Electrostatic calculations have predicted that the partial negative charge associated with D575PsaB plays a significant role in modulating the midpoint potentials of the A1A and A1B phylloquinones in photosystem I. To test this prediction, the side chain of residue 575PsaB was changed from negatively charged (D) to neutral (A) and to positively charged (K). D566PsaB, which is located at a considerable distance from either A1A or A1B, and should affect primarily the midpoint potential of FX, was similarly changed. In the 575PsaB variants, the rate of electron transfer from A1A to FX is observed to decrease slightly according to the sequence D/A/K. In the 566PsaB variants, the opposite effect of a slight increase in the rate is observed according to the same sequence D/A/K. These results are consistent with the expectation that changing these residues will shift the midpoint potentials of nearby cofactors to more positive values and that the magnitude of this shift will depend on the distance between the cofactors and the residues being changed. Thus, the midpoint potentials of A1A and A1B should experience a larger shift than will FX in the 575PsaB variants, while FX should experience a larger shift than will either A1A or A1B in the 566PsaB variants. As a result, the driving energy for electron transfer from A1A and A1B to FX will be decreased in the former and increased in the latter. This rationalization of the changes in kinetics is compared with the results of electrostatic calculations. While the altered amino acids shift the midpoint potentials of A1A, A1B, and FX by different amounts, the difference in the shifts between A1A and FX or between A1B and FX is small so that the overall effect on the electron transfer rate between A1A and FX or between A1B and FX is predicted to be small. These conclusions are borne out by experiment.

Hydrogen bonding between the protein and one or both of the two 1,4-quinone carbonyl groups of a benzo-of naphtho-quinone constitutes a significant protein-cofactor interaction in photosynthetic reaction centers. The redistribution of charge and spin density due to a particular H-bonding scheme leaves the largest hyperfine couplings (hfc) at the highest density positions, i.e., the nuclei of the carbonyl groups directly involved in H-bonding. The spin density changes at the ring carbon positions ate accessed experimentally via electron paramagnetie resonance-determined hfc tensor elements of selective ~3C isotope labels in one of the two carbonyl groups. Complete hfc tensor data are presented for each of the ~3C positions in the functional charge-separated state in reaction centers of photosystem I (PS I) isolated from cyanobacteria. A highly asymmetric H-bonding scheme for the A t quinone binding site due to a single dominant H-bond to one carbonyl group is confirmed. A comparison to other weU-studied quinone binding sites of other protein-cofactor systems with more complex H-bonding schemes reveals the uniqueness of the PSI site. The single-sided A~ quinone site provides ah ideal test case for the various sets of density functional theory (DFT) calculations that are currently available. While the overall agreement between experimental and calculated data is quite satisfactory, a significant discrepancy is found for the high-spin-density t3C position associated with the H-bonded carbonyl. The dominant hfc component (and spin density) is underestimated in the DFT calculations, not only for the high-asymmetry case in PS I, but also for other quinone binding sites with less asymmetry that result from more complex H-bonding schemes. The consequences and potential relevance of this finding for biological function are discussed.

Photosystem I (PS I) contains two molecules of phylloquinone that function as electron transfer cofactors at highly reducing midpoint potentials. It is therefore surprising that each phylloquinone is hydrogen bonded at the C(4) position to the backbone -NH of a Leu residue since this serves to drive the midpoint potential more oxidizing. To better understand the role of the H-bond, a PS I variant was generated in which L722(PsaA) was replaced with a bulky Trp residue. This change was designed to alter the conformation of the A-jk(1) loop and therefore change the strength of the H-bond to the PsaA-branch phylloquinone. Transient EPR studies at 80 K show that the A(1A) site in the PS I variant is fully occupied with phylloquinone, but the absence of methyl hyperfine couplings in the quinone contribution to the P(700)(*+)A(1)(*-) radical pair spectrum indicates that the H-bond has been weakened. In wild-type PS I, reduction of F(A) and F(B) with sodium dithionite causes a approximately 30% increase in the amplitude of the P(700)(*+)A(1)(*-) transient EPR signal due to the added contribution of the PsaB-branch cofactors to low temperature reversible electron transfer between P(700) and A(1A). In contrast, the same treatment to the L722W(PsaA) variant leads to a approximately 75% reduction in the amplitude of the P(700)(*+)A(1)(*-) transient EPR signal. This behavior suggests that A(1A) has undergone double reduction to phyllohydroquinone, thereby preventing electron transfer past A(0A). The remaining 25% of the P(700)(*+)A(1)(*-) radical pair spectrum shows an altered spin polarization pattern and pronounced methyl hyperfine couplings characteristic of asymmetric H-bonding to the phylloquinone. Numerical simulations of the polarization pattern indicate that it arises primarily from electron transfer between P(700) and A(1B). The altered reduction behavior in the L722W(PsaA) variant suggests that the primary purpose of the H-bond is to tie up the C(4) carbonyl group of phylloquinone in a H-bond so as to prevent protonation and hence lower the probability of double reduction during periods of high light intensity.