New Methodologies and Instrumentation
The array of structural and spectroscopic methods available in the department has been described in the previous report. It is the aim of our group to further develop both the methodology and instrumentation, in particular in the field of EPR and related spectroscopies. Fig. 1 summarizes the main techniques used in the department to study radicals, radical pairs, triplet states and transition metal complexes in biology and in model systems.

Figure 1.Selection of pulse EPR techniques, pulse schemes, spectroscopic observables and general fields of applications.

During the last 3 years advances have been made by installation of new spectrometer control and data acquisition software (SpecMan) for practically all instruments. Spectral simulation and fit programs are now available for most experiments (see report by Dr. Reijerse). The ENDOR facilities were extended (high power amplifier, extension to TRIPLE resonance), an OPO laser (range 410 -2400 nm, power 26 - 2 mJ, resp.) was acquired for in-situ illumination of samples, and our home-built Q‑band cw/pulse ENDOR resonator was slit for light access. First pulsed electron-electron double resonance (PELDOR) experiments at Q‑band have successfully been performed. All these modifications led to new experiments which are documented in the list of publications.

For systems that cannot be initiated by light a stopped flow system has been developed using a novel dielectric ring resonator with a time resolution of 300 µs. (Laßmann et al., 2005). The rapid freeze quench system has also been improved (non-aqueous samples).

The most recent exciting development is the installation of a home-built high field EPR machine working at 122 and 244 GHz using a quasi-optical bridge and a cryogen-free 12 T magnet (for a detailed description see report by Dr. Reijerse). First cw EPR experiments have been done on model systems (radicals and metal complexes) and also on proteins (e. g. the tyrosine in ribonucleotide reductase). In Fig. 2 a comparison of spectra at 34 and 244 GHz are shown, demonstrating the increase in Zeeman (g tensor) resolution (Reijerse et al., 2007). The extension to pulse operation and electron-nuclear and electron-electron double resonance experiments is under way.

(i) DFT and ab initio calculations of spectroscopic parameters
DFT and ab initio calculation are performed in our laboratory to verify the measured spectroscopic parameters and obtain reliable electronic and geometrical structures, e. g. of reaction intermediates, see example in Fig. 3.This is a prerequisite to understand the functional details and mechanisms of the proteins studied in our group. The calculations were performed using the ADF and ORCA program packages (collaboration with Dr. Neese) and the Gaussian 03 Program on the Linux cluster or the SGI Altix SMP server of the institute.

Figure 2.cw EPR at 34 and 244 GHz of the tyrosyl radical in mouse RNR at T = 30K.

In the report period the following main results were obtained.

• For the activation and the catalytic cycle of [NiFe] hydrogenase a reaction mechanism was proposed (Stein and Lubitz, 2004).
• The oxygen inhibition of [NiFe] hydrogenase was studied theoretically in detail (van Gastel et al., 2005).
• The first exchange coupled dinuclear manganese center in a protein was treated theoretically and all relevant magnetic resonance parameters (exchange coupling J; g tensor, isotropic and anisotropic hyperfine (hfc) and nuclear quadrupole couplings (nqc)) were calculated (Sinnecker et al., 2005). Good agreement with the experiment was obtained (Teutloff et al., 2005). Calculations of Mn clusters of higher nuclearity are under way (e. g. the tetranuclear manganese complex of the oxygen evolving complex (OEC) in PS II).
• Hydrogen bonding between the protein and quinone cofactor radicals was theoretically investigated for bacterial and plant photosynthetic reaction centers (RCs). Thereby relationships between experimental data (g, hfc and nqc tensors) and the lengths, geometries and strengths of such H-bonds in the protein environments could beestablished. Furthermore, the tuning of the physical properties of the quinones and possible light-induced structural changes were addressed (Sinnecker et al., 2004, 2006).

Figure 3.Spin density plot (BLYP/DZVP) of a model for the intermediate state Ni-C of the [NiFe] hydrogenase catalytic cycle carrying a hydride bridge between Ni and Fe (Foerster et al., 2005). Note delocalization of spin density (thiolate ligands). Contour value 0.005 e/a₀³

(iii)Hydrogenase
Knowledge of the structure and function of the enzyme hydrogenase is of central importance for a future biologically based hydrogen production technology. In our department both the [NiFe] and the [FeFe] hydrogenases are studied in a combined effort of several groups (Reijerse: [FeFe] hydrogenase, Lubitz: [NiFe] hydrogenase, Gärtner: molecular biology and genetics of hydrogenases).

Most of the hydrogenases are isolated from bacteria grown in the institute. Our group works with the sulfate reducing bacterium Desulfovibrio (D.) vulgaris Miyazaki F; the enzyme is isolated, purified, the activity tested, and it is also crystallized. For spectroscopic studies isotope labeling is performed routinely (⁶¹Ni, ⁵⁷Fe, ¹⁵N, ²H) (Dissertation Goenka) and samples are prepared of the various states; see also report by Prof. Gärtner (molecular biology / genetics).

Since crystal structures only exist for [NiFe] hydrogenases from the same class of bacteria (Desulfovibrio) we started to grow the photosynthetic bacterium Allochromatium (A.) vinosum in a 1100 l tank. The isolation and purification of the catalytic [NiFe] hydrogenase has been established and first single crystals have recently been obtained, which at present diffract to ~ 10 Å resolution (Dissertation Kellers). Furthermore, the bacterium D. desulfuricans is grown and the [FeFe] hydrogenase isolated and purified (Dissertation Wenk).

For the functional characterization of the [NiFe] hydrogenases spectroscopic methods are employed, in particular EPR and FTIR techniques. In the report period the following important results were obtained for the enzyme from D. vulgaris Miyazaki F

• The vibrational spectra of the intermediate states of the enzyme were determined using room and low temperature FTIR (Dissertations Fichtner, Kellers), see report by Dr. Bothe.
• Using spectroelectrochemistry the redox transitions of all states were followed and the midpoint potentials determined both for the enzyme activation process and the catalytic cycle (Fichtner et al. 2006), see also report by Dr. Bothe.
• The enzyme was labeled with ⁶¹Ni (>95%) and the metal hfc studied by EPR and ELDOR-detected NMR (see Fig. 4) in all paramagnetic states.

Figure 4. EPR and ELDOR-detected NMR of ⁶¹Ni-labeled Ni-A in D. vulgaris Miyazaki F.

• The central intermediate Ni‑C was structurally characterized using HYSCORE and ENDOR spectroscopy. It carries a hydride bridge between Ni and Fe as observed earlier by us for the regulatory hydrogenase from Ralstonia (R.) eutropha (Foerster et al., 2005).
• The bridging ligand in the oxidized ready state Ni‑B was unequivocally determined to be a hydroxide ion, which sheds light on the activation process of the enzyme (van Gastel et al., 2006).
  
• The oxygen inhibition was studied on the Ni‑A species in hydrogenase single crystals using ENDOR in combination with H/D exchange; the data were corroborated by DFT calculations (van Gastel et al., 2005). The inhibition of the enzyme by CO is currently under intensive investigation using FTIR and EPR techniques (Dissertation Pandelia).
• The interaction of the active center with the protein was studied and a hydrogen bond detected to a highly conserved histidine residue, which is assigned a functional role in the enzyme (Goenka et al., 2006).

Based on the spectroscopic data and concomitant DFT calculations a general model for both the activation/deactivation and the catalytic cycle of the [NiFe] hydrogenases has been proposed (for details see Stein and Lubitz, 2004;Lubitz et al. 2007).

Recently, a new [NiFe] hydrogenase from an oxygen-tolerant bacterium (A. ferrooxidans) has been characterized by us using different methods including FTIR and EPR spectroscopies(Schröder et al., 2007)

Work done on the [FeFe] hydrogenase using for exampleENDOR and HYSCORE on ⁵⁷Fe and ¹³C labeled hydrogenase (Dissertation Silakov) is described by Dr. E. Reijerse in his report.

Our results on the bimetallic hydrogenase enzymes have let to invitations to write three review articles with different objectives, which will be published in 2007 (Lubitz et al, Met. Ions Life. Sci.; van Gastel and Lubitz, Biol. Magn. Res. Lubitz et al., Chem. Rev.)

(iv) Water oxidase
In oxygenic photosynthesis light-induced water splitting takes place on a Mn₄CaO_x complex which is located in photosystem II. The structure of the manganese complex is difficult to determine from X-ray crystallography of PS II single crystals due to severe radiation damage of the cluster. However, recently a detailed structure has been obtained from X‑ray absorption spectroscopy (see report by Dr. Messinger). In our group we have studied two paramagnetic states (S_eff = ½) of the catalytic water splitting cycle by pulse EPR and ⁵⁵Mn ENDOR spectroscopy in two different frequency bands (X‑ and Q‑band), see Fig. 5.

Fig. 5⁵⁵Mn ENDOR at X- and Q-band of the S₂ state of PS II (BBY particles, 3% methanol) with simulations (hfc tensors of 4 coupled Mn nuclei); Top: Q-band FSE EPR spectrum of the S₂ state with simulation, (g and ⁵⁵Mn hfc data set is given).

The spectroscopy has been made possible by a series of advances in spectrometer design, including the probe head (resonator) and the data acquisition software.

From a careful analysis the g tensor and all four ⁵⁵Mn hfc tensors were obtained. This data set led to several important conclusions on the electronic structure of the water splitting complex (Kulik et al, 2007, submitted; see also report of Dr. Messinger).

• all 4 Mn ions are coupled in the complex, detailed coupling schemes have been obtained
• the S_o state has a Mn(III, III, III, IV), the higher oxidized S₂ state a Mn(III, IV, IV, IV) configuration, Mn(II) is not involved in the normal S‑state cycle
• with the help of other data specific oxidation states can be assigned to the 4 Mn ions in the structure
• a structural change in one of the bridges could be assigned for the S_o®S₁®S₂ transition.
• a model for binding of the 2 substrate water molecules is proposed.

Furthermore, a large array of manganese model compounds has been studied in collaboration with Prof. Wieghardt’s group (Teutloff et al. 2005), e.g. ⁵⁵Mn ENDOR experiments could be performed at 35 GHz both on dinuclear Mn(III)(Mn(IV) and Mn(II)Mn(III) complexes. (Epel et al., 2007, in preparation). Furthermore, a new Mn(II)MnII) complex in a protein was characterized by EPR (Epel et al., 2005).

Together with the group of F. Neese and L. Noodleman (Scripps Institute, La Jolla, CA) we succeeded to calculate for the first time all magnetic parameters of an exchange coupled Mn(III)Mn(IV) complex using a broken symmetry DFT approach (Sinnecker et al., 2004). The technique has also been applied to Mn‑catalase (Sinnecker et al., 2005). We are currently working on the application of this approach to the tetranuclear Mn cluster in PS II in collaboration with the group of Prof. Neese.

(v)Radicals, Radical Pairs and Triplet States in Photosynthesis
In the framework of a project in the Sfb 663 (University of Düsseldorf) time resolved EPR/ENDOR techniques are applied to study short-lived photoexcited states of pigment molecules (chlorophylls, carotenoids). The primary target are the cofactors in reaction centers of oxygenic photosynthesis. Fig. 6 shows the highly resolved orientation-selected Q‑band ENDOR of the triplet of the primary electron donor ³P680 in PS II. The RC triplet states show a strong polarization and a characteristic intensity pattern in the spectra resulting from the mechanism of triplet formation (radical pair recombination). Such and related experiments yield the electron spin distribution together with information about the zero field, g, hyperfine and nuclear quadrupole tensors. Time resolved experiments give detailed insight into triplet formation/decay and triplet transfer/delocalization and thus allow a full characterization of the system. This is valuable to understand excited state structure and dynamics in photosynthesis. Studies of the isolated pigments especially in single crystalline matrices (e. g. in a myoglobin matrix) are carried out in parallel to the in vivo studies.

Figure 6.Time resolved (pulse) ¹H ENDOR of the laser-induced triplet state ³P680 in PS II on two EPR field positions (extreme high/low field Z_I/Z_II) (Dissertation Niklas).

In a separate DFG project triplet states in bacterial photosynthetic RCs are studied (Dr. van Gastel). First results on native and mutant RCs were obtained. For example, the mechanism of triplet transfer from the primary donor to carotenoid has been studied and the first ENDOR of the very short-lived (~ 10 µs) ³Car could be detected (Dissertation Marchanka). Single crystal experiments are envisaged both on the RC and on the isolated pigments. Triplet states were also investigated from pigment molecules in de novo synthesized proteins (4-helix bundles), e. g. of Zn-chlorins (Dissertation A. Mennenga).

In photosynthetic RCs quinone molecules function as electron acceptors. Their properties are strongly influenced by interactions with the protein environment, in particular by hydrogen bonding. This effect has been studied for the phylloquinone acceptor in PS I (Teutloff et al., 2004, Pushkar et al., 2004, Niklas, dissertation) and together with the group of G. Feher (UC San Diego) for the ubiquinone acceptors in bacterial RCs (Flores et al., 2006, 2007) using EPR/ENDOR experiments together with DFT calculations performed on the radical anions created in the charge separation process. Our comprehensive theoretical analysis led to a profound understanding of such protein-cofactor interaction (Sinnecker et al., 2004, 2006).
High-field PELDOR experiments on the charge separated radical pair state  in Zn-substituted bacterial RCs demonstrate the power of this new technique to detect distances and in particular the relative orientations of the spin-carrying molecules (cooperation with Prof. K. Möbius, FU Berlin). This enables the detection of light-induced structural changes which are of central importance to understand the single electron transfer processes in photosynthesis and other systems (Savitsky et al., 2007).

Together with Prof. A. Holzwarth we have studied the electron transfer in PS I and postulated a new mechanism for the primary charge separation step in this type of RC (Holzwarth et al., 2005, 2006).

(vi) Ribonucleotide Reductase
The enzyme ribonucleotide reductase (RNR) produces all four deoxyribonucleotides, the basic building blocks of DNA. They are essential for DNA synthesis and repair in all organisms. RNR in mammals, especially human RNR, are under investigation as medical target either for cancer therapy or for a treatment against bacterial/viral infections. Mammalian RNR stores a stable tyrosyl radical that is required for enzymatic activity and can be destroyed by radical quenchers. Knowledge about the structure of the dimeric enzyme and the closer surrounding of the tyrosyl radical might thus help to design new drugs. Mouse RNR is nearly identical to human RNR but easier to obtain; a crystal structure is not yet available. We investigated the tertiary structure of the dimer by measuring the distance between the two tyrosyl radicals (Biglino et al., 2006), see Fig. 7. More detailed measurements at high magnetic field (180 GHz) to obtain the relative orientation of the two radicals are under way in collaboration with Dr. M. Bennati (Frankfurt/Main). Further we investigated the electronic structure of the tyrosyl radical in both mouse RNR and Mycobacterium tuberculosis RNR by pulse and cw ENDOR and HYSCORE experiments (Schmidt et al, to be published).

Figure 7.Pulse electron-electron double resonance (PELDOR) trace and distance distribution of the 2 tyrosine radicals in mouse RNR (R2-protein dimer) at T = 6 K. Distance between radicalsr = 3.25 nm (Biglino et al., 2006).

(vii)Protein Models
De novo synthesis of proteins has been pursued in our group to create small systems (“maquettes”) for studying protein-cofactor interaction and to learn about the minimal requirements to obtain functional systems (e. g. for electron transfer or catalysis) This project is carried out in cooperation with Prof. W. Gärtner. In the report period hemes, Zn‑chlorins and Zn‑pheophorbides were introduced into 4-helix bundles, the binding modes and geometries were characterized and the optical and magnetic resonance properties studied (Dissertation A. Mennenga), see Fig. 8. Light-induced electron transfer from the photoexcited pigments to exogenous quinone was obtained in liquid solution (Mennenga et al., 2006). Furthermore, model peptides for bacterial antenna systems (LH1 and LH2 from Rhodobactersphaeroides) were obtained by de novo synthesis.
Figure 8.Top: redox titration of protein maquette FePPIX-me-1 (64 amino acids, 2 cofactors, 4 helix bundle); Bottom: the ESI-MS spectrum (holoprotein and intact proteins with one and two cofactors bound are detected)
To model iron-sulfur centers several small peptides based on the sequence of the F_A, F_B centers in PS I were created and the [4Fe4S]-centers could be formed, which showed unusually low redox potentials. The [4Fe4S] clusters werecharacterized using different spectroscopic techniques (Dissertation Breitenstein; Antonkine et al., in prep.).