Principal Investigator Ronald J Pace Project q63

Department of Chemistry, Machine VP

Faculty of Science

Co-Investigator Karin A Ahrling

Department of Chemistry, Faculty of Science

Theoretical Studies of the Oxygen Evolving Site of Photosystem II

The oxygen evolving centre of photosystem II (PS-II) catalyses the oxidation of water to molecular oxygen and contains 4 Mn atoms in an as yet undetermined geometry. One intermediate oxidation state of the centre (S2) is EPR active and exhibits two characteristic signals which are believed to arise from different magnetic energy levels of the Mn complex. These are a structured g=2 signal (the multiline) and a broad signal around g=4.1.

In order to understand the interactions that give rise to the multiline signal, a program has been written to simulate its hyperfine structure from the exact solution of a model Hamiltonian. The optimised parameter set used to simulate the signal will aid in the interpretation of the structure of the Mn site. In particular it will indicate how many Mn are involved, as well as the nature of the interactions with the ligands. These interactions in themselves will give an indication of what type of ligand is involved, that is, the amino acid side chain to which it is attached. Preliminary indications are that one imidazole nitrogen (Histidine side chain) is a direct ligand to one Mn, and that the complex has a quasi axial symmetry overall.

The present study is the first to semi-quantitatively reproduce the detailed signal hyperfine structure. A key feature of the simulation is the incorporation, for the first time, of nuclear quadrupole hyperfine effects on both Mn within a bridged dimer.

What are the basic questions addressed?

What is the minimum chemically realistic set of Hamiltonian parameters which need be invoked to give a good simulation of the EPR spectra of the S2 state of the oxygen evolving centre of photosystem II (the so called multiline and 4.1 signals), at all experimental frequencies so far used to examine the problem? In addition, to what extent can the experimentally observed orientation dependence of these spectra be used to determine the arrangement of the Manganese atoms in the thylakoid membrane plane?

What are the results to date and the future of the work?

We now have optimised the fit of the simulated multiline spectrum to the experimental at the three experimental frequencies available to us. The resultant parameter set indicates that the hyperfine and quadrupole interactions arise from Mn ions in low symmetry environments, corresponding approximately to the removal of one ligand from an octahedral in both cases. The hyperfine parameters indicate a quasi axial anisotropy at MnIII, which while consistent with Jahn Teller distortion as expected for a d4 ion, corresponds here with the unpaired spin in the ligand deficient, z-direction of the molecular reference axis. The fitted parameters for MnIV are very unusual, showing a high degree of anisotropy not expected in a d3 ion. This degree of anisotropy can be qualitatively accounted for by a histidine ligand providing [[pi]] backbonding into the metal dxy orbital, together with a weakly bound or absent ligand in the x-direction. The inferred geometry is shown in figure 1. The above result is non-intuitive and to date no model compounds of this nature have been made. The above results will provide a new structure to study synthetically.

Future work will be focussed on the simulation of the g=4.1 signal and on simulating the spectra of one-dimensionally oriented particles. The latter simulations will provide information on the orientation of the catalytic site in the thylakoid membrane system.

FIGURE 1 Interpretation of the parameter set given by the simulation of the multiline EPR signal of the S2 state of the oxygen evolving site of photosystem II. The Mn-dimer is di-u-oxo, u-carboxylato bridged. The histidine is liganded to the MnIV. The dotted circle indicates a distant or very weakly bound ligand. MnIII is 5 coordinate.

What computational techniques are used and why is a supercomputer required?

The program generates a 72x72 Hamiltonian matrix for a large number (~250) of randomly generated molecular orientations. The matrix is evaluated for each orientation and hence the field positions and transition probabilities for the predicted ESR transitions are generated. The sum of these gives the required powder pattern spectrum. The derivative is calculated, point by point, for this simulated spectrum and the sum of the squared differences between the (derivative) experimental and simulated spectra ([[Sigma]](exp.-sim.)2) calculated. The simplex routine then calls the above subroutine, and varies the model Hamiltonian parameters to minimise [[Sigma]](exp-sim)2.

The g=4 program uses the same type of calculations as the multiline program, but it generates 3 smaller matrices. The first (a 4x4) finds the main transitions due to the Zeeman and fine structure terms, and the other two (6 x 6) calculates the hyperfine envelope as a perturbation on the main transitions. Calculations at a larger number of random orientations (~1000) are required to adequately simulate this spectrum.

Publications

Simulation of the S2 state multiline EPR signal of photosystem II - a multifrequency approach, Karin A Åhrling and R J Pace, Biophysical Journal (1995), in press.