Figure1. Schematic representation of solar water splitting into O₂ and H₂ by the enyzmes water-oxidase and hydrogenase. Shown are the structures of the active sites, particularly the Mn₄O_xCa cluster of photosystem II in oxygenic photosynthesis and the ‘H-cluster’ of [FeFe]-hydrogenase (Lubitz et.al., Energy Environ. Sci., 2008).

In this way insight into enzymatic processes is gained, for example into photosynthetic water splitting and (bio)hydrogen production which are the main topics of our research efforts (Fig. 1).

(i) New Methodologies and Instrumentation
One of the aims of our group is to further develop instrumentation and methodology, in particular in the field of multifrequency EPR and related spectroscopies. The special techniques, instruments and frequency bands currently available in our laboratory are summarized in Fig. 2. These span the frequency range from 2 to 244 GHz at fields between 0 and 12 T; for many bands both CW and pulse operation is available. Several instruments allow light access for in situ illumination of samples; recently a new Nd-YAG Quanta-Ray/OPO laser system (Newport Spectra Physics/GWU) has been acquired.

During the last 3 years our home-built CW/pulse Q-band ENDOR resonator has been further optimized and a new Q-band pulse bridge with higher microwave pulse power (5-10 W) was set-up (see report E. Reijerse). In addition to the newly acquired Bruker W-band EPR instrument a specialized self-constructed W-band spectrometer has recently been obtained (from FU Berlin, AG Möbius); both instruments are currently being set-up in our laboratories. Their designs are complimentary and open the possibility to measure a large variety of different samples with high sensitivity and spectral resolution using CW/pulse EPR, ENDOR and ELDOR. This includes also the future investigation of small metalloprotein single crystals.

Figure 2. EPR instruments available in the Institute: basic technique, frequency band, console, temperature range and magnet (for high field only). Note that two W-band and two pulse Q-band instruments exist with different designs and applications.

(ii) DFT and ab initio Calculations
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. The approach is illustrated in Fig. 3. The calculations are performed mostly using the ORCA program package in close collaboration with Prof. Frank Neese (Univ. Bonn and Max Planck Fellow at the institute). The computing facilities at the institute to perform such calculations are being extended (see report F. Neese).

During the report period the focus has been on calculations of

• The spin density distribution and hyperfine couplings of triplet states of chlorophylls and carotenoids (Niklas et al., 2007; Marchanka et al., 2009).
• Hydrogen bond geometries and their impact on the magnetic resonance parameters of radical ions in photosynthesis (Flores et al., 2007; Niklas et al., 2009).
• Magnetic resonance properties and structures of manganese complexes (Zein et al., 2008; Pantazis et al., CEJ, 2009).
• Structure of the water oxidizing complex in PS II (Pantazis et al. PCCP, 2009).
• Calculations on the [NiFe]- and [FeFe] hydrogenases (Dissertation M. Kampa).

Further details are found in the report by F. Neese.

Figure 3.Comparison of spectroscopically obtained and theoretically derived data leading to conclusions about the structure of intermediates and reaction mechanisms.

(iii) Hydrogenase
Structure and function of the enzyme hydrogenase is of central importance for a future biologically based hydrogen production technology and for the design and synthesis of bioinspired model systems for hydrogen conversion catalysis. In our department both the [NiFe] and the [FeFe] hydrogenases are studied in a combined effort by several groups (Reijerse: [FeFe] hydrogenase, Lubitz: [NiFe] hydrogenase, Gärtner: molecular biology and genetics of hydrogenases, Ogata: X-ray crystallography of hydrogenases).

The first crystal structure of a [NiFe] hydrogenase from a photosynthetic bacterium (Allochromatium vinosum) has been obtained to 2.0 Å resolution (Dissertation Kellers, 2008). The structure is similar to that of other known standard hydrogenases from sulfate reducing bacteria except for differences found for the oxidized state Ni-A and the proximal iron-sulfur cluster (see report H. Ogata).

The work on the sulfate-reducing standard hydrogenase Desulfovibrio (D.) vulgaris Miyazaki F, which is also isolated and crystallized in our laboratory, has been continued. In particular the carbon monoxide inhibited states and the light-sensitivity of all intermediates have been measured (Dissertation Pandelia 2009; Pandelia et al., 2009) by monitoring the CO/CN-vibrations of the active site via FTIR; an example is shown in Fig. 4.

Experiments using surface enhanced infrared absorption (SEIRA) spectroscopy have been successfully performed on this enzyme in cooperation with P. Hildebrandt (TU Berlin), (Millo et al., 2009) and first protein film voltametry on hydrogenases has also been carried out in our laboratory. Further details of the electronic structure were obtained on hydrogenase samples of D. vulgaris that are labeled with ⁶¹Ni (Flores et al., 2008) or ⁵⁷Fe (Ogata et al., in prep.). Based on these data a reaction scheme for the activation/deactivation of the enzyme, the catalytic cycle, the inhibition by CO and the light sensitivity has been set up (see Fig. 5).

Figure 4. FTIR difference spectra (light-dark) for the Ni-C→Ni-L conversion at 150 K, the second panel shows the time course of dark adaption. Bottom: Kinetic traces for the above conversion process in H₂O (H₂) and D₂O (D₂). The kinetic isotope effect is ~6; the activation barrier~46 kJ/mol (Kellers et.al, PCCP 2009).

The molecular biology work on D. vulgaris (Dissertation A. Kaur, 2009) is described in the report of W. Gärtner. Interestingly, a new [NiFeSe] hydrogenase has been found in this organism that is now being studied in detail in our laboratory. Furthermore, the sulfur metabolism of this species has been investigated and two key enzymes, the adenylsulfate reductase and the dissimilatory sulfite reductase have been crystallized (see reports of W. Gärtner and H. Ogata).

A particular problem for employing hydrogenases in biotechnological processes is the oxygen sensitivity of most enzymes. Our group has therefore started, in collaboration with M.-T. Guidici-Orticoni (Marseille), to investigate a [NiFe] hydrogenase from the hyperthermophilic oxygentole rant bacterium Aquifex (A.) aeolicus (Dissertation Pandelia, 2009). By FTIR spectroelectrochemistry in solution four redox intermediate states were characterized; they show significantly higher midpoint potentials than those measured for standard hydrogenases. Carbon monoxide was shown to bind to the active site of this hydrogenase much weaker than to that of D. vulgaris, i.e. the enzyme of A. aeolicus is less susceptible to CO inhibition. By EPR redox titrations the types and midpoint potentials of all iron sulfur centers of the electron transfer chain could be determined.

Based on these results a model for the oxygen tolerance of the hydrogenase I of A. aeolicus was proposed that is based on the unusual electronic properties of the ironsulfur cluster proximal to the catalytic NiFe center. EPR experiments performed on the paramagnetic states of this hydrogenase showed that no Ni-A (oxygen-inhibited state) exists, whereas the enzyme shows signals for the oxidized Ni-B and the reduced Ni-C state. By examining the interaction of the active site with the substrate in the catalytically active Ni-C state using HYSCORE and ENDOR spectroscopy, a weakly bound hydride was found that is lost upon illumination. It is proposed that the strength of the hydride bond maybe related to the activity of this enzyme.


Figure 5.Activation process and catalytic cycle of standard [NiFe]hydrogenases; the CO inhibition and light sensitivity of specific states is also shown. The formal oxidation states of Ni and Fe and the proposed bridging ligand X between the metals are depicted. Paramagnetic states are red, EPR-silent states blue (Lubitz et.al., 2007; Pandelia et.al., 2009).

(iv) Water Oxidase
Light-induced water splitting takes place in photosystem II of all organisms performing oxygenic photosynthesis. The locus of water oxidation is a Mn₄O_xCa cluster whose exact structure turned out to be difficult to determine by X-ray crystallography due to radiation damage (photo reduction of the Mn ions). Based on structures derived from XAS performed on PS II single crystals, hyperfine and g tensor data from Q-band ⁵⁵Mn ENDOR/EPR and relaxation data we have developed a model for the electronic structure of the cluster in two states (S₀, S₂) of the water splitting cycle (Kulik et al., 2007). This includes the spin and oxidation states of the 4 Mn ions as well as the exchange coupling between them and the electronic ground state of the coupled cluster. The function of the calcium ion has been probed by replacement of Ca with Sr, which affected only two of the ⁵⁵Mn hyperfine couplings. At present the effect of Ca removal is examined.

All experiments on the water oxidase have been done in collaboration with Johannes Messinger (now at the University of Umea, Sweden), more details are found in his report.

In all catalytic states (S₀ … S₄) of the water splitting cycle no Mn²⁺ has been found. This is, however, expected for the reduced S-states, e.g. the S_-2 state that exhibits a well resolved multiline EPR signal. These states can be prepared via chemical reduction using NO, NH₂OH, or NH₂NH₂. For the analysis of the planned experiments on S_-2 and for corroborating the conclusions on the other S-states we have studied a series of mixed valence dinuclear manganese complexes together with the group of Karl Wieghardt, including Mn^IIMn^III systems, using pulse EPR and ⁵⁵Mn ENDOR spectroscopy at Q-band, Fig. 6. Further model studies comprise manganese complexes with mixed/equal valences and clusters with ¹⁷O labeled bridges.

The data obtained from the various pulse EPR and ENDOR experiments on the water oxidase are important constrains for setting up structural models for the Mn₄O_xCa cluster including its amino acid surrounding. Together with the group of Frank Neese (Universität Bonn, MPG Fellow) we have started to explore these possibilities for manganese complexes of different nuclearity and also developed first models for the water oxidase (Pantazis et al., PCCP 2009).

The final goal of this project is to fully understand the geometrical and electronic structure of the manganese cluster in its various S-states - including water binding, proton release and dioxygen formation, i. e. to unravel the mechanism of light-induced water splitting in Nature. At present our knowledge about this very important process is still very incomplete.
Figure 6. (a) Structure of the Mn(II)Mn(III) complex pivOH and EPR spectrum with simulation; (b) ⁵⁵Mn ENDOR spectra at Q-band recorded at different EPR positions (1-3) with simulations (red) (Epel et.al., to be published).

(v) Radicals, Radical Pairs and Triplet States in Photosynthesis
Time resolved EPR/ENDOR techniques were applied to study short-lived photoexcited states of pigment molecules (chlorophylls, carotenoids). This project has been embedded in the Sonderforschungsbereich 663 (University of Düsseldorf).

The primary targets were the cofactors in reaction centers of oxygenic photosynthesis. In a different DFG project related species in bacterial photosynthesis have been studied (see report of van Gastel).

Figure 7. (a) Field swept echo-detected EPR (Q-band) of the triplet peridinin in the PCP antenna of A. carterae; (b) Davis ENDOR pulse sequence used; (c) single-crystal-like ¹H ENDOR spectrum (Q-band)yielding at least 12 hyperfine couplings including signs; (d) and result of DFT calculation of the spin density distribution (Niklas et al., JACS, 2007).

Further experiments on the unidirectionality of the electron transfer in the bacterial reaction center using various mutants and triplet spectroscopy have been performed (Marchanka et al., 2007) and a basic understanding of the spin density distribution of the triplet states of the isolated pigments has been developed based on an orbital mixing model (Marchanka et al., 2009). Investigations of pigment states in a myoglobin matrix (Dissertation Rangadurai) are under way. For further details see the report by Maurice van Gastel.

Quinone molecules function as electron acceptors inphoto synthetic RCs. 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 where it was shown that a single strong hydrogen bond to this molecule largely determines its structural and functional properties (Niklas et al., 2009). Together with the group of G. Feher (UC San Diego) a detailed study of the strengths and geometries of the hydrogen bonds to the primary quinone acceptor in bacterial reaction centers has been published (Flores et al., 2007) showing the power of Q-band ENDOR on ¹H/²H nuclei together with quantum chemical calculations to get insight into these important protein-cofactor interaction. In cooperation with the group of K. Möbius (FU Berlin) experiments on the radical pairs created in the charge separation process in bRC were performed using high field dipolar EPR spectroscopy (Savitsky et al., 2007), from which not only distances but also relative orientations of the radicals can be obtained in the range of ~10 to 40 Å. This enabled a detailed study of light-induced structural changes in the reaction center (Flores et al., in preparation).

(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 is thus under investigation as medical target either for cancer therapy or for a treatment against bacterial/viral infections. 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 enzyme and the closer surrounding of the tyrosyl radical might therefore help to design new drugs. We investigated the tertiary structure of the mouse RNRdimer by measuring the distance and relative orientation between the two tyrosyl radicals using high field pulsed electron-electron double resonance (PELDOR) in a cooperation with Dr. M. Bennati (University Frankfurt/Main; MPI for Biophysical Chemistry, Göttingen), see Fig. 8 (Denysenkov et al., 2008).
Figure 8. Determination of distance and orientation of the two tyrosine radicals in the R2 dimer of mouse RNR via high field PELDOR (Denysenkov et al.; Angew. Chem., 2008).
Recently an RNR has been described from Chlamydia that contains an FeMn complex, instead of a dinuclear iron cluster, that generates the tyrosyl radical in RNR. Interestingly, a tyrosine radical is not present in this RNR. We have studied together with the group of G. Auling (Hannover) the RNR of Corynebacterium ammoniagenes and found a MnMn complex in this species by X-ray crystallography (Fig. 9, Ogata et al., 2009).

The analysis of the complex manganese EPR spectra, in which also a tyrosine radical is involved, is currently underway in our laboratory.

Figure 9. X-ray crystallographic structure of the Mn-RNR (R2F) from C. ammoniagenis at 1.4 Å resolution showing the metal centers with surrounding amino acids and the active tyrosine radical.

(vii) Protein Models
De novo synthesis of proteins has been pursued in our group to create small systems (“maquettes”) for study in gprotein-cofactor interactions 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 the focus has been on iron-sulfur clusters. It could be shown that clusters spontaneously form in solution with small peptides (Koay et al., 2008). The Fe₄S₄ centers F_A and F_B in PS I were modelled and characterized with various techniques; they showed unusually low redox potentials (Antonkine et al., 2009).

In cooperation with D. Noy (Weizmann Institute, Rehovot,Israel) Fe₄S₄clusters were synthesized and characterized embedded in a novel non-natural α-helical coiled coil protein fold (Grzyb et al., submitted). In a different approach attempts have been made to determine the pathway of Fe₃S₄ cluster formation and to distinguish it from that of Fe₄S₄ (Dissertation Hoppe).

In a project related to “artificial photosynthesis” we have participated in characterizing an engineered reaction center based on modified bacterioferritin that contains a light-excitable Zn-chlorin and a dinuclear manganese cluster coupled by a tyrosine residue (cooperation with T. Wydrzynski, Canberra). It could be shown that photocatalytic oxidation of the Mn site occurs, most probably via the redox active tyrosine (see Fig. 10).

Figure 10. (a) Ribbon diagram of the E. coli bacterioferritin homodimer with two identical subunits each housing a dimetal center (2 Mn ions)and a Zn-chlorin at the interface; (b) estimated distances between the Mn₂, tyrosine and ZnCe₆ compared to (c) the arrangement in the PS II reaction center (Conlan et.al., BBA 2009).