The Negatively Acting Factors EID1 and SPA1 Have Distinct Functions in Phytochrome A-Specific Light Signaling

Abstract

EID1 (empfindlicher im dunkelroten Licht) and SPA1 (suppressor of phytochrome A[phyA]-105) function as negatively acting components in phyA-specific light signaling. Mutants in the respective genes led to very similar phenotypes under weak-light conditions. To examine whether both genes are functionally redundant, detailed physiological and genetic analyses were performed with eid1 and spa1 mutants isolated from the same wild-type background. Measurements of hypocotyl elongation, anthocyanin accumulation, and Lhcb1-transcript accumulation under different light treatments demonstrated that SPA1 has a strong influence on the regulation of very low fluence responses and a weaker influence on high-irradiance responses. In contrast, EID1 severely altered high-irradiance responses and caused almost no change on very low fluence responses. Analyses on eid1 phyA-105 double mutants demonstrated that EID1 could not suppress the phenotype of the weak phyA allele under continuous far-red light. Measurements on eid1 spa1 double mutants exhibited a strong interference of both genes in the regulation of hypocotyl elongation. These results indicate that EID1 and SPA1 are involved in different but interacting phyA-dependent signal transduction chains.

Phytochromes are a family of plant photoreceptors that are adapted to sense red and far-red light. The functional photoreceptor molecules are dimers of approximately 125-kD subunits, whereby each monomer carries a phytochromobilin chromophore, a linear tetrapyrrole, covalently linked to a conserved Cys residue. The holoprotein is synthesized in its red-light-absorbing Pr form. Upon absorption of red light, the molecule can be photoconverted into the far-red-light-absorbing Pfr form. Similarly, the Pfr form can be photoconverted back into the Pr form when far-red light is absorbed. Because Pfr acts as a positive effector for physiological responses, phytochromes can function as photoreversible light switches when exposed to treatments with pulses of red and far-red light (Furuya and Schäfer, 1996; Whitelam and Devlin, 1997; Neff et al., 2000).

Phytochromes are encoded by a small multigene family in all plant species analyzed (Mathews and Sharrock, 1997). In Arabidopsis, five genes have been described named PHYTOCHROME A (PHYA) to PHYE. The gene products of PHYB to PHYE are light stable in Arabidopsis, whereby the PHYB gene shows the highest expression. The analyses of null mutants and overexpression studies demonstrated that light-stable phytochromes predominantly regulate light responses under continuous red and white light. They are also responsible for the so-called low fluence responses (LFR) that exhibit the classical red/far-red photoreversible characteristic of phytochrome function. LFR are normally induced by red-light photon fluences above approximately 1 μmol m⁻² (Furuya and Schäfer, 1996; Whitelam and Devlin, 1997; Neff et al., 2000; Weller et al., 2001).

The product of the PHYA gene is light labile and accumulates to very high levels in the dark (Clough and Viestra, 1997). Studies using phyA mutants and plants overexpressing phyA indicate that this photoreceptor is responsible for the very low fluence responses (VLFR) and for the far-red-light-dependent high-irradiance responses (HIR; Furuya and Schäfer, 1996; Whitelam and Devlin, 1997; Neff et al., 2000; Weller et al., 2001). Both responses do not show the classical red/far-red photoreversibility. VLFR can be induced with extremely low-photon fluences between 0.001 and 1 μmol m⁻², and even far-red-light pulses are very often sufficient for induction. The extent of the HIR depends on the duration and the fluence rate of the irradiation, and even short intervening dark phases result in a breakdown of the responses. Thus, the HIR sensory system must be able to detect the amount of photons that reach the plant in a given time period. The action spectra of HIR show a maximum at approximately 720 nm, a wavelength that completely inhibits responses of the light-stable phytochromes (Mancinelli, 1994; Furuya and Schäfer, 1996; Neff et al., 2000; Dieterle et al., 2001).

Various screening strategies have been used to isolate Arabidopsis mutants, which exhibit alterations in light regulation (Hardtke and Deng, 2000; Neff et al., 2000). Until now, at least eight independent genetic loci have been identified that exhibit reduced phyA responses. The target genes for fhy3 and fin2 are not yet identified (Whitelam et al., 1993; Soh et al., 1998). So far, FHY1, FAR1, FIN219, PAT1, and HFR1/REP1/RSF1 have been characterized at the molecular level (Hudson et al., 1999; Bolle et al., 2000; Fairchild et al., 2000; Hsieh et al., 2000; Soh et al., 2000; Spiegelman et al., 2000; Desnos et al., 2001). HFR1, REP1, and RSF1 are allelic and code for a basic helix-loop-helix transcription factor that can form heterodimers with PIF3, another basic helix-loop-helix transcription factor involved in phytochrome signal transduction (Ni et al., 1998; Fairchild et al., 2000). The heterodimer binds to promoter elements present in many light-regulated genes. Most probably, the heterodimer also interacts with the Pfr form of phyA, which might lead to an induction of gene expression (Fairchild et al., 2000). The biochemical function of FHY1, FAR1, FIN219, and PAT1 is not yet defined.

Detailed physiological and genetic analyses could dissect phyA signal transduction in at least two branches that correspond to VLFR and HIR signaling. Yanovsky et al. (1997) could demonstrate that the Columbia (Col) ecotype does not respond to very low-light fluences. They identified two quantitative trait loci, vlf1 and vlf2, that are responsible for the expression of VLFR. In contrast, fhy3 mutants only exhibited defects in HIR, whereas mutations in the FHY1 gene resulted in alterations in VLFR and HIR (Cerdán et al., 1999; Yanovsky et al., 2000).

Compared with the large number of mutants with reduced phyA responses, only two mutants, spa1 and eid1, showed enhanced phyA-specific light responses (Hoecker et al., 1998; Büche et al., 2000; Dieterle et al., 2001). Both mutants accumulate normal levels of phyA, and phyA degradation in the mutants was not altered. Genetic and physiological results indicate that the encoded proteins function as negatively acting components of phyA signaling. SPA1 contains WD40 repeats similar to COP1 and exhibits some weak homology with protein kinases (Hoecker et al., 1999). Detailed analysis by Parks et al. (2001) demonstrated that light-induced growth promotion by SPA1 counteracts phytochrome-mediated growth inhibition during de-etiolation. EID1 is an F-box protein. F-box proteins are components of so-called SCF (Skp1/Cdc53/F-box protein) complexes that function as ubiquitin ligases. Thus, EID1 most probably acts by targeting activated components of phyA-signaling pathway to ubiquitin-dependent proteolysis (Dieterle et al., 2001).

The described phenotypes of spa1 and eid1 mutants were very similar. They exhibited hypersensitivity in all analyzed HIR under continuous far-red light, and both showed an increased sensitivity under weak continuous red light (Hoecker et al., 1998; Büche et al., 2000). However, it remained unclear whether these factors are functionally redundant and whether they are involved in independent or identical signal transduction chains. In a screening for novel hypersensitive mutants in phyA-signaling, we could isolate several eid1 and spa1 alleles in a Wassilewskija (WS) wild-type background. Because these alleles were present in the same ecotype, they could be used to compare different light responses of both mutations very accurately. The results obtained for different spa1 and eid1 mutants clearly indicate that both factors are not functionally redundant and that they are involved in different but interacting phyA-dependent signal transduction chains.

RESULTS

Characterization of Mutant Lines

A far-red-light field with low fluence rates was used to screen for hypersensitive mutants in phyA-dependent light signaling. Under these light conditions, the wild type remained almost completely etiolated, whereas hypersensitive mutants should exhibit a clear photomorphogenic development with expanded cotyledons, an open hook, and a reduced hypocotyl elongation (Dieterle et al., 2001). In total, eight independent mutant lines were isolated after screening approximately 20,000 T-DNA and approximately 30,000 EMS lines obtained in a WS background and approximately 10,000 T-DNA lines in a Col background. By complementation analysis, three mutants were identified as novel alleles of spa1 and five mutants as novel alleles of eid1. Thus, all isolated mutants could be assigned to known genes.

The isolation of eid1 and the spa1 alleles from WS enabled a direct comparison of the function of these two genes in the same wild-type background. The spa1 allele used in this study carries an artificial UAA stop codon instead of a CAA triplet that encodes for Gln₁₃₁. The resulting SPA1 protein should be terminated at this position instead of reaching its full length of 1,029 amino acids. The phenotype of the spa1 allele was compared with the phenotypes of the strong eid1-3 and the weak eid1-5 allele (Dieterle et al., 2001). The eid1-3 allele carries an insertion of a single basepair that leads to a frame shift at the 3′ end of the open reading frame. In the eid1-5 allele, a Trp codon was mutated to a stop codon. The resulting truncated EID1 protein should consist of 156 instead of 336 amino acids.

Photoreversibility in eid1 and spa1 Mutants

A unique property of the phytochrome system is its photoreversibility. To test whether the mutations in SPA1 and EID1 lead to a change in photoreversibility, seedlings were treated with multiple hourly light pulses of strong red light, strong RG9 light (extreme far-red light; λ_max = 760 nm), and strong red immediately followed by RG9 light. The hourly pulse treatments were repeated for 3 d after the induction of germination before hypocotyl lengths were measured. The red/RG9 pulse program was established to screen for mutants the exhibit a loss of photoreversibility (Kretsch et al., 2000).

The eid1 mutants and the wild type did not respond to hourly far-red-light pulse treatments, whereas multiple red-light pulse treatments induced the strongest effects (Fig. 1). Seedlings treated with red/RG9 pulses showed a weak photoreversion, but they never reached the hypocotyl length of plants treated with RG9-light pulses alone.

Figure 1.

Analysis of hypocotyl elongation after treatments with hourly light pulses. Seedlings were either kept in darkness or were treated with hourly light pulses of red (30 s of 40 μmol m⁻¹ s⁻¹) or RG9 light (extreme far-red; 3 min of 36 μmol m⁻² s⁻¹), or red-light pulses immediately followed by photoreverting RG9-light pulses. Pulse treatments were administered for 3 d after the induction of germination. Error bars represent se.

With spa1 seedlings, a weak inhibition of hypocotyl elongation was detected even with the RG9-light pulse treatments, but hourly red-light pulses remained more effective (Fig. 1). Interestingly, repetitive red/RG9-light pulse treatments induced the strongest effects leading to reduction in hypocotyl elongation even below the level obtained for seedlings irradiated with multiple red-light pulses.

Fluence Rate Response Curves for the Regulation of Hypocotyl Elongation under Continuous Irradiation

To analyze light sensitivity under continuous irradiation with different wavelengths, fluence rate response curves for hypocotyl elongation were measured for WS wild type, spa1, eid1-3, and eid1-5 (Fig. 2). The seedlings were irradiated at different fluence rates for 3 d after the induction of germination before hypocotyl lengths were determined.

Figure 2.

Fluence rate response curves for the inhibition of hypocotyl elongation under continuous light of different wavelengths. Hypocotyl lengths were measured 3 d after the induction of germination. The relative lengths were determined in relation to the length of dark-grown seedlings for each line. The hypocotyl lengths of dark controls correspond to those shown in Figure 1. se of the individual measurements were between 0.05 and 0.1. Monochromatic light was obtained by interference filters. A, 628-nm DIL interference filter. B, 640-nm DIL interference filter. C, 672-nm DIL interference filter. D, 692-nm DEPIL interference filter. E, 699-nm DEPIL interference filter. F, 708-nm DEPIL interference filter. G, 716-nm DIL interference filter. H, 720-nm DEPIL interference filter. I, 737-nm DEPIL interference filter.

Fluence rate response curves of the WS wild type exhibited a typical sigmoid shape for most of the analyzed wavelengths, except for 692- and 699-nm light (Fig. 2, D and E). At these wavelengths, saturating irradiation was not achieved. In red light (between 628 and 672 nm), seedlings reached a relative hypocotyl length of about 0.4 at saturating photon fluence rates. Far-red light (708–737 nm) showed strongest inhibition of hypocotyl elongation at saturating photon fluence rates with relative hypocotyl length of approximately 0.2.

Compared with the WS wild type, the strong eid1-3 and the weak eid1-5 allele exhibited an increased light sensitivity at all wavelengths analyzed (Fig. 2). Hypocotyl length of eid1 seedlings always reached the level of the dark controls or at least the levels observed for the wild type at very low-light fluence rates. In the red light (628–672 nm), both eid1 alleles exhibited biphasic fluence rate response curves with an increase in relative hypocotyl length between approximately 0.1 and 0.01 μmol m⁻² s⁻¹ (Fig. 2, A–C). As demonstrated in a preliminary study, this local minimum in light sensitivity is caused by the onset of phyA degradation. Light effects below this fluence rates were clearly phyA dependent, whereas the responses at higher fluence rates could be attributed to phyB function (Büche et al., 2000). At wavelengths above 672 nm, fluence rate response curves exhibit a normal sigmoid shape (Fig. 2, D–I).

Seedlings of the spa1 mutants always showed an increased light sensitivity when compared with the wild type (Fig. 2). The biggest difference between spa1 and the other plants was seen at very low-light fluence rates. Whereas WS and eid1 seedlings always reached the hypocotyl length of the respective dark controls in weak light, relative hypocotyl length of spa1 seedlings did not increase above a level of approximately 0.7 to 0.8 (mean 0.72) even at the lowest photon fluence rates applied. Compared with eid1, spa1 showed a higher sensitivity at very low fluence rates, but a lower sensitivity at intermediate or high fluence rates at wavelength below 737 nm. Thus, spa1 seems be involved in a regulatory mechanism that can be saturated even with very low amounts of photons.

Action Spectra for the Regulation of Hypocotyl Elongation

The fluence rate response curves in Figure 2 were used to calculate action spectra for hypocotyl elongation at wavelengths between 692 to 737 nm. Fluence rates leading to a relative hypocotyl length of 0.6 were determined using linear regression. To construct the action spectra, the value obtained for the wild type at 716 nm (0.166 μmol m⁻² s⁻¹) was set to 1 and the relative photon effectiveness for the wild type and the mutants at different wavelengths were calculated accordingly. For spa1 a second, corrected action spectrum was calculated to eliminate possible errors caused by the differences in the shapes of its fluence rate response curves. To correct for the increased level of hypocotyl inhibition at the onset of the HIR in spa1, the mean of all values obtained for hypocotyl inhibition at the plateau level of the fluence rate response curves was determined. This mean (0.72) was set to 1 and fluence rate response curves were recalculated. Photon fluence rates that lead to a relative hypocotyl length of 0.6 were again determined as described above.

The wild type exhibited a typical HIR action spectrum with a maximum at 716 nm (Fig. 3A). The uncorrected and the corrected spa1 action spectra exhibited an increase in light sensitivity compared with the wild type (Fig. 3B). Nevertheless, the overall shapes of the action spectra were very similar to the wild type and the maximum of the action spectra still remained at 716 nm.

Figure 3.

Action spectra for hypocotyl elongation. Fluence rate response curves shown in Figure 2 were used to determine the fluence rate that induced a relative hypocotyl length of 0.6. The value obtained for WS wild type at 720 nm (0.166 μmol m⁻² s⁻¹) was set to 1, and the relative photon effectiveness for the mutants and the wild type at different wavelength was calculated accordingly. A, Action spectrum for the WS wild type. B, Action spectra for spa1. The corrected action spectra [spa1(corr)] was determined to eliminate errors that might occur because of the different shape of the spa1 fluence rate response curves (see Fig. 2). To correct for the increased level of hypocotyl inhibition at the onset of HIR in spa1, the mean of all values obtained for hypocotyl inhibition at the plateau level of the fluence rate response curves was determined. This mean (0.72) was set to 1, and fluence rate response curves were recalculated. Photon fluence rates that led to a relative hypocotyl length of 0.6 were again determined as described above. C, Action spectra for eid1-3 and eid1-5.

The shape of the action spectra of the eid1 alleles clearly differs from the action spectra of WS and spa1 (Fig. 3C). Even though eid1-3 and eid1-5 revealed 50- and 17-fold higher relative photon effectiveness at 716 nm, their action spectra do not exhibit a typical HIR action spectrum with a maximum at 716 nm. Former analyses with eid1 phyB double mutants demonstrated that the respective action spectrum is very similar to the absorption spectrum of the Pr form of phytochromes (Dieterle et al., 2001).

Inhibition of Hypocotyl Elongation under Multiple Far-Red-Light Pulse Treatments

A characteristic property of the HIR is its dependence on continuous irradiation. Even short, intervening dark phases normally lead to a loss of the response (Mancinelli, 1994; Büche et al., 2000). To study this effect, we administered repetitive far-red-light pulses (716-nm DAL interference filter light, 6 μmol m⁻² s⁻¹, 2.5 min) over 3 d after the induction of germination, whereby the duration of intermitting dark phases was varied. Hypocotyl length was measured at the end of the light treatments.

With WS wild type, even intermitting dark phases of 1 min resulted in a clear reduction in the response. WS seedlings reached a level for the relative hypocotyl length of about 0.9 at intermitting dark phases between 30 min and 2 h. The relative hypocotyl length of the dark control was reached when far-red-light pulses were given every 4 h (Fig. 4A).

Figure 4.

Hypocotyl elongation and anthocyanin accumulation under multiple far-red-light pulse treatments. Seedlings were irradiated with multiple far-red-light pulses (716-nm DAL interference filters) of 6 μmol m⁻² s⁻¹ for 2.5 min, varying the duration of the dark phases between the light pulses. Hypocotyl length and anthocyanin accumulation were determined after applying the pulse program for 3 d after the induction of germination. A, Relative hypocotyl length for WS, spa1, eid1-3, and eid1-5. The hypocotyl lengths of the dark controls correspond to those shown in Figure 1. B, Relative hypocotyl lengths for WS, Col, and eid1-6. Hypocotyl length of dark controls were 10.0 ± 0.2 mm for Col and 9.1 ± 0.2 mm for eid1-6 (mean ± se). C, Relative anthocyanin contents of WS, spa1, and eid1-5 seedlings. Anthocyanin content was determined spectroscopically after extraction of 50 seedlings. The absolute absorption values obtained for continuous irradiation were 0.091 for WS, 0.094 for spa1, and 0.081 for eid1-5. Error bars represent se.

The spa1 mutant exhibited an increased persistence in the response toward multiple light treatments. The relative hypocotyl length increased up to an intermitting dark phase of 57 min and 30 s when it reached a plateau level similar to the results determined with fluence rate response curves. Even dark phases up to 4 h did not lead to a further decrease in the inhibition of hypocotyl elongation (Fig. 4A).

The weak eid1-5 exhibited a clearly increased persistence toward multiple pulse treatments. Even intermitting dark-phases of up to 2 h were still effective (Fig. 4A). In contrast to spa1 and similar to the WS wild type, eid1-5 reached the level of its dark control when intermitting dark-phases were extended to 4 h. Comparable but much stronger effects were obtained for the eid1-3 allele (Fig. 4A).

In principle, the light responses induced by multiple far-red pulse treatments can either be caused by signaling cascades inducing an HIR or a VLFR. Yanovsky et al. (1997) demonstrated that the Col ecotype is unable to respond to very low-light fluences. Therefore, experiments were performed with eid1-6 and its respective Col wild type to test whether the increased persistence toward multiple far-red-light pulses in eid1 can be suppressed by the loss of VLFR in this ecotype.

The Col and WS wild types exhibited very similar escape kinetics, even though WS seedlings always seemed to be more sensitive (Fig. 4B). Col seedlings reached the hypocotyl length of their dark controls when the intermitting dark phase is extended to 17 min and 30 s, whereas WS seedlings reached the level of the dark controls only with and intermitting dark phase of 4 h. This differences between Col and WS most probably reflect differences in the VLFR in both ecotypes. Nevertheless, eid1-6 showed decay kinetic similar to eid1-5. Thus, the loss of VLFR in Col could not eliminate the increased persistence toward multiple far-red-light pulse treatments observed for hypocotyl elongation in eid1.

Anthocyanin Accumulation under Multiple Far-Red-Light Pulse Treatments

Anthocyanin accumulation in Arabidopsis seedlings can be induced by a phyA-dependent HIR in far-red light, whereas hourly red-light pulses and continuous irradiation with strong red light did not result in an anthocyanin accumulation in WS, spa1, and eid1-3 (data not shown). These results demonstrate that anthocyanin accumulation in the WS ecotype cannot be obtained under LF and VLF conditions. Thus, measurements of anthocyanin accumulation would be an unambiguous proof to see whether an increased persistence toward multiple far-red-light pulses is caused by an alteration in the HIR decay kinetics.

Only a weak difference in the persistence of the HIR was obtained between spa1 and the WS wild type. In contrast, a much stronger effect was detected with eid1-5 (Fig. 4C). These results demonstrate that mutations in EID1 can alter the decay kinetics of the HIR signaling cascade involved in anthocyanin accumulation.

Lhcb1 Transcript Accumulation under VLF and LF Conditions

Accumulation of Lhcb1 transcripts shows a VLFR and a LFR in Arabidopsis (Cerdán et al., 2000). Furthermore, Lhcb1 transcript accumulation can be induced by one single light pulse, which should eliminate problems that might occur upon multiple pulse treatments. Therefore, this response was chosen to analyze differences in VLFR and LFR between the two mutants and their WS wild type. Three-day-old etiolated seedlings were irradiated with single 1-min red-light pulses of various intensities and plants were harvested after an additional dark-incubation period of 4 h to allow maximum Lhcb1 transcript accumulation.

In wild-type seedlings, Lhcb1 transcript accumulation above the dark level was first detectable in the VLF range at photon fluences of about 0.01 to 0.1 μmol m⁻² (Fig. 5). A strong increase was also seen in the LF range at light fluences above 10 μmol m⁻².

Figure 5.

Lhcb1 transcript accumulation under very low- and low fluence conditions. WS wild-type, eid1-5, and spa1 seedlings were grown in darkness for 3 d after the induction of germination. Etiolated seedlings were treated with 1-min red-light pulses of variable photon fluence rates to obtain different light fluences. Plants were harvested 4 h after pulse treatment, and total RNA was isolated. Lhcb1 transcript levels were measured by hybridization of the corresponding DNA probe to blots of 5 μg of total RNA. Fluorescent signals of 26S rRNA stained with ethidium bromide are shown as a loading control.

The onset of Lhcb1 transcript accumulation in eid1-5 seedlings was weakly shifted to higher light fluences in the VLF range, whereas the onset of LFR remained more or less unaltered (Fig. 5). Thus, mutations in the EID1 gene seem to reduce the sensitivity of this VLFR.

Compared with wild type and eid1-5, spa1 exhibited an extreme shift in light sensitivity under VLF conditions (Fig. 5). Light-dependent Lhcb1 transcript accumulation was even detectable with the weakest red-light pulse applied (0.001 μmol m⁻²). Compared with WS and eid1-5, Lhcb1 transcript in spa1 was clearly increased with light pulses of 0.01 to 10 μmol m⁻².

Far-Red-Light Responses of the phyA-105 eid1-3 Double Mutant

The spa1 mutant was originally screened as a suppressor of the weak phyA-105 allele under continuous far-red light (Hoecker et al., 1998). To test whether eid1 can also suppress the phyA-105 phenotype, eid1 phyA-105 double mutants were isolated, and hypocotyl elongation was analyzed under continuous far-red light.

Both the Rschew (RLD) and the WS background lines exhibited similar fluence rate response curves, whereas the eid1-3 mutant showed an increased light sensitivity (Fig. 6A). Neither the phyA-105 mutant nor the phyA-105 eid1-3 double mutant responded to continuous 725-nm light under the applied light fluence rates. Thus, eid1-3 cannot suppress the phyA-105 phenotype under the observed light conditions.

Figure 6.

Fluence rate response curves for the inhibition of hypocotyl elongation for different wild types, single, and double mutants under continuous far-red light. Hypocotyl lengths were measured 3 d after the induction of germination. Relative lengths were calculated in relation to the length of the respective dark controls. Hypocotyl length of dark controls for WS, eid1-3, eid1-5, and spa1 correspond to those shown in Figure 1. se of individual measurements were between 0.05 and 0.1. Monochromatic light was obtained by interference filters. A, Fluence rate response curves for WS and RLD wild type, eid1-3, phyA-105 (RLD background), and the eid1-3 phyA-105 double mutant obtained with a 725-nm DEPIL interference filter. Hypocotyl length of dark controls were 9.2 ± 0.2 mm for RLD, 8.0 ± 0.3 mm for phyA-105, and 9.4 ± 0.2 mm for the double mutant (mean ± se). B, Fluence rate response curves for WS wild type, eid1-5, spa1, and the eid1-5 spa1 double mutant obtained with a 720-nm DEPIL interference filter. Hypocotyl length of dark controls were 7.8 ± 0.3 mm for the double mutant (mean ± se).

Far-Red-Light Responses of the eid1-5 spa1 Double Mutant

To test for an interaction of eid1 and spa1 function by genetic means, hypocotyl elongation of the single mutants and an eid1-5 spa1 double mutant was analyzed under continuous far-red light (720-nm DIL filter). The eid1-5 spa1 double mutant exhibited an extremely increased light sensitivity (Fig. 6B). The comparison of the fluence rate response curve of eid1-5, spa1, and eid1-5 spa1 demonstrated that the light response in the double mutant is not simply additive. Reduction of hypocotyl elongation in eid1-5 spa1 remained saturated at a fluence rate of 0.013 μmol m⁻² s⁻¹. Under the same fluence rate, eid1-5 exhibited only a weak-light response and spa1 reached its plateau level that remained unaltered even at much lower fluence rates. Furthermore, the double mutant showed an inhibition of hypocotyl elongation below the level of spa1 even at 0.0056 and 0.0023 μmol m⁻² s⁻¹ that had nearly no effect with the eid1-5 mutant. These results clearly indicated that EID1 and SPA1 do not function independently.

DISCUSSION

The data indicate that SPA1 functions as a negative factor in phyA-specific light signaling cascades regulating VLFR. Loss of SPA1 led to increased levels of Lhcb1 transcripts under VLF conditions. An altered responsiveness in VLFR signaling also explains the observed hypersensitivity for the inhibition of hypocotyl elongation at extremely low-light fluence rates and upon multiple pulse treatments with far-red and extreme far-red (RG9) light. Furthermore, the loss of a negative regulator for VLFR might also be the reason for the observed hypersensitivity of spa1 seedlings under multiple red/RG9 pulse treatments. Seedlings with increased VLFR would respond to both, red- and RG9-light pulses. The increased level of phyA under red/RG9 pulse treatments should further amplify this effect. Hourly short red-light pulses normally induce a nearly complete degradation of phyA because of the formation of light labile Pfr, whereas photoreversion from Pfr back to Pr with RG9 light prevents complete proteolysis (data not shown).

In contrast to SPA1, EID1 seems not to be involved in the regulation of VLFR or has even a positive effect on the respective light responses. Fluence rate response curves for hypocotyl inhibition in eid1 mutants always reach the level observed for the wild type at the lowest fluence rates applied. Mutations in EID1 did not change the responsiveness to hourly RG9 pulse treatments and light sensitivity to multiple red/RG9, because Pfr levels adjusted by the extreme far-red-light pulses (Pfr/Ptot < 1%; Beggs et al., 1981) remain to low to induce VLFR. Furthermore, eid1 seedlings lost their ability to respond to multiple 716-nm light pulse treatments (Pfr/Ptot ≈ 5%; Beggs et al., 1981; Mancinelli, 1994) that can induce VLFR when the intermitting dark phase was increased to 4 h. Thus, eid1 mutants behave like WS wild type under these light pulse treatments. Finally, Lhcb1-transcript accumulation is even reduced in eid1-5 under VLF conditions.

SPA1 also alters the sensitivity of HIR (Hoecker et al., 1998; this paper), but its influence is only weak compared with the strong eid1-3 and the weak eid1-5 alleles. The different influences on HIR obtained for spa1 and eid1 became most obvious regarding the action spectra for hypocotyl elongation in far-red light. Mutations in EID1 altered the spectral sensitivity completely and caused an extraordinary increase in sensitivity. The action spectra for spa1 were very similar to the wild type and maximum sensitivity remained at 716 nm. The results of anthocyanin accumulation under multiple pulse treatments also exhibited only a rather weak influence of SPA1 on this HIR, whereas mutations in EID1 had a much stronger effect on this light response.

The observed differences between eid1 and spa1 cannot be attributed to differences in the strength of the alleles used for the experiments. The analyzed spa1 allele carries a stop codon close to the N terminus of the protein that is localized in front of the strongest alleles analyzed by Hoecker et al. (1998, 1999). Furthermore, the two additional spa1 alleles isolated in the WS background during our screenings exhibited the same alterations under all analyzed light conditions (data not shown). Finally, differences in the strength of the alleles cannot explain why the analyzed spa1 allele has a stronger influence on VLFR but a reduced effect on HIR when compared with both eid1 alleles. Thus, these observations clearly indicate that eid1 and spa1 have a distinct mode of action in phyA-specific light signaling. This interpretation is further underlined by the finding that even the strong eid1-3 allele cannot suppress the weak phyA-105 mutant, which carries an amino acid transition (Ala to Val) at position 893 of the apoprotein. The observed inability to suppress the phyA-105 allele can also explain why Hoecker et al. (1998) could not isolate eid1 mutants but five independent spa1 alleles during their screening for extragenic suppressers of phyA-105.

Our data indicate that SPA1 mainly functions as a negative effector on VLFR, whereas EID1 is only involved in signaling cascades regulating HIR. With respect to their influence on VLFR and HIR, spa1 and eid1 resemble the loss-of-function mutants fhy1 and fhy3, respectively. Mutations in the FHY1 gene led to reduction in VLFR and HIR, whereas mutations in FHY3 are mainly involved in HIR (Cerdán et al., 1999; Yanovsky et al., 2000). To explain their data with the fhy mutants, Yanovsky et al. (2000) proposed a model that assumes that FHY1 functions upstream of FHY3 because it is involved in both types of phyA-specific responses, VLFR and HIR. VLFR1, VLFR2, and FHY3 should function downstream of FHY1 in independent branches regulating VLFR or HIR. This model can be adapted to our findings by assuming that SPA1 functions together with FHY1 upstream of EID1 and FHY3 (Fig. 7A). This modified model would also explain why SPA1 is involved in VLFR and HIR signaling, why EID1 is unable to suppress the weak phyA-105 allele, and why spa1 eid1 double mutants exhibited a strong nonadditive interaction between both factors.

Figure 7.

Proposed models for phyA signaling. Independent signaling cascades are given as large arrows. Lines indicate positive (→) or negative () interactions between the proposed independent signaling cascades. A, Model assuming that SPA1 function upstream of EID1. B, Model assuming that EID1 and SPA1 function on two branches divided early in phyA-specific light signaling.

Nevertheless, this view on EID1 and SPA1 function is contradicted by the observed qualitative differences in spa1 and eid1 mutants with respect to HIR. If SPA1 functions upstream of EID1, spa1 loss-of-function mutations should always lead to the same changes in HIR as seen with eid1 mutants. But spa1 did not exhibit a biphasic shape of fluence rate response curves in red light, it did not change the HIR action spectra like eid1, and its influence on HIR was always lower when compared with eid1. This data indicate that branching in VLFR and HIR light signaling takes place very early in phyA-specific signal transduction. The differences between spa1 and eid1 mutants concerning the suppression of the phyA-105 allele indicates that branching might even occur on the level of the photoreceptor molecule. According to our model, an EID1/FHY3-dependent pathway should be mainly involved in the regulation of the HIR under strong, continuous far-red light (Fig. 7B). This branch of phyA-dependent light signaling might have evolved to measure photon fluence rates, i.e. the amount of photons that reach the plant in a given time period (Büche et al., 2000). The SPA1/FHY1-dependent signaling cascade should be involved in the regulation of VLFR and other inductive phyA-dependent light responses, which depend on the photon fluence, i.e. the absolute amount of photons that reach the plant. The interaction between both branches of phyA light signaling should occur at later steps of the signaling cascades (Fig. 7B).

MATERIALS AND METHODS

Plant Material

For genetic crossing and physiological analyses, the following ecotypes and photomorphogenic mutants of Arabidopsis were used: Col, WS, RLD, phyA-105 (ecotype RLD; Xu et al., 1995), eid1-3, eid1-5, spa1 (all ecotype WS), and eid1-6 (ecotype Col-7). All seeds were propagated in a phytochamber (Büche et al., 2000). The isolation of the mutants was described by Dieterle et al. (2001). Upon request, seeds of mutants will be made available in a timely manner for noncommercial research purposes. No restrictions or conditions will be placed on the use of any materials described in this paper that would limit their use in noncommercial research purposes.

Isolation of Double Mutants

To isolate double mutant lines, the respective parental lines were crossed and F₁ plants were allowed to self-fertilize. Individual F₂ plants were analyzed by PCR-based markers to identify homozygous double mutants and approximately 20 seedlings of the F₃ generation were used to verify the result obtained for the F₂ plants. Seeds of the F₃ generation were used for further analyses. The eid1-3 mutant (Dieterle et al., 2001) was crossed with phyA-105 (Xu et al., 1995). eid1-3 phyA-105 double mutants were identified by derived cleaved amplified polymorphic sequences marker analysis (Neff et al., 1998). For detection of the phyA-105 mutation, PCR was performed using oligonucleotides 5′-ACTGGACAGGGAAGGTGTAGTGACAG-3′ and 5′-TACTTGACTTGTGGAAGCAGTCAACAC-3′. The PCR productwas analyzed for the presence (wild type) or absence of a TspRI restriction site (phyA-105). For the detection of the eid1-3 mutation, a second PCR reaction was performed using oligonucleotides 5′-CTCGCTGTTT-GTTGTCCTGGTCTCTTCC-3′ and 5′-GTAAAGCAGTCC-AAGCACCAGAGACAGGAC-3′. The PCR product was digested with AvaII, which cuts the wild-type PCR product but not the eid1-3 fragment. eid1-5 spa1 double mutants were analyzed by cleaved amplified polymorphic sequences marker analysis (Konieczny and Ausubel, 1993). Genomic EID1 fragments were amplified using the oligonucleotides 5′-CCGTGGAAGAAGAATGAAAGACCTGTCCT-3′ together with 5′-CTCGCTGTTTGTTGTCCTGGTCTCTTCC-3′. BsaI could digest only DNA fragments obtained from eid1-5. To identify the spa1 mutation, PCR was performed using the oligonucleotides 5′-AGTTTGAGCATCTTTATCGTTTGGC-3′ together with 5′-GGCCTCCTCCTATTCAAATCTTCGT-3′. Mutation in the spa1 allele created an MboI restriction site that is absent in the wild-type allele.

Seedling Growth and Light Sources

Seeds were sown on four layers of filter paper circles (595, Schleicher & Schüll, Dassel, Germany), which were placed in 94/16 petri dishes (Greiner, Kremsmünster, Austria) supplemented with 5 mL of distilled water. The standard sowing procedure was followed by 2 d of cold treatment at 8°C in the dark and 1 d of red-light induction of germination at 25°C before irradiation with different light treatments for 3 d. Induction of germination was performed with a standard light field (3.9 μmol m⁻² s⁻¹; Heim and Schäfer, 1982). For all other light treatments, modified Prado 500-W universal projectors (Leitz, Wetzlar, Germany) were used as light sources with Xenophot longlife lamps (Osram, Berlin). To measure fluence rate response curves, light was passed through narrow-banded DIL and DEPIL interference filters (Schott, Mainz, Germany). To analyze photoreversion, Arabidopsis seedlings were treated with a multiple pulse program of 30 s of red light (40 μmol m⁻² s⁻¹) followed by 5 min of extreme far-red light (36 μmol m⁻² s⁻¹) and 56 min and 30 s of darkness (Kretsch et al., 2000). Red light for all pulse treatments was obtained by passing the light beam through KG65 filters (Balzers, Liechtenstein, Germany). To get extreme far-red light, 8 mm-thick RG9 cut-off filters (λ_1/2 = 756 nm) were used (Schott). RG9 cut-off filters together with the Osram Xenophot longlife lamps resulted in a peak emission of the extreme far-red light at 760 nm. Far-red-light pulse treatments were performed with 716-nm DAL interference filters (Schott).

Measurement of Hypocotyl Elongation and Anthocyanin Accumulation

Hypocotyl length was measured manually against a ruler. All data represent the mean of at least 40 seedlings analyzed in at least two independent experiments. Anthocyanin extraction and spectroscopic measurements were performed as described by Büche et al. (2000). All data represent the mean of at least five independent experiments.

Northern Analysis

After harvesting, samples were frozen immediately in liquid nitrogen and stored at −70°C. Total RNA was isolated with an RNeasy Plant Mini Kit (Qiagen, Hilden, Germany) using the RLC buffer for extraction. Northern blots were prepared according to standard procedures using 5 μg of total RNA and Duralon-UV membranes (Stratagene Europe, Amsterdam). A fragment of the Arabidopsis Lhcb1-2 gene cloned into the pBluescriptsKS vector (Stratagene) was used as a probe. The probe was labeled with DIG-dUTP (Boehringer, Mannheim, Germany) by a PCR using the primers 5′-CGTCTAGATCAATGGCCGCCTCAACAATGGC-3′ and 5′-CGAATTCGCTCACTTTCCGGGAACAAAGTTGG-3′. Prehybridization, hybridization, washing, and detection were done as described in the manual of the DIG nucleic acid detection kit (Boehringer). The membrane was washed twice with 2× SSC/0.1% (w/v) SDS and twice with 0.1× SSC/0.1% (w/v) SDS at 68°C for 15 min.