Direct Observation of a Photoinduced Radical-Pair Intermediate in a Cryptochrome DASH Blue-Light Photoreceptor

Proteins from the photolyase/cryptochrome family share their three-dimensional fold, sequence homology, and the redox-active flavin adenine dinucleotide (FAD) cofactor, but exhibit diverse activities.^[1] In response to blue or UV-A light, they function physiologically in DNA repair, entrainment of the circadian clock, or other processes.^[1-3] Members of the photolyase/cryptochrome family have been identified in various organisms ranging from bacteria to plants, animals and humans.^[1] Within this family, a phylogenetic cluster of genes originally identified from Arabidopsis and Synechocystis encode cryptochrome-like proteins, which are distinct from previously characterized “classic” plant (represented by Arabidopsis HY4) or animal (represented by Drosophila and Homo sapiens) cryptochromes, yet more closely resemble the latter.^[4] Remarkably, the cryptochromes from this new cluster (Cry-DASH) have now been found through all kingdoms of life.^[5] While multiple biological functions have been discussed, the availability of stable, recombinantly expressed, Cry-DASH proteins from diverse species provides the means of deciphering cryptochrome protein chemistry. Results from recent experiments point to the direction that Cry-DASH could work as a transcriptional regulator,^{[4, 5]} as well as a DNA repair enzyme for single-stranded DNA.^[6] Other experimental results suggest the participation of Cry-DASH in circadian input pathways.^{[7, 8]}

Redox reactions are proposed to play a key role in light-responsive activities of cryptochromes.^{[9, 10]} Both in vitro and in vivo experiments suggest that the FAD redox state is changed from fully oxidized (FAD^ox) to the radical form when adopting the signaling state.^{[11, 12]} The results agree with the redox activity of photolyases.^[13] In the latter, when starting from FAD^ox, photoinduced electron-transfer (ET) produces a radical pair (RP), comprised of an FAD and either a tyrosine or a tryptophan radical, directly observable by electron paramagnetic resonance (EPR).^[14-16]

ET-generated RPs are proposed to function as compasses for the geomagnetic field in a large and taxonomically diverse group of organisms.^[17-19] In cases, where magnetic orientation was found to be light-dependent,^[20] a sensor based on magnetic-field sensitive RP chemistry could be in effect, however, also other mechanisms of magnetoreception, e.g., based on iron-containing particles,^[21] can not be ruled out. Cryptochrome’s potential for forming RPs upon blue-light excitation, by analogy with photolyases, and also its presence in the eye of migratory birds,^[22] make it a candidate photoreceptor-based magnetoreceptor.^{[18, 23]} In principle, a compass based on RP photochemistry can be realized by (i) generation of a spin-correlated RP with coherent interconversion of its singlet and triplet states, (ii) modulation of this interconversion by Zeeman magnetic interactions of the two electron spins with the geomagnetic field, and (iii) sufficiently small inter-radical exchange and dipolar interactions such that they do not override the Zeeman interactions. Spin-correlated flavin-based RPs may be able to fulfill these criteria.^{[18, 24, 25]} Hence, understanding the suitability and potential of cryptochromes for magnetoreception requires identification of RP states and their origin, and the detailed characterization of magnetic interaction parameters and kinetics. Here, we use Xenopus laevis Cry-DASH (XlCry-DASH) as a paradigm system to test the hypothesis that spin-correlated RPs can be induced in cryptochromes by blue light.

A common motif of all structurally characterized members of the photolyase/cryptochrome family is a conserved chain of three tryptophans (Trp) for ET from the protein surface to the FAD.^{[26, 27]} From sequence alignment of XlCry-DASH with other family members of known structure,^{[4, 26]} we identified the putative Trp triad for ET (see Fig. 1). In E. coli DNA photolyase, a single photoinduced ET step from a nearby Trp reduces the FAD.^[28] The resulting radical state on the Trp is subsequently transferred to the terminal Trp (W306).^[13] Provided that the FAD is initially fully oxidized, ET generates a short-lived flavin-based RP species.^[14-16] Subsequent proton release or uptake has been shown to eventually result in a RP state from the neutral radicals, [W306^•⋯FADH^•]. Because of the high structural conservation, we assume a similar ET mechanism for DASH cryptochromes.

Figure 1.

The conserved Trp triad of XlCry-DASH. (a) Three-dimensional protein structure homology model from the SWISS-MODEL repository (UniProt ID: CRYD_XENLA). (b) Sequence alignment of five members of the photolyase/cryptochrome family. The conserved Trp residues of the putative ET chain in E. coli photolyase (PHR_ECOLI), Syn. Cry-DASH (CRYD_SYNSP), XlCry-DASH (CRYD_XENLA), garden warbler Cry-1a (CRY1a_GW), and A. thaliana Cry-1 (CRY1_AT) are marked with green triangles. Columns with an alignment score >0.7 are surrounded with a blue frame and the conserved amino acids colored red on white background. If the residues are strictly conserved, they are colored white on red background. The alignment was performed with MultAlin^[42] and further processed with ESPript 2.2.^[43]

Transient EPR (TREPR), with its time resolution reaching up to 10 ns allows real-time observation of short-lived RP states generated by pulsed laser excitation.^[29] We compared the RP signals of the wild-type (WT) XlCry-DASH protein with that of a mutant (W324F) lacking the terminal Trp residue of the conserved putative ET chain, see Fig. 1. In Cry-DASH proteins, the FAD’s isoalloxazine moiety may assume three different redox states, fully reduced (FADH⁻), one-electron oxidized (FAD^•− or FADH^•), or two-electron oxidized (FAD^ox), identifiable by their characteristic optical absorptions (Fig. 2). The pronounced absorbance near 380 nm is due to the second chromophore, methenyl tetrahydrofolyl polyglutamate.^[5] XlCry-DASH with homogeneous FAD^ox can be prepared from a mixture of the three redox states by treatment with potassium ferricyanide.^[14] FAD^ox has been chosen as the initial state because (i) it was found to be the physiologically relevant dark state in plant and animal cryptochromes,^[30-32] and (ii) from FAD^ox, potential spin-polarized RP intermediates can be generated by photoinduced ET.

Figure 2.

Optical absorption spectra of XlCry-DASH recorded at 273 K show the FAD cofactor in different oxidation states: FAD^ox (solid line), FADH^• (dotted line), and FADH⁻ (dashed line). The inset shows XlCry-DASH with the FAD^ox cofactor before illumination (solid line), and XlCry-DASH reoxidized by aerial oxygen after 12 hours of blue-light illumination (dotted line). This is to demonstrate that the protein remains intact in terms of its cofactor contents even upon intensive light illumination conditions.

In Fig. 3, the TREPR signal of WT XlCry-DASH recorded at physiologically relevant temperature (274 K) is depicted in a three-dimensional representation, as a function of the magnetic field B₀, and the time t after pulsed laser excitation at 460 nm. In contrast to conventional continuous-wave EPR, which requires magnetic-field modulation to improve the signal-to-noise ratio, TREPR is recorded in a direct detection mode, so as not to constrain the time resolution of the experiment. Consequently, positive and negative signals indicate the enhanced absorptive (A) or emissive (E) electron-spin polarization of the EPR transitions, respectively.

Figure 3.

Complete TREPR data set of XlCry-DASH measured at 274 K. To control for potential shape changes in the TREPR signal caused by gradual sample degradation, spectra were recorded from low to high magnetic field followed by detection in the opposite magnetic-field direction. Each time profile is the average of 120 acquisitions recorded with a laser pulse repetition rate of 1.25 Hz, a microwave frequency of 9.68 GHz, and a power of 2 mW at a detection bandwidth of 100 MHz. A: enhanced absorption; E: emission.

Upon photoexcitation, WT XlCry-DASH readily forms a spin-polarized paramagnetic species, which we assigned to a RP based on the spectral shape and narrow width of the signal. (A spin-polarized flavin triplet state detected under comparable experimental conditions would span >150 mT due to the large spin-spin interactions between the two unpaired electrons.^[33]) The time evolution of the TREPR signal reveals that the RP state lives for several tens of microseconds, the exponential signal decay being determined by relaxation of the spin polarization to the Boltzmann equilibrium populations. The spectrum of XlCry-DASH recorded after 500 ns (Fig. 4) resembles those obtained recently from light-induced transient RP species (comprised flavin and amino-acid raicals) resulting from FAD photoreduction of photolyases.^{[15, 16]} In the E. coli enzyme, the TREPR-observed RP state was assigned to [W306^•⋯FADH^•], W306 being the terminal Trp residue of the Trp triad. To unravel the origin of the RP signal in XlCry-DASH, we examined the W324F mutant, that lacks the equivalent terminal Trp (W324) (see Fig. 1). Under identical experimental conditions, the W324F protein does not exhibit any TREPR signal, see Fig. 4. We therefore conclude, that W324 is either the ultimate electron donor in ET to the flavin or constitutes an integral part of the Cry-DASH ET pathway leading from the protein surface to the FAD.

Figure 4.

TREPR spectrum of WT (solid blue curve) and W324F (solid green curve) XlCry-DASH recorded 500 ns after pulsed laser excitation. Experimental parameters were as in Fig. 3. The dashed curve shows a spectral simulation of the WT protein EPR data using the following parameters: g_FAD = (2.00431, 2.00360, 2.00217), g_Trp = (2.00370, 2.00285, 2.00246), Ω(g_FAD⋯g_Trp) = (126.5°, 76.5°, 246.5°), D = −0.36 mT, E = 0, Ω(g_FAD⋯D) = (0°, 109.9°, 110.5°), J = +0.24 mT.

Further information on the RP state [W324^•⋯FADH^•] in XlCry-DASH was obtained from spectral simulations performed on the basis of the correlated-coupled RP mechanism,^{[34, 35]} outlined in more detail in the Supporting Information. The calculations were performed using published g-tensor parameters for FAD^[36] and Trp^[37] neutral radicals. The relative orientations of the principal axes of both g-tensors and the dipole-dipole coupling tensor were taken from the homology model (see Fig. 1) and kept fixed. The strength of the dipolar coupling (D = −0.36 mT, E = 0) between FADH^• and W324^• was estimated based on the point-dipole approximation, D(r)/mT = −2.78/(r/nm)³, assuming an inter-radical distance r of 2.0 nm between the points of highest unpaired electron-spin density, C(4a) and C(3), in FADH^• and W324^•, respectively. Different overall inhomogeneous spectral line widths of Gaussian shape were considered in the calculations for both radicals. We restricted our simulations to TREPR spectra observed at very early time points to avoid spectral alterations due to anisotropic spin relaxation. The good agreement between calculated and experimental TREPR spectra (Fig. 4) supports our hypothesis that W324 is the terminal electron donor to FAD. However, satisfactory simulation of the overall E/E/A/A polarization pattern of the WT spectrum required two simultaneous assumptions: (i) a pure electronic singlet-state as RP precursor, which is consistent with findings from optical spectroscopy on the equivalent RP state in photolyase,^[38] and (ii) a positive value for the exchange interaction J, i.e., the triplet configuration of the RP is energetically favored by 2J over the singlet RP configuration. J is assumed to fall off exponentially with the inter-radical distance r, J(r) = J₀ exp(−βr),^[39] where a β-value of (14 ± 2) pm⁻¹ was proposed for ET in proteins.^[40] In our simulations, we obtained the best fit at J = +0.24 mT, which is larger than the value reported for the primary RP in bacterial photosynthesis (|J| = 0.9 mT at r = 1.8 nm^[41]) when scaled to the same inter-radical distance. Given that the bridging aromatic residues W377 and W400 in XlCry-DASH might be conducive for an efficient J coupling between FADH^• and W324^•, our value appears reasonable, but could well be in error by a factor of 2 to 4 due to uncertainties in model geometry and possible reorientations of FAD with respect to W324 in the RP as compared to the ground state. The large radical-radical interactions in the XlCry-DASH RP seemed, at first, to preclude a sufficiently strong response to the Earth’s weak magnetic field, yet the exchange and dipolar interactions might substantially cancel each other, as recently suggested by Efimova and Hore.^[25] In this case, the geomagnetic field could affect product yields in cryptochrome, allowing it to function as magnetoreceptor in the avian compass.

In conclusion, we demonstrated that Cry-DASH readily forms a RP species upon blue-light excitation. We proved the RP’s spin-correlated nature by directly observing electron-spin polarized EPR transitions in real time. Our observations support the conservation of this photo-induced reaction and its biological relevance among cryptochrome/photolyase proteins. We furthermore present the first spectral simulations for a flavin-based spin-correlated RP that allowed us to extract the exchange interaction parameter that is difficult to estimate based solely on three-dimensional structure data. Our simulations suggest (i) RP formation from a singlet-state precursor, and (ii) exchange interactions of significant magnitude, such that they may not be neglected when considering RPs of the type of [W324^•⋯FADH^•] as candidate spin states in a RP mechanism of geo-magnetoreception. Thus, the studies presented here show that the RPs of cryptochromes have fundamental properties appropriate for a magnetic compass.

Experimental Section

Protein Preparation

XlCry-DASH was expressed and purified in the dark, as described previously.^[5] For TREPR studies, protein samples stored in buffer (0.3 M NaCl, 0.1 M Tris•HCl, pH 8.0, 30–50% (v/v) glycerol) were supplemented with 5 mM potassium ferricyanide (PF) and incubated over night to ensure homogeneity of the FAD oxidation state. After removal of excess PF by ultrafiltration, samples were supplemented with 5 mM PF and 35% (v/v) glycerol and used for TREPR, or supplemented with 10 mM EDTA, and subsequently illuminated at 273 K with blue light (Halolux 100HL, Streppel, Wermelskirchen-Tente, Germany) selected with a 430–470-nm band filter to generate reduced states of the FAD. Concentrations of the individual FAD redox states were estimated based on their published absorbance coefficients using a Shimadzu UV-1601PC spectrophotometer.

EPR Instrumentation

Time-resolved detection of EPR following pulsed laser excitation was performed using a laboratory-built spectrometer described in more detail elsewhere.^[15] Pulsed optical excitation of the samples was provided by a Nd:YAG laser (Spectra Physics GCR-11) pumping an optical parametric oscillator (Opta BBO-355-vis/IR, Opta GmbH, Bensheim, Germany) tuned to a wavelength of 460 nm (pulse width, 6 ns; pulse energy, 4 mJ).