Lifetimes of Arabidopsis cryptochrome signaling states in vivo


For correspondence (e-mail


One crucial component in light signaling is the quantity of photoreceptor present in the active signaling state. The lifetime of the signaling state of a photoreceptor is limited because of thermal or otherwise back reversion of the chromophore to the ground state, and/or degradation of the photoreceptor in the light-activated state. It was previously shown that the lit state of plant cryptochromes contains flavin-neutral semiquinone, and that the half-lives of the lit state were in the range of 3–4 min in vitro. However, it was unknown how long-lived the signaling states of plant cryptochromes are in situ. Based on the loss of degradation of cry2 after prolonged dark incubation and loss of reversibility of photoactivated cry1 by a pulse of green light, we estimate the in vivo half-lives of the signaling states of cry1 and cry2 to be in the range of 5 and 16 min, respectively. Based on electron paramagnetic resonance measurements, the lifetime of the Arabidopsis cry1 lit state in insect cells was found to be ~6 min, and thus very similar to the lifetime of the signaling state in planta. Thus, the signaling state lifetimes of plant cryptochromes are not, or are only moderately, stabilized in planta.


Plants adapt to their light environment by using a large number of photoreceptors. The UV–A/blue light region is monitored by cryptochromes (Liu et al., 2010; Chaves et al., 2011), phototropins and members of the Zeitlupe family (Briggs and Christie, 2002; Demarsy and Fankhauser, 2009); phytochromes operate in the red and far-red region of the light spectrum (Rockwell et al., 2006). The WD–40/β–propeller protein UVR8 has been identified as a specific UV–B receptor (Rizzini et al., 2011; Christie et al., 2012; Wu et al., 2012). With the exception of the UV–B receptor, all other photoreceptors are encoded by small gene families in Arabidopsis thaliana. Together, these photoreceptors enable the plant to use light in an optimal way for photosynthesis and to avoid light damage.

Cryptochromes, including cry1 and cry2 from Arabidopsis thaliana, are UV–A/blue light photoreceptors that regulate plant photomorphogenesis. The cry1 mutant shows a severe phenotype in seedling development specifically under blue light, leading to elongated hypocotyls (Ahmad and Cashmore, 1993) and the strongly decreased formation of anthocyanins (Ahmad et al., 1995; Jackson and Jenkins, 1995), as a result of a lack of transcriptional induction of flavonoid and anthocyanin biosynthetic genes, such as chalcone synthase (CHS) (Christie and Jenkins, 1996; Fuglevand et al., 1996). cry2, the second cryptochrome in Arabidopsis has a prominent role in photoperiodic flowering under long-day conditions (Guo et al., 1998; El-Assal et al., 2001), but only a minor function in de-etiolation (Lin et al., 1998). In contrast to cry1, cry2 is light-labile, at least during the transition of seedlings from darkness to blue light, with strong and rapid degradation at fluence rates above 1 μmole m−2 s−1 (Ahmad et al., 1998; Lin et al., 1998; Yu et al., 2007; Weidler et al., 2012; Zuo et al., 2012). A third cryptochrome-like protein (cry3) exists in Arabidopsis that belongs to the so-called cryDASH family (Brudler et al., 2003; Kleine et al., 2003). cry3 has repair activity for UV–B lesions (cyclobutane pyrimidine dimers) in single-stranded DNA (Selby and Sancar, 2006) and loop structures of duplex DNA (Pokorny et al., 2008). Cryptochromes exist in plants, animals, fungi and bacteria (Chaves et al., 2011), and are related to DNA photolyases that repair UV lesions in DNA using photoexcited and fully reduced FADH* as the catalytic chromophore (Sancar, 2003). Like DNA photolyases, cryptochromes contain FAD as the chromophore but normally lack DNA-repair activity. They behave in a similar way as DNA photolyase to photoreduce FAD; however, plant cryptochromes accumulate the neutral FAD semiquinone (FADH°) instead of fully reduced FADH. This form of the photoreceptor is believed to represent the lit state (Chaves et al., 2011). For animal cryptochromes, it has likewise been postulated that the oxidized flavin represents the ground state based on direct in vivo measurements of radical accumulation in response to blue light (Hoang et al., 2008), whereas other workers postulate the photoexcited anionic radical (FAD·¯*) represents the lit state (Ozturk et al., 2011). Upon irradiation, phosphorylation of Arabidopsis cry1 and cry2 in planta has been detected within minutes. This phosphorylation in planta was only observed under blue light illumination, and does not depend on the presence of phytochromes (Shalitin et al., 2002, 2003). The cryptochrome C termini (CCT) are targets for phosphorylation and are constitutively phosphorylated when expressed in plants without the N–terminal photosensory domain (Shalitin et al., 2002). CCT overexpression causes a constitutive photomorphogenic phenotype (Yang et al., 2000). Thus, only the phosphorylated form of cryptochrome is biologically active. Moreover, structural changes in plant cryptochromes were monitored upon photoexcitation (Partch et al., 2005; Kondoh et al., 2011). Taken together, the formation of the lit state is the prerequisite to form the signaling state, which is distinguished from the lit state by further biochemical and structural alterations. In the case of cry2, phosphorylation is probably also the initial signal for the proteasome-mediated degradation that occurs in the nucleus (Shalitin et al., 2002; Yu et al., 2007, 2009; Weidler et al., 2012; Zuo et al., 2012). When Arabidopsis plants are transferred from blue light to darkness, dephosphorylation of cry1 and cry2 is observed with essentially no phosphorylated photoreceptors left after about 15 min (Shalitin et al., 2002, 2003).

Cryptochromes interact with other proteins such as the E3 ubiquitin-ligase COP1 (Wang et al., 2001; Yang et al., 2001), the bHLH transcription factor CIB1 (Liu et al., 2008) and the members of the COP1-related SPA family (Lian et al., 2011; Liu et al., 2011; Zuo et al., 2011; Weidler et al., 2012). The interaction of cry2 with CIB1, and of cry1 and cry2 with members of the SPA quartet, is induced by blue light (Liu et al., 2008, 2011; Lian et al., 2011; Zuo et al., 2011). Initially, the interaction of COP1 with cryptochromes was thought to be independent of light. Recent data showed, however, that cry2 binds COP1 more strongly under blue light in the presence of SPA1, whereas the SPA1–COP1 interaction is alleviated by cry1 under blue light (Liu et al., 2011; Zuo et al., 2011). As cry2 interacts with its photosensory N–terminal domain with CIB1 and SPA1 (Kennedy et al., 2010; Zuo et al., 2011), this domain is not only important for signal perception but also for signal transduction.

The lifetime of a photoreceptor in its signaling state determines the pool size of active photoreceptor, and thus the extent of a photoresponse. For phytochromes, the ratio between the amounts of physiologically active phytochrome (Pfr) and total phytochrome (Ptot) depends on the fluence rate and the spectral composition of the light, because of the light-induced photoconversion of the inactive red light absorbing (Pr) form to Pfr, with peak absorption at around 666 nm, followed by the back-conversion of Pfr to Pr by the photoexcitation of Pfr, with peak absorption at 730 nm (Rockwell et al., 2006). Furthermore, thermal conversion of the 15E isomer of the phytochromobilin chromophore in the Pfr state to the 15Z isomer in the Pr state reduces the level of phytochrome in its signaling state. This process is also known as dark reversion, with half-lives of the PfrPr heterodimer of 19 min for phyA and 3 min for phyB determined for the recombinant Arabidopsis receptors in vitro at 4°C (Eichenberg et al., 2000). Notably, the half-lives of PfrPr heterodimers can differ strongly from those of the PfrPfr homodimers. For example, the dark reversion of the Arabidopsis phyB PfrPfr homodimer has a half-life of >12 h, in contrast to the 3 min of the heterodimer (Eichenberg et al., 2000). The signaling state of phyB is stabilized by interaction with the response regulator 4 (ARR4; Sweere et al., 2001). Therefore, the measured lifetime of the lit state of a photoreceptor in vitro or in a heterologous expression system does not necessarily reflect its real lifetime in situ. Furthermore, the concentration of a photoreceptor's signaling state can be regulated by protein degradation. Well-defined examples in plants are phyA, with a half-life of about 22 h in dark-grown seedlings and 30 min in red light-treated seedlings (Hennig et al., 1999, 2000), and cry2, as mentioned above.

The half-life of the semireduced FADH° in plant cryptochromes, which represents the lit state, is in the range of 3–4 min, as determined in vitro for insect cell-expressed cry1 and cry2 (Banerjee et al., 2007; Bouly et al., 2007; Müller and Ahmad, 2011). However, essentially no data exist on the lifetimes of the signaling states of cryptochromes in situ. The only example is Drosophila melanogaster cry, which is rapidly degraded in vivo upon photoexcitation, like Arabidopsis cry2 (Koh et al., 2006). Based on the kinetics of decay a half-life of ~27 min was calculated for its signaling state (Ozturk et al., 2011). Physiological studies on UV–B-induced chalcone synthase (CHS) induction in Petroselinum crispum (parsley) cell cultures showed that a single pulse of blue light enhances this process strongly. This effect of blue light was very long lasting, for at least 20 h (Ohl et al., 1989). These data suggest that either a stable signaling component is formed by the action of a cryptochrome-type photoreceptor that enhances UV–B signaling, or that the signaling state of this cryptochrome photoreceptor is very stable in situ. To discriminate between these possibilities and to obtain reliable data on the lifespan of the signaling states of plant cryptochromes in situ we studied the decay time of the signaling state of Arabidopsis cry2 by measuring the degradable fraction of photoexcited cry2 in living Arabidopsis cells after various dark periods on ice, implying that these conditions allow normal photochemistry but blocked biochemistry. These studies revealed a signaling state lifetime for cry2 of 16 ± 5 min. The loss of reversibility of a pulse of blue light by a pulse of green light on the induction of CHS transcription, which is mediated by cry1, was observed for dark intervals between the two pulses with a half-life in the range of 5 ± 1 min. A half-life of the lit state of Arabidopsis cry1 in insect cells in the range of ~6 min was determined by EPR spectroscopy. These findings led us to conclude that the signaling states of plant cryptochromes are short-lived, both in vitro and in vivo, with essentially no or only little stabilization in planta.


Lifetime of the cryptochrome 2 signaling state in Arabidopsis cell culture

The Arabidopsis cry2 protein is rapidly degraded in etiolated seedlings upon exposure to blue light. There is compelling evidence that under blue fluence rates above 10 μmoles m−2 s−1 no other photoreceptor than cry2 is involved in this response (Yu et al., 2007; Weidler et al., 2012). Thus, the degradable fraction of cry2 represents roughly the fraction of cry2 in the signaling state. To determine the half-life of the cry2 signaling state in vivo we used an approach applied before to determine the lifetime of the signaling state of Drosophila cry (Ozturk et al., 2011), which is also light labile. This approach is based on the assumption that the photochemistry of cry2 including dark reversion occurs normally on ice, whereas the enzymatic modifications, such as proteolytic degradation, are inhibited on ice. Instead of etiolated seedlings we used dark-adapted Arabidopsis cell cultures for these experiments to allow the rapid cooling and warming of the samples. Immunoblot analysis showed that under continuous blue light irradiation at 25°C, essentially no cry2 protein was detectable after 60 min (Figure 1b). Thus, the cell culture behaves similarly to etiolated seedlings regarding cry2 degradation (Lin et al., 1998; Yu et al., 2007; Weidler et al., 2012; Zuo et al., 2012). Next, pre-cooled cell cultures were treated with a 15–min pulse of saturating blue light followed by dark incubation on ice for various periods of time. Cells were then shifted to 25°C for 60 min in darkness to allow complete degradation of this cry2 fraction, which remained in the signaling state after the dark period on ice (see Figure 1a for the experimental set-up). These pulse experiments revealed that the prolongation of the dark period on ice after the blue light pulse led to a successive increase in the level of cry2, and thus to a successive decrease in the level of degradable cry2 protein, as expected (Figure 1b). A first-order exponential fit of the quantitative immunoblot data revealed a half-life of the cry2 signaling state of 16 ± 5 min (Figure 1c).

Figure 1.

Decay of the cry2 signaling state in vivo, determined by the degradable fraction of cry2 remaining after dark periods of various duration. Pre-cooled dark-adapted Arabidopsis thaliana cell culture cells were illuminated with blue light (471 nm, 50 μmol m−2 s−1) for 15 min at 0°C. After illumination, samples were kept for various time periods on ice in darkness before transfer to 25°C for 60 min in darkness. Protein samples were analyzed by western blotting using cry2-specific antibodies and antibodies against α–tubulin. (a) Experimental set-up. The dotted arrows indicate dark incubation of various durations on ice after the blue light pulse; the continuous arrows indicate incubation at 25°C for 60 min in darkness to allow full cry2 degradation. (b) Results of a representative immunoblot. Samples cD (continuous darkness) and cB (continuous blue light) were kept on ice for 35 min in darkness or blue light, and then shifted to 25°C for 60 min in darkness or blue light, respectively. (c) Quantitative data of cry2 degradation. The cry2 levels were normalized against the α–tubulin signal in the same blot and lane using fluorophore-labelled secondary antibodies and the Li-COR Odyssey infrared imaging system. Mean values ± SE from three biological replicates are presented.

To test our supposition that degradation of cry2 is not taking place when the cells are cooled on ice, the level of cry2 protein was analyzed with western blots. Cells were kept on ice or at 20°C and irradiated with continuous blue light (471 nm, 50 μmol m−2 s−1) for up to 60 min. Figure S1 shows a rapid and strong decrease in the level of cry2 protein for cells incubated at 20°C. In contrast, no decrease in the cry2 signal could be detected for cells kept on ice. Quantification and normalization of the cry2 signal against the tubulin signal revealed a constant cry2 level in the cells kept on ice, whereas the cry2 level decreased to ~25% after 60 min of blue light irradiation at 20°C. We conclude from these data that the cooling of the cells to 0°C does prevent the de novo synthesis and degradation of cry2.

Lifetime of the cryptochrome 1 signaling state in Arabidopsis seedlings

In contrast to cry2, the protein level of cry1 is essentially not affected by light (Ahmad et al., 1998, 2002). We confirmed the light stability of cry1 for the light conditions applied in the following experiments, and observed the same cry1 level in darkness, and in samples that were treated for 8 h with continuous blue light or with dichromatic blue plus green light. A very small and roughly twofold higher level of cry1 protein was detected in seedlings irradiated with green light for 8 h (Figure S2).

We considered cry1-controlled gene induction as a suitable and relatively fast readout system to analyze the lifetime of the cry1 signaling state. cry1 signaling is known to be attenuated by green light, most likely because the flavin-neutral semiquinone absorbs green light followed by photoconversion to the short-lived and inactive fully reduced flavin state (Banerjee et al., 2007; Bouly et al., 2007; Sellaro et al., 2010). Thus, we wanted to use the green light reversibility of the cry1 response to detect the level of cry1 remaining in the signaling state after photoexcitation followed by dark periods of various durations. To this end, we analyzed CHS transcript levels in Arabidopsis seedlings, including three wild-type accessions and the cry1, cry1 cry2, phyA phyB and cry1 cry2 phyA mutants. These studies revealed a strong ~60–fold upregulation of CHS transcripts in the wild types compared with dark-grown seedlings, which is mainly controlled by cry1 (Figure S3). However, we also observed a strong dependency of this blue-light response on phytochromes, similar to the results of Ahmad and Cashmore (1997) who demonstrated weak accumulation of anthocyanin under blue light in the phyA phyB mutant. Thus, we considered endogenous CHS expression as not suitable to get information about the lifetime of the cry1 signaling state.

To avoid the complexity caused by the requirement of phytochrome for the full action of cry1 on the induction of the endogenous CHS gene, we decided to use another readout system for analyzing the lifetime of the cry1 signaling state, which is independent from phytochromes. This is particularly important because blue and green light establish different Pfr/Ptot ratios, and thus complicate further the interpretation of data from studies where the effect of a pulse of blue light should be reverted by a pulse of green light. We have shown before that the blue light-induced expression of the reporter β–glucuronidase under the control of a ~1–kbp fragment of Sinapis alba (white mustard) CHS-promoter (CHS::GUS) in Arabidopsis is controlled by cryptochrome independent from phyA and phyB (Batschauer et al., 1996). To this end, we first studied the effect of green light on blue light-induced expression of this chimeric gene construct under continuous irradiations. Monochromatic blue light led to a significant induction of CHS-promoter driven GUS expression (Figure 2). Green light alone had only a very low inductive effect. Most importantly, the upregulation of CHS::GUS expression by continuous blue light was strongly reduced by simultaneously applied green light. This confirms the previously described effect of green light on the lit state of cry1 (Banerjee et al., 2007; Bouly et al., 2007; Sellaro et al., 2010), and allowed us to replace continuous irradiation by repetitive light-pulse treatments. For these experiments, etiolated seedlings were treated every 30 min with a 5–min pulse of blue light, followed by a green light pulse of the same fluence rate and duration, without or with intermitting dark phases of between 5 and 20 min. These pulse treatments were applied for 8 h. From these seedlings protein extracts were made and GUS activities were determined (Figure 3a). A complete loss of reversibility by the green light pulse was observed when the dark intervals between the blue and green light pulse were longer than 15 min. Plotting reversibility against the dark intervals allowed a first-order exponential fit (Figure 3b), from which the half-life of green light reversibility was calculated to be ~5 ± 1 min.

Figure 2.

β–Glucuronidase activity in transgenic CHS::GUS Arabidopsis seedlings grown under continuous irradiation. Transgenic seedlings carrying a Sinapis alba (white mustard) chalcone synthase (CHS) promoter::β-glucuronidase (GUS) reporter fusion were kept for 38 h at 25°C in darkness, followed by irradiation for 8 h with LEDs emitting blue (cB), green light (cG), blue plus green light (cBG), each with a fluence rate of 10 μmol m−2 s−1, or kept in darkness (cD). Spectra of the LEDs are shown in Figure S4. GUS activity is presented as formed pmol 4–methyl-umbilliferone (MU) mg protein−1 min−1.

Figure 3.

Loss of green light reversibility of blue light-induced CHS::GUS expression. The same material of transgenic CHS::GUS seedlings grown as described in Figure 2 were treated over a period of 8 h with 5–min blue light pulses (λmax 471 nm, 10 μmol m−2 s−1) applied every 30 min. (a) Relative GUS expression normalized against the sample that received only blue light pulses (B), set as 100%. The other probes were either kept in constant darkness (cD) or received a 5–min green light pulse (λmax 570 nm, 10 μmol m−2 s−1) directly after the blue pulse (BG0) or after dark intervals of 5 min (BG5), 10 min (BG10), 15 min (BG15) and 20 min (BG20). (b) Monoexponential fit of data from (a) showing the fraction of blue light-induced GUS activity that can be suppressed by a pulse of green light given directly after the blue light pulse, or after various lengths of dark intervals between the two pulses. Mean values ± SEs of three biological replicates are presented.

Lifetime of the cryptochrome 1 lit state in insect cells

We have previously shown that the formation of the flavin radical state of cry1 can be monitored in insect cells by EPR spectroscopy (Bouly et al., 2007). We made use of this system to determine the lifetime of the cry1 lit state in an artificial in vivo system that roughly resembled the natural redox potential environment of plant cells, but lacked the plant-specific components required for photoreceptor phosphorylation and dephosphorylation, as well as interacting proteins. Blue light treatment of dark-adapted insect cells that overexpressed cry1 led to a rapid formation of the radical state with a time constant of 15 s (Figure 4). The exponential fit of the decrease of the EPR signal in darkness revealed a half-life of the radical state of 360 ± 20 s (Figure 4). Thus, the lifetime of the cry1 lit state in insect cells is very close to the value calculated from the loss of reversibility of the cry1 signal by green light in Arabidopsis seedlings.

Figure 4.

Formation and dark conversion of the flavin radical of cry1 in insect cells. Insect cells overexpressing Arabidopsis cry1 were kept at 8°C and treated with a 5–min pulse of blue light (λmax 450 nm; 440 μmoles m−2 s−1), followed by dark incubation. Electron paramagnetic resonance (EPR) signals of the FAD radical state were determined and normalized. Exponential fits for radical formation and dark adaptation resulted in half-lives for the FADox state of 15 s in blue light and 6 min (solid line) for the radical state in darkness, respectively.


This work was performed to obtain information about the lifetimes of the signaling states of Arabidopsis cry1 and cry2. In the following discussion we differentiate between the photoexcited state or lit state of the photoreceptor, which contains the neutral semiquinone radical of FAD (Chaves et al., 2011), and the signaling state, which is obtained from the lit state by further biochemical reactions, including phosphorylations (Shalitin et al., 2002, 2003; Bouly et al., 2003) and structural changes (Partch et al., 2005; Kondoh et al., 2011).

Lifetime of the signaling state of cry2

The signaling state of cry2 rapidly degrades, detectable within minutes of blue light exposure of etiolated Arabidopsis seedlings (Ahmad et al., 1998; Yu et al., 2007). We show here that dark-adapted Arabidopsis cell cultures have a cry2 degradation response very similar to etiolated Arabidopsis seedlings (Figures 1 and S1). The rapidity of cry2 degradation makes this response an ideal system to study the lifetime of the cry2 signaling state. We adopted the experimental set-up established by Ozturk et al. (2011), which is based on the probable assumption that at 0°C photochemical reactions in cryptochrome continue to occur more or less normally, whereas the biochemical reactions are blocked. This block also includes de novo synthesis of the photoreceptor protein. Shifting the cells to 25°C then allows the biochemical reactions to occur. We confirmed this proposition for Arabidopsis cells by comparing the degradation of cry2 in cells that were kept either on ice or at 20°C (Figure S1). With this set-up a half-life of 27 min for the Drosophila cry signaling state was calculated (Ozturk et al., 2011). In our studies we used cell cultures instead of seedlings because temperature shifts are more rapid and traceable in the suspension culture. After 60 min of incubation at 25°C an almost complete degradation of the remaining cry2 in the signaling state was assured (Figure 1b, cB). Based on quantitative immunoblot data, a half-life of 16 ± 5 min for the signaling state of cry2 was calculated. This value may have some experimental uncertainties for the following reasons: signaling state formation requires the phosphorylation of the lit state. As this reaction is likely to be temperature dependent, there is some delay during the warming up of the sample. It is essentially impossible to estimate the rate of transition from the lit state to the signaling state during the temperature shift. Moreover, the kinetics of cry2 phosphorylation in dark-to-light shifts showed a peak at around 15 min (Shalitin et al., 2002). These studies were performed using a fluence rate of blue light of 5 μmol m−2 s−1, which probably does not result in the complete conversion of the cry2 pool from the ground state to the lit state. We used a 10–fold higher fluence rate of blue light, but applied it for just 15 min. This light treatment did not convert all of the cry2 molecules to the signaling state because even continuous blue-light treatment does not result in the complete phosphorylation of all cry2 molecules (Yu et al., 2007; Zuo et al., 2012). This is also seen in our study by the fact that only ~65% of cry2 was degraded when cells were shifted directly after the pulse of blue light to 25°C (Figure 1c). The kinetics of cry2 dephosphorylation in light–dark transitions is not analyzed in detail. Data published by Shalitin et al. (2002) show rapid dephosphorylation of cry2 in darkness clearly visible after 10 min. A similar reversion rate half-life of t1/2 ~ 5.5 min was observed for the cry2–CIB1 interaction in mammalian cells (Kennedy et al., 2010). Thus, the kinetics of cry2 dephosphorylation and reversion of interaction of cry2 with CIB1 in darkness fit quite well with the half-life of the signaling state determined here for cry2.

Lifetime of the signaling state of cry1

Cry1 is a light-stable photoreceptor, at least in Arabidopsis (Ahmad et al., 1998). Thus, the assay used here to determine the lifetime of the signaling state of cry2 could not be applied for cry1. Instead, we have chosen the cry1-controlled CHS induction for a relatively rapid response (Feinbaum and Ausubel, 1988; Feinbaum et al., 1991; Christie and Jenkins, 1996; Fuglevand et al., 1996; Hartmann et al., 1998) to determine how long the cry1 signal lasts when blue light is switched off. The experimental set-up we applied included repetitive pulses of blue light every 30 min, followed by pulses of green light with intermitting dark periods of increasing duration. This approach ensured that all samples had received identical doses of blue light and green light. Green light was used to inactivate the fraction of cry1 remaining in the signaling state (Banerjee et al., 2007; Bouly et al., 2007; Sellaro et al., 2010).

The initial analysis of expression of the endogenous CHS revealed that its induction by blue light was clearly reduced in the phyA phyB mutant (Figure S3). Thus, cry1 requires phytochrome for the full induction of the endogenous CHS. Ahmad and Cashmore (1997) described the same dependency of cry1 on phytochromes, in particular on phyB, for anthocyanin accumulation under blue light. CHS is the first enzyme in the pathway specific for flavonoids and anthocyanins (Martin, 1993). Thus, the co-action of cry1 with phyB for this pathway may occur at the level of transcriptional control of CHS, and possibly other genes in this pathway. This dependency of cry1 on phytochrome complicates green light reversibility studies because the calculated Pfr/Ptot ratios for the applied light conditions were 45% for blue light and 73% for the dichromatic blue and green light treatments. Considering the importance of phytochromes, in particular of phyB for cry1-mediated CHS expression (Figure S3) and anthocyanin formation (Ahmad and Cashmore, 1997), the higher concentration of Pfr in the blue plus green light conditions could counteract the decreased level of cry1 in its signaling state. This would be particularly important if the calculated Pfr/Ptot of 45% under blue light was insufficient for the full action of the phytochrome.

Considering these complications it is obvious that an experimental system is needed where the cry1 response does not depend on phytochrome. Indeed, the transgenic Arabidopsis line carrying the white mustard CHS-promoter::GUS gene fusion fulfills this criterion because it showed the same induction by blue light for wild type and the phyA phyB mutant (Batschauer et al., 1996). Moreover, we show here that green light significantly reduces the blue light-induced GUS expression in this line, even under continuous dichromatic illumination (Figure 2). We conclude from these data that the lack of effect from green light on endogenous CHS expression is probably a result of the counteraction between decreased levels of cry1 signaling state and elevated Pfr levels. Why the expression of the CHS::GUS construct in blue light is uncoupled from phytochrome is not clarified.

The data presented in Figure 3 show a loss of reversibility of CHS::GUS expression by a pulse of green light given after a pulse of blue light with intermitting dark intervals between the light pulses. The half-life for green light reversibility was calculated to be 5 ± 1 min. Loss of reversibility experiments have been applied before to study phytochrome responses (e.g. Lange et al., 1971). Loss of reversibility includes two components: dark reversion of the signaling state of the photoreceptor that is no longer accessible to the reverting light; and coupling of the light signal to downstream components that are likewise not amenable to the reverting light. If the decay of the signaling state in darkness is much faster than the coupling of the light signal to the downstream component, then the kinetics of the loss of reversibility would approximately reflect the kinetics of dark reversion. As the decay of the signaling state or the lit state of cry1 cannot be measured by spectroscopic techniques in the plant, we made use of insect cells overexpressing cry1. In these cells the level of cry1 is high enough to detect its lit state formation and decay by EPR spectroscopy. The data presented in Figure 4 show that the normalized EPR signal reaches a plateau after a saturating pulse of blue light with a time constant of 15 s, reflecting fast lit-state formation. The decay of the lit state in darkness has a time constant of ~6 min. This value is very similar to the one calculated for the half-life for loss of reversibility in situ. We conclude from these data that lit state and signaling state decay is much faster than the time required for docking the cry1 signal to downstream components. The nearly identical values determined for the decay of the lit state and the signaling state indicate that the biochemical reactions occurring on cry1 during the light-to-dark transition are very fast. Indeed, Shalitin et al. (2003) showed that essentially all phosphorylated cry1 disappeared within 15 min of plants being shifted from blue light to darkness.

Analysis of the dark reversion of the Pfr state of various phytochromes expressed in yeast cells revealed half-lives in the range of 6–10 min at 4°C for light-stable Arabidopsis phytochromes, whereas the Pfr dark reversion half-life of Arabidopsis phyA is slightly slower (30 min), and that of Oryza sativa (rice) phyA (>12 h) is much slower, than for light-stable phytochromes (Eichenberg et al., 1999, 2000). Thus, the kinetics of dark reversion of cry1 and cry2 is in the same range as determined for light-stable phytochromes; however, it is important to note that the half-lives of Pfr states determined for isolated recombinant phytochromes or the ones expressed and present in yeast cells do not necessarily reflect the situation in planta. Indeed, there are examples for acceleration (Vierstra and Quail, 1985) as well as deceleration (Sweere et al., 2001) of phytochrome dark reversion in planta. This may include interactions with plant-specific factors. One example is the interaction of phyB Pfr with the pseudo-response regulator ARR4 that stabilizes the signaling state of phyB (Sweere et al., 2001). Comparing the kinetics of dark reversion of the lit states (neutral flavin semiquione) of cryptochromes determined in vitro with the values calculated here for the lifetimes of the signaling states in vivo indicates that these values are in the same range for cry1: 2.6 min in vitro (Burney et al., 2009) and 5 ± 1 min in planta (Figure 3). Thus, the signaling state of cry1 is not or is only moderately stabilized in planta. This stabilization does not necessarily require the interaction with partner proteins. In the case of cryptochromes it is well known that reduced concentrations of oxygen (Müller and Ahmad, 2011) and the addition of ATP (Immeln et al., 2007; Burney et al., 2009) stabilize the lit state. The somewhat slower decay of the cry1 lit state in insect cells compared with the isolated protein could be caused by the presence of ATP in the cells and/or lower oxygen concentration within the cells compared with the solution. Indeed, the minor differences observed in the EPR spectra of isolated cry1 and cry1 within the insect cells (Figure S5) could result from ATP-binding and or autophosphorylation (Bouly et al., 2003; Shalitin et al., 2003) in the cells, which does not occur in vitro in the absence of ATP.

For cry2 the decay of the lit state in vitro with a time constant of 250 ± 40 s has been published (Banerjee et al., 2007). This is about four times faster than the 16 ± 5 min determined here for the half-life of the signaling state (Figure 1). The signaling state of cry2 in vivo could also be stabilized by ATP and by lower oxygen concentration. Moreover, the assay used to determine the half-life of the cry2 signaling state included warming up the samples from 0°C to 25°C to let the biochemical reactions occur. This could also have caused some delay in the dephosphorylation of cry2, resulting in the lengthening of the signaling state lifetime.

Finally, cryptochromes form dimers (Sang et al., 2005; Rosenfeldt et al., 2008). For dicot phytochromes it is known that the thermal stability of the PfrPfr homodimer is much higher than that of the PfrPr heterodimer (Eichenberg et al., 1999, 2000). Like phytochromes, the cryptochromes already dimerize in darkness. Thus, the concentrations of ground state/ground state, ground state/signaling state and signaling state/signaling state dimers depend on the wavelength and the fluence rate of the light applied. In the case of cryptochromes, whether just the signaling state homodimers or even heterodimers are biologically active, whether cry1 and cry2 behave in this regard equally, and whether dark reversion is affected by the concentration of signaling state homodimers and heterodimers is unknown. These questions remain important topics for future research.

Experimental Procedures

Plant materials, growth conditions and light treatments

Arabidopsis thaliana wild types [Landsberg erecta (Ler), Columbia (Col–0) and Wassilewskija (WS)] and photoreceptor mutants (cry1, cry1 cry2, phyA phyB in the Ler background, and cry1 cry2 phyA in the Col–0 background) were used, as well as a transgenic line carrying a CHS::GUS fusion construct in the WS background (Kaiser and Batschauer, 1995). Arabidopsis thaliana seeds were surface-sterilized by treatment with 70% ethanol for 2 min, followed by incubation in 5% sodium hypochlorite, 0.05% Tween 20 for 15 min. Seeds were rinsed five times in sterile distilled water and dispersed on filter paper (Bio-Rad, After stratification for 3 days at 4°C in darkness, germination was induced by exposing seeds to white light for 4 h at 22°C. For analyzing GUS expression, plates were kept for 38 h in darkness at 22°C and exposed afterwards to continuous monochromatic blue, green or dichromatic blue and green light for 8 h. For quantitative real-time PCR measurements, wild-type and mutant seedlings were treated the same way as for the GUS assay. Light treatments were performed in light chambers (CLF Plant Climatics, equipped with LEDs [emission maxima at λmax 471 nm (blue light) and λmax 570 nm (green light); for emission spectra of LEDs see Figure S4]. Fluence rates were measured with an Optometer P–2000 spectroradiometer (Gigahertz Optik,, and set to fluence rates as indicated. Plant material was harvested under very dim red light and immediately shock-frozen in liquid nitrogen.

For analyzing cry2 degradation, a green Arabidopsis cell suspension culture (background Col0) was used, grown in 250–ml Erlenmeyer flasks under a 16–h white light/8–h dark regime at 25°C on a shaker set for 100 rpm. The volume of the medium in each flask was 50 ml. Subculturing and medium (MS medium with double strength B5 vitamins) were essentially as described (Mathur and Koncz, 1998), except that the medium contained the following hormones: 0.1 mg l−1 naphthaleneacetic acid (Sigma-Aldrich, and 0.5 mg l−1 kinetin. Flasks were transferred to darkness for 3 days (4 days after subculturing) on a shaker set for 100 rpm. Afterwards, flasks were pre-incubated for 2 h in complete darkness on ice and the temperature was controlled in the medium down to 0°C. In the experiments presented in Figure S1, a second sample was kept at 20°C under otherwise identical conditions. Cells were transferred to Petri dishes under dim red light and kept continuously on ice or at room temperature, as indicated. Afterwards cells were exposed to blue light (50 μmol m−2 s−1) for 15 min on ice in the light chamber. After this blue light pulse, samples were kept for various times on ice in darkness before transferring them to 25°C for 60 min in darkness. For comparison of cry2 degradation in cells kept at 0°C or 20°C, cells were exposed to blue light (50 μmol m−2 s−1) for up to 60 min. A 2–ml portion of each sample was centrifuged (5 min at 20 000 g) and the pellet was immediately frozen in liquid nitrogen.

Protein extracts, gel separation and immunoblot analysis

Frozen cells from the cell culture were ruptured using an MM200 cell mill (Retsch, in 50–μl 2.5× SDS sample buffer (0.5% w/v SDS, 0.05% w/v bromphenol blue, 50 mm EDTA, 25% v/v glycerol), and afterwards boiled for 25 min. Samples were centrifuged (5 min at 20 000 g) and supernatants were used for SDS-PAGE, followed by immunoblots. Protein concentrations were determined by the amido black protein assay (Schaffner and Weissmann, 1973). Equal quantitites (40 μg) of protein extracts were subjected to 10% reducing Laemmli mini-SDS-PAGE (Laemmli, 1970). Proteins were blotted to Parablot NCP nitrocellulose membranes (Macherey-Nagel, For immunodetection, polyclonal rabbit anti-CCT2 antibody (against cry2) and monoclonal mouse anti-α–tubulin antibody (Sigma-Aldrich) in 1 : 1000 dilutions were applied. After three 10–min washing steps in TBS–T buffer (20 mm Tris-HCl, pH 7.5, 150 mm NaCl, 0.1% v/v Tween 20) membranes were incubated with the secondary fluorophore-labeled antibodies (IRDye 680 goat anti-mouse IgG and IRDye 800CW goat anti-rabbit IgG; LI-COR, diluted 1 : 10 000 in TBS–T. The membranes were washed three times in TBS–T and scanned using the Odyssey infrared imaging system (LI–COR). Signal intensities of cry2 were calculated by the odyssey software. The cry2 signal was normalized against the α–tubulin signal in the same lane. The level of cry2 in the dark sample of each genotype was set as 100%. For cry1 immunodetection, 25 μg of total protein was used and the membrane was incubated with anti-CCT1 (against cry1) primary antibody, as described above for cry2.

Gene expression analysis

RNA was extracted from Arabidopsis seedlings using the Plant RNeasy Plant Mini Kit (Qiagen, A 1–μg sample of RNA was used for the digestion of remaining genomic DNA and reverse transcription using the QuantiTect Rev. Transcription Kit (Qiagen). Control PCR reactions were performed to exclude genomic DNA contaminations. For real-time PCR the SensiFAST SYBR No-ROX Kit (Bioline, was used, following the supplier's instructions. As a reference, the gene actin (ACT) was used. Primers used were as follows: CHS-forward, 5′–GCATGTGCGACAAGTCGAC–3′; CHS-reverse, 5′–CCAGAGAAGGAGCCATGTAAGC–3′; ACT-forward, 5′–ACCCTGTTCTTCTTACCGAGGCTCC–3′; ACT-reverse, 5′–GACACACCATCACCAGAATCCAGC–3′. The cycling parameters used were 95°C for 5 min followed by 40 cycles of 95°C for 42 s, 59°C for 35 s and 72°C for 30 s, followed by a final extension for 10 min at 72°C.

Glucuronidase assays

Fluorimetric GUS assays were performed essentially as described by Jefferson (1987). In brief, frozen seedlings were ruptured using a MM200 cell mill (Retsch) in 300 μl of extraction buffer (0.1 m potassium phosphate buffer, pH 7.8, 1 mm DTT), and the lysate was cleared by centrifugation (15 min, 20 000 g at 4°C). A 50–μl portion of the protein extract was mixed with 500 μl of substrate solution (1 mg ml−1 4–methyl-umbelliferyl-β-d-glucuronide in extraction buffer) and incubated at 37°C for 30 min. A 100–μl portion of each sample was thereafter mixed with 900 μl of stop buffer (0.2 m Na2CO3). Fluorescence was measured with an RF–5301 PC spectrophotofluorimeter (Shimadzu, with λEx = 365 nm and λEm = 455 nm. Values were normalized against the protein content of each sample and a calibration standard of 4–methyl-umbelliferone.

Insect cell expression of cry1 and EPR studies

Full-length Arabidopsis cry1 cDNA cloned in vector pBakPak9 (Clontech, was expressed in Sf21 insect cells and purified as described previously (Bouly et al., 2003, 2007). EPR spectroscopy on living insect cells expressing cry1 and on purified recombinant cry1 was performed as described by Hoang et al. (2008).

Calculation of half-lives

The half-lives (t1/2) of cry2 protein degradability and of the green light reversibility of the blue light-induced GUS activity were calculated as t1/2 = τ × ln2 from the time constants (τ) of the mono-exponential decay fits of the data presented in Figures 1c and 3b, respectively. Fits followed the formula y–axis value = y0 + A × e–(x–axis value/τ), where the other two parameters y0 and A are the offset (asymptote) and the amplitude of the decay (the difference in y–axis values for zero and infinity x–axis values), respectively.

Calculation of Pfr/Ptot ratios

Pfr/Ptot ratios (φλ) were calculated using the formula and constants described by Mancinelli (1994).

GenBank accession numbers

GenBank accession numbers for the genes analyzed in this study are: Arabidopsis thaliana CRY1 (AT1G04400), CRY2 (AT4G08920), CHS (AT5G13930), ACT2 (AT3G18780); Sinapis alba CHS (X16437).


Funding was provided by the Deutsche Forschungsgemeinschaft (DFG-Project BA985/12-1 to AB and the DFG-Project Bi464/8-3 to RB), and grants HFSP-RGP0045/2008-C and NSF 33701 to MA. We thank A. Damm and E. Happel for technical support, Cs. Koncz for the gift of the Arabidopsis cell culture and R. Pokorny for calculating the exponential fits.