Phytochromes differentially regulate seed germination responses to light quality and temperature cues during seed maturationpce_

Transcription

1 Plant, Cell and Environment (29) 32, doi:.1111/j x Phytochromes differentially regulate seed germination responses to light quality and temperature cues during seed maturationpce_ JENNIFER M. DECHAINE 1 *, GARY GARDNER 2 & CYNTHIA WEINIG 3 Departments of 1 Plant Biology, 2 Horticultural Science, University of Minnesota, St. Paul, MN 558, USA and 3 Department of Botany, University of Wyoming, Laramie, WY 8271, USA ABSTRACT The ratio of red to far-red light (R : FR) experienced by seeds during maturation affects germination, but the genetic regulation of this effect is poorly understood. In Arabidopsis thaliana, responses to R : FR are governed by five phytochrome photoreceptors, PHYA PHYE. PHYA, PHYB and PHYE mediate germination, but their roles in germination response to the seed maturation environment are largely unknown. Seeds of A. thaliana phytochrome mutants and natural accessions were matured in a factorial combination of cold (16 C) and warm (24 C) temperatures and high (R : FR = 1) and low (R : FR =.6) R : FR environments, resembling sunlight and foliar shade, respectively. Germination was observed in resulting seeds. All five phytochromes mediated germination responses to seed maturation temperature and/or R : FR environment. PHYA suppressed germination in seeds matured under cold temperature, and PHYB promoted germination under the same conditions. PHYD and PHYE promoted germination of seeds matured under warm temperature, but this effect diminished when seeds matured under reduced R : FR. The A. thaliana natural accessions exhibited interesting variation in germination responses to the experimental conditions. Our results suggest that the role of individual PHY loci in regulating plant responses to R : FR varies depending on temperature and provide novel insights into the genetic basis of maternal effects. Key-words: Arabidopsis thaliana; foliar shade; maternal effects; maternal environment; natural variation; red to farred ratio (R : FR); temperature. INTRODUCTION Plants rely on environmental cues to direct life-history processes. In species that disperse seeds close to the maternal plant, the seed maturation environment may be a reliable indicator of the environmental conditions in subsequent Correspondence: J. M. Dechaine. Fax: ; jdechaine@plantbio.uga.edu *Present address: Department of Plant Biology, University of Georgia, Athens, GA 362, USA. generations (Donohue & Schmitt 1998). The timing of seed germination is highly sensitive to several aspects of the seed maturation (also termed maternal) environment, including water availability, soil nutrients, photoperiod, temperature and light quality (reviewed in Gutterman 2; Finch- Savage & Leubner-Metzger 26; Contreras et al. 28; Donohue et al. 28). Light quality, that is, the ratio of red (R) to far-red (FR) light (R : FR), experienced by the maternal plant during seed development is predictive of canopy conditions; a low R : FR indicates below-canopy or competitive conditions, because chlorophyll selectively absorbs light in the red region of the spectrum while transmitting FR light (Smith 1982). A high R : FR indicates open sunlit gaps, which are favourable for germination and plant growth. Canopy conditions during seed maturation influence germination timing in a number of species (Schmitt, Niles & Wulff 1992; Donohue & Schmitt 1998; Gutterman 2; Orozco-Segovia et al. 2), and these effects have recently been shown to be adaptive in the wild (Galloway & Etterson 27). In Arabidopsis thaliana, responses to R : FR are governed by a family of five phytochrome photoreceptors, PHYA PHYE (Sharrock & Quail 1989). Their gene products fall into two classes: Type 1 includes phya, which increases to high levels in dark-grown seeds and seedlings and then rapidly declines in light. All remaining phytochrome proteins are Type 2 and are relatively stable in R light (Sharrock & Clack 22). The phytochrome protein is synthesized in the inactive, R light absorbing form (Pr), which then photoconverts to the active, FR light absorbing form (Pfr) in response to R light. Pfr reverts to Pr in FR light or over a dark period. Unique functions for each phytochrome species have been identified at several stages of plant development, such as seed germination, growth and flowering (reviewed in Quail 22). Photoreversible phytochrome effects on seed germination were identified several decades ago (Borthwick et al. 1952). More recently, roles for specific phytochromes in regulating germination responses have been demonstrated. In A. thaliana, PHYA induces germination under very low fluence (VLFR; <1 mmol m -2 ) light ranging in wavelength from 3 to 78 nm (Shinomura et al. 1996) and under heavy canopy shade (R : FR <.5) in the field (Botto et al. 29 Blackwell Publishing Ltd 1297

2 1298 J. M. Dechaine et al. 1996). Under continuous FR light, this VLFR is also dependent on the presence of PHYE (Hennig et al. 22). PHYA and PHYB interact to mediate germination under low fluence R light (LFR, > mmol m -2 )ina. thaliana (Botto, Sanchez & Casal 1995; Botto et al. 1996; Shinomura et al. 1996). In tomato, PHYB appears to be solely responsible for LFR, whereas PHYA inhibits germination under continuous R or FR light (Appenroth et al. 26). Phytochromes are not only sensitive to the germination environment, but they may also mediate the effects of light quality during seed maturation on subsequent germination. A. thaliana seeds matured under a high R : FR germinate in darkness, but seeds that matured under environments rich in FR light require light to germinate, and this effect is R : FR reversible (McCullough & Shropshire 197; Hayes & Klein 1974). A possible mechanism for this effect is that the photostationary state of the phytochrome protein (Pr or Pfr) in the dry seed is set by the R : FR environment during seed maturation (Cresswell & Grime 1981). Seeds show little response to photoreversible light cues when dry but become increasingly responsive over imbibition, which suggests that the phytochrome state in the dry seed is relatively stable and may determine the amount of light that is required for germination (Kendrick & Frankland 1969; Kendrick & Spruit 1977; Cresswell & Grime 1981). Accordingly, interspecific differences in a light requirement for seed germination are partially attributable to the timing of chlorophyll breakdown in the maternal tissues relative to seed maturation (Cresswell & Grime 1981). Species that require light for germination retain chlorophyll in the maternal tissues surrounding the seeds, and therefore, their seeds experience a low R : FR throughout maturation. Effects of light quality during seed maturation on germination are likely phytochrome-mediated, but the roles of individual phytochrome loci in these effects are unknown. Phytochrome effects are also sensitive to temperature. For example, a temperature reduction of 6 C revealed novel roles for PHYA, PHYD and PHYE in flowering (Halliday & Whitelam 23). Phytochrome function also varies depending on the temperature seeds experience during germination and maturation. Heschel et al. (27) showed that although PHYB influences germination across a range of germination temperatures, the effects of PHYA and PHYE become more significant at warmer and cooler temperatures, respectively. In addition, phytochromes differentially regulate seed germination depending on the temperature during seed maturation (Donohue et al. 28). For example, PHYB and PHYD are important for breaking coldinduced dormancy, whereas PHYA contributes to maintaining dormancy. In addition to using mutants to characterize the roles of the individual phytochrome loci, recent work has shown that natural allelic variation in phytochrome genes affects hypocotyl elongation and flowering time (Aukerman et al. 1997; Maloof et al. 21; Balasubramanian et al. 26; Filiault et al. 28). These studies suggest that natural selection may target phytochrome genes, and that allelic variation in phytochrome loci likely contributes to phenotypic differences among natural populations of A. thaliana. Natural allelic variation in phytochrome genes may also affect seed germination, but this has not yet been examined. In this study, we compare the germination responses of multiple phytochrome mutant plants to their wild-type background in order to investigate which, if any, phytochromes regulate the effects of light quality and temperature cues during seed maturation on subsequent germination. In addition, we examine germination responses to the same environmental cues in several A. thaliana natural accessions in order to test for genetic variation and more generally investigate the effects of the seed maturation treatments on germination in A. thaliana. MATERIALS AND METHODS Plant material In order to investigate the role of each phytochrome in mediating the effects of seed maturation environment on germination, we compared seed germination in A. thaliana wild-type genotypes Landsberg erecta [, Arabidopsis Biological Resource Center (ABRC) stock number: CS2], Columbia (Col-, CS6673) and Wassilewskija (Ws, CS915) to mutants that exhibit loss of function in one or more phytochrome genes. Single mutants of phya, phyb and phye were generated in the and/or Col- backgrounds through forward and backward genetic screens (Koornneef, Rolff & Spruit 198; Nagatani, Reed & Chory 1993; Reed et al. 1993, 1994; Devlin, Patel & Whitelam. 1998). We included multiple phya and phyb deficient mutant alleles in the background; among these, phya-23 (CS6221), phya-25 (CS6222), phyb-4 (CS6212) and phyb-7 (CS6215) have been shown to exhibit weak phytochromemediated phenotypic responses compared with phya-21 (CS6219), phya-22 (CS622), (CS6211), phyb-5 (CS6213) and phyb-9 (CS6217) (Reed et al. 1993, 1994). The phyd-1 mutation is a natural null allele that was identified in Ws and crossed into the background (Aukerman et al. 1997). Mutants deficient in phyc were identified from T-DNA insertion lines in the Col- (phyc-2) and Ws (phyc-3) backgrounds (Monte et al. 23). Ws is naturally deficient in phyd, and therefore, phyc-3 lacks phyc and phyd. Double and triple mutants were obtained by crossing (Devlin et al. 1998, 1999). All single phya and phyb mutants were obtained from ABRC. PhyC-2, phyc-3, phyd-1 phya-21/phyb-5, phya-21/phyd-1, / phyd-1 and phya-21//phyd-1 were obtained from Robert A. Sharrock. PhyE-1, phya-21/phye-1, / phye-1, /phyd-1/phye-1, phya-21//phye-1 and phya-21//phyd-1/phye-1 were obtained from Gary C. Whitelam. We also included six additional A. thaliana natural accessions (for a total of nine accessions, six plus Col-, and Ws) in order to preliminarily examine natural variation in A. thaliana germination responses to the seed maturation environments tested in this study. All accessions were obtained from the ABRC and included: Col-4 (CS933);

3 Phytochromes regulate effects of the seed maturation environment 1299 Ga-2 (CS6715); HOG (CS6178); JI-3 (CS6745); Lm-2 (CS6784); and Me- (CS1364). These accessions were chosen in two steps. Firstly, 25 accessions were chosen from a set of ~18 that were part of a larger association mapping experiment (C. Weinig, unpublished data) and exhibited significant variation in adult shade-avoidance phenotypes. From these, accessions were chosen because they clustered either with phya or phyb in hypocotyl responses to red or far-red light (Maloof et al. 21). The additional 15 were randomly chosen to represent the latitudinal range of A. thaliana. In a pilot experiment, these accessions were matured under simulated foliar-shade and neutral-shade treatments (see following methods) in a greenhouse and germinated under the same treatments. For further investigation, we chose the natural accessions that expressed the greatest variation in germination percentage when seeds had been matured under simulated foliar shade. One accession, Bla-11, did not set seed in all four seed maturation treatments and was excluded from analyses. All mutants and natural accessions are highly inbred and were allowed to self-pollinate. Maternal, paternal and direct genetic effects cannot be differentiated in this experiment. Seed maturation and germination conditions For seed maturation, four sets of plants were established by randomly planting two replicate individuals of each genotype into every other pot cavity (5 cm diameter 5 cm deep) of two Araflats (Betatech, Belgium) in Sunshine Mix, LP5 soil (Sun Gro Horticulture,Vancouver, Canada). Plants were stagger-planted based on the estimated days to seed maturation for each combination of genotype and seed maturation treatment. For example, in a prior experiment, plants had taken on average of 1 week longer to mature seeds in cold/neutral shade than in warm/neutral shade treatments.therefore, seeds were planted 1 week earlier in cold/neutral shade than in the warm/neutral shade. Stagger-planting ensured that seed collection could be done simultaneously and that after-ripening would not vary among genotypes or across seed maturation treatments. Seeds were cold-stratified at 4 C for 4 d and then placed into two Conviron E7/2 growth chambers (Controlled Environments Ltd,Winnipeg, Canada) under a 12 h photoperiod and constant 2 C. Growth chambers were fitted with eight white fluorescent lights (F38/871, 32W, Sylvania Octron, Danvers, MA, USA), two infrared lights (Geneva Scientific, Williams Bay, WI, USA) and four incandescent lights (Sylvania Double Life, Soft White, 12V/75W). In the treatments, total PAR = 2 mmol m -2 s -1 and R : FR = 1.1, as measured using a LI-25A light sensor (Li-Cor, Lincoln, NE, USA) and SKR 1, 66/73 nm sensor (Skye Instruments Ltd, Powys, UK), respectively. At bolting, growth chambers were reset to cold (17 C day/14 C night) and warm (26 C day /22 C night) temperatures, and white (#214, Full Touch Spun) and green (#73, Liberty Green) filters (Lee filters, Burbank, Canada) were placed over two flats in each growth chamber to impose simulated foliar shade (R : FR =.6, PAR = 17 mmol m -2 s -1 ) and neutral shade (R : FR = 1.1, PAR = 175 mmol m -2 s -1 ) light environments. This experimental set-up generated four seed maturation treatments: cold/simulated foliar shade (C/FS); cold/ neutral shade (C/NS); warm/simulated foliar shade (W/FS); and warm/neutral shade (W/NS). The temperature treatments in this study simulate the average cold and warm temperatures that wild A. thaliana plants would likely experience in early or late spring, respectively. The simulated foliar shade seed maturation environment simulates R : FR conditions typical of a competitive environment or under a plant canopy.the treatments were applied at bolting for two reasons. Firstly, A. thaliana plants in the wild germinate early in the spring when there is often little plant competition, and begin to experience shading at bolting when neighboring A. thaliana individuals are also bolting and flowering and have large rosettes. Secondly, several studies provide strong evidence that effects of maternal light quality on seed germination are perceived during seed maturation, possibly by the seed itself (McCullough & Shropshire 197; Hayes & Klein 1974; Cresswell & Grime 1981; Contreras et al. 28), and that changes in light quality perceived by the maternal plant before seed maturation have little or no direct effect on seed germination. Seeds of mutant plants and natural accessions were matured and collected under the four seed maturation treatments. To control for variation in seed maturation, six apical siliques at roughly the same maturity level were collected from the primary inflorescence of each replicate plant and pooled by genotype and seed maturation treatment. Seeds were stored in the dark at room temperature for 1 month and then six replicates of 2 seeds for each combination of genotype by seed maturation treatment were planted, under a green safe light, onto.5% agar in 4.5-cm-diameter plastic Petri dishes. Seeds imbibed water for 24 h in darkness and were placed into the same growth chambers in a 2 C, 12 h photoperiod, and the same neutral shade treatment as the seed maturation generation. Seeds were checked for germination every day for d. Germination was scored when a seed s cotyledons were green and had begun to unfold. We scored germination in this manner for ecological relevance. In several instances, the radical would emerge 1 2 mm from a seed but no more by the end of the d (and the seedling appeared to be no longer viable);these seeds would not have established in a natural environment. One caveat of this method is that we cannot explicitly separate the effects of radicle emergence and greening, although we do not expect our results to differ greatly for these two germination measures. In fact, mean germination percentages differed by at most 1% if seeds that germinated but failed to open cotyledons were included in the analyses versus if they were not (results not shown). Seed mass was also determined by weighing two groups of seeds for each genotype by seed maturation treatment combination. To evaluate the repeatability of our results, the entire experiment (seed maturation and germination) was completed twice, each time in two different Conviron E7/2 growth chambers. The six additional A. thaliana natural accessions, as well as phyc-3, phya-211 and

4 13 J. M. Dechaine et al. phya-211/phyb-9 were only included in the second experimental trial. Therefore, the results of the second trial are reported with discussion of the consistencies and differences between trials. Data analysis Germination percentage was calculated for each replicate by dividing the number of seeds germinated by the number of seeds planted. This value is presented for all genotype by seed maturation treatment combinations that exhibited >9% germination. If the germination percentage of a genotype by seed maturation treatment combination was <9%, we estimated maximum possible germination by germinating two replicates of 2 seeds on.5% agar supplemented with mm GA 4+7. All other experimental conditions were as described earlier. GA 4+7 promotes germination in A. thaliana; seeds that do not germinate with GA 4+7 are unlikely to be viable. GA 4+7 has the additional advantage of testing for possible suppression of germination by our seed maturation treatments, because GA regulates germination downstream of the phytochromes (Yamaguchi et al. 1998). Germination with GA 4+7 was >9% for all genotype by seed maturation treatment combinations tested except phya/phye (C/FS), phya/phyb/phye (C/FS, C/NS), phya/phyb/phyd (C/NS) and phya-211/ phyb-9 (Col-) (W/FS). Their germination percentages ranged from 65 to 85%. These results suggest that most seeds were viable and capable of germination and that the seed maturation treatments did not suppress germination in a manner independent of the mutant alleles. Germination percentage data were adjusted using the results of the GA 4+7 experiment as the maximum possible germination (instead of %). The GA 4+7 adjustment had very little effect on the results, because generally, seeds that germinated to <9% with GA 4+7 exhibited very low germination (<%) in all experiments. Germination day was calculated by assigning a germination day to each seed that germinated and calculating mean days to germination for each replicate. Germination percentage displayed constant variance. Days to germination violated the analysis of variance (anova) assumption of homoscedasticity. As suggested by a Box Cox procedure (PROC TRANSREG, SAS 21), germination day was transformed by taking the inverse. After transformation, germination day displayed constant variance, and the transformed values were used in subsequent analyses. To test for differences in the effects of seed maturation treatment on germination between wild-type genotypes and phytochrome mutants, we used an anova model (PROC GLM) in which we split the analysis by seed maturation treatment (light quality and temperature) and then tested for the fixed effects of growth chamber (during seed germination) and genotype on germination percentage and germination day. All mutants in a wild-type background were included in a model and a priori contrasts between each mutant genotype and its respective wild-type genotype and between specific mutant pairs were applied. A significant effect of genotype indicates that the two genotypes differed in their response to the specified seed maturation treatment combination (Donohue et al. 28). Bonferroni corrections for multiple tests, based on the total number of tests in one mutant background and seed maturation treatment, were applied to the results. We used a second anova model (PROC GLM) to test for effects of seed maturation treatment on seed mass, germination percentage and germination day among the nine A. thaliana natural accessions. This model included genotype and temperature and light quality during seed maturation, and all possible interactions as fixed effects. The three-way interaction between genotype, seed maturation temperature and light quality was not significant for germination percentage or day, and was removed from the model for those traits. RESULTS Phytochrome A PhyA mutant seeds derived from all seed maturation treatments except warm/neutral shade germinated to higher percentages than wild-type, although the significance of these effects differed by mutant allele (Fig. 1; Table 1). PhyA-21 and phya-23 exhibited significantly higher germination than when seeds matured under the cold seed maturation treatment. PhyA-22, phya-25 (only in neutral shade) and phya-211 exhibited the same trend, but they were not significant after Bonferroni correction (Figs 1 & 2). The one exception was that phya-25 seeds matured under cold and simulated foliar shade germinated to a lower percentage than. PhyA-25 is a documented weak mutant (Reed et al. 1994), and this result was not repeatable, that is, in the first experimental trial all phya mutants (including phya- 25) germinated to a higher percentage than when seeds were matured in cold temperature. Germination in phya deficient seeds ranged from 12 to 25% higher than wild-type genotypes in seeds matured in cold and neutral shade and 22 55% higher in seeds matured in cold and simulated foliar shade (excluding phya-25). Germination day was accelerated by 2 5 d over wild-type genotypes in phya mutant seeds matured under warm temperature and simulated foliar shade (Fig. 3; Table 2). Two phya mutants, phya-21 and phya-23, also exhibited accelerated germination in seeds matured under cold temperature and either light quality treatment (Table 2). Similar patterns of germination were observed in the first experimental trial, with the exception that phya-211 was not tested. Phytochrome B In general, mutants deficient in phyb exhibited reduced germination when compared with, and this was most significant when seeds matured under neutral shade (Fig. 1; Table 1). These trends were weaker in documented weak mutants, phyb-4 and phyb-7 (Reed et al. 1993) than and phyb-5. Seed germination was also reduced in phyb-9 (Col- background) seeds matured in cold temperature and

5 Phytochromes regulate effects of the seed maturation environment 131 Simulated foliar shade Neutral shade Germination percent phya-21 phya-22 phya-23 phya-25 phyb-4 phyb-5 phyb-7 phya-21/phyb-5 phya-21 phya-22 phya-23 phya-25 phyb-4 phyb-5 phyb-7 phya-21/phyb-5 Figure 1. Least-squared means 1 standard error of germination percentages for PHYA and PHYB deficient Arabidopsis thaliana seeds in the Landsberg erecta () wild-type background. PhyA-23, phya-25, phyb-4 and phyb-7 are documented weak mutants (Reed et al. 1993, 1994). Experimental conditions are as described in Table 1. LSmeans 1 SE were generated from the analysis of variance model described in Table 2. neutral shade, as well as warm temperature and simulated foliar shade (Fig. 2). This difference between phyb-9 and phyb mutants in the background under warm treatments is largely attributable to differing germination patterns between Col- and ; the Col- genotype exhibited % germination in seeds matured under warm/simulated foliar shade (Fig. 2). In addition, germination was accelerated by d in, phyb-4 and phyb-5 seeds matured under simulated foliar shade and either temperature treatment (Fig. 3; Table 2). Differences in germination between phyb alleles and wild-type genotypes followed the same trends in the first experimental trial (data not shown). Germination was reduced in phya/phyb ( and Col- backgrounds) mutant seeds matured under all four seed maturation treatments (Figs 1 & 2; Table 1). In addition, phya-21/phyb-5 seeds matured under warm temperature (and either light quality treatment) germinated 2 d earlier that (Fig. 3; Table 2). These results were repeated in both experimental trials. Phytochrome C In the Col- background, PHYC had no effect on germination in seeds matured in the warm temperature, that is, differences between phyc-2 and Col- were marginal in the second experimental trial and were not detected in the first experimental trial (Fig. 2; Table 1). In contrast, phyc-3 seeds germinated to a higher percentage than Ws when matured under warm temperature or simulated foliar shade (Fig. 2; Table 1). PHYC had some effect on germination in Col- seeds matured in cold temperature, but the direction of the effect differed between experimental trials, making it difficult to draw any conclusions. The phya/phyb/phyd/phye mutant showed some germination (2 5%) in seeds matured under all treatments except in cold/neutral shade (Fig. 4), suggesting that phyc plays some role in germination, or alternatively, A. thaliana seeds can germinate without phytochrome. Phytochrome D PHYD had weak effects on germination percentage in seeds matured under cold temperature and phyd seeds matured under cold temperature or simulated foliar shade germinated d earlier than. More specifically, PhyD seeds matured in cold temperature did not differ significantly in germination percentage from, and germination differences were not significant between phya/ phyd and phya, phyb/phyd and phyb, and phya/phyb/ phyd and phya/phyb (Fig. 5, Table 1). In contrast, phyd germination was significantly lower than in seeds matured under warm temperature, and this effect was most pronounced in seeds also matured under neutral shade. Seeds matured under warm temperature and either light quality treatment germinated to a lower percentage for phyb/phyd than phyb or phyd, and for phya/phyb/phyd than phya/phyb (Fig. 5; Table 1). PhyD results were consistent across both experimental trials (data not shown). Phytochrome E Seeds deficient in phye germinated significantly less than when matured under neutral shade and either temperature (Fig. 4; Table 1) and significantly later than when matured under neutral shade and warm temperature (Fig. 3; Table 2). PHYE also appeared to counteract the effects of PHYA across all seed maturation treatments, except warm/simulated foliar shade, that is, phya/phye germination was >75% less than phya germination. Mutant seeds matured in cold temperature that were deficient in phyb and phye (phyb/phye, phyb/phyd/phye, phya/ phyb/phye, and phya/phyb/phyd/phye) exhibited very

6 132 J. M. Dechaine et al. Table 1. Tests of significance between A. thaliana phytochrome mutants and wild-type genotypes for germination percentage Seed maturation treatment temperature temperature Simulated Neutral Simulated Neutral foliar shade shade foliar shade shade F ratio F ratio F ratio F ratio Mutants versus wild-type genotypes phya *** 12.4*** 8.47**.12 phya phya *** 7.8** 4.28*.3 phya * 13.48** * 78.89*** 27.2*** 46.13*** phyb *** 8.47**.2 phyb * 85.88*** *** phyb *** * phyd ** 67.71*** phye *** *** phya-211 (Col-) * phyb-9 (Col-) ***.83 phyc-2 (Col-) * ** phyc-3 (Ws) 27.55***.5* 73.34*** 25.51*** phya-21/phyb ** 82.32*** 52.57*** 81.65*** phya-211/phyb-9 (Col) 4.38* *** Multiple mutant comparisons phya-21/phyd-1 versus phya *.72 /phyd-1 versus ** phya-21//phyd-1 versus phya-21/phyb * 1.97 phya-21/phye-1 versus phya *** 124.1*** 28.35*** 58.1*** /phye-1 versus *** 16.54*** phya-21//phye-1 versus phya-21/phyb * 4.18* /phyd-1/phye-1 versus phyb-5/phyd Seeds were matured in growth chambers under a factorial combination of cold (17 C day /14 C night ) and warm (26 C day/22 C night) temperature treatments and simulated foliar shade (R : FR =.6, PAR = 17 mmol m -2 s -1 ) and neutral shade (R : FR = 1.1, PAR = 175 mmol m -2 s -1 ) light quality treatments. Seeds were collected under those treatments, stored for 1 month in darkness at room temperature, and germinated at 2 C in neutral shade (same conditions).a constant photoperiod of 12 h was maintained in both generations. Germination was scored when a seed s cotyledons were green and partially unfolded. Germination percentage was calculated after d of germination by dividing the total number of germinated seeds by the maximum possible germination (either seed total or germination percentage on GA 4+7). Germination percentage data were split by temperature and light quality during seed maturation and then tested for significant effects of growth chamber and mutant genotype using analysis of variance. A significant genotype effect indicates that the genotypes differ in their response to the indicated seed maturation treatment. The upper part of the table shows tests between mutants and their wild-type background, which is unless otherwise indicated in parentheses. The lower part of the table shows tests between specific mutant pairs. Comparisons significant after Bonferroni correction are highlighted in bold. *P <.5, **P <.1, ***P <.1. low germination when compared with (Fig. 4). In addition, phyb/phye seeds matured under warm temperature and neutral shade germinated >3% less than phyb or phye under the same conditions. Patterns of germination in phye seeds were consistent across both experimental trials (data not shown). Natural accessions maturation temperatures reduced germination percentages by almost half when compared with seeds matured under warm temperatures (Table 3). Light quality during seed maturation had no overall effect on germination percentage, but natural accessions differed in their responsiveness to seed maturation temperature and light quality for both germination percentage and day (Figs 6 & 7; Table 3a). Note that, for instance, germinates to a higher percentage under neutral than simulated foliar shade, and Col- shows the opposite pattern. Temperature and light environment, as well as their interaction, had a strong effect on seed mass, and accessions differed in the response of seed mass to these treatments (Table 3a). In general, seeds produced in cold temperature and/or simulated foliar shade were heavier than seed produced in warm temperature and/or neutral shade (Table 3b). DISCUSSION In this study, we examined how temperature and light quality during seed maturation influenced germination in

7 Phytochromes regulate effects of the seed maturation environment 133 Simulated foliar shade Neutral shade 9 Germination percent Col- phya-211 phyb-9 phyc-2 phya-211/phyb-9 Ws phyc-3 Col- phya-211 phyb-9 phyc-2 phya-211/phyb-9 Ws phyc-3 Figure 2. Least-squared means 1 standard error of germination percentages for PHYA, PHYB and PHYC deficient A. thaliana seeds in the Columbia- (Col-) background (phyc-2) and PHYC deficient seeds in the Wassilewskija (Ws) background (phyc-3). Ws and phyc-3 are naturally deficient in PHYD. PhyA-211, phya-211/phyb-9 and phyc-3 were included in only one (of two) experimental trials. A. thaliana phytochrome mutants and natural accessions. Our results suggest that all five A. thaliana phytochromes partially mediate germination in response to seed maturation temperature and/or light quality. In general, functional PHYA inhibited germination; whereas PHYB, PHYD and PHYE generally promoted germination, but the phenotypes of individual phytochrome mutants differed by seed maturation treatment. In addition, we demonstrated significant natural variation in germination responses to the tested seed maturation environments. Functional PHYA inhibited and slowed germination in seeds matured under cold temperature or simulated foliar shade. The only mutants not to show this response were phya-25 (under cold/simulated foliar shade) and phya- 211 (under warm/simulated foliar shade). PhyA-25 is a documented weak mutant (Reed et al. 1994); and Col exhibited % germination among seeds matured under warm temperature and simulated foliar shade, making it hard to discern effects of phya under that treatment. Also, the results of phya-211 should be interpreted with caution, because it was included in only one experimental trial. The inhibitory effect of PHYA is surprising, because previous research has almost exclusively shown that PHYA promotes germination in A. thaliana (Botto et al. 1996; Shinomura et al. 1996; Heschel et al. 27). The contrasting results are likely caused by our study s focus on the effects of the seed maturation environment on germination, whereas most previous work examined only the environment during seed germination. Consistent with this hypothesis, the only other study to investigate the effects 11 Simulated foliar shade Neutral shade 9 Germination day phya-21 phya-22 phya-23 phya-25 phyb-4 phyb-5 phyb-7 phya-21/phyb-5 phyd phye phya-21 phya-22 phya-23 phya-25 phyb-4 phyb-5 phyb-7 phya-21/phyb-5 phyd phye Figure 3. Least-squared means 1 standard error of germination day for single phytochrome mutants and the phya/phyb double mutant in the Landsberg erecta () background. Germination day was calculated as described in Table 2.

8 134 J. M. Dechaine et al. Table 2. Tests of significance between A. thaliana phytochrome mutants and Landsberg erecta () for germination day Seed maturation treatment temperature temperature Simulated foliar shade Neutral shade Simulated foliar shade Neutral shade F ratio F ratio F ratio F ratio Mutants versus phya * 8.44** 32.65***.3 phya ***.2 phya *** 8.16** 6.67*** 5.29* phya ***.5.96** ***.34 phyb * ***.66 phyb *** 6.39* 34.28***. phyb ***.2 phyd *** 13.35*** 27.24***. phye ** phya-21/phyb ***.24** Germination was checked every day for d. Each seed was assigned an appropriate germination day, and mean days to germination was calculated for a replicate. Only seeds that germinated were included in the mean calculations. Germination day data were inverse transformed for homoscedasticity and analysed as described in Table 1. F ratios for effects of the mutant genotype are shown. Mutants were contrasted to their wild-type background,. Comparisons significant after Bonferroni correction are highlighted in bold. *P <.5, **P <.1, ***P <.1. of seed maturation environment on germination in A. thaliana phytochrome mutants likewise demonstrated that PHYA inhibited germination in seeds matured in cold temperature, although only in a phyd background (Donohue et al. 28). We found that functional PHYC inhibited germination in a phyd deficient background (Ws) in seeds matured in warm temperature or simulated foliar shade. This suggests a new role for PHYC in regulating seed germination that is more similar to PHYA than to PHYB, PHYD, or PHYE. Although caveats include the fact that the phyc-3 results were not replicated and that mutants in Ws are not directly comparable with. This relationship between PHYC and PHYA is not completely unexpected, because PHYA and PHYC arose from a duplication event and share more sequence similarity with each other than with PHYB, PHYD or PHYE (Donoghue & Mathews 1998). Functional PHYB and PHYE acted antagonistically to PHYA to promote germination. All phyb mutants in the background germinated to a significantly lower percentage than wild-type in seeds matured under cold and neutral shade.although not significant, this pattern was also Germination percent Simulated foliar shade Neutral shade phya-21 phye phya-21/phye /phye /phyd/phye phya-21/phyb-5 phya-21/phyb-5/phye phya-21/phyb-5/phyd/phye phya-21 phye phya-21/phye /phye /phyd/phye phya-21/phyb-5 phya-21/phyb-5/phye phya-21/phyb-5/phyd/phye Figure 4. Least-squared means 1 standard error of germination percentages for PHYE deficient A. thaliana seeds in the Landsberg erecta () background.

9 Phytochromes regulate effects of the seed maturation environment 135 Simulated foliar shade Neutral shade 9 Germination percent phya-21 phyd phya-21/phyd /phyd phya-21/phyb-5 phya-21/phyb-5/phyd phya-21 phyd phya-21/phyd /phyd phya-21/phyb-5 phya-21/phyb-5/phyd Figure 5. Least-squared means 1 standard error of germination percentages for PHYD deficient A. thaliana seeds in the Landsberg erecta () background. observed in phyb-9 (Col). Germination was also significantly inhibited in seeds of and phyb-5 matured under cold and warm temperatures and neutral shade. PhyB-4 and phyb-7 seeds exhibited more variable germination responses, but these mutants are known to exhibit only weak phenotypic effects (Reed et al. 1993). Germination in the phye single mutant was also significantly inhibited when seeds were matured under neutral shade. Also, mutants deficient in phyb and phye counteracted the effects of phya; that is, double and triple mutants deficient in phya and phyb or phye germinated to lower percentages than wild-type genotypes and the phya single mutants. PHYB has previously been shown to have an antagonistic effect on PHYA in germination, but in that example PHYB Seed Germination Germination (a) mass (mg) percent day Source d.f. F ratio F ratio F ratio Table 3. Tests of significance and summary statistics for the A. thaliana natural accessions Replicate 1.35 Genotype (G) *** 65.44*** 25.97*** Temperature (T) *** ***.16 Light quality (LQ) 1 4.4*** *** T*LQ ***.13.2 G*T *** 7.97*** 3.38** G*LQ 8 4.7** 3.59*** 6.65*** G*T*LQ * (b) Seed maturation treatment Least-squared means 1SE /simulated foliar shade /neutral shade /simulated foliar shade /neutral shade (a) Analysis of variance tests for significant effects of temperature and light quality during seed maturation on seed mass, germination percentage, and germination day in nine A. thaliana natural accessions. F ratios and degrees of freedom (d.f.) are given for each test. LQ, seed maturation light quality treatment (simulated foliar shade or neutral shade);t, seed maturation temperature treatment (warm or cold). Treatments are as described in Table 1. The G*T*LQ term was non-significant for germination traits and was removed from the model. (b) Least-squared means 1 standard error (SE) for seed maturation temperature by light quality effects on seed mass and germination. *P <.5, **P <.1, ***P <.1.

10 136 J. M. Dechaine et al. 1 Simulated foliar shade Neutral shade 9 Germination percent Col- Col-4 Ga-2 HOG Jl-3 Lm-2 Me- Ws Col- Col-4 Ga-2 HOG Jl-3 Lm-2 Me- Ws Figure 6. Least-squared means 1 standard error for germination percentage in nine A. thaliana natural accessions. LSmeans were generated from the analysis of variance model described in Table 3. Only Landsberg erecta (), Col- and Lm-2 were included in both experimental trials. inhibited PHYA-mediated germination in FR (Hennig et al. 21). Our phyb results are consistent with Donohue et al. (28), who also found that mutants deficient in phyb generally exhibited lower germination than wild-type genotypes. However, that study found no significant effect of seed maturation treatment on germination in phye, possibly because light quality effects were not examined. Functional PHYD also increased germination percentage in seeds matured under warm temperatures but did not counteract the effects of PHYA. Donohue et al. (28) also found that PHYD promoted germination in seeds that matured under warm temperatures and were subjected to a warm stratification treatment; our seeds received no stratification, but were matured under warm temperatures. Interestingly, the differences in germination percentages between wild-type genotypes and seeds deficient in phyb, phyd and/or phye were generally most pronounced when the seeds were matured under neutral shade. It is unlikely that these results are solely attributable to differences in the state of the phytochrome proteins in the seeds, because under that hypothesis one would expect that seeds matured under a low R : FR, as in the simulated foliar shade treatment, would have less Pfr than seeds matured under the neutral shade treatment. Therefore, seeds matured under simulated foliar shade should germinate more slowly and to a lower percentage than seeds matured under neutral shade. This pattern is observed in the wild-type background, but the opposite is observed in mutants deficient in phyb, phyd and/or phye. A possible explanation for the effects of phyb, phyd and phye seeds matured under a low R : FR is that the light quality during seed maturation influences the total amount of phytochrome in the dry seed, therefore, altering the seed s sensitivity to light. A. thaliana seeds matured under a high R : FR germinate to a much higher percentage in dark than seeds matured under a very low R : FR, but they also Simulated foliar shade Neutral shade 9 Germination day Col- Col-4 Ga-2 HOG Jl-3 Lm-2 Me- Ws Col- Col-4 Ga-2 HOG Jl-3 Lm-2 Me- Ws Figure 7. Least-squared means 1 standard error for germination day in nine A. thaliana natural accessions.

11 Phytochromes regulate effects of the seed maturation environment 137 require over a 3 larger dosage of R light than seeds matured under low R : FR to reach 5% germination (McCullough & Shropshire 197). Seeds that matured under high R : FR retain viability and regain germinability to the level of seeds matured under low R : FR after 9 months of storage. These results suggest that seeds matured under a high R : FR express a reduced sensitivity to R light, which gradually increases over dark storage. Although differences in R : FR between seed maturation treatments are less severe in our study than in the aforementioned work, they should be enough to alter total phytochrome levels in seeds and may explain why seeds deficient in phyb, phyd and phye exhibited lower germination when matured under a high R : FR than when matured under a low R : FR. It remains to be tested if and how total phytochrome levels of the various phytochrome genes were affected by seed maturation treatment. Although we found overall patterns in the effects of PHYA, PHYB, and PHYE on seed germination, there was also considerable variation in phytochrome function in seeds matured under different treatments (e.g. warm versus cold). Phytochromes, particularly phyb, phyd and phye, exhibit conditional redundancy for several phenotypes (Aukerman et al. 1997; Devlin et al. 1998, 1999; Hennig et al. 21). Phytochromes bind to the same interacting partners (Ni, Tepperman & Quail 1998; Oh et al. 24), and environmental variation in phytochrome function may reflect differences in optimal binding conditions among phytochrome species (Halliday & Whitelam 23). Environmental variation in phytochrome function could also be caused by phytochrome effects on the gibberellic acid (GA) and abscisic acid (ABA) biosynthesis pathways. In response to R light, phytochromes interact with PIF3-like 5 (PIL5) to increase the amount of bioactive GAs in germinating seeds by increasing the transcription of GA-oxidase genes that convert inactive GA precursors to active GAs (Yamaguchi et al. 1998). Phytochromes also downregulate AtGA2ox2, a gene that reduces bioactive GA (Oh et al. 26). stratification increases the transcription of GA-oxidase genes by the transcription factor SPATULA (SPT) (Penfield et al. 25). Red light irradiation also reduces ABA levels in the seed by downstream members of the phyb pathway and interactions between GA and ABA (Finch-Savage & Leubner-Metzger 26; Seo et al. 26; Oh et al. 27). In addition, light conditions (in this case, photoperiod) during seed maturation have been shown to influence the amount of ABA in seeds and the seeds sensitivity to ABA in germination (Contreras et al. 28).Thus, the variability in seed germination response to temperature and light cues during seed maturation is likely caused by the complex interactions among the hormone signalling pathways and their environmental inputs (Holdsworth, Bentsink & Soppe 28). In addition to demonstrating phytochrome action in effects of seed maturation environment on germination, we examined the same germination responses in nine A. thaliana natural accessions, in order to test for ecologically relevant natural variation under our experimental conditions. One caveat to the following results is that only three of the accessions ; Col-; and Lm-2 were included in both experimental trials. We found highly significant genetic variation for germination response to the seed maturation treatments. For example, seed germination was suppressed in when seeds were matured under simulated foliar shade and suppressed in Col- when seeds were matured under neutral shade. Several other accessions exhibited no effect of seed maturation treatment on germination percentage (HOG, Jl-3) or only temperature effects (Lm-2, Me-, Ws). Likewise, germination was accelerated in seeds matured under neutral shade and warm temperature for, but accelerated in seeds matured under simulated foliar shade for HOG and Me-O. The large effect of the simulated foliar shade maturation treatment on HOG seeds accounts for the significant overall treatment effect of light quality on germination day (Table 3). In addition to significant interactions between genotype and seed maturation treatment, we found overall effects of the seed maturation treatments on seed mass and germination percentage. Seeds were heavier and germination percentage was severely reduced in seeds matured under cold temperatures. -induced dormancy has been found in a number of species and likely prevents seeds from germinating too early or late in the season (Lacey 1996; Gutterman 2; Donohue et al. 28). For the same reason, plants may produce heavier seeds under unfavourable conditions, such as cold temperatures and simulated foliar shade, if these seeds stay viable longer in the soil (although seed mass and germination percentage were not correlated in our study). Accordingly, lettuce plants produce seeds that are heavier, more dormant and stay viable longer in storage under long-day (which, similar to cold temperatures, predicts unfavourable conditions for germination) than short-day photoperiods (Contreras et al. 28). In conclusion, we demonstrated that A. thaliana phytochromes are important regulators of germination, and that the relative role of each phytochrome is influenced by temperature and light quality during seed maturation. We found an unexpected role for PHYA (and possibly PHYC) in inhibiting germination and showed that functional PHYB and PHYE are necessary for high germination, especially in seeds matured under neutral shade. In addition, using natural accessions of A. thaliana, we found ecologically relevant patterns of phytochrome-mediated germination responses to the tested experimental conditons. ACKNOWLEDGMENTS The authors thank two anonomyous reviewers for insightful comments that improved upon an earlier version of this manuscript. We are grateful to R.A. Sharrock and G.C. Whitelam for providing the phytochrome mutant seeds. We also thank D. Brinkman, K. Dorn, A. Schutte and E. Berkas for advice on experimental techniques and help in planting and maintaining experimental plants. We also thank the National Science Foundation for grant DBI-2273 (to C.W.).

12 138 J. M. Dechaine et al. REFERENCES Appenroth K.J., Lenk G., Goldau L. & Sharma R. (26) Tomato seed germination: regulation of different response modes by phytochrome B2 and phytochrome A. Plant, Cell & Environment 29, Aukerman M.J., Hirschfeld M., Wester L., Weaver M., Clack T., Amasino R.M. & Sharrock R.A. (1997) A deletion in the PHYD gene of the Arabidopsis Wassilewskija ecotype defines a role for phytochrome D in red/far-red light sensing. The Plant Cell 9, Balasubramanian S., Sureshkumar S., Agrawal M., Micheal T.P., Wessinger C., Maloof J., Clark R., Warthmann N., Chory J. & Weigel D. (26) The phytochrome C photoreceptor gene mediates natural variation in flowering and growth responses of Arabidopsis thaliana. Nature Genetics 38, Borthwick H.A., Hendricks S.B., Parker M.W., Toole E.H. & Toole V.K. (1952) A reversible photoreaction controlling seed germination. Proceedings of the National Academy of Sciences of the United States of America 38, Botto J.F., Sanchez R.A. & Casal J.J. (1995) Role of phytochrome B in the induction of seed-germination by light in Arabidopsis thaliana. Journal of Plant Physiology 146, Botto J.F., Sanchez R.A., Whitelam G.C. & Casal J.J. (1996) Phytochrome A mediates the promotion of seed germination by very low fluences of light and canopy shade light in Arabidopsis. Plant Physiology 1, Contreras S., Bennett M.A., Metzger J.D. & Tay D. (28) Maternal light environment during seed development affects lettuce seed weight, germinability, and storability. Hortscience 43, Cresswell E.G. & Grime J.P. (1981) Induction of a light requirement during seed development and its ecological consequences. Nature 291, Devlin P.F., Patel S.R. & Whitelam G.C. (1998) Phytochrome E influences internode elongation and flowering time in Arabidopsis. The Plant Cell, Devlin P.F., Robson P.R.H., Patel S.R., Goosey L., Sharrock R.A. & Whitelam G.C. (1999) Phytochrome D acts in the shadeavoidance syndrome in Arabidopsis by controlling elongation and flowering time. Plant Physiology 119, Donoghue M.J. & Mathews S. (1998) Duplicate genes and the root of angiosperms, with an example using phytochrome sequences. Molecular Phylogenetics and Evolution 9, Donohue K. & Schmitt J. (1998) Maternal environmental effects in plants: adaptive plasticity? In Maternal Effects as Adaptations (eds T.A. Mousseau & C.W. Fox) pp Oxford University Press, New York, NY, USA. Donohue K., Heschel M.S., Butler C.M., Barua D., Sharrock R.A., Whitelam G.C. & Chiang G.C.K. (28) Diversification of phytochrome contributions to germination as a function of seed-maturation environment. New Phytologist 177, Filiault D., Wessinger C.A., Dinneny J.R., Lutes J., Borevitz J.O., Weigel D., Chory J. & Maloof J. (28) Amino acid polymorphisms in Arabidopsis phytochrome B cause differential responses to light. Proceedings of the National Academy of Sciences of United States of America 5, Finch-Savage W.E. & Leubner-Metzger G. (26) Seed dormancy and the control of germination. New Phytologist 171, Galloway L.F. & Etterson J.R. (27) Transgenerational plasticity is adaptive in the wild. Science 318, Gutterman Y. (2) Maternal effects on seeds during development. In SEEDS: The Ecology of Regeneration in Plant Communities, 2nd edn (ed. M. Fenner) pp Redwood Press Ltd., Melksham, UK. Halliday K.J. & Whitelam G.C. (23) Changes in photoperiod or temperature alter the functional relationships between phytochromes and reveal roles for phyd and phye. Plant Physiology 131, Hayes R.G. & Klein W.H. (1974) Spectral quality influence of light during development of Arabidopsis thaliana plants in regulating seed-germination. Plant and Cell Physiology 15, Hennig L., Poppe C., Sweeve U., Martin A. & Schäfer E. (21) Negative interference of endogenous phytochrome B with phytochrome A function in Arabidopsis. Plant Physiology 125, Hennig L., Stoddart W.M., Dieterle M., Whitelam G.C. & Schäfer E. (22) Phytochrome E controls light-induced germination of Arabidopsis. Plant Physiology 128, Heschel M.S., Selby J., Butler C., Whitelam G.C., Sharrock R.A. & Donohue K. (27) A new role for phytochrome in temperaturedependent germination. New Phytologist 174, Holdsworth M.J., Bentsink L. & Soppe W.J.J. (28) Molecular networks regulating Arabidopsis seed maturation, afterripening, dormancy, and germination. New Phytologist 179, Kendrick R.E. & Frankland B. (1969) Photocontrol of germination in Amaranthus caudatus. Planta 85, Kendrick R.E. & Spruit C.J.P. (1977) Phototransformations of phytochrome. Photochemisty and Photobiology 26, Koornneef M., Rolff E. & Spruit C.J.P. (198) Genetic control of light-inhibited hypocotyl elongation in Arabidopsis thaliana (L) Heynh. Zeitschrift fur Pflanzenphysiologie, Lacey E.P. (1996) Parental effects in Plantago lanceolata L. I.: a growth chamber experiment to examine pre- and postzygotic temperature effects. Evolution 5, McCullough J.M.W. & Shropshire W. Jr (197) Physiological predetermination of germination responses in Arabidopsis thaliana (L) HEYNH. Plant and Cell Physiology 11, Maloof J.N., Borevitz J.O., Dabi T., et al. (21) Natural variation in light sensitively of Arabidopsis. Nature Genetics 29, Monte E., Alonso J.M., Ecker J.R., Zhang Y., Xin L., Young J., Austin-Phillips S. & Quail P.H. (23) Isolation and characterization of phyc mutants in Arabidopsis reveals complex crosstalk between phytochrome signaling pathways. The Plant Cell 15, Nagatani A., Reed J.W. & Chory J. (1993) Isolation and initial characterization of Arabidopsis mutants that are deficient in phytochrome-a. Plant Physiology 2, Ni M., Tepperman J.M. & Quail P.H. (1998) PIF3, a phytochromeinteracting factor necessary for normal photoinduced signal transduction, is a novel basic helix-loop-helix protein. Cell 95, Oh E., Kim J., Park E., Kim J.I., Kang C. & Giltsu C. (24) PIL5, a phytochrome-interacting basic helix-loop-helix protein, is a key negative regulator of seed germination in Arabidopsis thaliana. The Plant Cell 16, Oh E., Yamaguchi S., Kamiya Y., Bae G., Chung W.I. & Choi G. (26) Light activates the degradation of PIL5 protein to promote seed germination through gibberellin in Arabidopsis. The Plant Journal 47, Oh E., Yamaguchi S., Hu J., Yusuke J., Jung B., Paik I., Lee H.S., Sun T., Kamiya Y. & Giltsu C. (27) PIL5, a phytochromeinteractive bhlh protein, regulates gibberellins responsiveness by binding directly to the GAI and RGA promoters in Arabidopsis seeds. The Plant Cell 19, Orozco-Segovia A., Brechu-Franco A.E., Zambrano-Polanco L., Osuna-Fernandez R., Laguna-Hernandez G. & Sanchez- Coronado M.E. (2) Effects of maternal light environment on germination and morpholoical characteristics of Sicyos deppei seeds. Weed Research 4,