Simultaneous Binding of Fluoride and NO to the Nonheme Iron of Photosystem II: Quantitative EPR Evidence for a Weak Exchange Interaction between the Semiquinone QA and the Iron-Nitrosyl

Simultaneous Binding of Fluoride and NO to the Nonheme Iron of Photosystem II: Quantitative EPR Evidence for a Weak Exchange Interaction between the Semiquinone QA and the Iron-Nitrosyl
of 10
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
  Simultaneous Binding of Fluoride and NO to the Nonheme Iron of Photosystem II: Quantitative EPR Evidence for a Weak ExchangeInteraction between the Semiquinone Q A - and the Iron-NitrosylComplex Yiannis Sanakis, † Doros Petasis, † Vasili Petrouleas, ‡ and Michael Hendrich* ,† Contribution from the Department of Chemistry, Carnegie Mellon Uni V  ersity, 4400 Fifth A V  enue,Pittsburgh, Pennsyl V  ania 15213, and Institute of Material Science, NCSR “Democritos” 15310 AghiaParaske V  i Attikis, Greece Recei V  ed February 19, 1999. Re V  ised Manuscript Recei V  ed July 19, 1999 Abstract:  The effects of NO and fluoride on the iron-quinone complex of the acceptor side of PhotosystemII (PSII) are examined by X- and Q-band EPR spectroscopy. It is found that the EPR signal of the iron-nitrosyl complex changes upon addition of F - . The change is determined to be due to a superhyperfine interactionbetween the electronic spin ( S   )  3  /  2 ) and the fluorine nuclear spin (  I   )  1  /  2 ), indicating that both F - and NOare bound to the same nonheme iron. To the best of our knowledge, this is the first report of simultaneousbinding of both F - and NO to the same iron of a mononuclear nonheme protein. On the basis of an analysisof the hyperfine interaction, a cis F - Fe - NO coordination is indicated. Upon illumination of (NO, Cl - )- or(NO, F - )-treated PSII membranes at 200 K, new X- and Q-band EPR signals are observed in  B 1 || B  and  B 1 ⊥ B modes. Quantitative simulations of these signals provide an unambiguous assignment to the iron-quinonecomplex, Q A - { FeNO } , 7 of the acceptor site of PSII. The exchange interaction between the iron-nitrosyl complex( S  ) 3  /  2 ) and the semiquinone Q A - radical ( S  ) 1  /  2 ) is determined to be  J  )+ 0.5 cm - 1 (F - ) and + 1.3 cm - 1 (Cl - ) for  H  ex )  J  S 1 S 2 . A distribution in the exchange coupling is required to satisfactorily simulate the EPRspectra. This distribution is correlated to small structural variations of the iron-quinone acceptor side of PSII.The terminal electron acceptor of Photosystem II (PSII) 1 isan iron-quinone complex. On the basis of ample spectroscopicevidence 2 and sequence homologies with the structurallycharacterized photosynthetic bacteria, 3 it is generally acceptedthat this complex contains two quinone molecules, Q A  and Q B ,separated by a nonheme iron. The two quinones operate assequential electron acceptors: Q A  is a one- and Q B  a two-electron acceptor. The iron is coordinated by four histidines,two from each of two subunits. The 5 and 6 coordinationpositions of the iron differ for bacteria and PSII. In bacteria,these positions are occupied by a bidentate glutamate residue, 3 whereas in PSII at least one of these positions is occupied bybicarbonate. 4 FTIR difference spectroscopy indicates that bi-carbonate binds as a bidentate ligand in the Fe 2 + and amonodentate in the Fe 3 + state. 4c Various molecules have beenshown to bind in competition with bicarbonate, resulting inreversible deceleration of the electron transfer in PSII. 2 A direct demonstration of exogenous ligand binding at theiron site of PSII has been provided by treatment with NO. 4a,b Binding of NO in competition with bicarbonate results in theformation of an  { FeNO } 7 complex ( S   )  3  /  2 ) giving rise to acharacteristic EPR signal near  g  )  4.0. 5 In addition, cyanideand carboxylates bind at the nonheme iron site in apparentcompetition with NO or bicarbonate. 6,7 The studies of the effectsof the cyanide on the nonheme iron suggest that up to 3 sites ator in the vicinity of the iron are accessible by exogenousligands. 8 In the present study, we provide unequivocal evidencethat both F - and NO will simultaneously bind to the same iron.Illumination of PSII at 200 K induces single electron transferand formation of the Q A - Fe 2 + state at the electron acceptor site.The state has a half-integer spin configuration that gives rise toEPR signals at  g  e  2.0. These EPR signals are attributed to amagnetic interaction between Q A - and the nearby Fe 2 + ( S  ) 2)ion and are strongly affected by the nature of the labile ligandsof the iron. 8,9,10 The analogous signal at  g ) 1.82 in the reactioncenter of the photosynthetic bacterium  R. Sphaeroides 11 has beensimulated with a weak anisotropic exchange interaction (  J  iso ≈ * To whom correspondence should be addressed. Phone: 412-268-1058.Fax: 412-268-1061. E-mail: hendrich@andrew.cmu.edu. † Carnegie Mellon University. ‡ Institute of Material Science.(1) Abbreviations: PSII, Photosystem II; BBY membranes, thylakoidmembrane fragments enriched in PSII; DCMU, 3-(3,4-dichlorophenyl)-1,1-dimethylurea; Q A , Q B , the primary and secondary plastoquinone electronacceptors of PSII, respectively; EPR, electron paramagnetic resonance; MES,4-morpholineethanesulfonic acid; Tris, 2-amino-2-hydroxymethylpropane-1,3-diol; EDTA; ethylenediaminetetraacetate.(2) Diner, B. A.; Babcock, G. T. In  Ad  V  ances in Photosynthesis:Oxygenic Photosynthesis: The Light Reactions ; Ort, D. R., Yocum, C. F.,Eds; Kluwer Academic Publishers: Dordrecht, The Netherlands, 1996; Vol.4, pp 213 - 247. (b) Diner, B. A.; Petrouleas, V.; Wendoloski, J. J.  Physiol.Plant   1991 ,  81 , 423 - 436.(3) Deisenhofer, J.; Michel, H.  EMBO J.  1989 ,  8  , 2149 - 2169.(4) Petrouleas, V.; Diner, B. A.  Biochim. Biophys. Acta  1990 ,  1015 , 131 - 140. (b) Diner, B. A.; Petrouleas, V.  Biochim. Biophys. Acta  1990 ,  1015 ,141 - 149. (c) Hienerwadel, R.; Berthomieu, C.  Biochemistry  1995 ,  34 ,16288 - 16297.(5) The  { FeNO } 7 nomenclature refers to the covalent FeNO complexwhich has seven d-electrons.(6) Deligiannakis, Y.; Petrouleas, V.; Diner, B. A.  Biochim. Biophys. Acta  1994 ,  1188  , 260. (b) Petrouleas, V.; Deligiannakis, Y.; Diner, B. A  Biochim. Biophys. Acta  1994 ,  1188  , 271.(7) Vermaas, W. F. J.; Rutherford, A. W.  FEBS Lett.  1984 ,  175 , 243.(8) Koulougliotis, D.; Kostopoulos, T.; Petrouleas, V.; Diner, B. A.  Biochim. Biophys. Acta 1141 ,  1993 , 275 - 282. (b) Sanakis, Y.; Petrouleas,V.; Diner, B. A.  Biochemistry  1994 ,  33 , 9922 - 9928. 10.1021/ja990533b CCC: $18.00 © xxxx American Chemical SocietyPAGE EST: 9.1Published on Web 00/00/0000  0.3 cm - 1 ) between the semiquinone radical ( S   )  1  /  2 ) and thehigh-spin ferrous ion ( S   )  2), 12a or alternatively by a weakerexchange interaction plus a dipolar term. 12b The  g  )  4 EPR signal from the  { FeNO } 7 complex of PSIIvanishes upon illumination at 200 K, but photodissociation of NO cannot account for this elimination. 4a,b Instead, the loss of the  g  )  4 signal has been attributed to the light-inducedformation of the state Q A - { FeNO } , 7 which is an overall evenelectron configuration. No EPR signals attributable to this statehave been previously observed. In the present study, we havediscovered new signals from the light-induced Q A - { FeNO } 7 state, which establishes the presence of this state and itscorresponding electronic properties. Experimental Methods PSII-enriched thylakoid membranes were isolated from marketspinach by procedures described elsewhere. 13 Some experiments weredone in Mn-depleted PSII membranes after treatment with Tris. 14 Samples for EPR measurements were suspended in 0.4 M sucrose, 15mM NaCl, and 50 mM MES, pH 6.5, at 4 - 6 mg/mL of chlorophyll.Prior to NO treatment the samples were incubated with variable amountsof NaX (X ) Cl, I, Br, F) at 0  ° C in the dark. The NO treatment wascarried out anaerobically at 0  ° C in EPR tubes by slowly bubbling 5mL of a mixture of NO and N 2  of a given ratio, typically 2/3 or 1/4.This resulted in reproducible concentrations of NO of 0.5 - 0.7 or 0.1 - 0.15 mM, respectively, as measured by the characteristic EPR peak of NO in the  g ) 2 region. 4a,b Illumination of the samples was performedat 200 K in a dry ice/acetone with a 340 W projection lamp filteredthrough a solution of CuSO 4 . The samples were then returned to liquidnitrogen temperature prior to spectral measurements.X-band EPR spectra were recorded on a Bruker spectrometerequipped with an Oxford ESR 910 cryostat for low-temperaturemeasurements and a Bruker bimodal cavity for generation of themicrowave fields parallel and transverse to the static field. Q-band EPRspectra were recorded on a Bruker spectrometer equipped with aspecially designed low-temperature microwave probe and cryogenicsystem of our own construction. 15 For both instruments, the microwavefrequency was calibrated by a frequency counter and the magnetic fieldwith a gaussmeter. All experimental data were recorded with nonsat-urating microwave power unless otherwise noted. Analysis of Spectra.  In this section we collect the relevant termsfor the analysis of the spectra. While many parameters are introduced,we will demonstrate that the richness of the spectra allows anunambiguous determination of parameters, first for the  { FeNO } 7 complex and then for the Q A - { FeNO } 7 complex. For the interpretationof the EPR spectra, we use the spin Hamiltonian,where  J   is the isotropic exchange coupling between the iron-nitrosyland semiquinone radical,  D Fe  and  E  Fe  are the axial and rhombic zero-field parameters for the iron-nitrosyl site, and  g Fe  and  g R  are the intrinsicg-tensors of the iron-nitrosyl and semiquinone, respectively. The iron-nitrosyl g-tensor may vary from 2 due to second-order spin - orbitinteractions. The g-tensor for the quinone is known to be isotropic  g R )  2. 8b,16 The last term is the hyperfine contribution from the fluorinenucleus to the iron-nitrosyl site,  H  F ) S Fe A F I F , where  A F  is the fluorinehyperfine coupling tensor.Simulations of the EPR spectra are calculated from diagonalizationof eq 1. In dark-adapted samples, the quinone is oxidized ( S  R ) 0) andtherefore eq 1 is simplified by the absence of the first and fifth terms.We assume that the weak perturbation of the illumination, which causesthe transfer of an electron to the quinone, does not modify theparameters  D Fe  and  E  Fe . However, these parameters are modified bybinding of the fluorine. The nuclear hyperfine term is treated withsecond-order perturbation theory. The powder pattern is generated fora uniform spherical distribution of the magnetic field vector  B . Thetransition intensities are calculated from Fermi’s Golden rule using theeigenfunctions given by the diagonalization. The spectral line shape isgenerated for distributions of the parameters  J   and  E   /   D , specified asone standard deviation  σ   J   and  σ   E   /   D , respectively. A residual line width σ  B  in the magnetic field is specified, which is comparable to themagnitude of the magnetic field modulation. Gaussian spin-packets,properly normalized for field-swept spectra, 17 are folded into thespectrum at each resonance field position. Our treatment of the lineshape is unique in that it is based on the physical description of thespin system given by eq 1, rather than on the common practice of phenomenological line widths in the magnetic field. Thus, fewerparameters are required and these parameters have physical significance.The simulations are generated with careful consideration of allintensity factors, both theoretical and instrumental, to allow directscaling of spectra to sample concentrations. The only unknown variablequantity, relating spin concentration to signal intensity, is an instru-mental factor that depends on the microwave detection system. Thisfactor is then determined by a spin standard, CuEDTA, for which thecopper concentration was accurately determined from plasma emissionspectroscopy.We present below the predictions of eq 1 in graphic form for thecase  J   /   D  <  0.1, which, as will be shown, applies to the present data.For an  S  Fe  )  3  /  2  site (e.g.,  { FeNO } 7 ) and an  S  R  )  1  /  2  site (e.g., thesemiquinone radical, Q A - ), and an exchange interaction that is weakrelative to the zero-field splitting (  J  ,  D ), the lowest four energy levelscan be calculated with perturbation theory to obtain analytic expressionsfor the observed  g -values of the transitions as a function of thedimensionless parameter  J   /   f  , where  f   is the microwave frequency. Theexpressions for the energies of the lowest four levels are derived inAppendix 1. Figure 1 shows a plot of the experimentally observed g -values for magnetic fields along the principal axis directions and for  E   /   D ) 0.020. This latter value is selected to correspond approximatelyto the present spectra. Small variations in  E   /   D  do not significantly affectthe data of Figure 1. This plot provides insight into the interpretationof resonance patterns, which is not always clear from simulations alone.There are three groups of resonances with significant intensity labeledon the plot (I, II, III), with subscripts referring to the principal magneticfield directions. Observed resonance positions can be read off of thisplot; for example at X-band (9.5 GHz, 0.317 cm - 1 ) with  J  ) 0.4 cm - 1 ,resonances would be observed near  g ) 2.5 (272 mT), 9 (85 mT), and5 (153 mT) from groups I, II, and III, respectively. ResultsEffects of F - on the  { FeNO } 7 EPR Signal.  PSII membranespreviously treated with 25 mM NaCl or NaF in the dark for 1h were further incubated with NO for 45 min at 0  ° C. Figure 2shows X-band (9 GHz) EPR spectra of the Cl - (A) and F - (B)samples. As previously reported in the presence of Cl - , 4a,b weobserve a nearly axial signal with  g  y  )  4.10 and  g  x  )  3.97 (9) Rutherford, A. W.; Zimmerman, J. L.  Biochim. Biophys. Acta  1984 , 767  , 168 - 175. (b) Nugent, J. H. A.; Diner, B. A.; Evans, M. C. W.  FEBS  Lett.  1981 ,  124 , 241 - 244. (c) Nugent, J. H. A.; Doetschman, D. C.;MacLachlan, D. J.  Biochemistry  1992 ,  31 , 2935 - 2941.(10) Petrouleas, V.; Sanakis, Y.; Deligiannakis, Y.; Diner, B. A. In  Research in Photosynthesis ; Murata, N., Ed.; Kluwer Academic Publish-ers: Dordrecht, The Netherlands, 1992, Vol. II, pp 119 - 122.(11) Dutton, P. L.; Leigh, J. S.; Reed, E. W.  Biochim. Biophys. Acta 1973 ,  292 , 654 - 664.(12) Butler, W. F.; Calvo, R.; Fredkin, D. R.; Isaacson, R. A.; Okamura,M. Y.; Feher, G.  Biophys. J  .  1984 ,  45 , 947 - 973. (b) Dismukes, G. C.;Frank, H. A.; Friesner, R.; Sauer, K.  Biochim. Biophys. Acta  1984 ,  764 ,253 - 271.(13) Berthold, D. A.; Babcock, G. T.; Yocum, C. F.  FEBS Lett.  1981 , 134 , 231 - 234. (b) Ford, R. C.; Evans, M. C. W.  FEBS Lett  .  1983 ,  160 ,159 - 164.(14) For a review on the treatments affecting the Mn-complex of PSII,see Debus, R. J.  Biochim. Biophys. Acta  1992 ,  1102 , 269 - 352.(15) Petasis, D. T.; Hendrich, M. P.  J. Magn. Reson.  1999 ,  136  , 200 - 206.(16) Klimov, V. V.; Dolan, E.; Shaw, E. R.; Ke, B.  Proc. Natl. Acad.Sci. U.S.A.  1980 ,  77  , 7227 - 7231.(17) Pilbrow, J. R.  Transition Ion Electron Paramagnetic Resonance ;Clarendon Press: Oxford, U.K., 1990; pp 211 - 259, 341, 631 - 638.  H  S )  JS  Fe S  R +  D Fe ( S   z 2 - S  ( S  + 1)/3) Fe +  E  Fe ( S   x 2 - S   y 2 ) Fe +  β  B g Fe S  Fe +  β  B g R S  R +  H  F  (1) B  J. Am. Chem. Soc. Sanakis et al.  from the  S   )  3  /  2  FeNO complex. 18 This signal remainsunaffected over a broad range of Cl - concentrations (0.02 - 100 mM). In the presence of F - , the signal changes to givetwo absorption peaks and two derivative features. Figure 2C,Dshows Q-band (35 GHz) spectra of the same samples in thepresence of Cl - or F - , respectively. The spectra are plotted forequal  g -value scales at the two frequencies. The Q-bandspectrum of the chloride sample is essentially the same, whereasthe Q-band spectrum of the fluoride sample collapses to showa typical rhombic species, with  g -values of   g  y ) 4.20 and  g  x ) 3.90.Overlaid on the data of Figure 2 are simulations obtained bysimultaneous least-squares fitting of the signals at X- and Q-bandto eq 1 using only the iron-nitrosyl terms for  S  Fe  )  3  /  2 . Theparameters derived for the chloride sample are  E   /   D  )  0.013, σ  E/D ) 0.005, and  g Fe ) (2.017, 2.017, 2.000). The line shapesof the spectra are well modeled with a Gaussian distribution in  E   /   D  of width  σ   E   /   D . The parameters derived for the fluoridesample are  E   /   D ) 0.025,  σ   E   /   D ) 0.004, and  g Fe ) (2.026, 2.026,2.000). Both the X- and Q-band simulations for the fluoridesample also include an anisotropic superhyperfine interactionto an  I  ) 1  /  2  nucleus of   A  x ) 101 ( 5 and  A  y ) 146 ( 5 MHz. 18 From these simulations, the concentration of the iron-nitrosylcomplex was found to be 15  µ M. The  D -value can bedetermined from the temperature dependence of the intensityof the signals. A fit of the intensity versus temperature to aBoltzmann population of levels (not shown) gave  D ) 10 ( 2cm - 1 in the presence of Cl - and  D ) 8 ( 2 cm - 1 in the presenceof F - . Within the level of uncertainty, we conclude that theeffects of F - on the  D -value are minimal., whereas the  E   /   D value in the presence of F - is significantly larger than thecorresponding value in the Cl - control samples, indicating ahigher rhombic distortion for F - . The splitting of the  g  )  4signal at X-band in the presence of F - could possibly beattributed to a superposition of two species with different  E   /   D values. Clearly, however, this is not the case since the spectraat X- and Q-band would then be identical.The simultaneous match of the simulations to both the X-and Q-band data unequivocally requires a superhyperfineinteraction to the  I   )  1  /  2 19 F nucleus. The magnitude of the  A -value indicates that F - is bound to the FeNO to form anFFeNO complex. To the best of our knowledge, this is the firstreport of both F - and NO bound to the same iron of amononuclear nonheme protein. In the presence of 0.5 - 0.7 mMNO, the  K  d  for fluoride binding at the iron-nitrosyl complexwas approximately 1 mM. This  K  d  did not depend on the NaClconcentration, suggesting that Cl - and F - do not compete forbinding at the iron-nitrosyl complex. We also investigated theeffects of other halides on the iron-nitrosyl  g  )  4 signals.Superhyperfine splittings have been observed in EPR spectraof Br (  I   )  3  /  2 ) adducts of ferric heme proteins and syntheticanalogues. 19a The  g  )  4 signals obtained in PSII membranespretreated with 25 mM NaBr or NaI gave signals similar tothose of the Cl - spectrum, suggesting that these ions do notbind at the iron-nitrosyl complex. Furthermore, these resultsindicate that the fluoride effects are specific and not related tochanges in the protein environment in the vicinity of the iron-quinone complex induced by the magnitude of the ionic strength. Light-Induced Signals from the Q A - { FeNO } 7 Complex. The iron-nitrosyl signal at  g ) 4 is eliminated upon illumination, (18) The axes  x ,  y , and  z  are defined by the spin Hamiltonian eq 1, wherethe largest component of the zero-field splitting tensor is assumed to bealong the  z -axis. For the  | ( 1  /  2 >  state of a  S   )  3  /  2  system,  g  y  >  g  x  >  g  z .The hyperfine axis system is assumed coincident with the electronic system.The  g  z  feature near  g ) 2 is obscured by other signals and not shown; thus  A  z  is also not measurable.(19) Van Camp, H. L.; Scholes, C. P.; Mulks, C. F.; Caughey, W. S.  J. Am. Chem. Soc.  1977 ,  99 , 8283 - 8290. (b) Morimoto, H.; Kotani, M.  Biochim. Biophys. Acta  1966 ,  126  , 176 - 178. Figure 1.  Dependence of the observed  g -values,  g obs , versus  J   /   f   fortransitions I, II, and III for  E   /   D  )  0.02. Vertical lines indicate theassignments to the Q A - { FeNO } 7 complex in the presence of F - (solidline) and Cl - (dashed line) at X- and Q-band. The subscripts on thetransition labels refer to the magnetic field orientation. Figure 2.  X-band (A, B) and Q-band (C, D) EPR spectra (solid lines)of PSII membranes treated with 25 mM NaCl (A, C) or NaF (B, D)for 1 h at 0  ° C in the dark and further incubated with approximately0.5 mM NO. The simulations (dashed lines) are derived fromsimultaneous least-squares fits to the X- and Q-band data to an  S  Fe ) 3  /  2  center. For Cl - (A, C) D  )  10 cm - 1 ,  E   /   D  )  0.013,  σ   E   /   D  )  0.005, g  )  (2.017, 2.017, 2.000). For F - (B, D) D  )  8 cm - 1 ,  E   /   D  )  0.026, σ   E   /   D ) 0.004,  A  x ) 101 ( 5,  A  y ) 146 ( 5 MHz,  g ) (2.026, 2.026,2.000). EPR conditions: (A, B)  T  , 12 K; microwaves, 2 mW at 9.62GHz;  B mod , 12 G pp ; (C, D) T, 6 K; microwaves, 0.5 mW at 34.03 GHz;  B mod , 5 G pp .  Binding of Fluoride and NO to Photosystem II J. Am. Chem. Soc.  C  but no new EPR signals were observed at X-band. 4a,b This wasbelieved to be due to a magnetic interaction of   { FeNO } 7 ( S  Fe )  3  /  2 ) with the semiquinone Q A - ( S  R  )  1  /  2 ). Here we presentthe detection of signals at X- and Q-band that srcinate fromthis interaction.Figure 3 shows X-band EPR spectra before (dashed line) andafter (solid line) illumination at 200 K in the presence of Cl - (A) or F - (B). Illumination causes a significant decrease in theintensity of the  g  )  4 signals. Illumination induces no newsignals in the chloride-treated sample at X-band, while Figure3B shows new signals appearing at  g  )  2.5 and 9.3 for thefluoride sample. Figure 4A shows the resulting difference spectrafor the fluoride sample (light minus dark, solid line). All of thedashed lines in Figure 4 are simulations, which are discussedbelow. A new light-induced signal also appears near  g  )  9 inthe fluoride sample for microwave fields,  B 1 , oscillating parallelto the applied field  B . The difference spectrum is shown inFigure 4B. Finally, a new light-induced signal is also observedat Q-band near  g ) 3.1 in the fluoride sample. Figure 4C showsthe difference spectrum at Q-band. The light-induced signalshave frequency-dependent  g -values; thus the X- and Q-band of data Figure 4 are not plotted on an equal  g -value scale.A number of control experiments were performed to deter-mine the composition of the species giving the light-inducedsignals. The light-induced signals are invariably observed inuntreated samples or Tris-treated BBY samples which lack theMn oxygen-evolving cluster. These preparations also show weakEPR signals at  g ) 2.95 and 2.25 from cytochrome b 559 . 20 Nomodification of these signals occurs, indicating that cytochromeb 559  is not responsible for the light-induced signals.The inhibitors DCMU or Atrazine bind at the Q B  site andare known to affect the signals associated with the iron-quinonecomplex 7,21 as well as the redox and binding properties of thenonheme iron. In the presence of 0.25 mM NO, the addition of 0.1 mM DCMU reduces the  { FeNO } 7 signal by approximately50%, while similar concentrations of Atrazine slightly changethe rhombicity of the signal, as shown in Figure 5A. 22 Thus,Atrazine perturbs the electronic structure of the iron-nitrosylwithout affecting the binding of NO, whereas DCMU raises  K  d for NO binding. The light-induced signals in the presence of either of these inhibitors show analogous effects. For Atrazine,the light-induced signal, Figure 5B, retains intensity but smallshifts occur. The signal at  g ) 9.3 shifts to lower fields by 100G, while the signal at  g  )  2.5 shifts to higher fields by 40 G.For DCMU, the light-induced signals are approximately 50%weaker than without DCMU. These observations support theassignment of the light induced signals to the iron-nitrosylcomplex.For samples in the presence of Cl - , no new X-band EPRsignals are observed in either parallel or perpendicular modesupon illumination. However at Q-band, a new signal is observedat  g ) 2.6 and the difference spectrum (light minus dark, solidline) is shown in Figure 6. We will demonstrate below that thelack of a signal at X-band and the presence of signals at Q-bandare due to the larger magnitude of the exchange interaction inthe presence of Cl - . (20) For a review on usual EPR signals from PSII, see Miller, A.-F.;Brudvig, G. W.  Biochim. Biophys. Acta  1991 ,  1056  , 1 - 8.(21) Diner, B. A.; Petrouleas, V.  Biochim. Biophys. Acta  1987 ,  893 , 138 - 148.(22) D. Koulougliotis, Diploma Thesis, NCSR Democritos, unpublishedresults. Figure 3.  The effect of illumination at 200 K (solid lines) on theX-band EPR spectra of NO-treated PSII membranes in the presence of 25 mM NaCl (A) or NaF (B). The dark spectra (dashed lines) in bothsamples show weak features in the low-field region which are not lightsensitive and are attributed to unidentified impurities. EPR conditions: T  , 5 K; microwaves, 2 mW at 9.41 GHz;  B mod , 25 G pp . Figure 4.  X-band (A, B) and Q-band (C) difference spectra (lightminus dark, solid lines) of NO-treated PSII membranes in the presenceof 25 mM NaF. Spectrum B is recorded with  B 1 || B . The simulations(dashed lines) are derived from simultaneous least-squares fits to thethree spectra for an  S  Fe ) 3  /  2  site exchange-coupled to an  S  R ) 1  /  2  site.The simulation parameters are  D  )  8 cm - 1 ,  E   /   D  )  0.025,  σ   E   /   D  ) 0.004,  J  ) 0.47 cm - 1 ,  σ   J  ) 0.05 cm - 1 ,  A  x ) 101, and  A  y ) 146 MHz.EPR conditions are given in Figure 2. The  g  )  4 signal has beenremoved for clarity. Figure 5.  Difference spectra showing the effect of the PSII inhibitorsAtrazine and DCMU (100  µ M) on the  g ) 4 iron-nitrosyl signals andthe light-induced signals in the presence of 25 mM NaF: (A) dark-illuminated at 200 K, and (B) illuminated at 200 K-dark. EPRconditions:  T  , 5 K; microwaves, 2 mW at 9.41 GHz;  B mod , 25 G pp . D  J. Am. Chem. Soc. Sanakis et al.  Assignment of Light-Induced Signals from Samples withFluoride.  We will first consider the signals from the fluoridesamples because the spectra are better resolved and observedat both X- and Q-band. Figure 4 shows difference spectra (lightminus dark, solid lines) of the FFeNO sample at X-band (paralleland perpendicular modes) and Q-band. Clearly, we mustconsider that the new light-induced signals may involve the iron-nitrosyl complex, since the  g ) 4 signal from that site vanishesupon illumination. The inhibitor experiments shown above alsosupport this assignment. In addition, many studies have shownthat a semiquinone radical, Q A - , species is formed in PSII uponillumination at cryogenic temperatures. On the basis of thesefacts, we consider an exchange coupling between a structurallyunmodified iron-nitrosyl ( S  Fe  )  3  /  2 ) and a semiquinone, Q A - ( S  R  )  1  /  2 ). The iron-nitrosyl site parameters are derived priorto illumination, and inspection of eq 1 indicates that the onlyunknown parameter is the exchange coupling parameter,  J  .The simulations overlaid on the data of Figure 4 (dashed lines)use the FFeNO site parameters given in Figure 2 prior toillumination. We now vary only the isotropic exchange param-eter  J   and the width  σ   J   of a distribution in  J  . We find, for  J  ) 0.47 cm - 1 and  σ   J   )  0.05 cm - 1 , a simultaneous match to allthree spectra of Figure 4. There are no adjustable parametersbetween any of these spectra, and in particular, this also appliesto the intensity factors. The simulations are truly quantitativeand allow determination of the spin concentration. To within10%, we find a quantitative conversion of the photosensitivefraction of the  g  )  4 signal of Figure 3 to the light-inducedsignals of Figure 4. The quantitative simultaneous match to allthree spectra allows an unambiguous assignment of these signalsto an  S  Fe  )  3  /  2  FFeNO species exchange coupled to the  S  R  ) 1  /  2  semiquinone radical.A significant distribution in  J   of 10% is required to matchthe substantial broadening of the transitions. Such a distributionwas considered also in the simulation of the EPR signal fromthe state Q A - Fe 2 + in the photosynthetic bacterium  R.sphaeroides. 12a The position and broadening of the signals canbe understood from inspection of Figure 1. For  J  ) 0.47 cm - 1 ,transition I gives rise to the derivative signals near  g  )  2.5(X-band) and  g ∼ 3.1 (Q-band). Transitions II and III  z  give therather weak signals at  g  >  8 (X-band). The broadening isprofound for transitions III  x  and III  y , which give a weak broadabsorption for 4 < g < 8. The dramatic effect on the line shapesupon inclusion of a distribution of the  J  -value is due to thedependence of the  g obs  on  J   /   f   in Figure 1. For 1.25  <  J   /   f   < 1.56, the value of   g obs  for transition I varies by only 4% overthis range, while  g obs  for the other transitions varies by 20 - 45%. Thus, the effect of a distribution,  σ   J  , on the broadeningof the EPR signals is significantly larger for the transitions IIand III than for I. Transition II is not observed in Q-band. Thisis mainly due to a 12-fold decrease in the transition probabilityfrom X to Q-band and partly due to the broadening of thetransitions.We have also considered the following alternative broadeningmechanisms, but they all gave unsatisfactory simulations in theabsence of a distribution in the  J  -value. (1) A distribution in  E   /   D . This parameter, although necessary for simulations of the g  )  4 signals in the dark, had an insignificant effect on thesimulation of the light-induced signals. (2) The superhyperfinetensor A of   19 F .  The splitting due to  19 F binding is not resolved,as it is wholly within the line width of the observed signals. (3)Anisotropic exchange interaction. Anisotropic interactions canhave pronounced effects on the spectra, but these effectscontribute to rather small shifts in the signals that do not properlybroaden the spectra. Nevertheless, we have investigated theeffect of this interaction in detail.The most plausible srcin for anisotropy in the present caseis the dipolar interaction. The dipolar interaction 17 is definedwith the vector  r , joining the two paramagnetic species, andpolar angles  θ r  and  φ r , which describe the orientation of   r  rela-tive to the  g -tensor of the  S  Fe ) 3  /  2  spin system. From the crystalstructure of the bacterial reaction center, the distance betweenthe iron and the center of the quinone ring is 8.5 Å. 23 We haveincluded the dipolar term into the spin Hamiltonian of eq 1 andinto the simulations for  r  in the range 7 - 10 Å and arbitrarypolar angles. While for certain dipolar distances and angles weobserve shifts in the simulations, within experimental accuracy,we are not able to detect a significant contribution from a dipolarinteraction in the spectral data for the F - case. This can berationalized by an appropriate choice of distance and angle thatminimizes the dipolar interaction. For example, for  r ) 8.5 Å, θ r ) 90 ° , and  φ r ) 55 ° , the simulations give good fits similarto those shown in Figure 4. For  r g 10 Å, no significant changesare observed in the simulations for any dipolar angles. Assignment of the Light-Induced Signals from Sampleswith Chloride.  Figure 6 shows the difference spectrum (lightminus dark, solid line) of the FeNO sample at Q-band. Again,we consider an exchange coupling between a structurallyunmodified iron-nitrosyl ( S  Fe  )  3  /  2 ) and a semiquinone, Q A - ( S  R  )  1  /  2 ). The simulation overlaid on the data (dashed line)uses the parameters derived for the FeNO site given in Figure2 prior to illumination. In addition, the simulation of Figure 6includes an isotropic exchange interaction of   J  ) 1.3 cm - 1 and σ   J  ) 0.3 cm - 1 . The signal-to-noise of the spectrum is relativelypoor due to complications from overlapping signals that do notsubtract out well. Nevertheless, the simulations indicate aquantitative conversion of the photosensitive fraction of the  g )  4 signal in Figure 3 to the light-induced signal of Figure 6.The quantitative agreement allows an assignment of this signalto an  S  Fe ) 3  /  2 { FeNO } 7 species exchange coupled to the  S  R ) 1  /  2  semiquinone radical. These signals arise from transition I.The inability to detect signals at X-band can be understood fromFigure 1. For  J   /   f   )  4.2, the only transitions occur at  g  )  1.93(  I   x ) and 1.65 (  I   y ). Simulations (not shown) indicate that thecorresponding signals for these transitions are overwhelmed bymuch stronger signals from other paramagnetic species in thisregion. At Q-band, transitions II and III are exceedingly broaddue to the distribution in  J   and are therefore difficult to detect.In contrast to the F - case, the best simulations of the signalin the presence of Cl - occur with inclusion of a dipolar (23) Allen, J. P.; Feher, G.; Yeates T. O.; Komiya, H.; Rees, D. C.  Proc. Natl. Acad. Sci. U.S.A.  1988 ,  85 , 8487 - 8491. (b) El-Kabbani, O.; Chang,C.-H.; Tiede, D.; Norris J.; Schiffer, M.  Biochemistry  1991 ,  30 , 5361 - 5369. (c) Emler, U.; Fritzsch, G.; Buchanan, S. K.; Michel, H.  Structure 1994,  2  925 - 936. Figure 6.  Q-band difference spectrum (light minus dark, solid line)of NO-treated PSII membranes in the presence of 25 mM NaCl. Thesimulation (dashed line) is derived from a least-squares fit to an  S  Fe ) 3  /  2  site exchange-coupled to an  S  R ) 1  /  2  site. The simulation parametersare  D  )  10 cm - 1 ,  E   /   D  )  0.013,  σ   E   /   D  )  0.005,  J   )  1.32 cm - 1 ,  σ   J   ) 0.35 cm - 1 . The simulation includes a dipolar interaction for  r  ) 7.0 Åand  θ r  )  45 °  (see text). EPR conditions are given in Figure 2.  Binding of Fluoride and NO to Photosystem II J. Am. Chem. Soc.  E
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks

We need your sign to support Project to invent "SMART AND CONTROLLABLE REFLECTIVE BALLOONS" to cover the Sun and Save Our Earth.

More details...

Sign Now!

We are very appreciated for your Prompt Action!