PERSPECTIVE

Chemical magnetoreception in birds: The radical pair mechanism

Christopher T. Rodgersa,b and P. J. Horea,1 aDepartment of Chemistry, Physical and Theoretical Chemistry Laboratory, University of Oxford, South Parks Road, Oxford, OX1 3QZ, United Kingdom; and bOxford Centre for Clinical Magnetic Resonance Research, Level 0, John Radcliffe Hospital, Oxford, OX3 9DU, United Kingdom

Edited by Nicholas J. Turro, Columbia University, New York, NY, and approved December 2, 2008 (received for review September 24, 2008)

Migratory birds travel vast distances each year, finding their way by various means, including a remarkable ability to perceive the Earth’s magnetic field. Although it has been known for 40 years that birds possess a magnetic compass, avian magnetoreception is poorly understood at all levels from the primary biophysical detection events, signal transduction pathways and neurophysiology, to the processing of information in the . It has been proposed that the primary detector is a specialized ocular photoreceptor that plays host to magnetically sensitive photochemical reactions having radical pairs as fleeting intermediates. Here, we present a physi- cal chemist’s perspective on the ‘‘radical pair mechanism’’ of compass magnetoreception in birds. We outline the essential chemical requirements for detecting the direction of an Earth-strength Ϸ50 ␮T magnetic field and comment on the likelihood that these might be satisfied in a biologically plausible receptor. Our survey concludes with a discussion of , the photoactive that has been put forward as the magnetoreceptor molecule.

ould a chemical reaction be The radical pair mechanism [and a pair model, or to comment on (neuro)- used to detect the direction of few rare variants (10)] is currently the physiological aspects of magnetorecep- the Earth’s ? At only plausible way in which weak mag- tion, all of which have been covered first sight, most physical chem- netic fields can affect chemical reactiv- authoritatively and extensively in recent istsC would be skeptical. The energy of ity. Most experimental studies have reviews (3–5, 8, 24, 25). Nor do we pre- interaction of a molecule with a Ϸ50-␮T been of photochemically formed organic sume here to debate the relative merits magnetic field is Ͼ6 orders of magni- radicals in solution. The effects are usu- of the radical pair hypothesis versus pro- tude smaller than the average thermal ally not dramatic—typically Ͻ50% posals based on biogenic energy kBT, which in turn is 10–100 change in reaction yields or radical life- (Fe3O4) (26–32). We close with a dis- times smaller than the strength of a times and, with few exceptions (14), are cussion of cryptochrome, the photore- chemical bond. It therefore seems in- observed for magnetic field intensities ceptor protein that is presently the only conceivable that the position of a (1 mT–10 T) considerably stronger than molecule under consideration as the chemical equilibrium or the rate of an the Earth’s (25–65 ␮T). Compared with magnetoreceptor (2), and make sugges- activated chemical reaction could be these chemical studies, there are rela- tions for further research. We hope that significantly altered by such a minuscule tively few reliable examples of magnetic the thoughts set out here will be of in- perturbation. Nevertheless, there is field effects on biological processes, the terest to both physical and biological growing evidence that the striking abil- most notable by far being photosynthetic scientists who may be inspired to con- ity of birds to detect the direction of the reaction center where light ab- tribute to the elucidation of this intrigu- geomagnetic field (1) is based on a sorption leads to radical pair formation ing and incompletely understood chemical reaction whose product yields by sequential electron transfer steps sensory mechanism. depend on the orientation of the reac- (reviewed in refs. 15–17). tants within the field (2) (for reviews, Radical pair reactions were first pro- Key Features of a Radical Pair see also refs. 3–8). posed as a magnetoreceptor by Schulten Magnetoreceptor It has been known since the 1970s (18, 19), prompted by the findings that Chemistry. Magnetically sensitive reac- that certain chemical reactions do in avian magnetic orientation is light- tions almost always involve radicals, i.e., fact respond to applied magnetic fields. dependent—suggesting a photochemical molecules that have an odd number of The key species are radical pairs—pairs process—and is responsive to the incli- electrons and consequently an unpaired of transient radicals created simulta- nation rather than the polarity of the electron spin that may be found in one neously, such that the 2 electron spins, geomagnetic field (20–23); unlike a of 2 spin states: 1 or 2. A radical pair one on each radical, are correlated. magnetic compass needle, birds main- is a short-lived reaction intermediate They have the unique properties that tain their orientation when the ambient comprising 2 radicals formed in tandem their chemical fate is largely controlled field vector is reversed (5). whose unpaired electron spins may be ϽϽ by weak ( kBT) magnetic interactions Our aim here is to survey the radical either antiparallel (12, a singlet state, via their spin correlation and that the pair mechanism in the context of the S) or parallel (11, a triplet state, T). electron spins remain far enough from avian magnetic compass. We identify As each electron spin has an associated thermal equilibrium for long enough the key features required for efficient that the kBT objection is irrelevant. Ex- magnetoreception, comment on the like- perimental and theoretical studies of the lihood that they could occur in a biolog- Author contributions: C.T.R. and P.J.H. wrote the paper. ‘‘radical pair mechanism’’ over the last ically plausible receptor, and evaluate The authors declare no conflict of interest. 30 years have given an extraordinary the evidence for and against the radical This article is a PNAS Direct Submission. variety of information on the magnetic pair hypothesis. Throughout, the focus is 1To whom correspondence should be addressed. E-mail: properties, kinetics, and dynamics of firmly on the primary reception mecha- [email protected]. radicals and their reactions (reviewed in nism: We make no significant attempt This article contains supporting information online at www. refs. 9–13). This field of endeavor has to review the behavioral studies that led pnas.org/cgi/content/full/0711968106/DCSupplemental. come to be called ‘‘spin chemistry.’’ to and have since supported the radical © 2009 by The National Academy of Sciences of the USA

www.pnas.org͞cgi͞doi͞10.1073͞pnas.0711968106 PNAS ͉ January 13, 2009 ͉ vol. 106 ͉ no. 2 ͉ 353–360 Downloaded by guest on September 24, 2021 S[A•+ B•− ] T[A•+ B•− ] steps (33). These and other possibilities f are described in the supporting informa- 3 kT tion (SI) Appendix together with com- k f 4 ∗ S ments on the nature of the reactions A B b C kS that could lead to C and subsequently 1 2 convert it back to A B. 5 A well-studied precedent for this kind A B of magnetically sensitive radical pair chemistry is provided by the initial Fig. 1. A simple reaction scheme that could form charge separation steps of bacterial pho- the basis of a compass magnetoreceptor. The spin- tosynthetic energy conversion, which correlated radical pair is depicted in red for the proceed via a series of radical ion pairs singlet state and blue for the triplet. See Key Fea- tures of a Radical Pair Magnetoreceptor: Chemis- formed by sequential electron transfers try for further details. along a chain of immobilized chlorophyll and quinone cofactors in a reaction cen- ter protein complex (15–17). Provided , the interconversion subsequent forward electron transfer is and chemical fates of the S and T states blocked, the recombination of the pri- can be influenced by internal and exter- mary radical pair responds to magnetic nal magnetic fields. fields in excess of Ϸ1 mT. In unblocked In chemical terms, the minimum re- reaction centers, spin correlation can be quirement for a radical pair reaction to transferred along the electron transport Fig. 2. Quantum mechanical spin dynamics sim- be sensitive to an external magnetic chain from the primary to the secondary ulations for the reaction in Fig. 1, performed as field is that at least one of the S and T radical pair (34–36), whose lifetime is described in ref. 60. (A) In the absence of the mag- states undergoes a reaction that is not also magnetically sensitive (37–39). Simi- netic field and with no recombination (green), the open to the other, usually as a conse- lar effects occur in plant photosystems. fraction of radical pairs that exist in the triplet quence of the imperative to conserve Other biological radical pair processes state, pT(t), oscillates at the frequency of the hy- perfine coupling (here 14 MHz). (B) When a weak spin angular momentum. Fig. 1 shows are much less well characterized (see SI Appendix). magnetic field is introduced (green), pT(t) shows an what is probably the simplest reaction additional, slower, modulation at the frequency of scheme possessing the essential chemical the Zeeman interaction (here 1.4 MHz). The radical Hyperfine Interactions. We now turn to features required to form the basis of a pair reactions cause pT(t) to be exponentially magnetic compass. Imagining for now the essential magnetic properties of a damped (red) and allow the reaction product (spe- that the radicals are completely immo- radical pair magnetoreceptor, focusing cies C in Fig. 1) to accumulate (blue). The applied bile, the individual reaction steps are first on detecting the presence of an ex- magnetic field (50 ␮T) results in an increased tran- as follows. ternal magnetic field and then its direc- sient conversion of the radical pair into the triplet state, causing C to be formed more rapidly and in A and B in Fig. 1 could be portions tion. An absolute requirement is that the 2 radicals, between them, have at higher yield. Faster recombination than shown of the same molecule or distinct mole- least 1 hyperfine interaction, i.e., an in- here would allow scant time for the slow modula- cules held in close proximity by their traradical coupling between the mag- tion arising from the Zeeman interaction to alter surroundings (e.g., 2 cofactors or a co- pT(t); the yield of C would then be much less af- netic moment of the unpaired electron factor and an amino acid residue in a fected by the field. For details of the calculation, and the magnetic moment of an atomic protein). Species C is either the signal- see SI Appendix. nucleus such as 1Hor14N. Almost all ing state or leads to the signaling state biologically occurring radicals have sev- via subsequent chemical transformations eral potentially suitable hyperfine inter- magnetic moments of the 2 electrons (which are not shown in Fig. 1). An ap- actions. Ample data on hyperfine tensors causes additional periodic S 7 T inter- plied magnetic field can alter the yield is available from electron paramagnetic conversion. In a 50-␮T field, this oscilla- of C by regulating the competition be- resonance (EPR) spectroscopy (40, 41) Ϸ tween its formation (from the S and T tion has a frequency of 1.4 MHz and a and ab initio density functional period of Ϸ700 ns. The calculated time states, step 4) and the regeneration of calculations (42). A B (exclusively from the S state, step dependence for a prototype radical pair Hyperfine interactions are crucial be- is shown in Fig. 2. 2). If the interconversion of S and T is cause they drive the interconversion of hindered by the external magnetic field, To act as a compass, a radical pair the S and T states of the radical pair reaction must respond to the direction then less C will be produced and corre- and allow it to be modified by an exter- of the field and not just its intensity. spondingly more radical pairs recombine nal magnetic field. S 7 T interconver- This is unlikely to occur for solution- directly to A B. The opposite follows if sion is a coherent quantum mechanical phase reactions where rapid molecular the field enhances S 7 T interconver- process: Radical pairs oscillate between sion. It is important to note that the their S and T states at a variety of fre- tumbling tends to average any anisotro- external magnetic field is far too weak quencies determined by the strengths of pic responses. The radicals must, at least to initiate new radical reactions. the hyperfine interactions. 1H and 14N partially, be immobilized and oriented Variants on the reaction scheme in hyperfine couplings in organic radicals and possess appropriate anisotropic Fig. 1 are possible, and although the are typically in the range 10–1,000 ␮T, magnetic interactions. As originally pro- details of the chemistry may differ, the corresponding to frequencies of 300 kHz posed by Schulten (18), the most likely principles remain the same. Some of the to 30 MHz (the conversion factor is 28 source of anisotropy is the hyperfine more likely alternatives include: an ex- kHz ␮TϪ1). The time scale for signifi- interaction, i.e., the dependence of the cited triplet precursor state; electron cant transformation of S into T and vice electron–nuclear coupling on the orien- transfer in the reverse direction (to form versa is thus typically 10 ns–1 ␮s. Sensi- tation of the molecule in an external A•ϪB•ϩ); and formation of the radical tivity to an external magnetic field arises magnetic field. Almost every hyperfine pair via sequential electron transfer because the Zeeman interaction of the interaction has an anisotropic compo-

354 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0711968106 Rodgers and Hore Downloaded by guest on September 24, 2021 nent: As a consequence, S 7 T inter- AB CD conversion, and therefore the reaction z yield, should vary with the direction of z z the external field. This suggestion has z been confirmed by numerical simula- tions (2, 33, 43–45). y y y y Anisotropic magnetic field effects x arising from hyperfine interactions have x x recently been observed experimentally x (14). Somewhat related observations at higher magnetic fields, for radical pairs in solids (46, 47) and in viscous liquids (48–50), appear to have their origins in electron–electron dipolar interactions and g-tensor anisotropy. Additionally, time-resolved EPR (51–53) and solid state NMR (54) spectroscopies provide clear evidence for the involvement of A B C=D anisotropic hyperfine interactions in the spin dynamics of photosynthetic radical Fig. 3. Spin dynamics simulations of anisotropic reaction yields for model radical pairs performed as described in refs. 58 and 66 using the reaction scheme of Fig. 1. (Upper) Polar plots. (Lower) The pairs. corresponding signal modulation patterns for a bird looking directly along the Earth’s magnetic field Ritz et al. (2) have suggested that the vector. The heights of the vertical scale bars in the upper images correspond to singlet yields of 2% (black) cells responsible for light-dependent or 0.2% (red). (A and B) The simulations demonstrate that a relatively simple orientation dependence of magnetoreception are distributed the reaction yield (A) can be obtained from radical pairs containing a small number of hyperfine around and aligned within the retina, in interactions or from more complex radicals when a few symmetry-related hyperfine interactions domi- the manner of the visual rod and cone nate (B). (C and D) More intricate anisotropy patterns (C) can be dramatically simplified if the radical pairs cells. In turn, the magnetoreceptor mol- are axially rotationally disordered (D). The signal modulation pattern for C is identical to that for D and ecules would be oriented within the re- is only shown once. Note that in all cases the reaction yield is invariant to exact reversal of the magnetic ceptor cells perhaps by attachment to field vector, i.e., the response is that of an inclination compass rather than a polarity compass. See SI Appendix for details of the calculations. cytoskeletal proteins (55). Thus, cells at different retinal locations would make different angles to the Earth’s magnetic This condition is met if the spin correla- achieve a high quantum yield of radical field vector and so respond differently tion of the 2 electrons persists for more pairs, a plausible lower limit on the for- according to the anisotropy of the reac- than Ϸ1 ␮s, a requirement that con- ward electron transfer rate would be tions of the radical pairs within them. 9 Ϫ1 strains both the reaction kinetics and Ϸ10 s , implying re Յ 1.0 nm for each The transduction of the magnetic com- the molecular dynamics. The rate con- of the electron donor–acceptor pairs pass information may ‘‘piggy-back’’ the b ϩ f f stants kS kS and kT (Fig. 1) must involved in the formation of the magne- visual reception pathway (2, 56), so that therefore, ideally, be Ϸ106 sϪ1 and fairly toreceptive radical pair. The derivation the bird literally sees the magnetic field similar to one another (say within a of these constraints is given in the SI as a ‘‘signal (or visual) modulation pat- factor of 10) to allow significant compe- Appendix. tern,’’ reflecting the anisotropy of the tition between the S and T reaction radical pair reaction, and perhaps remi- pathways. Such values are certainly Geometry. The separation of the 2 radi- niscent of a ‘‘heads-up’’ display in an achievable, e.g., in electron transfer pro- cals not only has a strong influence on aircraft (2, 57). teins where rates can be tuned over electron transfer rates but also deter- Fig. 3 shows the results of a few spin dynamics simulations of reaction yield many orders of magnitude via the reac- mines the strength of the (exchange and anisotropies and the corresponding sig- tion Gibbs energy, the reorganization dipolar) spin–spin interactions between nal modulation patterns. Unless the rad- energy and, especially, the donor– the 2 unpaired electrons (63–65). The acceptor distance (61). The change in interconversion of S and T states by a icals contain very few magnetic nuclei Ϸ ␮ or possess some degree of molecular the reaction yield produced by a 50- T magnetic field is only possible Ϸ ␮ symmetry or are favorably disordered, 50- T field grows sigmoidally as a when the interradical interactions are ␮ the shape of the anisotropy can be com- function of the radical pair lifetime, lev- not much stronger than 50 T. Given Ϸ ␮ plex. There is no clear picture of what eling out at 1 s (see SI Appendix). that the dipolar coupling has a magni- ␮ would constitute the optimum sensory There is therefore no advantage, and tude of 500 T at a radical–radical input for the bird; however, it seems probably serious disadvantages (as a re- center-to-center separation rc of 1.8 nm ␮ ϭ reasonable to suppose that strongly sult of spin relaxation, see below), in and only falls to 50 Tatrc 3.8 nm anisotropic but relatively simple direc- having a radical pair lifetime much in (58), it is not immediately obvious how tional information would be favored. excess of 1 ␮s. this requirement can be reconciled with Simulations and experiments on solution- Approximate structural constraints the much smaller separations just esti- phase reactions suggest a few simple can be derived from the simple theory mated from electron transfer rates. The- design features discussed in the section of electron transfer reactions (61, 62). ory suggests that the inhibitory effects on (43, 58, 59). First, for the lifetime of the radical pair of the exchange and dipolar interactions to be Ϸ1 ␮s (to allow enough time for may partially cancel for immobilized Kinetics. The principal kinetic require- S 7 T interconversion, without exces- protein-bound radicals when rc ϭ 2.0 Ϯ ment for a sensitive magnetoreceptor is sive spin relaxation), the edge-to-edge 0.2 nm (58). Such a center-to-center that the geomagnetic field has sufficient separation re of the 2 radicals in the separation would be compatible with the time to modulate the S 7 T intercon- magnetically sensitive radical pair should above constraint on the edge-to-edge version of the radical pair (Fig. 2) (60). be less than Ϸ1.5 nm. Second, to separation (re Յ 1.5 nm) provided the

Rodgers and Hore PNAS ͉ January 13, 2009 ͉ vol. 106 ͉ no. 2 ͉ 355 Downloaded by guest on September 24, 2021 radicals had diameters of more than enough to host a suitable radical pair quency arising from hyperfine and/or Ϸ0.5 nm. The relative orientation of the and would anyway lead to efficient Zeeman interactions (70–72). Such reso- radicals within the pair may also be im- averaging of anisotropic magnetic inter- nances are diagnostic for radical pairs: portant; simulations indicate that align- actions. Examples of slowly relaxing Magnetite particles of sufficient size to ment of anisotropic hyperfine interac- spin-correlated radical pairs at physio- function as a compass are too large to tions could be a factor in optimizing the logical temperatures are mentioned reorient in magnetic fields fluctuating magnitude and shape of the reaction- briefly in the SI Appendix. as rapidly as 1 MHz (73, 74). yield anisotropy (43, 66). The experi- In the light of the above, we may con- Linearly polarized radio frequency mental evidence that the effects of weak clude that the ordering and motional fields 100 times weaker than the Earth’s magnetic fields are quenched by strong constraints can be simultaneously satis- field (Ϸ500 nT), with frequencies of 7.0 interradical interactions is reviewed fied if the radical pairs are uniaxially MHz or 1.315 MHz, are sufficient to briefly in the SI Appendix. ordered with a high-order parameter disrupt the migratory orientation of and if all relevant motions are slower caged European robins (75, 76). More- Disorder and Motion. As mentioned above, than Ϸ1 ␮s. over, disorientation is found when the a critical requirement for a functional angle between the Earth’s field and the compass is that the radical pairs must Evidence for Radical Pair radio frequency field is 24° or 48°, but be aligned within the host receptor cells Magnetoreception not when the 2 fields are collinear. The which, in turn, must be spatially ordered We have presented a rather long list of birds are also disoriented by a broad- within the bird’s body (2). We focus here magnetic, structural, kinetic, and dy- band (0.1–10 MHz) 85-nT radio fre- on the former. Rotational disorder would namic requirements that must be ful- quency field (75). These remarkable ob- cause the anisotropic responses of differ- filled if a radical pair is to operate as a servations appear to exclude a number ently oriented radical pairs to cancel one compass sensor. Using well-founded the- of the more obvious artifacts: it seems another, potentially reducing the direc- ory and experimental precedents, we unlikely, for example, that the motiva- tional sensitivity of the receptor. However, have argued that all these conditions are tion of freely moving birds would be perfect ordering, as in a crystal, is not individually plausible, both chemically disturbed to different extents after rota- necessary. The compass function would and physically. The crucial question is tion of the radio frequency field axis. be preserved if the magnetoreceptor mol- whether these constraints can be satis- That the radio frequency field appears ecules tended to be aligned along a com- fied simultaneously in a biologically to have little effect on the birds’ orien- mon cell-fixed axis (the ‘‘director’’) and credible magnetoreceptor. We discuss tation when it is parallel to the Earth’s had a reasonably large order parameter. first the extent to which the available field can be understood if one of the Such alignment would largely preserve evidence supports a radical pair mecha- radicals has (effectively) no hyperfine components of the magnetic response that nism and then evaluate how well the interactions (14, 66). Bearing in mind are cylindrically symmetric with respect to ‘‘cryptochrome hypothesis’’ stands up that the periodic S 7 T interconversion the director axis, and cancel all others, so against the various criteria. induced by a Ϸ500-nT Zeeman interac- simplifying the orientational dependence Much of the evidence in favor of a tion has a period of Ϸ70 ␮s, sensitivity of the reaction yield and resulting, per- photochemical radical pair magnetore- to such a feeble radio frequency field haps, in a more efficient compass (Fig. 3). ceptor is circumstantial. Schulten’s implies either a surprisingly long-lived This point is illustrated by a recent study original proposal (18) was inspired by spin-correlation or highly efficient signal of an aligned radical pair whose lifetime 2 behavioral features of the avian mag- transduction or both. In a very recent was found to have a rather straightfor- netic compass: the requirement for light study, European robins exposed to radio ward dependence on the orientation of at wavelengths below Ϸ565 nm (23) and frequency fields of various intensities the molecule in an applied magnetic field, the invariance to reversal of the field and frequencies showed an extraordinar- as a consequence of static orientational direction (21). However, it is conceiv- ily sensitive response at 1.315 MHz in averaging around the long axis of the rod- able that light has an entirely different the local geomagnetic field (46 ␮T), like molecule (14). physiological effect: Rather than a direct which shifted to 2.630 MHz when the There are also strict constraints on involvement in magnetoreception, it birds had been preconditioned to a the dynamics of the radical pair. First, could possibly affect the birds’ motiva- 92-␮T field (77). These resonances, at rapid molecular rotation tends to aver- tion to act upon magnetic compass in- the frequency of the electron Zeeman age anisotropic magnetic interactions. formation derived from a separate, interaction, were interpreted in terms For hyperfine and Zeeman interactions light-independent mechanism (7). More- of a radical pair in which one of the of Ϸ500 ␮T and Ϸ50 ␮T, respectively, over, although radical pairs cannot form radicals was devoid of hyperfine motions with characteristic correlation the basis of a polarity compass, one can interactions. times shorter than Ϸ1 ␮s would seri- speculate about magnetite-based recep- A remarkable aspect of the avian ously attenuate the anisotropy of the tors that would be consistent with incli- compass is the ‘‘functional win- reaction yield. The other consequence nation compass behavior (29, 67, 68). dow:’’ a 20–30% increase or decrease of molecular motion is that it causes the Probably the most convincing experi- in the intensity of the ambient magnetic electron spins to relax, i.e., to equili- mental test of the radical pair model has field is sufficient to disorient caged birds brate thermally with their surroundings, involved subjecting caged migratory (78). In the context of radical pair mag- with concomitant loss of spin correla- birds to weak radio frequency fields (2). netoreception, this unusual property tion. The criterion for a significant mag- Magnetic fields oscillating at frequencies could be related to the field-dependent netic field effect is that relaxation times in the range 1–100 MHz, with or with- changes in S 7 T interconversion that should be longer than the radical pair out a static field present, are expected are known to arise from energy-level lifetime, which was estimated above to to alter the yields of radical pair reac- degeneracies (44). As an example, simu- be Ϸ1 ␮s. This requires rotational mo- tions by modifying the spin dynamics lations are shown in the SI Appendix tion either to be slower than Ϸ1 ␮sor (69). This has been observed for reac- that reveal a strong dependence of the faster than Ϸ100 ps. The latter, known tions of organic radicals in solution form and amplitude of the reaction yield as the ‘‘extreme narrowing limit,’’ seems when the frequency of the applied field anisotropy on the magnetic field inten- implausibly fast for a molecule large matches a S 7 T interconversion fre- sity in the range 10–100 ␮T (66). Rob-

356 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0711968106 Rodgers and Hore Downloaded by guest on September 24, 2021 ins are able to adapt relatively quickly (83–85). They contain 2 nonco- times (several milliseconds in vitro) that to magnetic fields outside their normal valently bound chromophores: a redox- could be compatible with magnetic sen- functional window (79). This ability active flavin adenine dinucleotide sitivity, although there is no direct evi- would have a natural interpretation in (FAD) and a light-harvesting cofactor. dence yet that these species are spin the radical pair model: The bird must Their functions, not all of which require correlated. Time-resolved EPR studies learn to recognize a new anisotropy light activation, include entrainment of of the photoactivation reaction in 2 pho- pattern. circadian clocks and, in plants, regula- tolyases have found long-lived (Ͼ10 ␮s A number of quantitative theoretical tion of growth and development. Unlike at 5 °C) electron spin polarization pat- studies, most of which have been men- their evolutionary ancestors, the DNA terns that are characteristic of spin-cor- tioned above, have tested selected as- photolyases (86), to which they show related FAD–Trp and FAD–Tyr radical pects of the radical pair mechanism (2, high sequence homology and structural pairs (86, 100, 101). Similar observa- 33, 43, 44, 57, 58, 66, 80). However, for similarity, cryptochromes do not repair tions, of a radical pair comprising the the most part, disorder, motion and spin double-stranded DNA. Here, we focus FADH• radical and the neutral radical relaxation have all been ignored. Nu- narrowly on the extent to which crypto- form of the terminal Trp residue of the merical simulations generally support chromes appear to meet the criteria out- Trp-triad, have recently been reported the idea of a radical-pair-based compass lined above. Others, better qualified for a Cry-DASH protein (102). Further- and have shed some light on the condi- than ourselves, have weighed the emerg- more, a small (3–4%) effect of a rela- tions required for significant reaction ing physiological and biological evidence tively strong magnetic field (39 mT) on yield anisotropies. In favorable cases, for their involvement in magnetorecep- the radical pair recombination kinetics the orientational dependence is usually tion (3–5, 55, 87, 88). has been found for Escherichia coli pho- no more than a Ϸ50% variation in the Compared with photolyases, relatively tolyase (103). The magnetic sensitivity reaction yield (43, 44, 58). Under more little is known about the photocycles of appears to arise from competition on a realistic conditions, changes of 1–10% cryptochromes, and most of that comes microsecond time scale between spin- seem more likely. Whether such a re- from studies of plant proteins (for re- selective back electron transfer in sponse is sufficient for a viable compass views, see refs. 83, 85, 87, and 89). In [FAD•Ϫ TrpH•ϩ] and spin-independent requires rather more insight into the contrast to photolyases, the photoactive deprotonation to produce the much signal transduction mechanisms than forms of cryptochromes seem to contain longer lived (Ϸ100 ␮s) radical pair currently exists. The best that can be the FAD chromophore in its fully oxi- [FAD•Ϫ Trp•] which is not magnetically done at this stage is to estimate very dized redox state (90–92). Absorption of sensitive presumably because of spin approximately how many magnetorecep- blue light leads to the formation of pro- relaxation. The structural and photo- tors would be required to overcome sto- tein-bound flavin and tryptophan (and chemical similarity to photolyases sug- chastic fluctuations in the number of sometimes tyrosine) radicals (90, 93–95) gests that cryptochromes may exhibit complexes formed between the signaling apparently by a route similar to photo- similar magnetic responses. state and its signaling partner(s) (81). activation in photolyase (95–97). In this Further evidence for the involvement Such calculations suggest that a reaction reaction, which appears to play no part of cryptochromes has come from experi- yield anisotropy of 1–10% might be in the function of photolyase, photoex- ments on plants (104). Enhanced crypto- adequate (58). cited FAD is reduced to the fla- chrome-mediated blue-light responses in Finally, it has recently been estab- vosemiquinone radical (FAD•Ϫ or its A. thaliana seedlings grown in a 500-␮T lished, as a proof-of-principle, that a protonated neutral form FADH•)bya magnetic field have been observed, in- photochemical reaction can act as a sequence of intraprotein electron trans- cluding a Ϸ30% inhibition of hypocotyl magnetic compass (14). Green-light irra- fers along a conserved chain of 3 trypto- (stem) growth. If plants, in which mag- diation of a triad model compound com- phan residues (the ‘‘Trp-triad’’) that cul- netoreception has no apparent function posed of linked carotenoid, porphyrin, minates in the oxidation of the terminal and which presumably lack highly and fullerene groups produced a radical Trp residue to form a TrpH•ϩ radical evolved receptors, are sensitive to mag- pair whose lifetime was both sensitive that can deprotonate to form the neu- netic fields, then other cryptochrome- to magnetic fields Ͻ50 ␮T and that re- tral Trp• radical (94, 98, 99). (See the SI containing species may be also. sponded anisotropically. Frozen solu- Appendix for the structures of the redox Evidence that the photochemistry may tions at low temperatures were used states of flavins and of the Trp-triad.) be more complex than presented above, in these experiments to immobilize the The FADH• (or FAD•Ϫ) form of the has come recently from a behavioral molecules, which were either aligned protein accumulates under continuous assay that shows that cryptochrome is in a liquid crystalline solvent or probed blue-light irradiation both in vitro and necessary for light-dependent magnetic anisotropically using photoselection with in cryptochrome-expressing insect cells, responses in Drosophila, although the plane-polarized light. Clearly, rather dif- and is thought to be the signaling state wavelengths required are compatible ferent means of restricting molecular (91, 92). Corresponding to species C in with the action spectrum of the reduced motion would be required for an in vivo Fig. 1, the flavosemiquinone state of the form of the protein (105). The notion magnetoreceptor. protein is presumably stabilized by inde- that insect cryptochromes do not con- pendent reduction of the Trp or Tyr tain a fully oxidized FAD in vivo (and Cryptochromes as Magnetoreceptors counterradical. The detailed mechanism may therefore be unable to generate It was the suggestion, in 2000, that cryp- of cryptochrome signaling is presently radical pairs) seems to be supported by tochromes could host magnetically sensi- unclear, but is presumed to involve con- mutation studies in which the Trp-triad tive radical pairs that set in motion the formational changes in the protein that is disrupted (106). Fluorescence and work reviewed in this Perspective (2). promote interaction with downstream EPR measurements, however, indicate Nine years later, they are still the only signaling partners (83, 87). that (human and Drosophila) contenders as chemical magnetorecep- The photochemically formed FAD– cryptochromes are photoreduced from tors. Cryptochromes are 50- to 70-kDa Trp radical pairs in cryptochromes from the fully oxidized state in insect cells, blue-light photoreceptor , both the model plant Arabidopsis thali- allowing accumulation of the fla- first identified in plants in 1993 (82) and ana (AtCry) and the migratory garden vosemiquinone radical state of the since found in , insects, and warbler Sylvia borin (gwCry) have life- proteins (107).

Rodgers and Hore PNAS ͉ January 13, 2009 ͉ vol. 106 ͉ no. 2 ͉ 357 Downloaded by guest on September 24, 2021 Theoretical considerations indicate mum spin correlation. Although the ex- prediction and interpretation of the re- that photoactive flavoproteins may be change and dipolar interactions are not sponse of radical pairs to weak magnetic suitable as magnetoreceptors. Although negligible in [FADH• W324•ϩ][Ϸ500 fields. These calculations are challeng- FAD–Trp radical pairs have Ϸ15 mag- ␮T (58)], they seem to be of appropriate ing, especially for biologically plausible netic nuclei (mostly 1H, and a few 14N) magnitudes partially to cancel one an- radicals most of which contain a pleth- with isotropic hyperfine coupling con- other’s effects, allowing the Earth’s field ora of internal magnetic interactions. stants larger than Ϸ100 ␮T, simulations to modulate S 7 T interconversion. Whereas in strong fields, semiclassical suggest that the anisotropy of the reac- Other, at present hypothetical, sources approaches allow reliable simulations to tion yield is determined principally by 2 of magnetic responses in cryptochromes be performed at a cost that rises only nitrogens (N5 and N10, see SI Appendix are discussed in the SI Appendix. polynomially with the number of mag- for numbering scheme) in the FAD rad- To conclude, the limited information netic nuclei, current approaches for ical that have relatively large, axial and currently available suggests that crypto- weak fields scale exponentially. A reli- collinear hyperfine tensors (43, 66). In chrome radical pairs could, in principle, able method for low-field calculations the absence of such dominant interac- form the basis of a compass magnetore- with polynomial scaling would be a sig- tions, it is likely that the anisotropic ceptor. Probably the physicochemical nificant development (109). Now that a contributions of the various hyperfine aspect about which there is most uncer- model system has been shown to re- Ͻ ␮ interactions would tend to cancel one tainty is the mechanism and degree of spond to 50- T fields (14), it ought to another, destroying much of the direc- molecular alignment and immobilization. be possible to explore experimentally as tional information. well as theoretically the design con- Three geometrical constraints were Future Work straints imposed on a chemical compass derived above from considerations of The riddle of the avian magnetic com- by molecular disorder and motion. Ef- the electron transfer kinetics and the pass is being teased apart, yet even after forts to produce a model system that likely magnitudes of the interradical in- 4 decades’ endeavor, a considerable functions at physiological temperatures are already under way. teractions: re Յ 1.5 nm and rc ϭ 2.0 Ϯ shroud of mystery remains. In the fol- 0.2 nm for the magnetically sensitive lowing, we make a few suggestions for Finally, experiments to test the cryp- tochrome hypothesis must continue. radical pair, and re Յ 1.0 nm for the further work. The list has a physical electron donor–acceptor pairs involved chemistry bias and is far from Spectroscopic measurements, both in in its formation. The distances between comprehensive. vitro (90) and in vivo (77, 107), will pro- vide insight into the photocycles of the the FAD and the 3 constituents of the Animal behavior experiments are of proteins and reveal the identity and Trp-triad electron transfer chain (W400, crucial importance. Radio frequency properties of radical pair intermediates. W377, W324) in AtCry1 (108) (no crys- fields have hitherto only been used to Characterization of the influence of tal structure is yet available for an ani- disrupt the magnetic compass sense. static and time-dependent magnetic mal cryptochrome) are given in the SI Such experiments cannot exclude the fields on isolated proteins (103) and Appendix. The 2 constraints on the ter- possibility that disorientation arises from Cry-mediated processes in living systems minal pair [FADH• W324•ϩ] are satis- effects on the animals’ motivation rather ϭ ϭ (105) will be equally important. fied (re 1.47 nm and rc 1.90 nm). than on the primary magnetoreception The physical chemistry underlying the Moreover, each of the donor–acceptor events. Theory predicts that weak static avian magnetic compass has matured pairs that precede it, [FAD W400], and radio frequency fields of similar •ϩ •ϩ much in the last 10 years, yet there re- [W400 W377] and [W377 W324], strength can induce similar changes in main many intriguing questions to be have edge-to-edge distances (0.45, 0.48, product yield. It might therefore be pos- addressed. We hope that this short sur- and 0.51 nm, respectively) well below 1.0 sible to condition animals to respond, in vey will encourage more physical scien- nm. However, these distances may differ the absence of the Earth’s field, to the tists to join us in this endeavor. in avian cryptochromes, and FAD-Tyr presence and orientation of a radio fre- magnetosensitive radical pairs, with dif- quency field of suitable strength. Such ACKNOWLEDGMENTS. We thank our colleagues, ferent re and rc, are also possible. observations would be difficult to ex- collaborators, and coworkers, without whom we With the center-to-center separations plain without the involvement of radical would not have been in a position to write this perspective: Margaret Ahmad, Olga Efimova, De- in AtCry1, both the primary and second- pairs. A skeptic would need to conjec- vens Gust, Sue-Re Harris, Kevin Henbest, Ilya Ku- • •ϩ ary radical pairs, [FADH W400 ] and ture a light-induced, highly magnetically prov, Miriam Liedvogel, Kiminori Maeda, Henrik [FADH• W377•ϩ], are expected to have sensitive radical pair reaction that did Mouritsen, Haruko Okamoto, Thorsten Ritz, Alex strong exchange interactions [estimated not supply the birds with directional Robinson, Erik Schleicher, Christiane Timmel, Ͼ Ͼ Nicola Wagner-Rundell, Stefan Weber, Roswitha at 50 T and 100 mT, respectively information but did affect their moti- Wiltschko, and Wolfgang Wiltschko. We are (58)] and hence no appreciable S 7 T vation, a possibility that seems to us grateful to the Engineering and Physical Sciences interconversion or response to the geo- somewhat less probable than a light- Research Council and the EMF Biological Re- magnetic field. Consequently, the ter- dependent radical pair magnetoreceptor. search Trust for financial support and to the Ox- • •ϩ ford Supercomputing Centre for CPU time. C.T.R. tiary pair [FADH W324 ] should be Theoretical and computational studies thanks Somerville College, Oxford, for a Junior formed in a pure spin state with maxi- have an important role to play in the Research Fellowship.

1. Wiltschko W (1968) U¨ ber den Einfluss statischer Mag- 6. Wiltschko W, Wiltschko R (2005) Magnetic orienta- 10. Steiner UE, Ulrich T (1989) Magnetic field effects in netfelder auf die Zugorientierung der Rotkehlchen tion and magnetoreception in birds and other ani- chemical kinetics and related phenomena. Chem Rev (Erithacus rubecula). Z Tierpsychol 25:537–558. mals. J Comp Physiol A 191:675–693. 89:51–147. 2. Ritz T, Adem S, Schulten K (2000) A model for photo- 7. Johnsen S, Lohmann KJ (2005) The physics and neu- 11. Timmel CR, Henbest KB (2004) A study of spin chem- receptor-based magnetoreception in birds. Biophys J robiology of magnetoreception. Nat Rev Neurosci istry in weak magnetic fields. Philos Trans R Soc Lon- 78:707–718. 6:703–712. don Ser A 362:2573–2589. 3. Johnsen S, Lohmann KJ (2008) Magnetoreception in 8. Ritz T, Dommer DH, Phillips JB (2002) Shedding light 12. Woodward JR (2002) Radical pairs in solution. Prog animals. Phys Today 61:29–35. on magnetoreception. Neuron 34:503– React Kinetic Mech 27:165–207. 4. Mouritsen H, Ritz T (2005) Magnetoreception and its use 506. 13. Brocklehurst B, McLauchlan KA (1996) Free radical in bird navigation. Curr Opin Neurobiol 15:406–414. 9. Brocklehurst B (2002) Magnetic fields and radical re- mechanism for the effects of environmental electro- 5. Wiltschko R, Wiltschko W (2006) Magnetoreception. actions: Recent developments and their role in Na- magnetic fields on biological systems. Int J Radiat Biol BioEssays 28:157–168. ture. Chem Soc Rev 31:301–311. 69:3–24.

358 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0711968106 Rodgers and Hore Downloaded by guest on September 24, 2021 14. Maeda K, et al. (2008) Chemical compass model of 39. van Dijk B, Carpenter JKH, Hoff AJ, Hore PJ (1998) 62. Moser CC, Dutton PL (1992) Engineering protein- avian magnetoreception. Nature 453:387–390. Magnetic field effects on the recombination kinetics structure for electron-transfer function in photosyn- 15. Hoff AJ (1981) Magnetic-field effects on photosyn- of radical pairs. J Phys Chem B 102:464–472. thetic reaction centers. Biochim Biophys Acta thetic reactions. Q Rev Biophys 14:599–665. 40. Gerson F, Huber W (2003) Electron Spin Resonance 1101:171–176. 16. Volk M, Ogrodnik A, Michel-Beyerle ME (1995) in Spectroscopy of Organic Radicals (Wiley-VCH, Wein- 63. Verhoeven JW (2006) On the role of spin correlation in Anoxygenic Photosynthetic Bacteria, eds Blankenship heim, Germany). the formation, decay, and detection of long-lived, RE, Madigan MT, Bauer CE (Kluwer, Dordrecht, The 41. Hoff AJ (1989) Advanced EPR. Applications in Biology intramolecular charge-transfer states. J Photochem Netherlands), pp 595–626. and Biochemistry (Elsevier, Amsterdam). Photobiol C 7:40–60. 17. Boxer SG, Chidsey CED, Roelofs MG (1983) Magnetic- 42. Improta R, Barone V (2004) Interplay of electronic, 64. O’Dea AR, Curtis AF, Green NJB, Timmel CR, Hore PJ field effects on reaction yields in the solid-state—an environmental, and vibrational effects in determin- (2005) Influence of dipolar interactions on radical pair example from photosynthetic reaction centers. Annu ing the hyperfine coupling constants of organic free recombination reactions subject to weak magnetic Rev Phys Chem 34:389–417. radicals. Chem Rev 104:1231–1253. fields. J Phys Chem A 109:869–873. 18. Schulten K, Swenberg CE, Weller A (1978) A biomag- 43. Cintolesi F, Ritz T, Kay CWM, Timmel CR, Hore PJ 65. Wasielewski MR (2006) Energy, charge, and spin netic sensory mechanism based on magnetic field (2003) Anisotropic recombination of an immobilized transport in molecules and self-assembled nanostruc- modulated coherent electron spin motion. Z Phys photoinduced radical pair in a 50-␮T magnetic field: A tures inspired by photosynthesis. J Org Chem Chem NF 111:1–5. model avian photomagnetoreceptor. Chem Phys 71:5051–5066. 19. Schulten K, Windemuth A (1986) in Biophysical Effects 294:385–399. 66. Rodgers CT (2007) D Phil thesis. University of Oxford of Steady Magnetic Fields (Springer, Berlin), pp 99– 44. Timmel CR, Cintolesi F, Brocklehurst B, Hore PJ (2001) (Oxford). 106. Model calculations of magnetic field effects on the 67. Davila AF, Fleissner G, Winklhofer M, Petersen N 20. Wiltschko W, Merkel FW (1966) Orientierung zugun- recombination reactions of radicals with anisotropic (2003) A new model for a magnetoreceptor in ruhiger Rotkehlchen im statischen Magnetfeld. Verh hyperfine interactions. Chem Phys Lett 334:387–395. pigeons based on interacting clusters of superpara- der Dtsch Zoolog Ges 59:362–367. 45. Canfield JM, Belford RL, Debrunner PG (1996) Calcu- magnetic magnetite. Phys Chem Earth 28:647–652. 21. Wiltschko W, Wiltschko R (1972) Magnetic compass of lations of Earth-strength steady and oscillating mag- 68. Solov’yov IA, Greiner W (2008) -mineral-based European robins. Science 176:62–64. netic field effects in coenzyme B-12 radical pair sys- magnetoreceptor in birds: Polarity or inclination com- 22. Merkel FW, Fromme HG (1958) Untersuchungen u¨ber tems. Mol Phys 89:889–930. pass? Eur Phys J D 10.1140/epjd/e2008–00118-y. das Orientierungsvermo¨gen na¨chtlich ziehender Rot- 46. Bube W, Michel-Beyerle ME, Haberkorn R, Steffens E 69. Timmel CR, Hore PJ (1996) Oscillating magnetic field kehlchen, Erithacus rubecula. Naturwissenschaften (1977) Sensitized charge carrier injection into organic- effects on the yields of radical pair reactions. Chem 45:499–500. crystals studied by isotope effects in weak magnetic Phys Lett 257:401–408. 23. Wiltschko W, Wiltschko R (2001) Light-dependent fields. Chem Phys Lett 50:389–393. 70. Woodward JR, Timmel CR, McLauchlan KA, Hore PJ magnetoreception in birds: The behaviour of Euro- 47. Groff RP, Suna A, Avakian P, Merrifield RE (1974) (2001) Radio frequency magnetic field effects on elec- Magnetic hyperfine modulation of dye-sensitized de- tron-hole recombination. Phys Rev Lett 87:077602. pean robins, Erithacus rubecula, under monochro- layed fluorescence in organic crystals. Phys Rev B 71. Henbest KB, Kukura P, Rodgers CT, Hore PJ, Timmel CR matic light of various wavelengths and intensities J 9:2655–2660. (2004) Radio frequency magnetic field effects on a Exp Biol 204:3295–3302. 48. Boxer SG, Chidsey CED, Roelofs MG (1982) Anisotropic radical recombination reaction: A diagnostic test for 24. Deutschlander ME, Phillips JB, Borland SC (1999) The magnetic interactions in the primary radical ion-pair the radical pair mechanism. J Am Chem Soc 126:8102– case for light-dependent magnetic orientation in an- of photosynthetic reaction centers. Proc Natl Acad Sci 8103. imals. J Exp Biol 202:891–908. USA 79:4632–4636. 72. Rodgers CT, Henbest KB, Kukura P, Timmel CR, Hore PJ 25. Frost BJ, Mouritsen H (2006) The neural mechanisms 49. Dutta R, Basu S, Chowdhury M (1991) Magnetic-field (2005) Low-field optically detected EPR spectroscopy of long distance . Curr Opin Neu- effect on the diphenylhexatriene dimethylaniline ex- of transient photoinduced radical pairs. J Phys Chem robiol 16:481–488. ciplex system - anisotropic effects in viscous media. A 109:5035–5041. 26. Kirschvink JL, Walker MM, Diebel CE (2001) Magne- Chem Phys Lett 182:429–434. 73. Weiss BP, et al. (2004) Ferromagnetic resonance and tite-based magnetoreception. Curr Opin Neurobiol 50. Sengupta T, Aich S, Basu S (1999) Isotropic and aniso- low-temperature magnetic tests for biogenic magne- 11:462–467. tropic magnetic field effect on the exciplex between tite. Earth Planet Sci Lett 224:73–89. 27. Walker MM (2008) A model for encoding of magnetic all-s-trans-1,6-diphenylhexa-1,3,5-triene and 1,4- 74. Kirschvink JL (1996) Microwave absorption by mag- field intensity by magnetite-based magnetoreceptor dicyanobenzene. J Phys Chem B 103:3784–3790. netite: A possible mechanism for coupling nonther- cells. J Theor Biol 250:85–91. ϩ Ϫ 51. Link G, et al. (2001) Structure of the P700 A1 radical mal levels of radiation to biological systems. Bioelec- 28. Walker MM, Dennis TE, Kirschvink JL (2002) The mag- pair intermediate in photosystem I by high time res- tromagnetics 17:187–194. netic sense and its use in long-distance navigation by olution multifrequency electron paramagnetic reso- 75. Ritz T, Thalau P, Phillips JB, Wiltschko R, Wiltschko W animals. Curr Opin Neurobiol 12:735–744. nance: Analysis of quantum beat oscillations. JAm (2004) Resonance effects indicate a radical-pair mech- 29. Kirschvink JL, Gould JL (1981) Biogenic magnetite as a Chem Soc 123:4211–4222. anism for avian magnetic compass. Nature 429:177– basis for magnetic-field detection in animals. Biosys- 52. Weber S, Ohmes E, Thurnauer MC, Norris JR, Kothe G 180. tems 13:181–201. (1995) Light-generated nuclear quantum beats—A 76. Thalau P, Ritz T, Stapput K, Wiltschko R, Wiltschko W 30. Fleissner G, et al. (2003) Ultrastructural analysis of a signature of photosynthesis. Proc Natl Acad Sci USA (2005) Magnetic compass orientation of migratory putative magnetoreceptor in the beak of homing 92:7789–7793. birds in the presence of a 1.315 MHz oscillating field. pigeons. J Comp Neurol 458:350–360. 53. Fursman CE, Hore PJ (1999) Distance determination in Naturwissenschaften 92:86–90. 31. Fleissner G, Stahl B, Thalau P, Falkenberg G, Fleissner spin-correlated radical pairs in photosynthetic reac- 77. Ritz T, et al. (2008) Magnetic compass of birds is based G (2007) A novel concept of Fe-mineral-based mag- tion centres by electron spin echo envelope modula- on a molecule with optimal directional sensitivity. netoreception: Histological and physicochemical data tion. Chem Phys Lett 303:593–600. Biophys J, in press. from the upper beak of homing pigeons. Naturwis- 54. Jeschke G, Matysik J (2003) A reassessment of the 78. Wiltschko R, Wiltschko W (1995) Magnetic Orienta- senschaften 94:631–642. origin of photochemically induced dynamic nuclear tion in Animals (Springer, Berlin). 32. Solov’yov IA, Greiner W (2007) Theoretical analysis of polarization effects in solids. Chem Phys 294:239–255. 79. Wiltschko W, Stapput K, Thalau P, Wiltschko R (2006) an iron mineral-based magnetoreceptor model in 55. Mouritsen H, et al. (2004) Cryptochromes and neuro- Avian magnetic compass: Fast adjustment to intensi- birds. Biophys J 93:1493–1509. nal-activity markers colocalize in the retina of migra- ties outside the normal functional window. Naturwis- 33. Solov’yov IA, Chandler DE, Schulten K (2007) Mag- tory birds during magnetic orientation. Proc Natl senschaften 93:300–304. netic field effects in Arabidopsis thaliana crypto- Acad Sci USA 101:14294–14299. 80. Wang KF, Ritz T (2006) Zeeman resonances for radical- chrome-1. Biophys J 92:2711–2726. 56. Heyers D, Manns M, Luksch H, Gu˝ntu˝rku˝n O, Moutit- pair reactions in weak static magnetic fields. Mol Phys 34. Hore PJ (1989) in Advanced EPR Applications in Biol- sen H (2007) A visual pathway links brain structures 104:1649–1658. ogy and Biochemistry, ed Hoff AJ (Elsevier, Amster- active during magnetic compass orientation in migra- 81. Weaver JC, Vaughan TE, Astumian RD (2000) Biolog- dam), pp 405–440. tory birds. PLoS ONE 2:e937. ical sensing of small field differences by magnetically 35. Hulsebosch RJ, Borovykh IV, Paschenko SV, Gast P, 57. Wang K, Mattern E, Ritz T (2006) On the use of mag- sensitive chemical reactions. Nature 405:707–709. Hoff AJ (1999) Radical pair dynamics and interactions nets to disrupt the physiological compass of birds. 82. Ahmad M, Cashmore AR (1993) HY4 gene of A- in quinone-reconstituted photosynthetic reaction Phys Biol 3:220–231. thaliana encodes a protein with characteristics of a centers of Rb. sphaeroides R26: A multifrequency 58. Efimova O, Hore PJ (2008) Role of exchange and di- blue-light photoreceptor. Nature 366:162–166. magnetic resonance study. J Phys Chem B 103:6815– polar interactions in the radical pair model of the 83. Lin CT, Todo T (2005) The cryptochromes. Genome 6823. avian magnetic compass. Biophys J 94:1565–1574. Biol 6:220. 36. Tang J, Utschig LM, Poluektov O, Thurnauer MC (1999) 59. Rodgers CT, Norman SA, Henbest KB, Timmel CR, Hore 84. Partch CL, Sancar A (2005) Photochemistry and pho- Transient W-band EPR study of sequential electron PJ (2007) Determination of radical re-encounter prob- tobiology of cryptochrome blue-light photopig- transfer in photosynthetic bacterial reaction centers. ability distributions from magnetic field effects on ments: The search for a photocycle. Photochem Pho- J Phys Chem B 103:5145–5150. reaction yields. J Am Chem Soc 129:6746–6755. tobiol 81:1291–1304. 37. Dzuba SA, et al. (1996) Control of radical pair lifetimes 60. Timmel CR, Till U, Brocklehurst B, McLauchlan KA, 85. Sancar A (2003) Structure and function of DNA pho- by microwave irradiation. Application to photosyn- Hore PJ (1998) Effects of weak magnetic fields on free tolyase and cryptochrome blue-light photoreceptors. thetic reaction centres. Chem Phys Lett 253:361–366. radical recombination reactions. Mol Phys 95:71–89. Chem Rev 103:2203–2237. 38. van Dijk B, Gast P, Hoff AJ (1996) Control of radical 61. Moser CC, Keske JM, Warncke K, Farid RS, Dutton PL 86. Weber S (2005) Light-driven enzymatic catalysis of pair lifetime by a switched magnetic field. Phys Rev (1992) Nature of biological electron transfer. Nature DNA repair: A review of recent biophysical studies on Lett 77:4478–4481. 355:796–802. photolyase. Biochim Biophys Acta 1707:1–23.

Rodgers and Hore PNAS ͉ January 13, 2009 ͉ vol. 106 ͉ no. 2 ͉ 359 Downloaded by guest on September 24, 2021 87. Partch CL, Sancar A (2005) Photochemistry and pho- 95. Aubert C, Mathis P, Eker APM, Brettel K (1999) Intra- 102. Biskup T, et al. (2009) Direct observation of a photo- tobiology of cryptochrome blue-light photopig- protein electron transfer between tyrosine and tryp- induced radical-pair intermediate in a cryptochrome ments: The search for a photocycle. Photochem Pho- tophan in DNA photolyase from Anacystis nidulans. DASH blue-light photoreceptor. Angew Chem Int Ed tobiol 81:1291–1304. Proc Natl Acad Sci USA 96:5423–5427. 48:404–407. 88. Mo¨ ller A, Sagasser S, Wiltschko W, Schierwater B 96. Aubert C, Vos MH, Mathis P, Eker APM, Brettel K 103. Henbest KB, et al. (2008) Magnetic-field effect on the (2004) Retinal cryptochrome in a migratory passerine (2000) Intraprotein radical transfer during photoacti- photoactivation reaction of Escherichia coli DNA pho- bird: A possible transducer for the avian magnetic vation of DNA photolyase. Nature 405:586–590. tolyase. Proc Natl Acad Sci USA 105:14395–14399. compass. Naturwissenschaften 91:585–588. 97. Lukacs A, Eker APM, Byrdin M, Brettel K, Vos MH 104. Ahmad M, Galland P, Ritz T, Wiltschko R, Wiltschko W 89. Losi A (2007) Flavin-based blue-light photosensors: A (2008) Electron hopping through the 15 Å triple tryp- (2007) Magnetic intensity affects cryptochrome- photobiophysics update. Photochem Photobiol tophan molecular wire in DNA photolyase occurs dependent responses in Arabidopsis thaliana. Planta 83:1283–1300. within 30 ps. J Am Chem Soc 130:14394–14395. 225:615–624. 90. Liedvogel M, et al. (2007) Chemical magnetorecep- 98. Byrdin M, et al. (2004) Intraprotein electron transfer 105. Gegear RJ, Casselman A, Waddell S, Reppert SM tion: Bird cryptochrome 1a is excited by blue light and proton dynamics during photoactivation of DNA (2008) Cryptochrome mediates light-dependent mag- and forms long-lived radical-pairs. PLoS ONE photolyase from E. coli: Review and new insights from netosensitivity in Drosophila. Nature 454:1014–1018. 2:e1106. an ‘‘inverse’’ deuterium isotope effect. Biochim Bio- 106. Kavakli IH, Sancar A (2004) Analysis of the role of intra- 91. Bouly JP, et al. (2007) Cryptochrome blue light pho- phys Acta 1655:64–70. protein electron transfer in photoreactivation by DNA toreceptors are activated through interconversion of 99. Kao YT, et al. (2008) Ultrafast dynamics and anionic photolyase in vivo. Biochemistry 43:15103–15110. flavin redox states. J Biol Chem 282:9383–9391. active states of the flavin cofactor in cryptochrome 107. Hoang N, et al. (2008) Human and Drosophila crypto- 92. Banerjee R, et al. (2007) The signaling state of Arabi- and photolyase. J Am Chem Soc 130:7695–7701. chromes are light activated by flavin photoreduction dopsis cryptochrome 2 contains flavin semiquinone. 100. Weber S, et al. (2002) Photoactivation of the flavin in living cells. PLoS Biol 6:e160. J Biol Chem 282:14916–14922. cofactor in Xenopus laevis (6-4) photolyase: Observa- 108. Brautigam CA, et al. (2004) Structure of the photol- 93. Giovani B, Byrdin M, Ahmad M, Brettel K (2003) Light- tion of a transient tyrosyl radical by time-resolved yase-like domain of cryptochrome 1 from Arabidopsis induced electron transfer in a cryptochrome blue- electron paramagnetic resonance. Proc Natl Acad Sci thaliana. Proc Natl Acad Sci USA 101:12142–12147. light photoreceptor. Nat Struct Biol 10:489–490. USA 99:1319–1322. 109. Kuprov I, Wagner-Rundell N, Hore PJ (2007) Polyno- 94. Zeugner A, et al. (2005) Light-induced electron trans- 101. Gindt YM, et al. (1999) Origin of the transient electron mially scaling spin dynamics simulation algorithm fer in Arabidopsis cryptochrome-1 correlates with in paramagnetic resonance signals in DNA photolyase. based on adaptive state-space restriction. J Magn vivo function. J Biol Chem 280:19437–19440. Biochemistry 38:3857–3866. Reson 189:241–250.

360 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0711968106 Rodgers and Hore Downloaded by guest on September 24, 2021