Variable Electron Transfer Pathways in an Amphibian Cryptochrome
TRYPTOPHAN VERSUS TYROSINE-BASED RADICAL PAIRS*
- Till Biskup‡,1,
- Bernd Paulus§,
- Asako Okafuji§,2,
- Kenichi Hitomi¶‖,
- Elizabeth D. Getzoff¶,
- Stefan Weber§ and
- Erik Schleicher§,3
- From the ‡Department of Chemistry, Physical and Theoretical Chemistry Laboratory, University of Oxford, Oxford OX1 3QZ, United Kingdom,
- the §Albert-Ludwigs-Universität Freiburg, Institute of Physical Chemistry, Albertstraße 21, 79104 Freiburg, Germany,
- the ‖Section of Laboratory Equipment, National Institute of Biomedical Innovation, 7-6-8, Saito-Asagi, Ibaraki, Osaka 567-0085, Japan, and
- the ¶Department of Molecular Biology and the Skaggs Institute for Chemical Biology, The Scripps Research Institute, La Jolla, California 92037
- ↵3 To whom correspondence should be addressed. Tel.: 49-761-203-6204; Fax: 49-761-203-6222; E-mail: erik.schleicher{at}physchem.uni-freiburg.de.
Abstract
Electron transfer reactions play vital roles in many biological processes. Very often the transfer of charge(s) proceeds stepwise over large distances involving several amino acid residues. By using time-resolved electron paramagnetic resonance and optical spectroscopy, we have studied the mechanism of light-induced reduction of the FAD cofactor of cryptochrome/photolyase family proteins. In this study, we demonstrate that electron abstraction from a nearby amino acid by the excited FAD triggers further electron transfer steps even if the conserved chain of three tryptophans, known to be an effective electron transfer pathway in these proteins, is blocked. Furthermore, we were able to characterize this secondary electron transfer pathway and identify the amino acid partner of the resulting flavin-amino acid radical pair as a tyrosine located at the protein surface. This alternative electron transfer pathway could explain why interrupting the conserved tryptophan triad does not necessarily alter photoreactions of cryptochromes in vivo. Taken together, our results demonstrate that light-induced electron transfer is a robust property of cryptochromes and more intricate than commonly anticipated.
- Electron Paramagnetic Resonance (EPR)
- Electron Transfer
- Flavoproteins
- Radicals
- Tyrosine
- Cryptochrome
Introduction
Cryptochromes (Crys)4 (1, 2), together with the highly homologous photolyases (PLs), form one of the three known families of flavoproteins that are sensitive to blue light; the other two groups being phototropins, proteins harboring a so-called LOV domain (light-oxygen-voltage), and BLUF (blue-light using FAD) proteins (3, 4). Members of the Cry/PL protein family share an overall three-dimensional fold, the flavin adenine dinucleotide (FAD) cofactor in a U-shaped conformation characteristic of and unique to these proteins, and a chain of three conserved tryptophans leading from the FAD to the protein surface (2, 5, 6). This “Trp triad” provides an electron transfer (ET) channel for photoreduction of FAD in several members of the protein family (7,–,10). Although this conserved Trp triad is missing in the class II cyclobutane pyrimidine dimer PLs, an alternative chain of three tryptophans has been proposed as an ET pathway in these proteins (11) and confirmed by crystal structures and mutational studies (12, 13).
Whereas Crys were originally identified and described as PL homologues lacking DNA repair activity, they are now known to take part in a diverse set of physiological functions, ranging from controlling developmental processes in plants, via the entrainment of the circadian clock in as distinct groups as cyanobacteria, algae, plants, insects, and mammals, to the control of gene expression by means of transcriptional modulation (14,–,16). At least for two members of cryptochromes of the DASH type, in vitro activity for repair of single-stranded, but not double-stranded DNA, could be detected (17, 18). Recently, Crys have gained additional interest as being a promising candidate for the primary sensor molecule of the magnetic compass of (but not restricted to) migratory birds (19,–,22).
For class I cyclobutane pyrimidine dimer PLs, mechanistic details of the primary processes after light excitation have been elucidated (23,–,25), whereas unquestionable mechanisms for the remaining members of the Cry/PL family are still missing (26, 27). The growing body of evidence that Crys perform highly diverse tasks in the distinct phyla of life points toward a similar diversity of the detailed primary processes taking place after the initial light excitation of the FAD at the molecular level. Some Crys likely perform their role without being directly light-activated, e.g. core components of the circadian clock (28, 29). In contrast, insect-type Crys are thought to be photoreceptors involved in the light-input pathway responsible for entrainment of the circadian clock (2, 30). In light of this diversity in function and activation of Crys it is tempting to speculate about multiple reaction mechanisms for Crys. Therefore, detailed spectroscopic investigations of various distinct members of the protein family that are aimed toward the elucidation of mechanistic details as well as unraveling differences between reaction mechanisms of specific Crys seem highly demanding.
ET (from the surface to the FAD) involving the Trp triad has been a common theme in PL research (7,–,9, 31, 32). This photoreduction reaction has been the subject of intense studies during the last decades, yet the biological relevance of this photoreaction in PLs is still under debate (33,–,35). A few mechanisms have been proposed for the primary signaling events in plant and insect Crys (27, 36) including light-induced switching of flavin redox states (involving the conserved Trp triad) (26, 37,–,39). In contrast to PLs, in which the reduced FAD is the active form, dark-adapted Crys contain fully oxidized FAD. As a consequence, a radical pair (RP) comprising an FAD radical and an amino acid radical is generated upon blue-light excitation. Such RP states in Crys, if spin-correlated, have been proposed to form the basis of a photochemical magnetoreceptor (40).
A growing number of mutational studies on the Trp triad have been reported, the majority of which favor the importance of this conserved chain (26, 33, 41). Nevertheless, several studies in which single Trp residues of the conserved triad were mutated showed no physiological effect in vivo and/or only minute effects in vitro (42,–,45). Therefore, single Trp mutations are not always sufficient to disrupt ET and consequently inhibit signaling state formation, assuming that light-induced switching of flavin redox states is part of the initial photochemistry.
Recently, we identified a partly conserved Trp nearby the Trp triad as an alternative terminal electron donor in Synechocystis sp. Cry DASH (SCry), despite its longer distance from the middle Trp of the fully conserved triad (46). This diversity in ET pathways in Crys could account for the apparent inconsistencies among previous studies concerning the biological significance of the Trp triad. Therefore, a thorough investigation of possible ET pathways including unequivocal assignment of RP partners (most likely either Trp or Tyr residues) in different Crys is essential for understanding the functional diversity of the Cry/PL family. As has been shown previously, surface-exposed Trp-324 is key to RP formation in wild-type (WT) Xenopus laevis Cry DASH (XCry) and is the terminal electron donor for the FAD cofactor (10). However, the ET cascade to FAD remained unidentified.
To further investigate the nature of ET pathways at the molecular level, it is essential to identify participating amino acids. Whereas optical spectroscopic methods are well suited due to their high time resolution comparable with that of the ET events themselves, they usually lack both the spatial resolution and the potential to unambiguously identify individual amino acids involved in ET. In contrast, EPR is particularly useful to study the identity and the surroundings of (transient) paramagnetic species that occur during the course of ET (47), whereas being restricted in time resolution for monitoring some of the fast primary photoexcitation events in a time interval from about 10 to 100 ns until the initial spin polarization is decayed. Therefore, we combined steady-state UV-visible spectroscopy with transient absorption (TA) and transient EPR (TREPR) spectroscopy to (i) investigate the FAD photoreduction process, (ii) directly observe the formation and kinetics of transient (optically detectable) species, and (iii) characterize spin-polarized RPs formed upon blue-light irradiation of the Crys under investigation. For these studies, we examined point mutants W377F and W400F in which the middle and inner Trps, respectively, of the conserved pathway were each replaced by a redox-inert phenylalanine, and compared our results with those of previously published studies of the equivalent point mutant (W324F) of the terminal Trp (10).
EXPERIMENTAL PROCEDURES
Sample Preparation
XCry was expressed and purified in the dark, as described previously (10). For TREPR studies, protein samples stored in a buffer (0.1 m Tris·HCl, pH 8.0, 0.3 m NaCl, and 30–50% (v/v) glycerol) were supplemented with 5 mm potassium ferricyanide, K3[Fe(CN)6], and incubated overnight to ensure homogeneity of the FAD oxidation state. After removal of excess K3[Fe(CN)6] by ultrafiltration, samples were either supplemented with 5 mm K3[Fe(CN)6] and 35% (v/v) glycerol and used for TREPR, supplemented with 10 mm EDTA for photoreduction measurements, or directly used for time-resolved optical studies. Protein concentration and homogeneity of the FAD-cofactor redox state were controlled with UV-visible spectroscopy (Varian Cary 100 Scan or Shimadzu UV-1601PC) using published absorbance coefficients (48).
UV-Visible Spectroscopic Studies
For optical spectroscopy, concentrated samples were diluted to a final absorbance of 0.5 at 450 nm. 10 mm EDTA was added as external electron donor (49). Samples were transferred into a cuvette (Hellma 105.250-QS) after degassing the sample solution with a turbo molecular pump. Subsequently, illumination was performed at 280 K with an LED (Philips Lumileds Lighting Company, LuXEON Rebel LXML PR01 0226) emitting light at 455 ± 5 nm with a spectral irradiance of 60 ± 6 μW cm−2 nm−1. The temperature was regulated to ±1 K by a temperature controller (Julabo F20-HC). UV-visible absorption spectra (typically ranging from 300 to 700 nm) of each sample were measured with a double-beam UV-visible spectrophotometer (Shimadzu UV2450). The absorption was recorded from the dark-adapted system and immediately after different periods of illumination. Samples were reoxidized by opening the cuvette in the dark and subsequently absorption spectra were recorded after different periods of time. Loss of the second chromophore methenyl tetrahydrofolyl polyglutamate (MTHF) was determined by scaling the spectra to same absorption at 470 nm (maximum of FADox) and comparing the absorption at 382 nm (maximum of MTHF absorption).
Analysis of Photoreduction Kinetics from UV-Visible Absorption Spectra
The two-step photoreduction process of the enzyme-bound FAD cofactor was analyzed using the reaction scheme in Equation 1,
where k1 and k2 are apparent rate constants for the (first-order) one-electron reduction steps of the fully oxidized flavin, FADox, and the neutral flavin radical, FADH•, respectively. In the absence of any exogenous oxidant (such as oxygen) in the buffer solutions, the corresponding reoxidation
rates, k−1 and k−2, are very small compared with k1 and k2 and have therefore been neglected in the analyses. Consequently, a kinetic analysis analogous to that performed by Okafuji
et al. (11) can be conducted.
Time-resolved Optical Studies
Time-resolved optical spectroscopy was performed at 279 K with a commercial laser-flash photolysis spectrometer (Edinburgh Instruments LP920) and recorded with a digital oscilloscope (Tektronix TDS-3012C). The protein sample was placed in a synthetic quartz (suprasil) semi-microcell (Hellma 108F-QS). The temperature was regulated to ±0.1 K by a temperature controller (Lauda Alpha RA 8). Optical excitation was carried out with an OPO system (Continuum OPO PLUS) pumped by a Nd:YAG laser (Continuum Surelite I) at a wavelength of 460 nm, a pulse width of ∼6 ns, and a pulse energy of 6 mJ. The repetition rate of the spectrometer was set to 0.016 Hz. To account for background signals, transients were measured alternately with and without optical excitation, and used for calculation of difference absorbance spectra with Beer-Lambert's law.
Global Analysis
To test for a specific kinetic model, a target analysis was performed using the software Glotaran (version 1.0.1) (50).
Time-resolved X-Band EPR Studies
Time-resolved detection of EPR following pulsed laser excitation was performed at 274 K using a laboratory built spectrometer in conjunction with a Bruker microwave bridge (ER041 MR) (10). The protein samples were placed in a synthetic quartz (Suprasil) sample tube (1.8 mm inner diameter) and irradiated in a dielectric ring resonator (Bruker ER 4118X-MD5) immersed in a laboratory built helium gas-flow cryostat. The temperature was regulated to ±1 K by a temperature controller (Lake Shore 321). A microwave frequency counter (EIP 548) was used to monitor the microwave frequency. Identical microwave power of 2 mW was used in all experiments. Optical excitation was carried out with an OPO system (Opta BBO-355-visible/IR) pumped by a Nd:YAG laser (Spectra Physics GCR-11) at a wavelength of 460 nm, a pulse width of ∼6 ns, and a pulse energy of 4 mJ. The repetition rate of the laser was set to 1.25 Hz. A transient recorder (Tektronix TDS520A) with a digitizing rate of 2 ns (resolution 8 bit) was used to acquire the time-dependent EPR signal. To eliminate the background signals from the laser, TREPR signals were accumulated at off-resonance magnetic field positions (background) and subtracted from those recorded on-resonance.
TREPR experiments of the XCry Y50F/W400F and W324F/W400F double mutants were performed at 270 K using a commercial EPR spectrometer (Bruker ESP380E) in conjunction with a Bruker microwave bridge (ER 046 MRT). The protein samples were placed in a synthetic quartz (Suprasil) sample tube (1 mm inner diameter) and irradiated in a dielectric ring resonator (Bruker ER 4118X-MD5), which was immersed in a helium gas-flow cryostat (Oxford CF-935). The temperature was regulated to ±0.1 K by a temperature controller (Oxford ITC-503S). A microwave frequency counter (Hewlett Packard HP 5352B) was used to monitor the microwave frequency. Identical microwave power of 2 mW was used in all experiments. Optical excitation was carried out with an OPO system (Opta BBO-355-visible/IR) pumped by a Nd:YAG laser (Spectra Physics Quanta Ray GCR 190-10) at a wavelength of 460 nm, a pulse width of ∼6 ns, and a pulse energy of 4 mJ. The repetition rate of the laser was set to 1 Hz. A transient recorder (LeCroy 9354A) with a digitizing rate of 2 ns (resolution 11 bit) was used to acquire the time-dependent EPR signal. To eliminate the background signals from the laser, TREPR signals, accumulated at off-resonance magnetic field positions were subtracted from those recorded on-resonance.
RESULTS
Optical Spectroscopy
To monitor the photoreduction and reoxidation behavior of XCry, UV-visible spectra were measured at 280 K from the dark-adapted system and immediately after different periods of blue-light illumination. Fig. 1A shows the blue-light-induced spectral changes in the optical absorptions of WT XCry in the presence of 10 mm EDTA under anaerobic conditions. The strong absorption with a peak maximum at 382 nm is assigned to the second cofactor, MTHF (14). Because the MTHF absorption overlaps with that of FADox (S0 → S2 transition), only one FADox absorption peak at around 450 nm (S0 → S1 transition) together with one of the vibrational side bands at 470 nm is clearly visible. One major difference in the photoreduction of WT XCry, compared with photoreduction of other Crys and PLs is the lack of significant radical absorption (from FADH• and/or FAD•̄) (11, 39, 51,–,53). The FADox absorption of WT XCry decreased with increasing illumination, however, the expected concomitant growth of a broad absorption between 500 and 650 nm (indicative of FADH•) and/or an absorption at around 400 nm and broad absorption above 500 nm (indicative of FAD•̄) were hardly observed. Thus, neither FADH• nor FAD•̄ accumulated sufficiently during photoreduction for unambiguous detection by conventional steady-state UV-visible absorption spectroscopy.
Blue-light induced spectral changes in the optical absorptions of WT and mutant XCrys. All measurements were performed in the presence of 10 mm EDTA (280 K). A, spectral changes in the optical absorptions of WT XCry after the indicated illumination times under anaerobic conditions. B, spectral changes in the reoxidation process by atmospheric oxygen in the dark. C, representative spectral changes in the optical absorption of XCry W400F after the indicated illumination times under anaerobic conditions. D, comparison of the photoreduction speed at 480 nm of WT XCry, W324F, W377F, and W400F performed under anaerobic conditions. Fits (black lines) were calculated using the reaction scheme described under “Experimental Procedures.”
Fig. 1B shows the spectral changes during recovery of FADox, after the sample was exposed to atmospheric oxygen in the dark. In this reoxidation process, the typical broad band between 500 and 650 nm indicative of FADH• could be clearly observed. Although spectral changes in the region where MTHF absorbs were also observed during photoreduction, these are most likely due to overlapping absorption changes of the flavin. After reoxidation was completed, the MTHF signal at about 380 nm exhibited almost full recovery (only 4% below initial absorption). Therefore, despite the long illumination time, significant sample damage could be excluded.
To elucidate the importance of the Trp triad in XCry photoreduction, the three Trp point mutants, W324F, W377F, and W400F, were photoreduced with 10 mm EDTA under anaerobic conditions, and the results were compared with those from WT XCry (see Fig. 1D). Comparison of UV-visible spectra from the mutant proteins with that of the WT protein reveals that all mutants investigated here appear to fold properly and bind FAD and MTHF similar to the WT (supplemental Fig. S1).
As compared with the WT, all examined mutants showed significantly decreased photoreduction rates and accumulations of the neutral FAD radical as is obvious from its characteristic absorption bands between 500 and 650 nm (Fig. 1C). Thus photoreduction is hampered by replacing any one of the three conserved Trps of the triad. Because EDTA is known to not change its reducing capacity over time, changes in the sample condition can be excluded even for the extremely long illumination times required for photoreduction of some of the mutants.
To extract quantitative photoreaction kinetics, reactions for all XCry samples were analyzed individually, assuming three redox states and a first-order two-stage reversible reaction scheme (cf. “Experimental Procedures”). Moreover, the reduction steps from FADox to FADH• and from FADH• to FADH− (k1 and k2, respectively) are light-inducible, whereas both reverse reactions (k−1 and k−2) do not depend on light but on the oxygen concentration in the sample, which is negligible under anaerobic conditions.
Photoreduction kinetics of the WT and all mutants under anaerobic conditions are compared in Fig. 1D. Calculated photoreduction rates and losses of MTHF after complete reoxidation are summarized in Table 1. MTHF losses after complete reoxidation increased with the very long illumination times required for some mutants (for example, see W377F in Table 1), thus effects due to MTHF loss in these mutants cannot be completely ruled out.
Analysis of the blue-light illumination kinetics for WT XCry and its mutants
Transient Optical Spectroscopy
The WT XCry protein was excited with a nanosecond laser pulse at 460 nm, and transient absorbance changes were monitored over a wavelength range from 370 to 700 nm for up to 10 μs, with a spectral resolution of 4 nm (see selected spectra and time traces in Fig. 2, A and B, respectively). Qualitatively, the spectrum can be divided into different parts and assigned accordingly. The negative band at about 450 nm is assigned to FADox ground-state bleaching. Positive difference bands detected at 375–415 and 500–700 nm are attributed to (i) excited flavin triplet states and (ii) RPs comprising a flavin and an amino acid radical. Difference absorbance (Fig. 2B) at ∼582 nm decreased within the first 2 μs, whereas difference absorbance at 500–550 nm changed more slowly.
Transient optical spectroscopy of WT XCry. A, difference absorption spectra of WT XCry in the wavelength region between 370 and 700 nm recorded at the indicated times after pulsed laser excitation. Each spectrum is the average of 10 time points (corresponding to an integration window of 50 ns for spectra taken at 150 ns and 2 μs and a window of 5 μs for spectra taken at 20, 100, and 200 μs). B, selected transients (410, 450, 510, and 582 nm) of WT XCry. Each time profile is the average of two acquisitions recorded with a shot repetition rate of 0.016 Hz. All spectra were recorded at 279 K. C, species-associated difference spectrum components of WT XCry corresponding to the kinetic model shown in D. For further details, see text.
By analysis and comparison with the reported spectra (8, 9, 54, 55) we attribute this changing spectral pattern to an initially generated tryptophanyl radical cation and a flavin radical. Only minute signals are observed from the excited triplet state of FAD, for which a broad absorption peaking around 650 nm is typically expected (56). Therefore, we performed a global analysis of our data in terms of a kinetic scheme with sequentially interconverting species (Fig. 2D), in which each species is characterized by a species-associated difference spectra (57). Two components were required for an adequate description of the time-resolved data, with lifetimes of 220 ± 3 ns and 52.4 ± 0.6 μs. The resulting species-associated difference spectrum are presented in Fig. 2C. The first species-associated difference spectrum (blue curve) has a broad maximum at 610 nm and can thus be assigned to the absorption of a Trp•+ radical species (8, 55) with a rate constant of kA = 4.54 ± 0.07 × 106 s−1. This tryptophanyl cation radical deprotonates, most likely involving a solvent water molecule, forming a neutral radical, which together with the FAD anion radical constitutes the secondary RP. This secondary RP decays with a rate constant of kB = 1.91 ± 0.02 × 104 s−1(see Fig. 2D). Neither protonation of the FAD anion radical (to form the neutral FADH• radical) nor any other species are observed within a millisecond.
It has to be noted, however, that the kinetic scheme illustrated in Fig. 2D is different from the chemically correct reaction scheme starting with the primary radical pair FAD•−···Trp•+ (8, 58). The scheme represents the model used to fit the data and, as global analysis is only sensitive to changes of species, cannot account for the non-changing FAD•̄ species (57).
TA spectroscopy was also applied to the three XCry single point mutants under otherwise identical conditions (see supplemental Fig. S2). Whereas XCry W324F showed virtually no TA signal, transient spectra obtained with XCry W400F and W377F samples basically resembled those of the WT protein (Fig. 2B), showing FAD ground state bleaching at 450 nm and buildup of absorption in the 400- and 500–600-nm regions. However, the decreased stability of the two mutant samples in combination with reduced quantum yield for ET (see also below) resulted in significantly reduced signal-to-noise ratios. Therefore, an unambiguous global analysis of the data of these mutants was not possible, and consequently, the type of amino acid involved in photo-induced ET could not be clarified by means of transient optical spectroscopy. Unfortunately, this includes identification of a Tyr radical, which typically absorbs in a region around 410 nm (59), where also the FAD anion radical shows optical transitions.
Given the identical laser excitation conditions (excitation wavelength, pulse energy, and pulse repetition rate) in all cases, however, the relative quantum yields could be extracted from the TA data depicted in Fig. 2 and supplemental Fig. S2 by comparing the relative absorption changes of the signals at 450 nm. The absorption of WT XCry was normalized to 100 ± 1.6%. The values for XCry W377F, W400F, and W324F were calculated as: 7 ± 2, 17 ± 3, and 0.2 ± 3.1%, respectively (see Table 1). Thus, the qualitative results obtained from steady-state spectroscopy could be further substantiated. The proximal Trp mutant (W400F) shows only a moderate decrease in quantum yield, whereas the ET rate of the middle and the terminal Trp mutants (W377F and W324F) are 14- and 500-fold decreased, respectively.
TREPR Spectroscopic Data
TREPR spectroscopy allows real-time observation of, e.g. short-lived RP and triplet states (47), generated by pulsed laser excitation. In contrast to conventional continuous wave EPR spectroscopy, which usually involves magnetic field modulation and lock-in amplification to improve the signal-to-noise ratio, TREPR data are recorded in a high bandwidth direct-detection mode, so as not to constrain the time resolution of the experiment. Consequently, positive and negative signal amplitudes in TREPR correspond to enhanced absorptive (A) and emissive (E) electron-spin polarization of the EPR transitions, respectively.
Upon blue-light photoexcitation, WT XCry with the FAD cofactor initially being in its fully oxidized redox state, exhibits a spin-polarized paramagnetic species at ambient temperatures, which we assigned to a RP based on the spectral shape and the narrow width of the signal (Fig. 3A) (60, 61). The RP signatures are not observed in the W324F mutant consistent with observation from TA spectroscopy (see Fig. 3B and supplemental Fig. S2). The disappearance of the RP TREPR signature in the W324F mutant, together with conclusions drawn from spectral simulations, led us to assign the TREPR signal of WT XCry to the RP state FAD•···Trp-324• (10).
TREPR signals generated by pulsed laser excitation (460 nm; 1.25/1 Hz pulse repetition rate; 4 mJ pulse energy) of XCry (A) WT, (B) W324F, (C) W400F and W377F (upper and lower green trace, respectively), (D) Y50F/W400F, and (E) W324F/W400F, recorded at 274/270 K centered at 500 ns after the laser pulse with an integration window of 200 ns in direct detection mode (integrated amplitudes with A, enhanced absorption, and E, emission). Please note that spectra A and B were taken from Ref. 10 and mutants Y50F/W400F and W324F/W400F were measured with a different setup and thus, their signal strength cannot be directly compared with that of the other XCry samples. Instrument settings: 9.69 GHz; microwave power: 2 mW. Each data point represents the average of 60 acquisitions (100 for Y50F/W400F and W324F/W400F, respectively) recorded with a detection bandwidth of 100 MHz (25 MHz for Y50F/W400F and W324F/W400F, respectively). The dashed gray curves show spectral simulations for the FAD•···Tyr-50• RP made with the following parameters: gFAD = (2.00431, 2.00360, and 2.00217); gTyr-50 = (2.00767, 2.00438, and 2.00219); dipolar interaction, D = −0.51 mT; exchange interaction, J = +0.24 mT. For details, see supplemental data and supplemental Table S2.
In this article we investigated the role of the remaining two residues of the conserved Trp triad, Trp-377 and Trp-400 (see Fig. 3C). In contrast to the terminal W324F mutant, following pulsed laser excitation both mutant proteins exhibit a TREPR signal comparable with that of the WT (Fig. 3A), but revealing some significant differences upon closer inspection. Due to the overall spectral shape and width, the signals can be assigned to a transiently formed spin-polarized RP. Compared with the WT signal, however, the signals of W400F and W377F are clearly shifted toward lower magnetic fields; thus a paramagnetic species with a g-value larger than that of a Trp radical forms a RP together with the FAD. Among the redox-active amino acids, only Tyr is known to form a radical with g-principal values larger than those of a Trp radical (62). Therefore, assigning the TREPR signal of the two mutant proteins to a FAD•···Tyr• RP state (which is identical in both cases) seems evident.
A structural model of XCry based on the SCry homolog (10) indicates that two Tyr residues, Tyr-50 and Tyr-397, bridge between FAD and the protein surface (see below and Fig. 4A) and are in close distance to the conserved Trp triad, thus providing a probable alternative ET pathway. The edge-to-edge distances between FAD and Tyr-397, Tyr-397 and Tyr-50, and FAD and Tyr-50, are 6, 8, and 14 Å, respectively. Therefore, surface-exposed Tyr-50 could mediate ET to exogenous electron donors, as proposed for the terminal Trp, Trp-324, of the Trp triad in the WT.
Proposed electron transfer pathways in WT and mutant XCry. A, selected residues of the XCry model structure showing the FAD cofactor (yellow) and the conserved Trp triad (blue). The proposed alternative ET pathway along Tyr-397 and Tyr-50 is shown in green. B to E, schematic representation showing the proposed ET pathways for each of the investigated XCry mutants (dashed pathway in panel C is proposed but could not be observed experimentally). Please note that the ET pathway in panel E remains highly ambiguous. For further details, see text.
To test this potential second ET pathway, we performed TREPR spectral simulations based on the correlated coupled RP mechanism (63,–,65), assuming that the RP in XCry is generated in a spin-correlated fashion from a singlet state precursor (10, 66). In a first approach, all parameters affecting the TREPR spectra (g-tensors, dipolar and exchange spin-spin couplings, hyperfine-governed and residual line width parameters) were taken from simulations of the WT TREPR data, except for the g-tensor components of the amino acid radical, where principal values for Tyr rather than Trp were used (supplemental Fig. S4 and supplemental Table S1). The results show a Tyr-based RP, FAD•···Tyr•. This TREPR signal is shifted to lower magnetic field values relative to a Trp-based RP, with the extent of the field shift depending on the orientation of the amino acid radical relative to FAD•.
For an unambiguous assignment of the TREPR signal to FAD•···Tyr-50• we performed additional simulations based on the following assumptions: RP formation starts with ET from Tyr-397 to FAD in its excited singlet state, thus forming the intermediate RP FAD•···Tyr-397•. In a subsequent ET step from Tyr-50 to Tyr-397, the terminal RP state FAD•···Tyr-50• is generated. By analogy to ET along the conserved Trp triad in the WT, the sequential ET steps in the mutant proteins should also be completed before significant singlet to triplet interconversion takes place, and too fast for resolution of the intermediate FAD•···Tyr-397• RP state by our TREPR instrumentation (time resolution of >200 ns). Therefore, the detected TREPR signal is assigned to the final RP state, which is initially also singlet configured. Recently, TREPR measurements of the RP formed in WT XCry performed at two microwave frequencies (X-band, 9.7 GHz, and Q-band, 34 GHz) together with spectral simulations have been reported, presenting clear evidence for the FAD•···Trp-324• RP originating from a pure singlet state precursor (66). In light of the obvious similarities among the TREPR signals of WT XCry and W400F and W377F mutants in terms of their overall polarization pattern, this assumption seems to be justified as well for simulating the TREPR spectra of these two mutant proteins. The calculation includes characteristic g-tensor principal values and principal axes orientations of a flavin radical and a Tyr radical, as well as the geometry of the radicals with respect to each other, as derived from a calculated model structure of XCry. The strength of the dipolar interaction (DRP = −0.51 mT) between the unpaired electron spins was derived from the point dipole approximation, by using a distance of r = 17.6 Å between the centers of maximum spin density in FAD and Tyr-50 (C4a and C4, respectively). Assuming characteristic EPR line shapes consistent with the typical hyperfine structures of the two radicals, the exchange interaction, J, was the only variable parameter in the computations. The exchange interaction was slightly enhanced (J = +0.24 mT) as compared with the WT to account for the shorter distance between Tyr-50 and FAD versus Trp-324 and FAD. The remarkably good overall agreement between experimental and simulated TREPR spectra confirmed our signal assignment to the FAD•···Tyr-50• RP state (Fig. 3C, dashed gray lines).
Our results from the spectral simulations are further corroborated by spectroscopic investigation of two double mutants. In the case of the XCry Y50F/W400F mutant, in which Tyr-50 and Trp-400 were replaced by the redox-inert phenylalanine (Fig. 3D), steady-state photoreduction kinetics (see Table 1), performed under identical experimental conditions as compared with the other XCry samples, lead to decreased rate constants comparable with those of the W400F mutant, and virtually no TA signal (supplemental Fig. S3). In contrast to the absent TREPR signal for the W324F mutant, the Y50F/W400F double mutant exhibited a TREPR signal that again strongly resembles that of the WT and W400F and W377F mutants, albeit much weaker in signal amplitude. Closer inspection reveals that the TREPR spectrum of the double mutant (Fig. 3D) is clearly shifted toward higher magnetic fields as compared with the signals of the W400F and W377F mutants (Fig. 3C), and closely matches the signal position of the WT protein. This indicates FAD•···Trp• RP formation in the Y50F/W400F double mutant.
Examination of the XCry model structure (see supplemental data) points toward the terminal Trp-324 as being the most likely candidate for RP formation with FAD, as in the WT. The weak signal amplitude, however, limits further analyses and spectral simulations in the Y50F/W400F mutant.
Steady-state photoreduction kinetics of the second double mutant, XCry W324F/W400F, performed under identical experimental conditions as compared with the other XCry samples, lead again to decreased rate constants comparable with those of the W400F mutant (see Table 1). In contrast to the FAD•···Tyr• TREPR signal of the W400F mutant, however, the W324F/W400F double mutant exhibited virtually no TREPR or TA signal (Fig. 3E and supplemental Fig. S3).
This rather unexpected result (one would again presume to see a FAD•···Tyr• RP) can be best rationalized by taking into account that the mutation of two close amino acids (one of them being highly conserved) could have negative impact on the overall stability of the protein and furthermore, on stabilization of the radical pair (please note that with the absorption spectra depicted in supplemental Fig. S1, only the correct binding of the FAD and to a minor extent, the MTHF cofactor can be probed). This is further corroborated by the amount of MTHF loss during photoreduction/reoxidation that is in the range of 10% despite rather fast photoreduction kinetics (in contrast to the W377F mutant). To shed more light on this finding, we produced another double mutant, Y397F/W400F, which unfortunately did not incorporate FAD correctly and thus, escaped further spectroscopic investigations (data not shown).
DISCUSSION
Steady-state Photoreduction
The present study examines ET pathways and kinetics of XCry WT and a series of Trp mutants. Steady-state photoreduction kinetics were recorded under anaerobic conditions. Analysis of the data shows that WT XCry harbors FADox as a stable species in the dark, which is converted to FADH− via FADH• by blue-light illumination (Fig. 1A). FADH− reverts to FADox in the dark with half-lives in the range of days (Fig. 1B). To more precisely quantify differences in the photoreaction kinetics between WT XCry and mutant proteins, anaerobic photoreduction reactions were analyzed assuming the presence of the three FAD redox states and a kinetic scheme outlined previously (11, 67). The resulting rate constants k1 and k2 (Fig. 1D and Table 1) of the different samples are as follows: (i) WT XCry has an apparent rate constant k1 of 14.5 × 10−3 min−1 and a significant value for k2 of 1310 × 10−3 min−1. As a result, WT samples are readily reduced to FADH−; (ii) XCry W377F, W400F, W324F/W400F, and Y50F/W400F mutants, on the other hand, exhibit 2–8 times smaller k1 values and have negligible values for k2 relative to the WT protein. Correspondingly, they cannot be fully reduced even after extended illumination periods and remain in their FADH• state; (iii) XCry W324F exhibits an even smaller apparent rate constant k1 of 0.058 × 10−3 min−1 and a negligible value for k2. The value for k1 is further decreased when MTHF decomposition is taken into account (Table 1). Thus, different mutations in the Trp triad appear to have diverse effects on the photoreduction kinetics, and only the mutation of the terminal, surface-exposed Trp-324 leads to a 250-fold deceleration as compared with the WT protein.
Radical Pair Formation and Decay
Fast laser spectroscopy is used extensively to examine electron transfer dynamics and the kinetics of light-active molecules. With this method, we were able to identify the chemical nature of the two RPs in WT XCry and extract accurate rate constants for their appearance. Within the first microsecond, the primary cationic tryptophanyl radical deprotonates and produces the secondary, better stabilized RP, comprised of an FAD anion radical and a neutral tryptophanyl radical, which recombines on a time scale of tens of microseconds. TA experiments were repeated with the three Trp mutants, W324F, W377F, and W400F (see supplemental Fig. S2), but rigorous data analyses turned out to be more difficult and could not be successfully performed. The TA spectrum recorded for the W324F mutant showed no analyzable signal, indicating that no RP is stabilized within a time range of a few nanoseconds. Besides the apparent poor quality of the spectra, W377F shows enhanced flavin triplet-state formation (absorbing between 550 and 700 nm with a lifetime of τ = 3.0 ± 0.4 μs and a rate constant of kT = 0.33 ± 0.04 × 106 s−1) obscuring the spectra. To a minor extent, similar observations were made with the W400F mutant. This hinders an unambiguous interpretation of the data via spectral deconvolution.
Although we were not successful in ascertaining the identity of the electron donating amino acid in the mutants, we compared their first-order monoexponential recombination rates (at 510 nm) of the secondary RP with that of the WT, for which a value of 70 μs was obtained. Both mutant samples show significantly decreased RP lifetimes with values of 3.9 ± 0.9 and 4.0 ± 0.7 μs for W400F and W377F, respectively. These values are more than 10-fold shorter than that of the WT and give a first hint that the molecules forming the RP in WT XCry and in these two XCry mutants are different.
Identification of Alternative ET Paths
To further analyze the ET pathways in XCry and unravel alternative electron donating amino acids, we used TREPR spectroscopy. TREPR spectroscopy provides information on the nature and identity of transient spin-polarized paramagnetic states, such as the RPs formed after blue-light photoexcitation and subsequent ET steps. Hence, it is well suited to gain insight into intraprotein ET processes and participating amino acids. TREPR spectra are quite sensitive to the g-values of the paramagnetic species involved. As Tyr and Trp exhibit sufficiently different average g-values to allow for radical discrimination even at conventional X-band microwave frequencies (≈9.5 GHz), TREPR can be used to distinguish FAD•···Trp• from FAD•···Tyr• RPs (see Fig. 3 and supplemental Fig. S4 and supplemental Table S1). Based on the spectral shifts, the TREPR spectra of the XCry W377F and W400F mutants can be clearly assigned to FAD•···Tyr• RPs.
Inspection of the available model structure of XCry (10) suggests an alternative pathway for sequential ET between the FAD and the protein surface involving key Tyr residues Tyr-397 and Tyr-50 (Fig. 4). Surface-exposed Tyr-50 could mediate ET to exogenous electron donors, as proposed for the terminal Trp-324 of the conserved Trp triad (Fig. 5). The edge-to-edge distance from Tyr-50 on the surface to FAD is ∼14 Å, with bridging Tyr-397 being 8 Å apart from Tyr-50 and 6 Å from FAD. Besides these two, there are no other evident Trp or Tyr residues found in the surroundings of the FAD cofactor that could potentially participate in light-induced ET.
Surface accessibility of the aromatic amino acids potentially acting as terminal electron donors. Trp-324 is depicted in blue and Tyr-50 is depicted in green.
Recently, Krapf et al. (68) showed that the charge-separated state consisting of FAD and the terminal Trp of the conserved Trp triad has a lower lying free enthalpy than those involving the two other Trps, thus stabilizing the charge separation between FAD and the terminal Trp. This is most probably due to the terminal Trp being exposed to the protein surface. By analogy, surface-exposed Tyr-50 could stabilize the charge separation of the putative FAD•···Tyr-50• RP, if it donates an electron to Tyr-397 (Fig. 5). Charge separation could be further stabilized by deprotonating Tyr-50, as has been shown for the terminal Trp in PLs (8, 69).
The notion of Tyr-50 being the terminal electron donor in XCry W400F or W377F is further corroborated by spectral simulations performed using the geometry of the RP state comprising FAD• and Tyr-50• (Fig. 3C, dashed gray lines and supplemental Table S1). The overall agreement between experimental and simulated TREPR spectra is remarkably good, especially when taking into account that the only adjustable parameter in the simulations was the exchange interaction, J. Finally, the TREPR signal detected in the Y50F/W400F double mutant, which is different from the spectra detected from the W377F and W400F mutants, clearly shows that Tyr-50 plays an essential role in RP formation leading to the TREPR signal of the W400F (and the W377F) mutant. As mentioned above, Tyr-50 is located at the protein surface with no other Trp or Tyr residues nearby, and easily accessible for proton acceptors in the solvent that could stabilize the RP by deprotonation (Fig. 5). All these findings strongly support the role of Tyr-50 as the terminal electron donor for FAD in the W377F and W400F mutants.
Relative Quantum Yields of Radical Pair Formation
The TREPR signal of the W400F mutant is considerably stronger than that of W377F, as can be seen from a qualitative comparison of the signal-to-noise ratios in Fig. 3C. This difference can be rationalized by taking into account that in W400F ET from Trp-400 to FAD is blocked and therefore only ET from the further distant Tyr-397 to FAD is feasible. In W377F, however, competition between Trp-400 and Tyr-397 for ET to FAD is conceivable. After ET from Trp-400 to FAD, the formed RP will most likely recombine very rapidly due to (i) the short distance between Trp-400 and FAD, and (ii) the absence of any nearby alternative redox partner, from which Trp-400 could abstract an electron. Light-induced ET between the equivalent redox partners in Escherichia coli PL (FAD to Trp-382) is complete within a few picoseconds (70) (which is well below the time resolution of our TA instrumentation), suggesting a comparable ET rate constant in XCry. This interpretation is further supported by the quantum yields obtained from TA datasets (see also above). Whereas XCry W400F exhibits a 6-fold decrease with respect to the WT, the quantum yield is further decreased by a factor of 2.5 in XCry W377F.
Although the Y50F/W400F double mutant lacks both the proximal Trp of the conserved chain and the putative terminal electron donor, Tyr-50, of the alternative ET chain, a weak TREPR signal was observed for Y50F/W400F after blue-light illumination. This signal was assigned to a FAD•···Trp• RP due to its spectral position on the magnetic field axis (Fig. 3). Although we cannot unambiguously assign the signal to a specific Trp residue, it is tempting to speculate that here we see, once again, the signature of the FAD•···Trp-324• RP, as in the WT protein. From inspecting the model structure, there is no other nearby Trp residue to act as electron donor. The exact pathway of ET, however, remains highly ambiguous and could involve both Tyr and Trp residues as well as helix α15 that has been proposed earlier for PLs (71).
Amino acid sequence alignments and additional manual inspections of the available crystal structures and model structures for members from the Cry/PL protein family show that the proposed terminal electron donor (Tyr-50) of the alternative ET pathway is not conserved even within the DASH-type class of proteins (15, 72). Only Cry-1 from Vibrio cholerae has a redox active amino acid at the position equivalent to Tyr-50 in XCry. The only other nearby residue highly conserved within the DASH-type cryptochromes is Tyr-397 in XCry, which, together with another conserved Tyr (Tyr-391 in XCry), bridges the distance between FAD and the second cofactor, MTHF. These two tyrosines have been proposed to play a key role in energy transfer between the antenna pigment and FAD exemplified in other DASH-type Crys (73, 74). Besides that, the whole region surrounding XCry Tyr-50 seems not to be conserved in other Crys or PLs.
In light of the other ET pathway in XCry described here, one might wonder why the XCry W324F mutant has a drastically decreased photoreduction rate and produces no detectable TREPR signal (see Figs. 1 and 3) (10). A likely explanation, consistent with the much weaker RP signal for W377F as compared with W400F, is that ET in XCry W324F proceeds from FAD to Trp-377 and is followed by a fast charge recombination. This is in line with recent theoretical and experimental investigations of the equivalent ET in E. coli PL showing that the first two ET steps are nearly isoenergetic, whereas the third ET step leading to oxidation of the terminal Trp is clearly favored in terms of the free energy of the system, thus stabilizing charge separation (58, 68). In E. coli PL, the ET along the Trp triad is completed within 30 ps (75). Therefore, recombination from the second ET step back to the ground state should be on a similar time scale. In addition, the FAD ← Trp-400 ← Trp-377 ET should be favored over ET via Tyr-397 ← Tyr-50, because Trp-400 is much closer to FAD than Tyr-397 (with distances of 3.4 and 5.6 Å, respectively).
Although the redox potentials for Tyr and Trp oxidation at neutral pH values are in the same range (0.9–1.1 V) (76), further ET from Trp-377 to, e.g. Tyr-50 is highly unlikely. This is because of the much larger distance between Trp-377 and Tyr-50 (7.9 Å) compared with Trp-377 and Trp-400 (4.6 Å) leading to a 100-fold decreased ET rate according to the “Dutton ruler” (77). The same argument holds for ET between Trp-324 and Tyr-50 with a distance of 8.9 Å.
But why is the XCry W324F mutant still photoreduced according to steady-state optical spectroscopy, although there is no detectable TREPR signal? To answer this question, two major aspects between the two different techniques must be considered. (i) Photon fluency rates are incomparable for the two experiments. (ii) Both experiments cover different time scales. Whereas TREPR has a sub-microsecond time resolution, monitoring spin-polarized states far from thermal equilibrium, steady-state UV-visible spectroscopy as used here for the observation of photoreduction, operates on the time scale of seconds to hours.
Excited flavins are known to be very potent oxidizing agents (78) that can abstract electrons from their surroundings. Therefore, photoreduction of the flavin will occur as a side reaction after photoexcitation of the FAD cofactor, if the intermediate tryptophanyl cation radical is reduced before RP recombination. However, the quantum yields for these processes are far too low to be detectable by time-resolved spectroscopic methods such as TA or TREPR.
Tyr involvement in FAD photoreduction has been suggested in some PLs and Crys (8, 37, 69, 79), however, the respective ET pathways involving a Tyr remained unknown. In this work, we demonstrate the formation of FAD•···Tyr• RPs in XCry mutants, and additionally elucidate the identity of the residues. By mutagenesis and different spectroscopic analyses we characterized the residues involved in ET and RP formation, and revealed an alternative pathway besides the canonical pathway facilitated by the Trp triad. This alternative pathway in XCry may well explain why disruption of the Trp triad does not necessarily abolish FAD photoreduction in some proteins in vivo.
Conclusions
In this contribution we describe how molecular spectroscopy corroborated by spectral simulations leads to a more detailed picture of the various ET pathways and processes in XCry upon blue-light excitation. Investigation of the overall photoreduction capabilities of the WT and a number of XCry mutants by means of steady-state UV-visible spectroscopy reveals clear differences, whereas time-resolved spectroscopy provides further insight. TA unravels aspects of the photochemistry and the rate constants for the decay and transformations of the radicals. TREPR, due to its high spatial resolution, is capable of distinguishing between Trp-based and Tyr-based RPs and allows identification of the exact nature of the RP partners by spectral simulations. Therefore, it is well suited to complement the frequently applied TA spectroscopy to address the identity of RP partners. Here we present, for the first time, a full assignment of the different radical species occurring in different XCry mutants after light excitation.
XCry exhibits robust photoreduction capabilities provided by multiple (partly intertwined) ET pathways, giving new insights into the apparently contradicting results in the literature with respect to mutations within the Trp triad. FAD in its excited state proves to be a very potent electron acceptor that abstracts electrons from its surrounding even when its most favorable RP partner is deleted. Thus, our data provide a solid basis to consider the photoreduction of the flavin cofactor as being one important step in the photoreaction of this protein. Taken together with recent results from another protein (SCry), cryptochrome photochemistry appears to be far more complex and diverse than commonly anticipated.
Acknowledgments
We thank Chiharu Hitomi for assisting in protein mutagenesis, John A. Tainer for encouragement and Jim Norris for critical reading of our manuscript. We (T. B., S. W., and E. S.) thank Robert Bittl (FU Berlin) for the use of some EPR equipment and for helpful discussions. We (B. P., S. W., and E. S.) thank Thomas Berthold for help with the EPR measurements.
Footnotes
-
↵* This work was supported, in whole or in part, by National Institutes of Health Grant GM37684 (to E. D. G.), Deutsche Forschungsgemeinschaft Grants BI 1249/1-1 and BI 1249/1-2 (to T. B.), and a grant from the Skaggs Institute for Chemical Biology (to K. H.).
-
↵
This article contains supplemental Figs. S1–S4 and Tables S1 and S2.
-
↵4 The abbreviations used are:
- Cry
- cryptochrome
- DASH
- Drosophila Arabidopsis Synechocystis Human
- ET
- electron transfer
- MTHF
- 5,10-methenyltetrahydrofolyl polyglutamate
- PL
- photolyase
- RP
- radical pair
- SCry
- Synechocystis sp. cryptochrome DASH
- TA
- transient absorption
- TREPR
- time-resolved EPR
- XCry
- X. laevis cryptochrome DASH.
- Received September 9, 2012.
- Revision received February 15, 2013.
- © 2013 by The American Society for Biochemistry and Molecular Biology, Inc.
REFERENCES
- 1.↵
- 2.↵
- 3.↵
- 4.↵
- 5.↵
- 6.↵
- 7.↵
- 8.↵
- 9.↵
- 10.↵
- 11.↵
- 12.↵
- 13.↵
- 14.↵
- 15.↵
- 16.↵
- 17.↵
- 18.↵
- 19.↵
- 20.↵
- 21.↵
- 22.↵
- 23.↵
- 24.↵
- 25.↵
- 26.↵
- 27.↵
- 28.↵
- 29.↵
- 30.↵
- 31.↵
- 32.↵
- 33.↵
- 34.↵
- 35.↵
- 36.↵
- 37.↵
- 38.↵
- 39.↵
- 40.↵
- 41.↵
- 42.↵
- 43.↵
- 44.↵
- 45.↵
- 46.↵
- 47.↵
- 48.↵
- 49.↵
- 50.↵
- 51.↵
- 52.↵
- 53.↵
- 54.↵
- 55.↵
- 56.↵
- 57.↵
- 58.↵
- 59.↵
- 60.↵
- 61.↵
- 62.↵
- 63.↵
- 64.↵
- 65.↵
- 66.↵
- 67.↵
- 68.↵
- 69.↵
- 70.↵
- 71.↵
- 72.↵
- 73.↵
- 74.↵
- 75.↵
- 76.↵
- 77.↵
- 78.↵
- 79.↵
