This was reminiscent of the outcompetition of rhythmic cyanobacterial lines by the arrhythmic line CLAb under continuous light (8). CCA1over-expression abolished circadian rhythms of CO 2fixation and stomatal opening in constant conditions. Fourier analysis (22) estimated period lengths of 24.1T 0.6hours and 23.5T 0.3hours for stomatal conductance and carbon assimilation, respectively, in Col-0wild type, but failed to detect circadian regulation for CCA1-ox. Similarly, the toc1-1and ztl1mu-tations cause respective shortening and exten-sion of the circadian period of CO 2fixation and stomatal opening rhythms that occur in continuous light (23, 24).
Under light-dark cycles, rhythmic stoma-tal opening and closure was restored in CCA1-ox, but anticipation of dawn and dusk was absent, which demonstrated that the clock remained stopped in CCA1-ox, even in light-dark cycles (Fig.2D). The stomata continued to open for the entire photoperiod in CCA1-ox, whereas in Col-0, stomatal open-ing ceased around midday. CCA1-ox, there-fore, had higher total transpiration than Col-0during the light period (Fig.2D). Thus, the clock allows stomata to anticipate dusk and might participate in enhancement of water-use efficiency.
Because photosynthesis increased when exogenous and endogenous periods were simi-lar (Fig.2) and because long-term carbon fixation is correlated to leaf chlorophyll con-tent (25), we reasoned that circadian resonance might increase vegetative growth. Col-0wild type grown under T20, T24, and T28had greatest vegetative biomass in T24(Fig.3, A
Fig. 3. Environmentally matched clock period enhances vegetative growth. (A and B ) Dry aerial biomass (A)and visible leaf area (B)in Col-0wild type after growth under T20, T24, and T28. (C and D ) Dry aerial biomass (C)and vis-ible leaf area (D)after growth of the long-period mutant ztl-1and short-period mutant toc1-1under T20and T28. (E and F ) Dry aerial biomass (E)and visible leaf area (F)after 35days’growth of Col-0wild type and arrhythmic CCA1-ox line in T24. Biomass was mea-sured after 32days (Aand C) or 35days (E).For all groups, n 05; data are means þor T SEM. (A,C, and E) In two-sample t tests comparing aerial biomass to the line with the clock period matched to the environment, significance of results:*P G 0.05, **P G
0.01.
and B). Under T20, aerial biomass was reduced by 47%relative to T24, and growth under T28resulted in 42%less biomass relative to T24. When the long-and short-period mutants ztl1and toc1-1were grown under T20and T28, ztl1had maximum aerial biomass and leaf area under T28. In toc1-1, aerial biomass and leaf area were maximal under T20(Fig.3, C and D). These measures of growth were lower in toc1-1than ztl1-1under all conditions. Sepa-rately, we compared growth of Col-0and CCA1-ox under T24. After 32days, the aerial biomass of CCA1-ox was 53%lower than that of Col-0(Fig.3, E and F). We did not compare growth of CCA1-ox and Col-0in continuous light, because wild-type plants become arrhyth-mic under extended continuous light (26). Therefore, circadian resonance enhances growth in wild-type plants and mutants with altered circadian period, and stopping the clock further reduces growth. Enhancement of biomass and photosynthesis by the circadian clock conse-quently indicates processes by which the clock increases fitness (14–17).
Fitness arising from circadian resonance might be reported by seedset (5). We compared seedset from short-and long-period mutants and wild type, under T24. In a single experi-ment, there were no large or consistent differ-ences in mean seed production (toc1-1, 16200seeds per plant; wild type, 15757seeds per plant; ztl-1, 15005seeds per plant; n 015). Because the clock determines flowering time, which becomes altered in period mutants (13), seedset is likely to be an ambiguous marker for the fitness implications of circadian resonance. However, we have demonstrated that circadian resonance increases the established fitness traits of photosynthesis and biomass (Figs.1to 3) (14–17). We performed reciprocal competitions be-tween long-and short-period plants (7), using two short-period mutants of TOC1E toc1-1and toc1-2, t , 20hours (27) ^and two long-period mutants of ZTL E ztl-1and ztl-27, t , 28hours (9, 10) ^. Mixed populations of toc1and ztl were grown under T20and T28, which generated a crowded lawn of plants. In these conditions, interactions among neighboring plants of dif-fering genotypes affect physiological outcomes, in addition to the interaction between each plant and the light-dark cycle that we tested previously. We grew monocultures of each line, to assess the importance of differing growth rates among neighbors compared with growth rates among lines. Testing for com-petitive advantage derived from circadian resonance, using reciprocal competition, re-quires a mixed population of two period-length genotypes; therefore, it cannot be assessed in just wild type. The period-length mutants were appropriate for this experiment because the growth disadvantage that occurred when T m t in wild type also occurred with the period mutants (Figs.1to 3).
In two separate competition experiments, using different ZTL and TOC1mutants, under T20, TOC1mutants grew more successfully than ZTL mutants, as indicated by multiple parameters including chlorophyll content, leaf number, rosette diameter, and aerial biomass (Fig.4). Conversely, under T28, growth of ZTL mutants was enhanced compared with TOC1. This was similar to results obtained when plants were grown without competition (Figs.1to 3). However, competition caused mortality of some plants, which did not occur in the absence of competition. Mortality was greater in ztl-27under T20, and greater in toc1-2under T28(Fig.4B). Circadian resonance, therefore, en-
63222JULY 2005VOL 309SCIENCE www.sciencemag.org
R E P O R T S
hanced growth and survival, and this was more pronounced under competition than during monoculture (Fig.4A). In two separate mono-culture experiments, there was no consistent pattern of T cycle–dependentmortality (Fig.4C). Therefore, both poor individual growth and outcompetition confer a disadvantage under dissonant T cycles. Our data underestimate the true growth advantage that occurs under competition, because physiological parameters were not measured in dead plants. Neither genotype had an advantage in all conditions, which implicated circadian effects of the mutations rather than secondary phenotypes. Competition between toc1-1and ztl-1, and between toc1-1or toc1-2and ztl-27, gave the same result, which discounts the likelihood of background mutations or allele-specific effects (28). Arabidopsis entrains stably to T cycles far from t (24, 29, 30), so the long-term growth advantage was likely due to correct phasing of rhythmic processes relative to the environment in one genotype, and an incorrect phase in its competitor (7). This suggests that a correctly matched circadian clock confers a competitive advantage, whereas the enhance-ment of two key fitness traits E biomass and photosynthesis; (14–17) ^by circadian reso-nance indicates that enhanced photosynthesis
is one mechanism by which the clock in-creases fitness.
Our experiments demonstrate that the circa-dian clock allows plants to increase photo-synthesis and that the clock underlies a doubling of Arabidopsis productivity. This may derive from correct anticipation of dawn and dusk, and synchronization of the synthesis of light-harvesting complex proteins and chlorophyll, both of which are unstable in their unbound state (18). Incorrect matching of en-dogenous rhythms to environmental rhythms reduced leaf chlorophyll content, reduced as-similation, reduced growth, and increased mor-tality. Optimization of these parameters by circadian resonance could represent one of the mechanisms that has selected for circadian clock function during plant evolution. We suggest that selective plant breeding for enhanced crop performance must be performed carefully, be-cause phase and period changes could arise from the close genetic linkage of phase and period loci (31) to the trait under selection, and cause alterations to clock function that might reduce vegetative yield. Clock manipulation could enhance food production during exploration of space and other planets, where the light-dark cycle may differ from the terrestrial 24-hour period. Circadian resonance is likely to provide
an advantage in all kingdoms, because reso-nance of the internal clock with the external light-dark cycle ensures an optimal phase relation between physiology and the day-night cycle and provides the basis for anticipation of changes in environmental conditions.
References and Notes
1. C. H. Johnson, M. Knight, A. Trewavas, T. Kondo, in Biological Rhythms and Photoperiodism in Plants , P. L. Lumsden, A. J. Millar, Eds. (BiosScientific, Oxford, 1998), pp. 1–34.
2. M. W. Young, S. A. Kay, Nat. Rev. Genet. 2, 702(2001).3. M. J. Yanovsky, S. A. Kay, Curr. Opin. Plant Biol. 4, 429(2001).
4. T. P. Michael et al. , Science 302, 1049(2003).
5. R. M. Green, S. Tingay, Z. Y. Wang, E. M. Tobin, Plant Physiol. 129, 576(2002).
6. H. R. Highkin, J. B. Hanson, Plant Physiol. 29, 301(1954).
7. Y. Ouyang, C. R. Andersson, T. Kondo, S. S. Golden, C. H. Johnson, Proc. Natl. Acad. Sci. U.S.A. 95, 8660(1998).
8. M. A. Woelfle, Y. Ouyang, K. Phanvijhitsiri, C. H. Johnson, Curr. Biol. 14, 1481(2004).
9. Materials and methods are available on Science Online. 10. D. E. Somers, T. F. Schultz, M. Milnamow, S. A. Kay,
Cell 101, 319(2000).
´, C. A. Strayer, N.-H. Chua, S. A. 11. A. J. Millar, I. A. Carre
Kay, Science 267, 1161(1995).
12. Z. Y. Wang et al. , Plant Cell 9, 491(1997).
13. R. Hayama, G. Coupland, Curr. Opin. Plant Biol. 6, 13
(2003).
14. A. M. Arntz, E. H. DeLucia, N. Jordan, Ecology 81,
2567(2000).
15. J. K. Ward, J. K. Kelly, Ecol. Lett. 7, 427(2004).16. P. S. Curtis, X. Wang, Oecologia 113, 299(1998).17. T. Mitchell-Olds, Evol. Int. J. Org. Evol. 50, 140
(1996).
18. S. L. Harmer et al. , Science 290, 2110(suppl.data)
(2000).
19. A. Prombona, J. Argyroudi-Akoyunoglou, Plant Sci.
167, 117(2004).
20. K. Kaasik, C.-C. Lee, Nature 430, 467(2004).
21. A. P. Withrow, R. B. Withrow, Plant Physiol. 24, 657
(1949).
22. J. D. Plautz et al. , J. Biol. Rhythms 12, 204(1997).23. A. N. Dodd, K. Parkinson, A. A. R. Webb, New Phytol.
162, 63(2004).
24. D. E. Somers, A. A. R. Webb, M. Pearson, S. A. Kay,
Development 125, 485(1998).
25. H. Griffiths et al. , in Physiological Plant Ecology , M. C.
Press, J. D. Scholes, M. G. Barker, Eds. (Blackwell,Oxford, 1998), pp. 415–441.
26. T. L. Hennessey, C. B. Field, Plant Physiol. 96, 831
(1991).
´et al. , Science 293, 880(2001).27. D. Alabadı
28. N. Salathia et al ., unpublished observations.
29. M. J. Yanovsky, S. A. Kay, Nature 419, 308(2002).30. L. C. Roden, H. R. Song, S. Jackson, K. Morris, I. A.
´, Proc. Natl. Acad. Sci. U.S.A. 99, 13313(2002).Carre
31. D. E. Somers, W.-Y. Kim, R. Geng, Plant Cell 16, 769
(2004).
32. Authors E.K., R.T., A.H., and A.J.M. identified ztl-27; N.S.
and A.J.M. conceived and performed competition and seedset experiments with A.H.; A.N.D., J.M.H., and A.A.R.W. performed all other experiments. We thank the U.K. Biotechnology and Biological Sciences Re-search Council (A.A.R.W.,J.M.H., and A.J.M.); the Royal Society of London and the Isaac Newton Trust (A.A.R.W.);and the Howard Hughes Medical Institute (F.N.).We are grateful to E. Tobin (UCLA)and S. Kay (Scripps)for collaboration and donation of CCA1-ox and toc1-2, respectively. Supporting Online Material
www.sciencemag.org/cgi/content/full/309/5734/630/DC1
Materials and Methods References and Notes
1April 2005; accepted 7June 2005
10.1126/science.1115581
Fig. 4. Correct circadian period enhances growth and survival. (A ) In two separate competition experiments using different mutant lines and one monoculture experiment, comparative growth and survival of toc1and ztl under T20and T28; n 049, except n 017to 20for dry biomass values; data are means T SEM. (B ) Representative individuals from a competition experiment. Mortality in short-and long-period lines after competition (B)or monoculture (C ) in T20and T28.
www.sciencemag.org SCIENCE VOL 30922JULY 2005633