Pulsed electron electron double resonance: beyond nanometre distance measurements on biomacromolecules

Transcription

1 Biochem. J. (2011) 434, (Printed in Great Britain) doi: /bj REVIEW ARTICLE Pulsed electron electron double resonance: beyond nanometre distance measurements on biomacromolecules Gunnar W. REGINSSON and Olav SCHIEMANN 1 Biomedical Sciences Research Complex, Centre of Magnetic Resonance, University of St Andrews, St Andrews KY16 9ST, Scotland, U.K. PELDOR (or DEER; pulsed electron electron double resonance) is an EPR (electron paramagnetic resonance) method that measures via the dipolar electron electron coupling distances in the nanometre range, currently nm, with high precision and reliability. Depending on the quality of the data, the error can be as small as 0.1 nm. Beyond mere mean distances, PELDOR yields distance distributions, which provide access to conformational distributions and dynamics. It can also be used to count the number of monomers in a complex and allows determination of the orientations of spin centres with respect to each other. If, in addition to the dipolar through-space coupling, a through-bond exchange coupling mechanism contributes to the overall coupling both mechanisms can be separated and quantified. Over the last 10 years PELDOR has emerged as a powerful new biophysical method without size restriction to the biomolecule to be studied, and has been applied to a large variety of nucleic acids as well as proteins and protein complexes in solution or within membranes. Small nitroxide spin labels, paramagnetic metal ions, amino acid radicals or intrinsic clusters and cofactor radicals have been used as spin centres. Key words: biomacromolecule measurement, dynamics, nucleic acid, pulsed electron electron double resonance (PELDOR or DEER), spin labelling. INTRODUCTION Molecular and structural biology are engaging ever larger biomacromolecular complexes either isolated or within membranes or whole cells. Thus biophysical methods are needed which can provide structural and dynamic information for these molecular architectures on the relevant nanometre scale. Fluorescence-based methods, such as FRET (fluorescence resonance energy transfer) are very powerful in this respect as they can be applied at the single-molecule level and provide realtime dynamics over several time scales [1]. However, FRET lacks precision when it comes to quantifying distances or distance changes. In contrast, a pulsed EPR (electron paramagnetic resonance) [2] method called PELDOR (or DEER; pulsed electron electron double resonance) yields reliably very precise distances in the range nm. Like FRET, PELDOR has no molecular mass restriction, but it usually requires micromolar concentrations and the sample has to be frozen. Thus both methods are complementary. Milov et al. [3] were the first to introduce the three-pulse version of PELDOR [3] (Figure 1A), and the groups of Spiess [4] and Jeschke [5] extended it to the dead-time-free four-pulse version (Figure 1B). A more specialized method is the DQC (double quantum coherence)-epr filter experiment from the laboratory of Freed [6]. Over the last few years several other review articles have been published [7 12]. More experimentally focused articles with detailed protocols for PELDOR measurements have also been published [13 17]. The present review will briefly describe different ways of spin-labelling biomacromolecules and then introduce the dipolar coupling between unpaired electrons. This is followed by a disussion about the signal the PELDOR pulse sequence generates. The information that can be collected from PELDOR experiments will be demonstrated by summarizing applications of PELDOR to model systems, proteins and nucleic acids. SPIN LABELLING In order to be able to apply EPR spectroscopic methods such as PELDOR, the biomacromolecule needs to have at least one unpaired electron or, in other words, it needs to have one or more paramagnetic centre. These can, for example, be metal ions (e.g. Cu 2+ or Mn 2+ ), metal clusters (e.g. FeS), organic cofactors (e.g. semiquinones or flavin radicals) or amino acid radicals (e.g. tyrosyl radicals) formed during catalytic cycles. In cases where the biomolecule is diamagnetic and has no such centre, paramagnetic labels carrying unpaired electrons can be introduced. Commonly, these labels are nitroxides [18,19]. For proteins, MTSSL (methanethiosulfonate spin label) [20] is mostly used, which reacts with the SH-group of cysteine residues and forms a covalent disulfide bridge (Figure 2). The site specificity for this reaction is introduced by site-directed mutagenesis by which unwanted cysteine residues are replaced by, e.g., serine residues, and new cysteine residues are introduced at positions where they are wanted [21]. In cases where cysteine residues cannot be removed without loss of structure or function, ligation techniques [22] or the incorporation of non-natural amino acids via newly coded trnas [23] can be used. Over the last decade, several strategies have been established to also site-specifically spin label oligonucleotides [24]. For example, nitroxides can be site-specifically attached to the bases [25 29], the phosphate [30] or sugar group [31]. The site specificity is introduced by incorporating a modified Abbreviations used: ABC transporter, ATP-binding-cassette transporter; cw, continuous wave; EPR, electron paramagnetic resonance; FRET, fluorescence resonance energy transfer; LHCII, light-harvesting chlorophyll a/b complex; MD, Molecular Dynamics; MTSSL, methanethiosulfonate spin label; PELDOR (or DEER), pulsed electron electron double resonance; PSII, photosystem II; RE, resonance echo; RNR, ribonucleotide reductase. 1 To whom correspondence should be addressed ( os11@st-andrews.ac.uk).

2 354 G. W. Reginsson and O. Schiemann If the distance between both electrons is small and thus the dipolar interaction strong, the line splitting is large and might be observed in a cw (continuous wave) EPR spectrum [40 42]. For nitroxides measured at X-band frequencies, the upper distance limit for a cw EPR observation of a dipolar splitting is approx nm. Above this distance the dipolar splitting is covered under the inhomogeneous linewidth. The advantage of the PELDOR experiment is that it suppresses hyperfinecoupling contributions and selects the electron electron coupling which enables the measurement of much longer distances. The longest experimentally verified distance obtained with PELDOR is currently 8 nm [43], the shortest 1.5 nm [44]. Banham et al. [45] investigated the distance regime in which both methods can be applied. Figure 1 Microwave pulse sequence for three- and four-pulse PELDOR (A) Three-pulse PELDOR pulse sequence. (B) Dead-time-free four-pulse PELDOR sequence. Reproduced with permission, from Reginsson, G. W. and Schiemann, O. (2011) Biochem. Soc. Trans., 39, c the Biochemical Society. Figure 2 MTSSL and its reaction with cysteine residues phosphoramidite into the oligonucleotide at the desired place in the sequence during automated oligonucleotide synthesis. The modification can be a nitroxide label or a chemical group, which reacts specifically with a corresponding group on a nitroxide. These strategies are all limited by the need to synthesize the oligonucleotide, which currently puts the length limit at approx. 80 bases. Obeid et al. [32] use enzymatic labelling to overcome this problem. In 2010, a method was published in which a spin-labelled cytosine analogue was non-covalently, but sitespecifically, attached to oligonucleotides [33]. The site specificity is achieved with the help of an abasic site opposite a guanine. With the possibility to buy oligonucleotide strands with abasic sites from various companies, this labelling strategy no longer requires an in-laboratory chemical step for labelling of nucleic acids. If the biomolecule contains diamagnetic metal ions, e.g. Mg 2+, these can be exchanged for paramagnetic ones, e.g. Mn 2+ [34 37]. In any case, labelling, mutagenesis and exchange of metal ions do lead to alteration of the biomolecule, thus control experiments with respect to the integrity of its structure and function are required. DIPOLAR COUPLING If a biomolecule contains two unpaired electrons their magnetic dipole dipole interaction gives rise to a splitting of their EPR lines. Within the point-dipole approximation, which is fulfilled for large distances and spin-localized systems [38], the dipolar coupling ν dd is described by eqn (1) [39]: v dd = μ 0 γ A γ B 8π 2 r 3 AB ( 1 3cos2 θ ) (1) where μ 0 is the vacuum permeability, is the Planck constant divided by 2π, γ A and γ B are the magnetogyric ratios of the two spins A and B, r AB is the distance between them and θ is the angle between the distance vector r AB and the external magnetic field B 0. PELDOR PULSE SEQUENCE AND SIGNAL Nowadays it is the dead-time-free four-pulse PELDOR experiment that is routinely used (Figure 1B) [4,5]. The detection sequence π/2-τ 1 -π-τ 1 -echo 1 -τ 2 -π is applied at a microwave frequency ν A. This pulse sequence creates after the time τ 2, after the last π pulse, an RE (refocused echo) from spins (spins A) in resonance with the microwave frequency ν A. The time position of the RE does not change during the experiment, since the time intervals of the detection sequence are not changed. Introduction of the inversion pulse at frequency ν B during the time interval T flips spins (spins B), which are in resonance with this second frequency. Using two different and phase-independent microwaves and keeping the RE at a fixed position in time strongly suppresses hyperfine contributions. Incrementing the pump pulse from t = 0tot = T leads to an oscillation of the amplitude V of the RE dependent on the time position t of the inversion pulse if spins A and B are coupled. The frequency of this oscillation is the dipolar coupling ν dd between both spins. The actual zero time of this oscillation is when the inversion pulse crosses the Hahn echo. Since the time axis t for the inversion pulse can be chosen to start before the Hahn echo, one actually records negative times for the dipolar evolution, but more importantly one has no dead time for the dipolar evolution. In the three-pulse PELDOR version (Figure 1A), a dead-time occurs because the inversion pulse cannot usually be applied on the π-pulse of the detection sequence [46] without introducing too strong artefacts at the beginning of the time trace. Since PELDOR is not a single-molecule experiment (it usually requires spin concentrations of 100 μm), the PELDORsignalV(t) (Figure 3a) is the product of the contribution from the interaction between the two spins in one molecule V(t) intra and the interaction between spins on different molecules V(t) inter (eqn 2) [47]: V(t) = V(t) intra V(t) inter (2) If the spins are randomly distributed in the sample, V(t) inter can be written in the form of a monoexponential decay function which depends on the radical concentration. Dividing V(t)bythis decay function leaves the wanted intramolecular signal V(t) intra (Figure 3b), which can be described by eqn (3) [10]: V(t) intra = 1 n n {1 λ B [1 cos (v dd t)]} (3) n A=1 B=1 B A ν dd isgivenbyeqn(1),n is the number of radicals per biomolecule, λ B is the fraction of B spins inverted by the pump pulse, t is the time delay of the pump pulse, and < > is the integration over all values of r AB and θ.

3 PELDOR beyond distances 355 Figure 3 The PELDOR signal (a) The green line is the convolute of intra- and inter-molecular interaction as obtained in a PELDOR experiment. The red line is the fitted intermolecular contribution. (b) The wanted intramolecular signal obtained after dividing the original time trace [green line in (a)] by the intermolecular contribution [red line in (a)]. (c) Pake pattern obtained after Fourier transformation of the signal in (b). Reprinted by permission from Macmillan Publishers Ltd: Nature Protocols [13], copyright (2007). Figure 4 Schematic representation of a spin-labelled DNA RNA duplex and the correlation between distances obtained from PELDOR and MD simulations on a series of such duplexes with different numbers of base pairs between the labelled sides Distances for RNAs are shown as and from DNAs as. Adapted with permission from [54], Copyright 2004 American Chemical Society; and Piton, N., Mu, Y., Stock, G., Prisner, T. F., Schiemann, O. and Engels, J. W. Base specific spin-labeling of RNA for structure determination. Nucleic Acids Res. 2007, 35, , by permission of Oxford University Press. Each molecule in a frozen sample has a fixed and specific orientation of r AB with respect to B 0 and thus a specific value for θ. However,sinceV(t) intra integrates over all molecules and thus all values of θ (if the molecules are not aligned), Fourier transformation of V(t) intra does not give a single line, but the socalled dipolar Pake pattern (Figure 3c). The peaks and edges of the Pake pattern correspond to molecules with θ = 90 = θ and θ = 0 = θ ll respectively. Reading off the frequency at θ = 90 and substituting it into eqn (4), the rearranged version of eqn (1) for nitroxides, yields the distance r AB between the two spins: r AB [nm] = 3 v (θ ) [MHz] Instead of Fourier transforming the time traces, they are commonly analysed with the program DeerAnalysis from G. Jeschke [48], which is based on Tikhonov regularization and the assumption that the full Pake pattern (all orientations of θ) is excited. The upper distance limit for PELDOR is determined by the need to resolve one full oscillation period for a precise distance determination. This requires that the time window t of the experiment is long enough to include a full oscillation, which in turn requires that the relaxation time T M of the system is long enough. For example, to measure the so-far longest distance of 80 Å (1 Å = 0.1 nm), it was necessary to dissolve the bisnitroxide model system in a special matrix to make T M long enough [43]. (4) Interestingly, a recent paper by Ward et al. [49] demonstrated that such long distances can even be accessed in aqueous buffer solutions and for as large protein complexes as histones, if the proteins and the buffer are fully deuterated. The first PELDOR distance measurements on proteins have been achieved in the laboratory of Kawamori using intrinsic radicals in PSII (photosystem II) [50,51]. Persson et al. [52] were the first to apply PELDOR to spin-labelled proteins, specifically the human carbonic anhydrase II. Schiemann et al. [53,54] established a PELDOR-based nanometer distance ruler for spinlabelled biomolecules in aqueous buffer solution using nucleic acids (Figure 4) [26,54]. Also on nucleic acids, the laboratory of Norman showed that PELDOR can resolve distances from up to five different doubly labelled DNAs in one sample [55]. Today a wide range of biomolecules have been studied with PELDOR. Below, a few examples are collected where PELDOR has been used to resolve the arrangement of protein domains or of proteins in complexes. Park et al. [56] applied PELDOR to the chemotaxis receptor kinase assembly of Thermotoga maritima to unravel the arrangement of the CheA CheW complex. The predicted structure, constructed from 40 distances, matches nicely with the crystal structure of this complex published in the same paper. Based on the structure and biochemical data, suggestions with respect to the functional mechanism of this complex have been made. The laboratory of Fajer has published several papers on troponin showing that the inhibitory region of the ternary

4 356 G. W. Reginsson and O. Schiemann Figure 5 Distance measurements on the KcsA ion channel using PELDOR (a) Orientation-selective X-band PELDOR time traces from spin-labelled KcsA overlaid with the simulated time traces. (b) Structure of the spin-labelled tetrameric KcsA. Reproduced by permission from [74]. Copyright 2009 American Chemical Society. complex is α-helical [57] and that Ca 2+ induces structural changes [58]. Meyer et al. [59] determined the structures of the dimeric full-length G-protein MnmE in the GTP and GDP states. The observed distance changes of up to 1.8 nm clearly indicate large conformational changes associated with the hydrolysis event. The laboratory of Jeschke used a combined approach of modelling and PELDOR to determine the structural arrangement of the physiologically relevant homodimer of the Na + /H + antiporter NhaA of Escherichia coli [60]. Grote et al. [61] used PELDOR to study the conformational changes of an ABC transporter (ATPbinding-cassette transporter). The doubly spin-labelled series of ABC maltose transporter variants (MalFGK 2 -E) showed that the substrate-binding protein MalE is bound to the transporter throughout the transport cycle and that not only does the ATPbinding side undergo three conformational changes during function, but also so does the periplasmic MalF-P2 loop. This is suggested to be an important mechanistic feature of receptor-coupled ABC transporters. Kay et al. [62] used the flavin cofactor bound to each of the two subunits in the homodimer of the human augmenter of liver regeneration. They reduced the two flavins to the neutral radical form and obtained well-resolved PELDOR time traces. The distances correlate well with those from the X-ray structure. Bowman et al. [63] managed to assess the large assembly of the tetrameric histone complex. In the field of nucleic acids, PELDOR has been used to follow the conversion of two complementary hairpin RNAs into a duplex RNA [25], ligand binding to the neomycin riboswitch [64] and magnesium(ii)-ion-induced folding of the tertiary stabilized hammerhead ribozyme [65]. Structures of damaged nucleic acids yielding bended DNA double helices have also been studied [46,66]. With respect to generating structures from PELDOR-derived distances, Jeschke recently published a program called MMM which is freely available on the Web [67], and Ward et al. [68] reported a strategy to refine NMR structures using PELDORderived distances as long-range constrains. MEMBRANES In contrast with the PELDOR data discussed above, studies on membrane-bound proteins show, in a lot of cases, only barely visible dipolar oscillations or just multi-exponential decays. In such cases, conclusions are mainly qualitative and should be drawn carefully. This might also be considered when debating whether α-synuclein adopts an extended [69] or horseshoe [70] conformation. The problem with spin-labelled proteins reconstituted into membranes is molecular crowding due to the confinement into two dimensions and experimentally imposed low lipid-to-protein ratios [71]. This leads to very close intermolecular contacts and thus very short T M relaxation times, which shortens the usable PELDOR time windows to a length of sometimes below 1 μs. In turn, this strongly limits the maximum distance that can be extracted and causes problems in determining the intermolecular background because the decay is no longer monoexponential. The fast echo decay also strongly diminishes the intensity of the PELDOR signal, which leads to PELDOR time traces with low signal-to-noise. Thus, although PELDOR is, in principle, ideally suited to study large membrane protein complexes in membranes because it is not size-restricted and can assess the necessary distance range, ways have to be found and sample preparation protocols established to overcome these challenges. The current state of knowledge is that molecular crowding can be reduced by using nanodiscs [72]. In these nanodiscs, the membrane is confined by proteins allowing the discs to host sometimes only one membrane protein per disc, but this has to be shown more widely. An additional possibility is to dilute the intermolecular interaction by adding non-labelled protein. Below, a few examples are given where the PELDOR data show modulations and are of high quality. Zou et al. [73] used PELDOR to define the amplitude and nature of the ATP-hydrolysis-induced conformational change in the ABC transporter Msb. They obtained high-quality PELDOR data for several doubly labelled mutants and distinguished them from those cases where they don t by taking the error into account. Based on their analysis they found that residues on the cytoplasmic side undergo a 2 3 nm closing motion, whereas a nm opening motion is observed on the extracellular side. The transmembrane helices undergo relative movement to create the outward opening consistent with the X-ray model. Endeward et al. [74] have been able to resolve the expected three distances for the trimeric potassium ion channel KcsA both in detergent and reconstituted into a lipid membrane (Figure 5). DYNAMICS The distance distribution is encoded in the damping of the modulation. The faster the modulation is damped the broader

5 PELDOR beyond distances 357 Figure 6 Folding of membrane protein monitored by PELDOR (a) Structure of monomeric LHCII with the spin-labelled residues indicated by numbers. (b) PELDOR time traces of double mutant 90/196 and the corresponding distance distribution. (c) Time course of the folding event monitored on the two double mutants 106/160 and 90/196. Reprinted from Proc. Natl. Acad. Sci. U.S.A., 106, Dockter, C., Volkov, A., Bauer, C., Polyhach, Y., Joly-Lopez, Z., Jeschke, G. and Paulsen, H., Refolding of the integral membrane protein light-harvesting complex II monitored by pulse EPR, , Copyright (2009), with permission from Elsevier. Figure 7 Spin-labelling nucleic acids Spin labels 1 [25], 2 [26] and 3 [27,28] are connected to the base moiety. Label 4 [30] is attached to the phosphate backbone and label 5 [31] to the 2 -sugar side. Label 6 [29] is a rigid label fused on to a cytosine base, and label 7 [33] is its non-covalent analogue. Reproduced with permission from Reginsson, G. W. and Schiemann, O. (2010) Biochem. Soc. Trans. 39, c the Biochemical Society. is the distance distribution and vice versa. As mentioned in the introduction, PELDOR experiments are performed in the frozen state, which means that real-time measurements of distance changes/dynamics are not possible. However, freezing a sample considerably rapidly freezes conformations of a biomolecule populated at the freezing point of the sample. This means that PELDOR does not access conformational distributions at 50 K (the usual measuring temperature), but actually the one populated at approx. 243K ( 30 C), the freezing point of aqueous buffers with cryoprotectant added. This is still not room temperature, but much closer to the biologically relevant temperature. It has also been shown repeatedly that the distance distribution can be correlated with conformational states and dynamics at room temperature as long as they lead to distance changes between the spin labels. The laboratory of Jeschke and colleagues demonstrated this nicely on end-labelled phenyl-acetylene model systems of various length and in matrixes with different freezing temperatures [75,76]. The dynamics of similar systems have also been analysed by the laboratory of Prisner using a simple dynamics model [77]. Applications to biological systems are still rare. A very convincing example is the study of Dockter et al. [78] on the folding of the major LHCII (light-harvesting chlorophyll a/b complex) of the photosynthetic apparatus. They used a combination of PELDOR and rapid freeze quench to monitor the folding of this complex by means of taking snapshots along the folding trajectory (Figure 6) [78]. An early correlation between MD (Molecular Dynamics) simulations and PELDOR-derived distances from doubly spin-labelled

6 358 G. W. Reginsson and O. Schiemann Figure 8 Orientation-selective PELDOR measurements on DNA (a) Schematic representation of a DNA duplex with the Ç spin labels coloured red. Next to it is a schematic representation of the orientation of the two labels with respect to each other and to r. (b) Field sweep spectrum of the duplex in (a) with the position of the inversion pulse (green profile) and of the detection pulses indicated by arrows. (c) The PELDOR time traces obtained by moving the detection pulses over the field sweep spectrum. Adapted from [83], with permission. DNA double helices has been performed by Schiemann et al. [54]. Marko et al. (A. Marko, V. Denysenkov, D. Margraf, P. Cekan, O. Schiemann, S. T. Sigurdsson and T. F. Prisner, unpublished work) recently studied the breathing dynamics of 12 DNA duplexes using the rigid spin label Ç (see Figure 7) and simulated the orientation-selective X- and G-band (180 GHz) PELDOR time traces with a correlated stretch twist breathing model in which the helix pitch is not changed, but the radius is. ORIENTATION SELECTION At X-band frequencies, the spectrum of a frozen nitroxide sample has a width of approx. 190 MHz, which is governed by the anisotropy of the hyperfine coupling tensor of the nitrogen in the NO-bond. The pulse lengths most commonly used for X-band PELDOR experiments are 16 ns and 32 ns for the π/2 and π pulses of the detection sequence respectively, and 12 ns for the inversion pulse. The pulses of the detection sequence and the pump pulse have thus an excitation bandwidth of 16 and 67 MHz respectively and do therefore only excite a fraction of the nitroxide molecules. Whether this fraction samples all values of θ depends on the flexibility of the spin-labelled biomolecule and the field position at which the pulses are applied. Usually, the inversion pulse is applied at the centre of the cavity and at the maximum of the central line where it excites all orientations although not equally weighted. For nitroxides, the detection pulses are usually applied at a MHz higher frequency than the inversion pulse. At a frequency offset of 90 MHz, the detection sequence excites predominantly the m I = 1component of A ZZ of the 14 N hyperfine tensor, whereas smaller frequency offsets select more of the off-diagonal elements and deselect A ZZ. At small frequency offsets between detection sequence and pump pulse, their excitation profiles overlap in the frequency domain, which means that some spins are excited by both the detection sequence and the pump pulse. These spins do not contribute to V(t) intra which diminishes the PELDOR effect. In addition, the frequency overlap increases the contribution of the unwanted hyperfine artefacts in the PELDOR time trace. Those molecules contribute to the intramolecular PELDOR signal V(t) intra in which one nitroxide spin has been excited by the detection sequence and the second by the pump pulse. If the doubly labelled biomolecule, or at least the nitroxides, are highly flexible then there is no correlation between the hyperfine and g-tensors of the individual nitroxides and also not between the nitroxides hyperfine/g-tensors and the distance vector r AB connecting both. Thus positioning the detection pulses at the low field edge where they excite those molecules with A ZZ parallel to the outer magnetic field B 0, they still do excite all orientations of θ because A ZZ and r AB have no correlation. However, if the labels and the biomolecule is rigid, then selecting a specific hyperfine tensor component selects at the same time a specific value of θ [80]. Acquiring PELDOR time traces with different frequency offsets and subsequently simulating them enables one to determine the orientation of the two nitroxides with respect to r AB. A second method for a direct extraction of orientations from the set of time traces is based on Tikhonov regularization and has been published by Marko et al. [81]. At X-band alone, one would mainly be able to resolve the orientation of A ZZ with respect to r AB ; however, including PELDOR measurements at higher fields/frequencies may allow resolving of all angles. For nitroxides, orientation selectivity has been shown first on bisnitroxide model systems using X- and S-band frequencies [44] and later also at W-band [82]. Although these studies were able to obtain orientation-selective PELDOR time traces, they were hampered by the rather large rotational degree of freedom of the nitroxides used. The laboratory of Sigurdsson recently designed the so-called rigid spin label Ç for nucleic acids (Figure 7, label 6 [29]). The nitroxide ring of this label is fused on to a cytosine base which base-pairs via three hydrogen bonds to an opposing guanine base, which strongly reduces its rotational freedom. Utilizing this label incorporated pairwise into DNA duplexes, Schiemann et al. [83] showed that strong orientation selectivity can already be obtained at X-band frequencies (Figure 8). The laboratories of Steinhoff [84] and Prisner [74] demonstrated that orientational information can also be obtained from X-band PELDOR measurements using the conventional spin label MTSSL if this label s rotational degree of freedom is inhibited by interactions with the protein. Beyond nitroxides, the power of high-field/high-frequency PELDOR in determining orientations has been shown on RNR (ribonucleotide reductase) [85,86]. This study was able to fully resolve the orientation of two tyrosyl radicals within the RNR by stepping the pump and detection pulses with a fixed frequency separation over the tyrosyl radical spectrum. It should be kept in mind that analysis of PELDOR time traces affected by orientation selectivity cannot be performed with the current version of DeerAnalysis (DeerAnalysis2010), since it assumes that all orientations of θ are excited [48].

7 PELDOR beyond distances 359 EXCHANGE COUPLING An indication for a non-zero J is that the dipolar frequencies ν dd for the perpendicular and the parallel orientation of r AB with respect to B 0 do not behave like 1:2. If the isotropic exchange coupling constant J has to be taken into account, the coupling between both spin centres is the sum of the dipolar contribution ν dd and J (eqn 5) [10]: v AB = v dd + J (5) Up to now, no example has been published where an exchange coupling needed to be taken into account for the interpretation of PELDOR data from a biomolecule. The reason for this is that the distances measured with PELDOR are usually above 2 nm. Since the through-space or direct orbital overlap contribution to the exchange coupling constant decays exponentially with the distance, this contribution can be neglected at such large distances. The through-bond contribution can be considerable even at such distances, but only if the connecting bonds provide a conjugated bridge [87 89]. For the biomolecules studied up to now this is not the case, but one might consider this carefully for nucleic acid structures [90]. If J is non-zero, a recent study on bisnitroxide model systems revealed that the magnitude and sign of J can be determined from the PELDOR data [91]. The effect of J on the Pake pattern is that the two halves of this pattern are shifted against each other. How much depends on the magnitude of J and the direction of the shift on the sign as predicted in silico previously [10] and as known from ENDOR (electron-nuclear double resonance) spectra. With the possibility of reading off two frequencies for ν AB at the maximum and the edge of the Pake pattern one can build two equations from eqn (5) which can than be rearranged into eqns (6) and (7), yielding ν dd and J respectively: v dd [MHz] = v v 3 J [MHz] = 2v + v (7) 3 However, as outlined in the same paper [91], this always yields two solutions which one might distinguish using simulations. Also here, a word of warning, the DeerAnalysis program in its current version does not account for J. MULTISPIN SYSTEMS Most of the studies described above have dealt with two coupled spin centres/two nitroxides per biomolecule. However, PELDOR is also suited to resolve distances between more than two nitroxides and it can be used to actually count the number of interacting spin centres. The latter can be very useful if one wants to know how many proteins are coming together to form a complex. In order to count the number of interacting spins in one complex, one is not interested in the frequency of the modulation, but the depth V λ of the modulation. V λ is the signal intensity after intermolecular background correction and at long times t where all modulation in the time trace is damped (Figure 3b) [47]. From the value of V λ, the number of spins coupled in one complex can be calculated according to eqn (8) as verified experimentally by Bode et al. [92] (Figure 9): n = InV λ In (1 λ B ) + 1 (8) (6) Figure 9 spins Dependence of the modulation depth on the number of coupled Monomer (green line), dimers (blue lines), trimers (red line) and tetramer (black line). Reproduced with permission from [92]. Copyright 2007 American Chemical Society. The only free parameter in the calculation of n is λ B, the fraction of spins inverted by the pump pulse. The value of λ B can be determined from an independent PELDOR experiment of a standard biradical measured under the same buffer and spectrometer conditions (pulse lengths, cavity coupling, etc.). From its V λ value, λ B is calculated as 1 V λ. In general, care has to be taken that: (i) all modulation is damped, this requires long time traces, which can be a problem for large distances; (ii) the intermolecular background has to be fitted carefully, as bad fits can give wrong V λ values; (iii) in the case of mixtures of oligomers, one has to work with the weighted ratios and account for dipolar relaxation differences [92]; and (iv) higher-order spin correlations are currently not included in DeerAnalysis2010, so using this program for multi-spin systems may give wrong distance distributions [93]. Milov et al. [94] were the first to use PELDOR to estimate that four trichogin GAIV molecules come together to form a membrane-spanning channel [94] and Banham et al. [95] used it to conclude that three von Willebrand A factors form the active complex [95]. A study on the octameric carbohydrate export channel WzA found good agreement between the modulation depth and the octameric state of the complex in the crystal [96]. Interestingly, for long time windows T, the observed modulation depth of 50% rather corresponded to a dimer. However, shortening the pulse sequence revealed an increase in V λ and, extrapolating V λ back to a pulse sequence length of zero, yielded the V λ of 0.98 expected for an octamer. The rationale for this behaviour was incomplete spin labelling and different relaxation times for octamers carrying different amounts of spin labels. METAL IONS PELDOR can be used to measure distances not only between organic radicals, but also between metal centres or metal centres and organic radicals. Metal centres can be challenging for PELDOR measurements because of (i) very short T M relaxation times leading to short PELDOR time windows and low signalto-noise ratios, (ii) very broad spectral widths due to large g- anisotropies, which decreases the signal-to-noise ratio further and can prevent, due to the limited bandwidth of the resonator, a free

8 360 G. W. Reginsson and O. Schiemann Figure 10 PELDOR time traces (light grey, original time trace; black background, corrected time trace) of protein p75icd labelled 2-fold with the gadolinium complex shown on the far right Adapted with permission from [108]. Copyright 2010 American Chemical Society. choice of pump and probe positions, (iii) high-spin states that can interfere with the PELDOR experiment, and (iv) delocalization of spin density into ligands, in which case the point dipole approximation has to be extended. Copper(II) ions have been used several times for PELDOR measurements because their T M values and spectral width are not too extreme and it is an electron spin S = 1 / 2 system. The laboratory of Huber was the first to measure Cu Cu distances in the covalently linked dimer of the electron transfer protein azurin, showing that such measurements are possible, and gaining information about the arrangement of the two monomers in the complex without spin labelling [97]. Kay et al. [98] published PELDOR measurements between two copper(ii) ions in the dicupric human serum transferrin and lactoferrin [98]. The high-quality data supported the arrangement of the two subdomains. Jeschke s laboratory published in 2002 a first study between two nitroxides and a copper(ii) ion in a model complex demonstrating that the nitroxide nitroxide distance can be individually measured and distinguished from the copper(ii) nitroxide distances by choosing adequate pump and probe positions [99]. Subsequent publications on copper porphyrin nitroxide model systems showed how to deal with spin delocalization into the ligand and how distributions of exchange coupling constants can be mistaken for a distribution in r AB [100,101]. This analysis also included dynamics of the model system and orientation selection. Another thorough analysis also including dynamics of copper(ii) ion materials has been published by Lovett et al. [102,103]. The laboratory of Bittl analysed PELDOR measurements between an iron sulfur cluster and a nickel iron centre in the hydrogenase from Desulfovibrio vulgaris Miyazaki F taking the spin projection factors for the individual iron centres in the cluster into account [104]. Already in the last century, Kawamori s laboratory did PELDOR measurements involving the manganese cluster in PSII in its S = 1 / 2 state [105,106]. The laboratory of Goldfarb is the first who successfully performed PELDOR measurements on a high-spin metal centre, a gadolinium(iii) complex [107]. They could select the m S = + 1 / 2 transition by performing the PELDOR measurements at 95 GHz. An interesting property of this complex is that the relaxation times are rather long and the EPR spectrum is dominated by a single line. This, together with the worked out derivatization of this complex with a thiol group, makes this gadolinium complex a promising candidate for a new spin label (Figure 10) [108]. CONCLUSIONS PELDOR has emerged as a powerful biophysical method and is applied to soluble and membrane proteins, nucleic acids and complexes thereof. The biomolecules can either be spin-labelled with nitroxides or one makes use of intrinsic radical cofactors, amino acid radicals or paramagnetic metal centres. PELDOR not only measures mean distances with high resolution and reliability, but also yields distance distributions, which enable access to dynamics and conformational distributions. It can count the number of singly spin-labelled biomolecules in a complex and, if an exchange coupling mechanism is present, it can be easily disentangled. Importantly it can also resolve the orientation of labels, which makes a correlation of the nitroxide nitroxide distance with the biomolecular structure easier. The latter two properties, together with the distance precision, small labels and that not two different labels are needed makes this method highly complementary to FRET. And indeed, the first papers are appearing where both methods are combined [109]. Also promising are recent technological advances that enable high field/highfrequency PELDOR measurements with high power/short pulses [110,111]. These new spectrometers have increased sensitivity, allowing concentrations as low as 1 μm to be measured [111], and enable highly orientation-selective measurements. FUNDING Work in our laboratory is supported by the Biotechnology and Biological Sciences Research Council [grant numbers BB/H017917/1, F004583/1]. O.S. thanks the Research Councils of the U.K. for a fellowship and G.W.R. thanks the School of Biology, University of St Andrews for a SORS (Scottish Overseas Research Students) Scholarship. REFERENCES 1 Lakowicz, J. R. (2006) Principles of Fluorescence Spectroscopy, 3rd edn, Springer, New York 2 Schweiger, A. and Jeschke, G. (2001) Principles of Pulse Electron Paramagnetic Resonance, Oxford University Press, Oxford 3 Milov, A. D., Salikhov, K. M. and Shirov, M. D. (1981) Application of the double resonance method to electron spin echo in a study of the spatial distribution of paramagnetic centers in solids. Sov. Phys. Solid State 23, Martin, R. E., Pannier,, M., Diederich, F., Gramlich, V., Hubrich, M. and Spiess, H. W. (1998) Determination of end-to-end distances in a series of TEMPO diradicals of up to 2 8 nm length with a new four-pulse double electron-electron resonance experiment. Angew. Chem. Int. Ed. 37,

9 PELDOR beyond distances Pannier, M., Veit, S., Godt, A., Jeschke, G. and Spiess, H. W. (2000) Dead-time free measurement of dipoledipole interactions between electron spins. J. Magn. Reson. 142, Borbat, P. P. and Freed, J. H. (1999) Multiple-quantum ESR and distance measurements. Chem. Phys. Lett. 313, Jeschke, G. (2002) Determination of the nanostructure of polymer materials by electron paramagnetic resonance spectroscopy. Macromol. Rapid Commun. 23, Schiemann, O. and Prisner, T. F. (2007) Long-range distance determinations in biomacromolecules. Q. Rev. Biophys. 40, Klare, J. P and Steinhoff, H. J. (2009) Spin labelling EPR. Photosynth. Res. 102, Milov, A. D., Maryasov, A. G. and Tsvetkov, Y. D. (1998) Pulsed electron double resonance (PELDOR) and its applications in free-radicals research. Appl. Magn. Reson. 15, Borbat, P. P. and Freed, J. H. (2007) Measuring distances by pulsed dipolar ESR spectroscopy: spin-labeled histidine kinases. Methods Enzymol. 423, Tsvetkov, Y. D., Milov, A. D. and Maryasov, A. G. (2008) Pulsed electron-electron double resonance (PELDOR) as EPR spectroscopy in nanometer range. Russ. Chem. Rev. 77, Schiemann, O., Piton, N., Plackmeyer, J., Bode, B. E., Prisner, T. F. and Engels, J. W. (2007) Spin labeling of oligonucleotides with the nitroxide TPA and uses of PELDOR, a pulse EPR method, to measure intramolecular distances. Nat. Protoc. 2, Jeschke, G. (2002) Distance measurements in the nanometer range by pulse EPR. ChemPhysChem 3, Schiemann, O. (2009) Mapping global folds of oligonucleotides with pulsed electron electron double resonance methods. Methods Enzymol. 469, Jeschke, G. and Polyhach, Y. (2007) Distance measurements on spin-labelled biomacromolecules by pulsed electron paramagnetic resonance. Phys. Chem. Chem. Phys. 9, Fajer, P. G., Brown, L. and Song, L. (2007) ESR spectroscopy in membrane biophysics. Biol. Magn. Reson. 27, Berliner, L. J. (ed.) (1998) Spin labeling the next millennium. Biological Magnetic Resonance, vol. 14, Plenum Press, New York 19 Likhtenshtein, G. I., Yamauchi, J., Nakatsuji, S., Smirnov, A. I. and Tamura, R. (2008) Nitroxides: Applications in Chemistry, Biomedicine, and Materials Science, Wiley-VCH, Weinheim 20 Berliner, L. J., Grunwald, J., Hankovszky, H. O. and Hideg, K. (1982) A novel reversible thiol-specific spin label: papain active site labeling and inhibition. Anal. Biochem. 119, Altenbach, C., Marti, T., Khorana, H. G. and Hubbell, W. L. (1990) Transmembrane protein structure: spin labeling of bacteriorhodopsin mutants. Science 248, Becker, C. F. W., Lausecker, K., Balog, M., Kálai, T., Hideg, K., Steinhoff, H.-J. and Engelhard, M. (2005) Incorporation of spin-labelled amino acids into proteins. Magn. Reson. Chem. 43, S34 S39 23 Fleissner, M. R., Brustad, E. M., Kálai, T., Altenbach, C., Cascio, D., Peters, F. B., Hideg, K., Peuker, S., Schultz, P. G. and Hubbell, W. L. (2009) Site-directed spin labeling of a genetically encoded unnatural amino acid. Proc. Natl. Acad. Sci. U.S.A. 106, Zhang, X., Cekan, P., Sigurdsson, S. T. and Qin, P. Z. (2009) Studying RNA using site-directed spin-labeling and continuous-wave electron paramagnetic resonance spectroscopy. Methods Enzymol. 469, Sicoli, G., Wachowius, F., Bennati, M. and Hoebartner, C. (2010) Probing secondary structures of spin-labeled RNA by pulsed EPR spectroscopy. Angew. Chem. Int. Ed. 49, Piton, N., Mu, Y., Stock, G., Prisner, T. F., Schiemann, O. and Engels, J. W. (2007) Base specific spin-labeling of RNA for structure determination. Nucleic Acids Res. 35, Jakobsen, U., Shelke, S., Vogel, S. and Sigurdsson, S. T. (2010) site-directed spin-labeling of nucleic acids by click chemistry: detection of abasic sites in duplex DNA by EPR spectroscopy. J. Am. Chem. Soc. 132, Ding, P., Wunnicke, D., Steinhoff, H.-J. and Seela, F. (2010) Site-directed spin labeling of DNA by the azide-alkyne click reaction: nanometer distance measurements on 7-deaza-2 -deoxyadenosine and 2 -deoxyuridine nitroxide conjugates spatially separated or linked to a da-dt base pair. Chem. Eur. J. 16, Barhate, N., Cekan, P., Massey, A. P. and Sigurdsson, S. T. (2007) A nucleoside that contains a rigid nitroxide spin label: a fluorophore in disguise. Angew. Chem. Int. Ed. 46, Popova, A. M. and Qin, P. Z. (2010) A nucleotide-independent nitroxide probe reports on site-specific stereomeric environment in DNA. Biophys. J. 99, Edwards, T. E., Okonogi, T. M., Robinson, B. H. and Sigurdsson, S. T. (2001) Site-specific incorporation of nitroxide spin-labels into internal sites of the TAR RNA: structure-dependent dynamics of RNA by EPR spectroscopy. J. Am. Chem. Soc. 123, Obeid, S., Yulikov, M., Jeschke, G. and Marx, A. (2008) Enzymatic synthesis of multiple spin-labeled DNA. Angew. Chem. Int. Ed. 47, Shelke, S. and Sigurdsson, S. T. (2010) Noncovalent and site-directed spin labeling of nucleic acids. Angew. Chem. Int. Ed. 49, Hunsicker-Wang, L., Vogt, M. and DeRose, V. J. (2009) EPR methods to study specific metal-ion binding sites in RNA. Methods Enzymol. 468, Kisseleva, N., Khvorova, A., Westhof, E. and Schiemann, O. (2005) Binding of manganese(ii) to a tertiary stabilized hammerhead ribozyme as studied by electron paramagnetic resonance spectroscopy. RNA 11, Kisseleva, N., Kraut, S., Jaeschke, A. and Schiemann, O. (2007) Characterizing multiple metal ion binding sites within a ribozyme by cadmium-induced EPR silencing. HFSP J. 1, Wolfson, A. D., Kisseleva, N., Khvorova, A., Westhof, E. and Schiemann, O. (2008) The different role of high- and low-affinity metal ions in cleavage by a tertiary stabilized cis hammerhead ribozyme from tobacco ringspot virus. Oligonucleotides 18, Riplinger, C., Kao, J. P. Y., Rosen, G. M., Kathirvelu, V., Eaton, G. R., Eaton, S. S., Kutateladze, A. and Neese, F. (2009) Interaction of radical pairs through-bond and through-space: scope and limitations of the point-dipole approximation in electron paramagnetic resonance spectroscopy. J. Am. Chem. Soc. 131, Slichter, C. P. (1980) Principles of Magnetic Resonance, Springer Verlag, Berlin 40 Hustedt, E. J. and Beth, A. H. (2000) Structural information from CW-EPR spectra of dipolar coupled nitroxide spin labels. In Distance Measurements in Biological Systems by EPR (Berliner, L. J., Eaton, S. S. and Eaton, G. R., eds), pp , Kluwer Academic/Plenum Publishers, New York 41 Rabenstein, M. D. and Shin, Y.-K. (1995) Determination of the distance between two spin labels attached to a macromolecule. Proc. Natl. Acad. Sci. U.S.A. 92, Fritscher, J., Beyer, M. and Schiemann, O. (2002) Synthesis, crystal structure and magnetic properties of a novel nitroxide biradical. Theoretical investigation of the exchange mechanisms. Chem. Phys. Lett. 364, Jeschke, G., Bender, A., Paulsen, H., Zimmermann, H. and Godt, A. (2004) Sensivity enhancement in pulse EPR distance measurements. J. Magn. Reson. 169, Weber, A., Schiemann, O., Bode, B. and Prisner, T. (2002) PELDOR at S- and X-band frequencies and the separation of exchange coupling form dipolar coupling. J. Magn. Reson. 157, Banham, J. E., Baker, C. M., Ceola, S., Day, I. J., Grant, G. H., Groenen, E. J. J., Rodgers, C. T., Jeschke, G. and Timmel, C. R. (2008) Distance measurements in the borderline region of applicability of CW EPR and DEER: a model study on a homologous series of spin-labelled peptides. J. Magn. Reson. 191, Kuznetsov, N. A., Milov, A. D., Koval, V. V., Samoilova, R. I., Grishin, Y. A., Knorre, D. G., Tsvetkov, Y. D., Fedorova, O. S. and Dzuba, S. A. (2009) PELDOR study of conformations of double-spin-labeled single and double-stranded DNA with non-nucleotide inserts. Phys. Chem. Chem. Phys. 11, Milov, A. D., Ponomarev, A. B. and Tsvetkov, Y. D. (1984) Electron electron double resonance in electron spin echo: model biradical systems and the sensitized photolysis of decalin. Chem. Phys. Lett. 110, Jeschke, G., Chechik, I. P., Godt, A., Zimmermann, H., Banham, J., Timmel, C. R. and Jung, H. (2006) DeerAnalysis2006: a comprehensive software package for analyzing pulsed ELDOR Data. Appl. Magn. Reson. 30, Ward, R., Bowman, A., Sozudogru, E., El-Mkami, H., Owen-Hughes, T. and Norman, D. G. (2010) EPR distance measurements in deuterated proteins. J. Magn. Reson. 207, Astashkin, A. V., Kodera, Y. and Kawamori, A. (1994) Distance between tyrosines Z+ and D+ in plant photosystem II as determined by pulsed EPR. Biochim. Biophys. Acta 1187, Astashkin, A. V., Hara, H. and Kawamori, A. (1998) The pulsed electron electron double resonance and 2+1 electron spin echo study of the oriented oxygen-evolving and Mn-depleted preparations of photosystem II. J. Chem. Phys. 108, Persson, M., Harbridge, J. R., Hammarstroem, P., Mitri, R., Mårtensson, L.-G., Carlsson, U., Eaton, G. R. and Eaton, S. S. (2001) Comparison of electron paramagnetic resonance methods to determine distances between spin labels on human carbonic anhydrase II. Biophys. J. 80, Schiemann, O., Weber, A., Edwards, T. E., Prisner, T. F. and Sigurdsson, S. T. (2003) Nanometer distance measurements on RNA using PELDOR. J. Am. Chem. Soc. 125, Schiemann, O., Piton, N., Mu, Y., Stock, G., Engels, J. W. and Prisner, T. F. (2004) A PELDOR-based nanometer distance ruler for oligonucleotides. J. Am. Chem. Soc. 126,