Eur. J. Biochem. 270, 2921–2928 (2003) FEBS 2003
doi:10.1046/j.1432-1033.2003.03691.x
PRIORITY PAPER
Novel ATP-binding and autophosphorylation activity associated with Arabidopsis and human cryptochrome-1 Jean-Pierre Bouly1, Baldissera Giovani1,2, Armin Djamei3, Markus Mueller3, Anke Zeugner1, Elizabeth A. Dudkin4, Alfred Batschauer3 and Margaret Ahmad1,4 1
Universite´ Paris VI, Paris, France; 2Service de Bioe´ne´rge´tique, Commissariat a` l’Energie Atomique Saclay, Gif-sur-Yvette, France; Plant Physiology, Phillips-Universitaet Marburg, Germany; 4Penn State University, Media, PA, USA
3
Cryptochromes are blue-light photoreceptors sharing sequence similarity to photolyases, a class of flavoenzymes catalyzing repair of UV-damaged DNA via electron transfer mechanisms. Despite significant amino acid sequence similarity in both catalytic and cofactor-binding domains, cryptochromes lack DNA repair functions associated with photolyases, and the molecular mechanism involved in cryptochrome signaling remains obscure. Here, we report a novel ATP binding and autophosphorylation activity associated with Arabidopsis cry1 protein purified from a baculovirus expression system. Autophosphorylation occurs on
serine residue(s) and is absent in preparations of cryptochrome depleted in flavin and/or misfolded. Autophosphorylation is stimulated by light in vitro and oxidizing agents that act as flavin antagonists prevent this stimulation. Human cry1 expressed in baculovirus likewise shows ATP binding and autophosphorylation activity, suggesting this novel enzymatic activity may be important to the mechanism of action of both plant and animal cryptochromes.
Cryptochromes are blue-light photoreceptors found in plants and animals implicated in multiple blue-light dependent signaling pathways [1]. These include de-etiolation responses such as inhibition of hypocotyl elongation and anthocyanin accumulation in plants, leaf and cotyledon expansion, transition to flowering, or regulation of bluelight regulated genes. In animal systems, cryptochromes have been shown to play a role in circadian rhythms, either directly as components of the circadian pacemaker in mouse [2,3] or, in Drosophila, more indirectly by feeding light information into the circadian clock [4]. The defining characteristics of cryptochromes are N-terminal domains with marked similarity to photolyases [4–6], a class of flavoprotein that catalyse repair of UV-damaged DNA via light-dependent electron transfer reactions [7]. Cryptochromes bind similar cofactors to photolyases, yet lack DNA repair activity [8–10], suggesting evolution of novel activities to explain their role in signaling. Interestingly, although sharing many sequence similarities and apparent functional analogy, plant and animal cryptochromes appear to have evolved independently from different ancestral photolyases, with animal cryptochromes sharing greater sequence similarity to type 6-4 photolyases and plant cryptochromes more similar to type I microbial photolyases [5].
A further defining characteristic of both plant and animal cryptochromes are C-terminal extensions, not found in photolyases, which are essential for a number of cryptochrome functions. Ectopic expression of the C-terminal domain of plant cry1, for example, results in a constitutive de-etiolation response in the absence of light, leading to the suggestion that cryptochromes may function via a light-dependent conformational change that renders these photoreceptors accessible to proteins implicated in cellular signaling pathways [11]. The identification of several such signaling molecules, in particular cop1, which binds to the C-terminal of cry1 both in vivo and in vitro, lends support to such a notion [12,13]. Animal cryptochromes also have been shown to interact directly with cellular signaling intermediates, notably components of the circadian clock [14,15]. Photolyases function by light-dependent electron transfer subsequent to excitation of the flavin cofactor, either involving the pyrimidine dimer of UV-damaged DNA or through a separate pathway involving intraprotein electron transfer [16]. It has recently been shown that a similar lightdependent intramolecular transfer reaction, involving both tyrosine and tryptophan radicals, also occurs in cryptochrome [17]. However, the means whereby such an intramolecular electron transfer reaction, lasting only milliseconds, can play a role in signaling and result in interaction of the photoreceptor with putative downstream signaling intermediates that are not permanently bound to the photoreceptor remains a puzzle. In this work we present a novel ATP-binding and autophosphorylation activity, not found in photolyases, but present in both plant and animal cryptochromes, and discuss this activity in light of a possible role in signaling of the cryptochrome photoreceptors.
Correspondence to M. Ahmad, Universite´ Paris VI, UMR-CNRS 7632, Tour 53 E 5, Casier 156, 4, Place Jussieu, 75252 Paris Cedex 05, France. Fax: + 33 144272916, Tel.: + 33 144272916, E-mail:
[email protected] (Received 4 February 2003, revised 8 April 2003, accepted 4 June 2003)
Keywords: cryptochrome; photolyase; blue light; photoreceptor; autophosphorylation.
FEBS 2003
2922 J.-P. Bouly et al. (Eur. J. Biochem. 270)
Experimental procedures
Quantification of ATP binding in terms of affinity and stoichiometry
Purification of Arabidopsis and human cry1 from insect cells Full-length Arabidopsis cry1 retaining an N-terminal His6 affinity tag was expressed and purified to apparent homogeneity from Sf21 insect cells on nickel columns as described [8]. The purified protein is a yellow protein that binds flavin in the oxidized form. When Co2+ affinity resin (Clontech Laboratories, Palo Alto, CA) was used instead of Ni2+, lysates turned bright yellow during the binding reaction due to release of flavin from the cry1 protein to the surrounding medium. The purified bound protein, in contrast, proved colourless and the absorption spectrum was obtained in a Beckmann DU7400 spectrophotometer. Human cry1 protein containing an N-terminal His6 affinity tag was expressed and purified over Ni2+ affinity columns by the same procedure. ATP-agarose affinity chromatography Assays involving purified cryptochrome from baculovirus were performed as follows: after two washes with 1 mL of buffer A (25 mM Hepes, pH 7.4; 150 mM NaCl; 1 mM dithiothreitol and 60 mM MgCl2), 0.1 mL of adenosine 5¢-triphosphate immobilized on cross linked 4% beaded agarose (Sigma, cat. no. A2767) were incubated with 4 lg of cryptochrome in buffer A for 2 h at 4 C. After incubation, the beads together with bound protein were washed three times with 1 mL of buffer B (buffer A containing 500 mM NaCl). The bound cryptochrome was eluted by incubation with 100 lL of buffer C (buffer A containing 20 mM ATP) for 1 h at 4 C. The eluted supernatant was analyzed by SDS/PAGE. For plant cryptochrome binding assays, total protein extracts were prepared from a dark-grown Arabidopsis cell culture: 1 mL (2 mg) of total protein extract in extraction buffer (50 mM Tris/HCl, pH 7.5; 10 mM NaCl; 5 mM MgCl2; 2.5% glycerol; plant protease inhibitor mix, Sigma) was incubated with 0.2 mL of ATP-agarose (Sigma, cat. no. A2767) for 1 h at 4 C. After incubation, the beads were washed four times with 1 mL of wash solution (50 mM Tris/HCl, pH 7.5; 100 mM NaCl; 5 mM MgCl2) and the protein eluted by incubation in 120 lL elution buffer (wash buffer containing 20 mM ATP) for 15 min at 4 C. Proteins from the different fractions were subjected to SDS/PAGE and immunoblot analysis with antibodies against cry1 and against histone as a negative control.
Cry1 protein (3 lg) was bound to the Ni2+ or Co2+ resin and subjected to stringent washes as for the purification procedure. Instead of eluting the bound protein, samples were subsequently incubated for 2 h at 20 C in 50 mM Hepes (pH 7.0), 20 mM MgCl2 and protease inhibitors with increasing concentrations of ATP containing [a-32P]ATP at constant specific activity. After five washes with buffer containing 50 mM Tris/HCl (pH 7.5), 500 mM NaCl and 10 mM imidazole, the amount of ATP bound either to the immobilized cry1 or remaining free in solution was deter1 mined by liquid scintillation counting and Kd have been determined. Phosphorylation reactions and phosphoamino acid analysis Phosphorylation reactions were carried out at 25 C in buffer containing 4 lg of cryptochrome, 5 lCi of [c-32P]ATP or [a-32P]ATP diluted to a final concentration of 200 lM in unlabelled ATP, 50 mM Hepes, pH 7.0, 20 mM MgCl2 or 20 mM MnCl2 or 20 mM CaCl2 for 1 h. The reactions were stopped by addition of SDS/PAGE sample buffer and labelled proteins were visualized on SDS gels by Coomassie staining followed by autoradiography. For phosphoamino acid analysis, 32P-labelled cryptochrome was treated as described by Hardin & Wolniak [18]. Light-dependent phosphorylation reactions Phosphorylation reactions were carried out at 25 C in buffer containing 2 lg of cryptochrome; 50 mM Hepes, pH 7.0; 5 lCi of [c-32P]ATP diluted to a final concentration of 2 lM in unlabelled ATP and 2 lM MgCl2. Samples were maintained in dark for 10 min, 10 min in white light, or alternatively illuminated for 5 min in the absence of substrate (ATP and MgCl2) followed by 10 min further illumination in the presence of ATP and MgCl2 (5 + 10min). The reactions were performed in the presence of 10 mM 2-mercaptoethanol, 1 mM KI or 0,003% of H2O2. Reactions were stopped by addition of SDS/PAGE sample buffer and labelled proteins were visualized on SDS gels by Coomassie staining. After cutting out the cry1 band, c-32P incorporation was determined by liquid scintillation counting; c-32P incorporated in the sample kept in the dark (with 2-mercaptoethanol) is taken as the reference. All cryptochrome preparations for light-induction studies were freshly purified and used after 24 h dark adaptation at 4 C without prior freezing of the sample.
Results
Direct photo-crosslinking of nucleotides to cry1 32
Cry1 protein samples were mixed with 40 lCi [a- P]ATP diluted to a final concentration of 2 lM in unlabelled ATP; 50 mM Hepes, pH 7.0; 20 mM MgCl2 and the reaction mixture (50 lL) was incubated on ice in the dark for 20 min. Twenty-five microlitres was removed and exposed to short-wavelength UV light for 10 min on ice. After treatment, the samples were subjected to SDS/PAGE and autoradiography.
To identify possible novel biochemical activities associated with cryptochrome, Arabidopsis cry1 (Atcry1) was expressed in a baculovirus system and purified to near homogeneity by Ni2+ affinity column chromatography as previously described [6]. In such preparations flavin is bound in the oxidized form and the loosely associated methenyltetrahydrofolate 2 (MTHF) secondary cofactor is apparently lost in the purification process [19]. We have determined that if Atcry1
FEBS 2003
Cryptochrome autophosphorylation (Eur. J. Biochem. 270) 2923
Fig. 1. Purification and phosphorylation of Arabidopsis cry1 from insect cells. (A) Twenty micrograms of Atcry1 protein purified to apparent homogeneity by either Ni2+ or Co2+ affinity column chromatography was resolved on SDS gels and stained with Coomassie Blue. (B) Absorption spectra of Atcry1 protein (0.7 mgÆmL)1) purified from either Ni2+ or Co2+ affinity columns. Bound flavin is in the oxidized form. (C) Phosphorylation of native or flavin-depleted cry1 protein. Samples were labeled with [c-32P]ATP as indicated in Experimental procedures and run on SDS polyacrylamide gels. The left panel represents the Coomassie stained protein samples, in the right panel the autoradiogram of these samples is shown.
protein expressed in Sf21 cell extracts is purified on Co2+ instead of Ni2+ affinity columns, both the flavin cofactor and 3 MTHF are lost, and the cryptochrome apoprotein depleted in both cofactors can be isolated (Fig. 1A,B). We have used these preparations to investigate possible cryptochromeassociated phosphorylation activity in vitro and observed strong labelling of a band corresponding to Atcry1 in the presence of MgCl2 and c-32P ATP. This activity is found only in cryptochrome preparations retaining flavin chromophore (native protein), but not in flavin-depleted protein preparations under the identical assay conditions (Fig. 1C). Cryptochromes have been identified as substrates for protein kinases in several systems [20,21] and have no homology to known protein kinases themselves. Therefore, to eliminate the formal possibility that trace quantities of contaminating kinases may copurify with native Atcry1 and thereby cause the observed labeling reaction, we have directly investigated the ATP binding activity of cryptochrome by several complementary approaches. In the first approach, purified Atcry1 protein was found to bind quantitatively and completely to ATP agarose affinity columns, no cry1 proteins were lost in the different columns washes and all the protein bound to the column could be completely and specifically eluted with ATP, providing evidence that Atcry1 has ATP binding site(s) (Fig. 2A). In another approach, UV cross linking studies were performed in the presence of a-32P labelled ATP (Fig. 2B), which can bind to cryptochrome but not radioactively phosphorylate it. Such photo cross-linking studies are a classic method for the identification of ATP binding sites in proteins, due to the tighter association of ATP to the protein upon photoactivation of the purine [22]. After incubation with Atcry1 protein, the majority of [a-32P]ATP is not retained after electrophoresis on denaturing SDS gels although, after sufficiently long exposure times, a faint band of residual bound ATP can be visualized on autoradiographs at the position of the Atcry1 protein. After UV-treatment, the degree of labeling is significantly elevated, consistent with tighter binding and crosslinking of the nucleotide directly to the Atcry1 protein (Fig. 2B).
Next, the relative affinity of cryptochrome for ATP was examined by quantitative methods under nondenaturing conditions, and the dissociation constant (Kd) was determined for ATP binding (Fig. 2C). Atcry1 protein from insect cell extracts was immobilized on Ni2+ or Co2+ affinity columns and subjected to stringent washing as in the purification procedure. However, instead of eluting the protein, immobilized samples were exposed to [a-32P]ATP at constant specific activity and varying concentrations of ATP, and the proportion of bound ATP ascertained by scintillation counting. The binding curve obtained shows that a higher amount of ATP is bound by native Atcry1 compared with the flavindepleted form of Atcry1 at all concentrations of ATP tested. Scatchard plot analysis is consistent with a single ATP binding site per cryptochrome. The proportion of ATP binding is calculated as 0.4 molecules ATP bound per molecule of native Atcry1, with a binding affinity of dissociation constant Kd of 19.8 lM. These data are not consistent with a minor contaminant protein being responsible for ATP binding or with nonspecific binding, both of these values being well within the range of such data obtained with known ATP binding protein with high and specific affinity for ATP [23,24]. Flavin-depleted Atcry1 showed somewhat reduced binding affinity and stoichiometry (Kd of 25.1 lM and a calculated 0.19 ATP molecules per molecule cry1 protein; values derived from data in Fig. 2C). Possibly the observed reduction in binding affinity may be due to a degree of cry1 protein misfolding in such flavin-depleted preparations, as crystal structure analysis reveals flavin is in contact with multiple amino acids throughout both Escherichia coli photolyase and a recently characterized novel cryptochrome protein [25], and may thereby help to stabilize the tertiary structure. Finally, to confirm that our findings are not limited to our recombinant cryptochrome preparations, native Atcry1 protein was assayed from crude extracts of Arabidopsis cell cultures and subjected to ATP agarose affinity column chromatography. It was found that all the native cry1 (and
2924 J.-P. Bouly et al. (Eur. J. Biochem. 270)
FEBS 2003
Fig. 2. ATP-binding activity of purified cryptochromes. (A) ATP agarose affinity purification of Arabidopsis cryptochrome visualized on Coomassiestained gels. Lane 1, purified protein before binding reaction; lane 2, supernatant after incubation with the ATP binding resin (unbound Atcry1); lane 3, supernatant after three washes with 500 mM NaCL; lane 4, eluted sample with 20 mM ATP; lane 5, remaining cryptochrome on resin after elution (as determined by boiling of the resin in SDS/PAGE sample buffer subsequent to the last elution). (B) Direct photocross linking of nucleotides. Atcry1 protein samples were electrophoresed before (–) and after (+) UV crosslinking treatment, visualized by Coomassie staining and subsequent autoradiography of dried gels. (C) Scatchard plot analysis for the binding of ATP to cry1 of data from the insert; insert binding of ATP to cry1 immobilized on Ni2+ or Co2+ affinity columns. (D) ATP binding activity of plant cry1. Total protein extracts from an Arabidopsis cell culture were incubated with ATP-agarose and bound protein eluted with 20 mM ATP. Protein fractions were subjected to SDS/PAGE and immunoblot analysis using antibodies specific for cry1 and for histone as a control. The following amounts of protein were loaded per lane: lane 1 (total protein extract), 200 lg; lane 2 (supernatant), 200 lg; lane 3 (eluate), 40 lg. After elution, no cry1 signal could be observed in the ATPagarose fraction.
also cry2, data not shown) protein bound quantitatively to ATP agarose affinity columns, and could be completely and specifically eluted with ATP (Fig. 2D). From these data it can be concluded that crypochrome quantitatively and specifically binds ATP, which is a necessary condition if it is to undergo an autophosphorylation reaction. To further characterize the phosphorylation reaction, it was determined that labelling of Atcry1 in the presence of [c-32P]ATP vastly exceeds that of [a-32P]ATP (Fig. 3A), indicating that there is transfer of the labeled c-phosphate of ATP to the Atcry1 protein and thereby a phosphorylation reaction as opposed to simply nucleotide binding. Labeling also occurs in the presence of [c-32P]GTP (not shown). The phosphorylation reaction requires magnesium and does not occur in the presence of either MnCl2 or CaCl2, neither is it stimulated by CaCl2 in the presence of MgCl2. The phosphotransfer reaction has a requirement that flavin is bound to the molecule (Fig. 3A); in a prior study of Atcry1 phosphorylation, the flavin proved not to be bound and for this reason autophosphorylation was not detected [21] (M. Ahmad, unpublished data). Because flavin stabilizes the conformation of the closely related E. coli photolyase in addition to participating in catalysis [26], possibly the
conformational and/or catalytic requirements for ATP binding are less stringent than those for the phosphotransfer reaction. Interestingly, none of several possible substrates for classic protein kinases were phosphorylated by cryptochromes, including histones, casein, and MBP (MAP kinase substrate) (not shown). Therefore cryptochrome may only be capable of phosphorylating itself and not other substrates. To identify the phosphorylated amino acid(s) of cryptochrome, native cryptochrome was labeled and separated on SDS/PAGE gels, which were subsequently immersed in acid, base, or neutral solutions prior to autoradiography. 4 The radioactivity in the cry1 band was retained under acid but not base conditions (not shown), indicating that the phosphate link is base labile (characteristic of phosphoserine and phospho-threonine but not phospho-tyrosine). To provide definitive evidence of the labeled residue, phosphorylated cryptochrome was submitted to phosphoamino acid analysis [18]. As shown in Fig. 3B, only labeled phospho-serine was detected, and the nonradiolabeled standards (phospho-serine, -threonine, and -tyrosine) were clearly visible as well defined spots by ninhydrin detection. Therefore, cryptochrome undergoes autophosporylation
FEBS 2003
Cryptochrome autophosphorylation (Eur. J. Biochem. 270) 2925
Fig. 3. Phosphorylation activity associated with Atcry1. (A) Phosphorylation reactions were performed as described in Experimental procedures under the indicated buffer and/or purification conditions. (B) Phosphoamino acid analysis. Phosphoamino acid standard spots visualized with ninhydrin are circled. The origin is the lowest spot. 7 (C) Time course of Atcry1 autophophorylation. The mean of three experiments of time course is shown. At equilibrium, the stoichiometry of Atcry1 autophopshorylation was between 0.1 and 0.15 mol phosphate per mol Atcry1.
exclusively at serine residue(s). In order to test the degree of phosphorylation, time course experiments have been performed. We found that the autophosphorylation reaction
was saturable after about 30 min under in vitro conditions (Fig. 3C). It has been reported that the Atcry2 protein is rapidly phosphorylated in vivo in response to blue-light irradiation [27], and also that cryptochrome responses appear to be under redox control in Arabidopsis cell culture systems [28]. To relate our in vitro phosphorylation data to these possible early signaling events of the photoreceptor in vivo, we investigated Atcry1 autophosphorylation in response to light and redox state. A short irradiation of purified cry1 protein previously maintained in darkness for 24 h significantly stimulated the autophosphorylation reaction (Fig. 4A). Interestingly, irradiation by light for 5 min prior to addition of substrate resulted in substantially increased phosphorylation as compared to simultaneous irradiation (Fig. 4A, 5+10min light), suggesting that preillumination results in a long-lived activation of the receptor consistent with changes in redox state of the flavin. Flavin antagonists such as KI that have been found to inhibit redox reactions mediated by other flavoproteins [29] or oxidizing agents such as H2O2 abolished light stimulation of the phosphorylation reaction (Fig. 4B,C), indicating that cryptochrome autophosphorylation is regulated in vitro by both redox state and light. It has been determined that, as for cry2 [24], a rapid bluelight dependent shift in gel mobility due to phosphorylation of cry1 protein occurs in Arabidopsis seedlings (A. Batschauer, unpublished observations). However, we have determined that both phosphorylated and unphosphorylated baculovirus-expressed cry1 protein migrate at the same mobility as dark-adapted cry1 from etiolated Arabidopsis seedlings (data not shown). Therefore, the shift in mobility resulting from blue-light dependent phosphorylation of cry1 (and presumably also cry2) in vivo is a result of labeling by external plant protein kinases, likely at different sites in the protein than those involved in the autophosphorylation reaction. Cryptochromes from plant and animal systems differ in apparent evolutionary origins, animal cryptochrome being most similar to 6-4 photolyases whereas plant cryptochromes apparently evolved from the class I CPD photolyases [5]. Nevertheless, there is considerable similarity between these two classes of signaling molecule. To establish whether animal cryptochrome may likewise contain ATP binding and autophosphorylation activity, human cry1 protein (Hscry1) was expressed in insect cell culture with an N-terminal His6 tag as for Arabidopsis cry1. Purified Hscry1 protein was isolated by nickel affinity column chromatography and the identity of the expressed protein confirmed by Western blot analysis to antibody specific for the C-terminal domain of mouse and human cry1 protein (Fig. 5). Like Arabidopsis cry1, Hscry1 protein was shown to bind to an ATP agarose affinity column and could be eluted specifically with ATP. Furthermore, Hscry1 protein is also phosphorylated in the presence of MgCl2 and [c-32P]ATP. Thus, ATP binding and autophosphorylation activity is apparently retained in both plant and animal cryptochromes.
Discussion We have characterized an ATP binding and autophosphorylation reaction that is intrinsic to the Arabidopsis cry1
2926 J.-P. Bouly et al. (Eur. J. Biochem. 270)
FEBS 2003
Fig. 4. Effect of light and redox agents on cry1 autophosphorylation. Samples were maintained in dark for 10 min, 10 min in white light, or alternatively illuminated for 5 min in the absence of substrate (ATP and MgCl2) following by 10 min further illumination in the presence of ATP and MgCl2 (5 + 10min). The reactions were performed (A) in the presence of 10 mM 2-mercaptoethanol. (B) In the presence 1 mM potassium iodide (KI). (C) In the presence 0.003% of H2O2. Error bars represent the SE of three independent experiments. Autoradiogram shows autophosphorylation activity of one experiment.
Fig. 5. ATP-binding and autophosphorylation activity associated with hCRY1. A cDNA clone comprising the entire coding region of human CRY1 gene with the addition of an N-terminal His6 tag-encoding linker was cloned and expressed in baculovirus expression system as 8 described for Atcry1 [8]. Subsequent to purification on Ni resin, a band of approximately 67 kDa was eluted which crossreacted by Western blot analysis with anti-mCRY1 Ig diluted 500· (made to a 21 amino acid peptide sequence within the C-terminus of CRY1) from Alpha Diagnostic International, San Antonio, TX, ref. CRY11-A. Purified Hscry1 was incubated together with ATP agarose resin and allowed to bind as described for AtCry. Bound and unbound fractions were visualized on polyacrylamide gels followed by Western blot analysis with anti-mCry1 Ig. Lane Pre, sample prior to addition of ATP agarose beads; Post, remaining Hscry1 protein in sample subsequent to incubation with ATP-agarose; Eluted, sample eluted specifically from the column, after the indicated washes (see Experimental procedures), with 20 mM ATP. An autoradiogram shows autophosphorylation activity associated with purified human cry1 under identical assay conditions to AtCry1.
blue-light photoreceptor. This phosphorylation activity is unexpected as cryptochrome shares no amino acid sequence similarity with known protein kinases or ATP binding proteins, although precedents exist for other signaling molecules with unpredicted ATP binding [23] and autophosphorylation activity [23,30]. Furthermore, cryptochromes do not phosphorylate classic substrates such as histones, casein, or MBP, and thereby do not appear to act as protein kinases. Photolyases, the apparent
evolutionary ancestors of cryptochromes, have not been reported to bind ATP; nor have we detected autophosphorylation activity in purified E. coli photolyase preparations (M. Ahmad, unpublished results). Because photolyases contain nucleotide binding pockets located in close association with the catalytic flavin cofactor [31], it seems plausible that a novel ATP-binding and autophosphorylation activity, regulated by light and redox state of the cofactor, could have evolved from such an enzyme. Nevertheless, as there is very little difference between the amino acid sequence of cryptochromes and the most closely related photolyases it is quite a puzzle how such a profound transformation in activity could result. We have determined that the N-terminal domain of Atcry1 with homology to photolyases by itself is sufficient to undergo the autophosphorylation reaction in vitro (M. Ahmad, unpublished data). Therefore, the transformation in enzymatic activity from photolyases to cryptochromes has occurred with just a few amino acid substitutions. Further experiments to explore this transformation should be extremely interesting from a structural, mechanistic, and evolutionary point of view for both function of photolyases and cryptochromes. How can this autophosphorylation reaction be reconciled with the novel signaling function of cryptochromes? It has been proposed that cryptochromes may function by undergoing intramolecular conformational changes that make them accessible to degradative enzymes [32,33] or expose C-terminal effector domains to substrate [11–13]. It has furthermore been shown that Atcry1 undergoes intramolecular light-dependent electron transfer reactions in vitro [17], and that reduction/oxidation reactions involving electron transfer through the conserved tryptophan pathway are critical for light dependent activities of CRY in organisms such as Drosophila, or Xenopus [34–36]. However, intramolecular redox reactions alone are too rapid to result in stable conformational changes of the photoreceptors, as they occur within fractions of a second and are unlikely to involve profound changes in structures of cofactors. Unless cryptochrome were permanently associated with its substrate (as is the case with photolyases), it is difficult to reconcile intramolecular electron transfer reactions with stable changes in photoreceptor conformation
FEBS 2003
Cryptochrome autophosphorylation (Eur. J. Biochem. 270) 2927
that would be long-lived enough to enable activated photoreceptor to find and interact with potential substrate molecules. In the present work we show that the autophosphorylation of Atcry1 is sensitive both to light and the redox state of the enzyme, and that it requires flavin. Hence intramolecular phosphorylation, regulated by light-induced changes in redox state, may be a mechanism to effect long-lived, covalent intramolecular conformational changes important for the signaling function of cryptochromes. The observation that this activity has been identified in both plant and animal cryptochromes, which apparently evolved from separate ancestors, lends support to the functional significance of this phenomenon. Further experiments to explore the signaling role of this autophosphorylation activity are currently in progress.
Acknowledgements We are indebted to Dr Paul Galland for valuable advice, Nabil Lounis for help with phosphorylation studies, Alain Picaud for the phosphoamino acid analysis, Andre´ Klarsfeld for the gift of HsCRY1 antibody, and to members of the plant science laboratory (LPDP) at the University of Paris for helpful discussions. This work was funded by a fellowship from the C.E.A to B. G., grants from the CNRS (Atipe Blanche and SdV) to M. A., and from the Deutsche Forschungsgemeinschaft (Ba985/7-2) to A. B.
References 1. Ahmad, M. (2003) Cryptochromes and flavoprotein blue-light photoreceptors. In Handbook of Photobiology and Biochemistry 5 (Nalwa, H.S., ed.). Academic Press, San Diego, in press. 2. Van der Horst, G.T.J., Muitjtjens, M., Kobayashi, K., Takano, R., Kanno, S., Takao, M., de Wit, J., Verkerk, A., Eker, A.P.M., van Leenen, D., Buijis, R., Bootsma, D., Hoeijmakers, J.H.J. & Yasui, A. (1999) Mammalian Cry1 and Cry2 are essential for maintenance of circadian rhythms. Nature 398, 627–630. 3. Stanewsky, R. (2002) Clock mechanisms in Drosophila. Cell Tissue Res. 309, 11–26. 4. Ahmad, M. & Cashmore, A.R. (1993) Hy4 gene of A. thaliana encodes a protein with characteristics of a blue-light photoreceptor. Nature 366, 162–166. 5. Kanai, S., Kikuno, R., Toh, H., Ryo, H. & Todo, T. (1997) Molecular evolution of the photolyase-blue-light photoreceptor family. J. Mol. Evol. 45, 535–548. 6. Deisenhofer, J. (2000) DNA photolyases and cryptochromes. Mut Res. 460, 143–149. 7. Carell, T., Burgdorf, L.T., Kundu, L.M. & Cichon, M. (2001) The mechanism of action of DNA photolyases. Curr. Opin. Chem. Biol. 5, 491–498. 8. Lin, C., Robertson, D.E., Ahmad, M., Raibekas, A.A., Jorns, M.S., Dutton, C.L. & Cashmore, A.R. (1995) Association of flavin adenine dinucleotide with the Arabidopsis blue-light receptor CRY1. Science 269, 968–970. 9. Malhotra, K., Sang-Tae, K., Batschauer, A., Dawut, L. & Sancar, A. (1995) Putative blue-light photoreceptors from Arabidopsis thaliana and Sinapis alba with a high degree of sequence homology to DNA photolyase contain the two photolyase cofactors but lack DNA repair activity. Biochem. 34, 6892–6899. 10. Hsu, D.S., Zhao, X.D., Zhao, S.Y., Kazantsev, A., Wang, R.P., Todo, T., Wei, Y.F. & Sancar, A. (1996) Putative human bluelight photoreceptors hCRY1 and hCRY2 are flavoproteins. Biochem. 35, 13871–13877.
11. Yang, H.Q., Wu, Y.J., Tang, R.H., Liu, D., Liu, Y. & Cashmore, A.R. (2001) The C termini of Arabidopsis cryptochromes mediate a constitutive light response. Cell 103, 815–827. 12. Wang, H., Ma, L., Li, J., Zhao, H. & Deng, X. (2001) Direct interaction of Arabidopsis cryptochromes with COP1 in light control of development. Science 294, 154–158. 13. Yang, H.Q., Tang, R.H. & Cashmore, A.R. (2001) The signaling mechanism of Arabidopsis CRY1 involves direct interaction with COP1. Plant Cell 12, 2573–2587. 14. Rosato, E., Codd, V., Mazzotta, G., Piccin, A., Zordan, M., Costa, R. & Kyriacou, C.P. (2000) Light–dependent interaction between Drosophila CRY and the clock protein PER mediated by the carboxy terminus of CRY. Curr. Biol. 11, 909–917. 15. Shearman, L.P., Sriram, S., Weaver, D.R., Maywood, E.S., Chaves, I., Zheng, B., Kume, K., Lee, C.C., van der Horst, G.T., Hastings, M.H. & Reppert, S.M. (2000) Interacting molecular loops in the mammalian circadian clock. Science 288, 1013–1019. 16. Aubert, C., Vos, M.H., Eker, A.P.M. & Brettel, K. (2000) Intraprotein radical transfer during photoactivation of DNA photolyases. Nature 405, 586–590. 17. Giovani, B., Byrdin, M., Ahmad, M. & Brettel, K. (2003) Lightinduced electron transfer in a cryptochrome blue-light photoreceptor. Nat. Struct. Biol. 10, 489–490. 18. Hardin, S.C. & Wolniak, S.M. (1998) Low-voltage separation of phosphoamino acids by silica gel thin-layer electrophoresis in a DNA electrophoresis cell. Biotechniques 24, 344–346. 19. Ahmad, M., Grancher, N., Heil, M., Giovani, B., Black, R.C., Galland, P. & Lardemer, D. (2002) Action spectrum for cryptochrome-dependent hypocotyl growth inhibition in Arabidopsis thaliana. Plant Physiol. 129, 774–785. 20. Eide, E.J., Vielhaber, E.L., Hinz, W.A. & Virshup, D.M. (2002) The circadian regulatory proteins BMAL1 and cryptochromes are substrates of casein kinase I epsilon. J. Biol. Chem. 277, 17248–17254. 21. Ahmad, M., Jarillo, J.A., Smirnova, O. & Cashmore, A.R. (1998) The CRY1 blue light photoreceptor of Arabidopsis interacts with phytochrome A in vitro. Mol. Cell. 1, 939–948. 22. Bayley, H. & Knowles, J.R. (1977) Photoaffinity labeling. Methods Enzymol. 46, 69–114. 23. Stephenson, K. & Hoch, J.A. (2001) PAS-A domain of phosphorelay sensor kinase A: a catalytic ATP-binding domain involved in the initiation of development in Bacillus subtilis. Proc. Natl Acad. Sci. USA 98, 251–256. 24. Guthapfel, R., Gueguen, P. & Quemeneur, E. (1996) ATP binding and hydrolysis by the multifunctional protein disulfide isomerase. J. Biol. Chem. 271, 2663–2666. 25. Brudler, R., Hitomi, K., Daiyasu, H., Toh, H., Kucho, K., Ishiura, M., Kanehisa, M., Roberts, V.A., Todo, T., Tainer, J.A. & Getzoff, E.D. (2003) Identification of a new cryptochrome class. Structure, function, and evolution. Mol. Cell. 11, 59–67. 26. Park, H.-W., Kim, S.-T., Sancar, A. & Deisenhofer, J. (1995) Crystal structure of DNA photolyase from Escherichia coli. Science 268, 1866–1872. 27. Shalitin, D. et al. (2002) Regulation of Arabidopsis cryptochrome 2 by blue-light-dependent phosphorylation. Nature 417, 763–767. 28. Long, J.C. & Jenkins, G.I. (1998) Involvement of plasma membrane redox activity and calcium homeostasis in the UV-B and UV-A/blue light induction of gene expression in Arabidopsis. Plant Cell 10, 2077–2086. 29. Short, T.W., Porst, M. & Briggs, W.R. (1992) A photoreceptor system regulating in vivo and in vitro phosphorylation of a pea plasma membrane protein. Photochem. Photobiol. 55, 773–781. 30. Nishiwaki, T., Iwasaki, H., Ishiura, M. & Kondo, T. (2000) Nucleotide binding and autophosphorylation of the clock protein
2928 J.-P. Bouly et al. (Eur. J. Biochem. 270) KaiC as a circadian timing process of cyanobacteria. Proc. Natl Acad. Sci. USA 97, 495–499. 31. Komori, H., Masui, R., Kuramitsu, S., Yokoyama, S., Shibata, T., Inoue, Y. & Miki, K. (2001) Crystal structure of thermostable DNA photolyase: Pyrimidine-dimer recognition mechanism. Proc. Natl Acad. Sci. USA 98, 13560–13565. 32. Ahmad, M., Jarillo, J.A. & Cashmore, A.R. (1998) Chimeric Proteins between cry1 and cry2 Arabidopsis blue light photoreceptors indicate overlapping functions and varying protein stability. Plant Cell 10, 197–207. 33. Lin, C., Yang, H., Guo, H., Mockler, T., Chen, J. & Cashmore, A.R. (1998) Enhancement of blue-light sensitivity of Arabidopsis
FEBS 2003 seedlings by a blue-light receptor cryptochrome 2. Proc. Natl Acad. Sci. USA 95, 2686–2690. 34. Zhu, H. & Green, C.B. (2001) A putative flavin electron transport pathway is differentially utilized in Xenopus CRY1 and CRY2. Curr. Biol. 11, 1945–1949. 35. Ishikawa, T., Hirayama, J., Kobayashi, Y. & Todo, T. (2002) Zebrafish CRY represses transcription mediated by CLOCKBMAL heterodimer without inhibiting its binding to DNA. Genes Cells 7, 1073–1086. 36. Froy, O., Chang, D.C. & Reppert, S.M. (2002) Redox potential: differential roles in dCRY1 and mCRY1 functions. Curr. Biol. 12, 147–152.
