Excitonic Energy Level Structure and Pigment−Protein Interactions in

Mar 18, 2011 - Institute of Molecular and Cell Biology, University of Tartu, Tartu, ... Phone: +49-30-31421067. ... To the best of our knowledge, this...
0 downloads 0 Views 2MB Size
ARTICLE pubs.acs.org/JPCB

Excitonic Energy Level Structure and Pigment-Protein Interactions in the Recombinant Water-Soluble Chlorophyll Protein. I. Difference Fluorescence Line-Narrowing J. Pieper,*,†,^,# M. R€atsep,‡,^ I. Trostmann,§ H. Paulsen,§ G. Renger,† and A. Freiberg‡,|| †

Max-Volmer-Laboratories for Biophysical Chemistry, Berlin Institute of Technology, Berlin, Germany Institute of Physics, University of Tartu, Tartu, Estonia § Institute of General Botany, Johannes Gutenberg University Mainz, Germany Institute of Molecular and Cell Biology, University of Tartu, Tartu, Estonia

)



ABSTRACT: Difference fluorescence line-narrowing spectroscopy at 4.5 K was employed to investigate electron-phonon and electron-vibrational coupling strengths of the lower exciton level of water-soluble chlorophyll-binding protein (WSCP) from cauliflower reconstituted with chlorophyll a or chlorophyll b, respectively. The electron-phonon coupling is found to be moderate with integral Huang-Rhys factors S in the order of 0.81-0.85. A weak dependence of S on excitation wavelength within the inhomogeneously broadened fluorescence origin band is attributed to a sizable contribution of nonresonant excitation that varies with excitation wavelength. The strongly asymmetric and highly structured one-phonon profile is characterized by a peak phonon frequency (ωm) of ∼24 cm-1 and further discernible peaks at 48 and 88 cm-1, respectively. A structural assignment of this unusual onephonon profile is proposed. As will be shown in the accompanying paper (part II) (DOI 10.1021/jp111457t), the parameters of electron-phonon coupling readily account for shape and position of the fluorescence origin bands at 666.1 and 683.8 nm for chlorophyll b- and chlorophyll a-WSCP, respectively. A rich structure of S1fS0 vibrational frequencies was resolved in the wavenumber range between 180 and 1665 cm-1 for both chlorophyll a- and chlorophyll b-WSCP. The corresponding individual Huang-Rhys factors fall in the range between 0.0011 and 0.0500. To the best of our knowledge, this is the first report of S-factors for vibrational modes of chlorophyll b. Most remarkable is the presence of two additional modes at 228 and 327 cm-1 compared with the vibrational spectrum of chlorophyll in solution. The additional modes can most likely be attributed to H-bond formation in the vicinity of the chlorophyll molecule bound by WSCP.

1. INTRODUCTION The initial steps of photosynthesis take place in pigmentprotein assemblies also referred to as antennae or light-harvesting complexes (LHCs). Light absorption is followed by efficient excitation energy transfer (EET) to reaction center (RC) complexes where the primary charge separation takes place, resulting in formation of a cation-anion radical pair. Subsequently, the charge separation is stabilized by further electron transfer steps across the membrane.1 For efficient light harvesting, excitation energy transfer, and photoprotection, antenna proteins bind a number of pigment molecules like chlorophyll (Chl) or bacteriochlorophyll and different carotenoids. The flow of excitation energy is determined by complex pigment-pigment and pigmentprotein interactions (for a review see, e.g., van Amerongen and Croce2). Although the structure of several antenna complexes has been resolved by X-ray diffraction crystallography to nearly atomic r 2011 American Chemical Society

resolution,3-6 a complete microscopic understanding of EET in antenna proteins has not yet been achieved. One major obstacle is the high complexity of native antenna complexes with a large number of bound chromophores, which precludes a straightforward determination of pigment site energies. Thus, it is instructive to verify the current concepts describing pigment-pigment and pigment-protein interactions by investigating a relatively simple native model system. Such a naturally abundant model pigment-protein complex is the water-soluble Chl-binding protein (WSCP), which can be isolated from Amaranthaceae, Chenopodiaceae, and Polygonaceae (so-called “class-I WSCP”) and from Brassicaceae (class-II WSCP). The two classes of WSCP can be distinguished according to their response to illumination. Received: December 2, 2010 Revised: January 27, 2011 Published: March 18, 2011 4042

dx.doi.org/10.1021/jp111455g | J. Phys. Chem. B 2011, 115, 4042–4052

The Journal of Physical Chemistry B Class-I WSCP exhibits a significant red shift upon light excitation, while no photoconversion is observed for class-II WSCP (for a review see ref 7). The class-II WSCPs (hereafter referred to as WSCP) are water-soluble proteins with a molecular weight of approximately 20 kDa, exhibit high thermal stability, and form a tetrameric assembly upon Chl binding. Two types of class-II WSCPs are distinguished based on their Chl a:b ratio. All classII WSCPs form tetrameric complexes containing one to four Chl molecules.7 A recombinant version of WSCP from cauliflower binds two Chl molecules (class-IIa),8,9,12 while WSCP from Lepidium virginicum contains four Chl molecules per WSCP tetramer (class-IIb).10 WSCP can be reconstituted with chlorophyll a (Chl a), chlorophyll b (Chl b), or chlorophyll d (Chl d). The functional role of WSCP is still unclear. However, because of the low number of pigment molecules, it appears unlikely that WSCP is involved in light harvesting or in protective dissipation of an excess population of excited states of the photosynthetic apparatus. One striking feature is the stimulation of WSCP formation under drought8, salt11, and heat12 stress. Therefore, it has been suggested that the complex might provide protection against degradative processes. Furthermore, it has been speculated that WSCP may act as a transporter of free Chl13,14 from the thylakoid membrane to the chloroplast envelope, where Chl catabolism is believed to take place.15 The crystal structure is so far available only for the class-IIb WSCP-Chl complex from Lepidium virginicum that has been investigated by X-ray diffraction.10 The latter WSCP complex is shown to be a homotetramer comprising four protein chains of 180 amino acids and four Chl binding sites. The Chls were found to form “open sandwich” dimers with an opening angle of 27 in the case of Chl a binding. All Chl molecules are tightly packed in a hydrophobic cavity at the center of the complex and thus isolated from bulk solvent. This structural motif is proposed to be the reason for the high photostability of WSCP.9 Although a crystal structure for the class-IIa WSCP is so far lacking, theoretical calculations of optical spectra of class-IIa WSCP indicate that the local binding motif of the “open sandwich” Chl dimer is very similar in both types of class-II WSCPs.16,17 The spectroscopic properties of class-IIa WSCP have been studied using a variety of experimental methods and theoretical simulations.9,16-18,21 The arrangement of the Chl molecules in the form of an open sandwich dimer was first suggested based on Gaussian fits of stationary absorption and magnetic circular dichroism (MCD) spectra measured at 1.7 K on recombinant class-IIa WSCP reconstituted with either Chl a or Chl d.18 Within this model, the tilt angle between the chlorin planes of the pigments was determined to be about 60. Later, a detailed theoretical analysis based on the stationary absorption and CD spectra of class-IIa WSCP led to the conclusion that the angle is approximately 30.16 This result turned out to be in better agreement with the structure gathered from X-ray diffraction crystallography (see above).10 It is also interesting to note that the structural arrangement of the excitonically coupled Chl molecules in WSCP leads to a larger oscillator strength of the upper exciton level compared with that of the lower one.16,18 Compare also the large intensity of the main absorption band with the weaker low-energy shoulder in Figure 1. In contrast to the latter phenomenon, strongly coupled Chl dimers, e.g., in reaction centers (“special pair” P in bacterial RCs, P700 in PS I, and P680 in PS II) as well as in antenna proteins, e.g., in LH1 and

ARTICLE

Figure 1. 4.5 K absorption spectra of Chl b-WSCP (blue line) and Chl a-WSCP (red line). Spectral features are labeled by their peak wavelengths in nanometer. The inset shows the bonds established between a Chl and the monomeric protein subunit as reported by Horigome et al. for type-IIb WSCP.

LH2, are characterized by a larger oscillator strength of the lower exciton level.19,20 Transient absorption spectroscopy with femtosecond time resolution (200 fs) at room temperature revealed characteristic biphasic kinetics with time constants of 400 fs and 7-8 ps in class-IIa WSCP containing Chl heterodimers.9 The 400 fs kinetics was attributed to relaxation within a strongly coupled Chl a-Chl b heterodimer, and the slower component of 7-8 ps was assumed to arise from F€orster type EET in heterodimers of weakly coupled Chl a- and Chl b-molecules.9,16 Time and space correlated picosecond fluorescence spectroscopy21 revealed at least three kinetic components with temperature-dependent time constants. The dominating decay kinetics is characterized by a time constant of 6-7 ns at 10 K, which almost linearly decreases with temperature to 5 ns at 300 K. In addition, a strongly temperature-dependent fluorescence component was found to exhibit a typical time constant of 80-120 ps at 10 K which is attributed to a slow protein relaxation within the lowest exciton level.21 So far, recombinant type IIa WSCP has mainly been investigated by time-resolved spectroscopy. Therefore, complementary studies on pigment-pigment and pigment-protein interactions were performed by using site-selective spectroscopy like fluorescence line narrowing (FLN) and spectral hole burning (SHB) at low temperature. The present report (part I) focuses on a detailed characterization of electron-phonon and electronvibrational coupling of excitonically coupled Chl dimers in type IIa WSCP using delta-FLN experiments. Briefly, a delta-FLN spectrum is obtained as the difference between two FLN spectra recorded before and after spectral hole burning.22,23,25,26 In contrast to “conventional” FLN, this subtraction technique provides ZPL virtually free from scattering artifacts of the excitation laser beam so that electron-phonon as well as electron-vibrational coupling strengths can be directly determined.25,26 The accompanying paper (part II, DOI 10.1021/jp111457t), reports corresponding SHB results, which provide new insight into excited state positions as well as homogeneous and inhomogeneous broadening of the excitonic energy levels of WSCP in the frequency domain. 4043

dx.doi.org/10.1021/jp111455g |J. Phys. Chem. B 2011, 115, 4042–4052

The Journal of Physical Chemistry B

2. MATERIALS AND METHODS Sample Preparation. Recombinant WSCP from cauliflower (Brassica oleracea var. botrytis) with an N-terminal hexahistidyl (His) tag was expressed as previously described.21 After cell lysis and centrifugation, the supernatant containing soluble WSCP was stored in aliquots at -20 C to be used for sample preparation. Purified Chl a and Chl b were prepared as described in Hobe et al.27 Aliquots of pigments were dried and stored at -20 C in an inert (nitrogen) atmosphere. Reconstitution and purification of soluble WSCP was performed as described in Renger et al.16 using purified Chl a or Chl b, respectively, for the pigment solution. Finally, WSCP was diluted in a glass-forming buffer solution containing 300 mM imidazole, 20 mM NaP (pH = 7.5), and 70% w/w glycerol. The total Chl concentration of samples used for hole burning and fluorescence measurements was 0.2 and 0.004 mg/mL, respectively (∼200 and 4 μM). Experimental Setup. The setup used for difference fluorescence line-narrowing (delta-FLN) has been described in detail in refs 24-26. Briefly, delta-FLN requires excitation of pre- and postburn FLN spectra as well as an intermittent hole-burning sequence, which were carried out with excitation/burn wavelengths between 638 and 688 nm using a Spectra Physics model 375 dye laser (line width of 1) is obtained by folding the one-phonon profile l1 R-times with itself, so that the form of the one-phonon profile determines the shape of the whole PSB. The dimensionless Huang-Rhys factor S is a measure for the linear electron-phonon coupling strength and characterizes the

ARTICLE

average number of phonons accompanying a particular electronic transition. The delta-FLN spectrum in the short burn time limit is widely identical to the homogeneously broadened spectrum,22,23,29 which for several phonon modes (k), one-phonon profiles, and arbitrary temperature transforms to30 -

LðωÞ ¼ e

∑k Sk ð2nk þ 1Þ Y k

¥

R-r

R

þ 1Þ ½Sk nk  ∑ ∑ ½Sk ðnk ðR - rÞ!r! R¼0 r¼0

 lR, r ½ω - Ω0 (

∑k ðR - 2rÞωk

r

ð2Þ

where hnk = [exp pωk/kT - 1]-1 describes the thermal occupation for phonons of mode k according to Bose statistics and - and þ correspond to absorption and fluorescence, respectively. As in eq 1, R (with R = 1, 2, ...) denotes the total number of phonon transitions regardless of creation and annihilation processes, while r gives the number of annihilated phonons (0 e r e R). By analogy to eq 1, the profile lR,r (R > 1) is obtained by folding the one-phonon profile l1,0 |R - 2r|-times with itself. Then, for l1,0 being a Gaussian with a width of ΓG, the profile lR,r becomes a Gaussian with a width of |R - 2r|1/2ΓG. If l1,0 is a Lorentzian with a width of ΓL the profile lR,r becomes a Lorentzian with a width of |R - 2r|ΓL.30 In this work, the onephonon profile is assumed to have Gaussian and Lorentzian shape at its low- and high-energy wings, respectively. Finally, inhomogeneously broadened absorption and fluorescence spectra are obtained by a convolution of eq 2 with the inhomogeneous distribution function (IDF) N(Ω - ωC)30 -

LðωÞ ¼ e

∑k Sk ð2nk þ 1Þ Y k

¥

Z 

R-r

R

½Sk ðnk þ 1Þ ½Sk nk  ∑ ∑ ðR - rÞ!r! R¼0 r¼0

dΩ0 NðΩ0 - ωC ÞlR, r ½ω - Ω0 -

∑k ðR - 2rÞωk 

r

ð3Þ

which is assumed to be a Gaussian with a width of Γinh centered at ωC. The vibronic coupling strength of the jth vibronic line observed in FLN or delta-FLN spectra is characterized by its Huang-Rhys factor Sj and can be determined according to31 expð - Sj Þ ¼

I0 - 0 I0 - 0 þ Ij

ð4Þ

where I0-0 is the intensity of the whole 0-0 band including its ZPL and PSB contributions, while Ij is the intensity of a particular vibronic band. This approach implies that the individual vibronic contributions to the spectrum are additive. A justification is given by the generally small values of each individual Sj that reduces the vibronic Franck-Condon progression to a single one-vibration term of the weight Sj exp(-Sj) ≈ Sj. In the multimode case, the number of such terms is equal to the number of local modes. The cubic frequency dependence of fluorescence31 was taken into account. In a similar way, the integrated vibrational coupling strength Svib can be calculated from nonline-narrowed fluorescence spectra following expð - Svib Þ ¼ 4044

I0 - 0 I0 - 0 þ Ivib

ð5Þ

dx.doi.org/10.1021/jp111455g |J. Phys. Chem. B 2011, 115, 4042–4052

The Journal of Physical Chemistry B

Figure 2. 4.5 K absorption (black dotted line) and non-line-narrowed fluorescence spectra (red line) of Chl a-WSCP labeled with their peak wavelengths. The blue line is a delta-FLN spectrum excited/burned at 682.0 nm and recorded with a fluence of 6.7 mJ/cm2, while the fluence applied for hole-burning was 161 mJ/cm2. The inset shows a magnification of the delta-FLN spectrum plotted on a wavenumber scale. The vertical numbers label selected vibrational frequencies in wavenumbers; see Table 1 for a complete list of all discernible vibrational ZPL. The frequencies labeled in blue have not been observed for Chl in solution.

where I0-0 is the integrated intensity of the fluorescence origin band including its ZPL and PSB contributions, and Ivib is the total intensity of all vibrational satellite bands.

3. RESULTS The 4.5 K absorption spectra of WSCP reconstituted with either Chl a or Chl b are shown as red and blue lines in Figure 1, respectively. For Chl a-WSCP, the main absorption band is found at 672.5 nm with a weaker shoulder at about 681 nm. In the case of Chl b-WSCP, the absorption spectrum is mainly composed of a pronounced band peaking at 655.4 nm and a shoulder at about 663 nm. Thus, the latter absorption spectrum appears to be mainly blue-shifted by about 17 nm compared to Chl aWSCP, and the intensity ratio between the main peak and the weaker shoulder is slightly different. In addition, both absorption spectra exhibit a number of satellite bands on the high-energy side of their main absorption bands, while a further very weak red-shifted band at 676.1 nm is present in Chl b-WSCP only. The 4.5 K absorption spectrum of Chl a-WSCP is quite similar to the low-temperature spectra reported by Hughes et al.,18 Renger T. et al.,16 and Theiss et al.,17 who identified the main absorption peak and the shoulder toward the red with the absorption bands of the upper and lower exciton levels of a strongly coupled dimer of Chl molecules. This interpretation is confirmed by the satellite hole structure observed in spectral hole-burning experiments that are described in the accompanying paper (see part II, DOI 10.1021/jp111457t). In the present work, delta-FLN spectroscopy is employed to investigate electron-phonon and electron-vibrational coupling of WSCP reconstituted with either Chl b or Chl a. Chl a-WSCP Vibrational Frequencies. The nonlinenarrowed fluorescence spectrum of Chl a-WSCP is shown as a red curve in Figure 2. It consists of a wide—most probably strongly inhomogeneously broadened—origin band with a peak wavelength of 683.8 ( 0.2 nm and a number of broad vibrational satellite bands at longer wavelengths. The peak position is close to the low-energy absorption shoulder (see dashed black curve in

ARTICLE

Figure 2) so that the lowest exciton level can most probably be identified with the fluorescence origin (see part II (DOI 10.1021/jp111457t) for a more detailed assignment). Much more structure is visible in delta-FLN spectra of Chl a-WSCP, which were selectively excited at several wavelengths within the inhomogeneously broadened fluorescence origin band ranging from 680 to 689 nm. A typical delta-FLN spectrum of Chl a-WSCP obtained with an excitation wavelength of 682 nm is shown as a blue curve in Figure 2. In contrast to the nonlinenarrowed fluorescence spectrum, the delta-FLN spectrum is composed of a sharp—resolution-broadened—ZPL at 682 nm, a broad low-frequency phonon wing peaking at about 24 cm-1, and a multitude of narrow vibronic lines at higher frequencies. In contrast to the delocalized vibrational modes constituting the phonon wing, most of the narrow lines at frequencies higher than ∼200 cm-1 correspond to localized vibrations, which are the fingerprint of the Chl a pigment molecule.32-34 A magnification of the vibronic region is displayed in the inset of Figure 2. The full set of vibrational frequencies resolved in this experiment is listed in Table 1. Since the ZPL intensity is directly available from the delta-FLN experiment, the individual Huang-Rhys factors can be determined for each vibrational mode according to eq 4. For comparison, Table 1 also shows the vibrational frequencies reported for the (widely localized) lowest energy Qy-state of the antenna complex CP29.26 A closer inspection of this data set reveals that within experimental uncertainty most of the vibrational frequencies and coupling strengths are very similar to those reported before for Chl a, while two new modes appear at 228 and 327 cm-1, respectively, and the relative intensity of the 1285 cm-1 mode is enhanced (see Discussion). Chl a-WSCP Electron-Phonon Coupling. Further deltaFLN spectra were selectively excited at several wavelengths within the fluorescence-origin band ranging from ∼680 up to ∼689 nm. The phonon (PSB) region of several of these deltaFLN spectra is shown as red curves in Figure 3, while the black and blue curves correspond to the respective pre- and postburn FLN spectra. As pointed out before, the ZPL in “conventional” FLN spectra are usually strongly contaminated by scattering light from the excitation laser, which can be eliminated by subtracting pre- and postburn FLN spectra within the delta-FLN approach. A further advantage of delta-FLN spectroscopy is clearly visible in Figure 3 by inspecting the curves obtained for an excitation wavelength of 680.1 nm, which is located at the high-energy wing of the fluorescence origin band. It is apparent that under these conditions the conventional FLN spectra (black and blue curves) have not reached full selectivity due to a large amount of nonresonant excitation because the peaks of these spectra are observed at ∼684 nm, close to the maximum of the nonlinenarrowed fluorescence spectrum. In contrast, the delta-FLN spectrum exhibits much higher selectivity with the ZPL accompanied by a typical PSB with a displacement of only ∼24 cm-1. The conventional FLN spectra display the same selectivity only at excitation wavelengths larger than ∼682 nm, i.e., for wavelengths close to and larger than the peak wavelength of the nonline-narrowed fluorescence spectrum. At these wavelengths, the lineshapes of FLN and delta-FLN spectra shown in Figure 3 are rather similar. As expected for this case, the peak of the PSB of ∼24 cm-1 is shifting with the excitation wavelength. There are two reasons for the generally higher spectral selectivity of deltaFLN: First, even in the case of a single inhomogeneously broadened fluorescence band, delta-FLN suppresses the pseudo-PSB contribution so that the delta-FLN spectrum becomes 4045

dx.doi.org/10.1021/jp111455g |J. Phys. Chem. B 2011, 115, 4042–4052

The Journal of Physical Chemistry B

ARTICLE

Table 1. Vibrational Frequencies νj ((2 cm-1) and Huang-Rhys Factors, Sj ((0.0005), of Chl a-WSCP and Chl b-WSCP in Selectively Excited Qy Differential Fluorescence Line-Narrowing Spectra at 4.5 K Chl a-WSCP νj, cm-1

Chl a-WSCP Sj

Chl b-WSCP

νj, cm-1

νj, cm-1 181 221 240 269 283 298 325 352 366 405 430 470 488 515 537 572 598 620 641 700

0.0173 0.0246 0.0182 0.0064 0.0036 0.0104 0.0112 0.0249 0.0112 0.0061 0.0050 0.0075 0.0061 0.0045 0.0157 0.0132 0.0036 0.0047 0.0033 0.0019

713 734 746 757 800

0.0025 0.0107 0.0112 0.0229 0.0022

834 863 887 922 977 998 1023 1045 1068 1108 1128 1150 1172 1186 1227 1243 1264 1288 1305 1326 1360

0.0140 0.0033 0.0019 0.0291 0.0110 0.0036 0.0022 0.0056 0.0050 0.0087 0.0011 0.0244 0.0121 0.0226 0.0249 0.0090 0.0126 0.0224 0.0093 0.0509 0.0093

1393 1443 1484 1522 1550 1573 1628 1654 1681 Sum:

0.0328 0.0121 0.0107 0.0185 0.0241 0.0182 0.0081 0.0135 0.0067 0.74

183 228

0.0078 0.0082

192

259 283

0.0160 0.0052

262 287

327 356 384 402 435

0.0055 0.0245 0.0147 0.0088 0.0085

481 520 547 574 584 609 656 690 703 718

0.0137 0.0157 0.0042 0.0072 0.0049 0.0075 0.0013 0.0013 0.0019 0.0039

467 517 545 573

742 754 803 816 840

0.0353 0.0212 0.0101 0.0066 0.0072

745

920 985 993 1024 1047 1067 1104

0.0251 0.0158 0.0129 0.0068 0.0215 0.0170 0.0176

918 989

1144 1184

0.0303 0.0460

1217 1243 1259 1285 1304 1329 1354 1375 1390 1441 1490 1532 1556 1591 1620 1649

0.0502 0.0176 0.0101 0.0193 0.0107 0.0389 0.0105 0.0094 0.0173 0.0232 0.0088 0.0330 0.0212 0.0058 0.0091 0.0114

Sum:

Chl b-WSCP

CP29 from ref 26

350 386 401 425

607 701

800 842

1049 1070 1109 1145 1186 1227

1262 1286 1307 1329 1353 1374 1387 1439 1490 1537 1556 1610 1665

0.80 4046

Sj

Chl b from ref 32 νj, cm-1

245 255 310 345 375 405 465 515 560 595 625 680 700 732 742 750

832 885 923 980 1003 1040 1070 1087 1120 1140 1190 1217 1253 1275

1330 1270 1390 1480 1515 1540

dx.doi.org/10.1021/jp111455g |J. Phys. Chem. B 2011, 115, 4042–4052

The Journal of Physical Chemistry B

Figure 3. Typical 4.5 K delta-FLN spectra (lower red curves) of Chl aWSCP excited/burned in the wavelength range between 680 and 690 nm. The actual excitation/burn wavelength is given in each frame, and the ZPLs are cut off at ∼14% of their full intensity. The black and blue curves show the corresponding pre- and postburn FLN spectra, respectively, recorded with a fluence of 6.7 mJ/cm2. The fluence applied for hole burning was 161 mJ/cm2.

similar to the single-site spectrum.22,23,29 Second, in the case of the presence of additional red-shifted emitters populated, e.g., by EET, hole burning will most likely be more efficient for the state burned selectively, so that the contribution of nonselectively burned fluorescent states is expected to be suppressed by deltaFLN. A comparison of the delta-FLN spectra obtained at 680.1 nm with those measured at the red edge of the fluorescence origin band reveals, however, that the former delta-FLN spectrum exhibits an enhanced contribution around 684 nm. Thus, even the delta-FLN spectra at the blue edge of the fluorescence origin band may contain some contribution due to nonselective excitation (see below). The phonon wing of the delta-FLN spectrum excited at 685.1 nm is shown in more detail as a black curve in Figure 4. Its shape is strongly asymmetric and exhibits an interesting substructure with a pronounced peak at ∼24 cm-1, a shoulder at ∼48 cm-1, and a further peak at ∼88 cm-1. Similarly broad and asymmetric one-phonon profiles have been observed in FLN/delta-FLN experiments on LHC II,35,36 LH2,22 LH1,37 and CP2926 as well as in single molecule experiments on LH238 and PS I.39 Except for the case of LH2,40 however, the one-phonon profiles are often widely featureless. Therefore, the origin of the

ARTICLE

highly structured one-phonon profile of Chl a-WSCP may lie either in a specific protein structure or in the excitonic nature of the fluorescing state (see Discussion). At first glance, the peak positions of ∼24, ∼48, and ∼88 cm-1 suggest that the higher-frequency peaks may be roughly identified as two- and three-phonon transitions of the main peak at ∼24 cm-1, and the delta-FLN spectrum can be described using a single one-phonon profile. In contrast, however, no Huang-Rhys factor S can be found that properly describes the intensity ratios between ZPL and the individual phonon transitions. Rather, the S factor that properly simulates the relative intensities of the phonon transitions turns out to be too high to fit the ZPL simultaneously. Thus, a satisfactory fit of this spectrum, shown as a red line in Figure 4, requires the use of eq 3 with the assumption of three individual one-phonon profiles composed of Gaussian and Lorentzian wings at their low- and high-energy sides (see Table 2 for parameters) and yields an integral Huang-Rhys factor S of 0.81 at 682 nm. Similar fits and corresponding integral HuangRhys factors S have been obtained for all delta-FLN spectra shown in Figure 3. The wavelength dependence of the HuangRhys factors S is shown in Figure 5 (blue diamonds) along with the Gaussian IDFs of the lower (fluorescent) exciton level of Chl a-WSCP which is determined by constant fluence HB as outlined in the accompanying paper (see part II, DOI 10.1021/ jp111457t). The data reveal a strong decrease of S with increasing excitation wavelength at the blue side of the IDF. In the center and at the red side of the IDF, the S-values are slowly converging toward 0.8 (see, e.g., 682 nm). This finding is in agreement with the observation of nonselectively excited contributions to the delta-FLN spectra at the blue side of the IDF (see Figure 3); i.e., a part of the delta-FLN intensity would be due to excitation of the higher exciton level, which leads to nonselectively excited fluorescence only after exciton relaxation to the lowest exciton level. The latter effect is also consistent with the high burn efficiency and concomitant fast saturation of hole A corresponding to the lower exciton level in the case of Chl a-WSCP (see part II, DOI 10.1021/jp111457t). This effect leads to a significant nonselective contribution to the delta-FLN spectrum and thus to a seemingly higher S factor. Chl b-WSCP Vibrational Frequencies. The delta-FLN results obtained for Chl b-WSCP exhibit many parallels to those presented above for Chl a-WSCP. The nonline-narrowed fluorescence spectrum of Chl b-WSCP peaking at 666.1 ( 0.2 nm is shown in Figure 6 as a red line. As for Chl a-WSCP, this position is close to that of the low-energy shoulder of the corresponding absorption spectrum (see black dashed line). More structure is visible in a Chl b-WSCP delta-FLN spectrum that is shown as a blue curve in Figure 6 for an excitation wavelength of 666.3 nm. Again, the spectrum is composed of a sharp ZPL coincident with the excitation wavelength, a broad low-frequency phonon wing peaking about 23 cm-1 to the red of the ZPL, and a well-resolved vibronic structure. A magnification of the vibronic region is shown in the inset of Figure 6, where the most intense vibronic lines are labeled by their frequencies in units of wavenumbers. All discernible vibronic lines corresponding to localized vibrations of the Chl b molecule and their respective S-factors are listed in Table 1 and compared to vibrations of Chl b in solution determined by Avarmaa and Rebane.32 Furthermore, the nonline-narrowed fluorescence spectrum of Chl b-WSCP exhibits a rather intense shoulder at ∼680 nm compared to that of Chl a-WSCP (cf. Figure 2), while the vibronic S-factors of the vibrations in the corresponding wavenumber range 4047

dx.doi.org/10.1021/jp111455g |J. Phys. Chem. B 2011, 115, 4042–4052

The Journal of Physical Chemistry B

ARTICLE

Figure 4. Simulation (red line) of the 4.5 K delta-FLN spectrum of Chl a-WSCP excited at 685.1 nm (black curve) calculated using a HuangRhys factor S of 0.81 and three one-phonon profiles with peak phonon frequencies of 24, 48, and 88 cm-1, respectively. See Table 2 for details.

Figure 5. Huang-Rhys factors S as a function of excitation wavelength as determined for Chl b- (red triangles) and Chl a-WSCP (blue diamonds). For comparison, the red and blue lines show the Gaussian IDFs of the lower (fluorescent) exciton level of WSCP reconstituted with either Chl b or Chl a (see part II, DOI 10.1021/jp111457t), respectively.

Table 2. Parameters of the One-Phonon Profiles Used to Fit the 4.2 K Delta-FLN Spectra of Chl a-WSCP Shown in Figure 4 Chl a

profile I

profile II

profile III

Huang-Rhys factor

Stotal = 0.81

peak phonon frequency

ωm [cm-1]

24

48

88

fwhm of Gaussian wing

Γg [cm-1]

25

20

24

fwhm of Lorentzian wing

ΓL [cm-1]

25

50

60

fwhm, one-phonon profile

Γ [cm-1]

25

35

42

0.45

0.20

0.16

(200 - 350 cm-1) are widely similar. Thus, it is very likely that the shoulder at ∼680 nm not only reflects vibrational satellite bands but also contains a fluorescence contribution originating from the weak absorption band at 676 nm. Chl b-WSCP Electron-Phonon Coupling. Delta-FLN spectra selectively excited at several wavelengths ranging from 662.1 to 683.2 nm are shown in Figure 7. The latter excitation wavelengths cover the spectral range of the low-energy absorption bands at 663 and 676 nm, respectively (see Figure 1). Both absorption bands are identified as low-energy fluorescent states of the bulk Chl b-WSCP sample by constant-fluence hole burning as outlined in the accompanying paper (see part II, DOI 10.1021/jp111457t). We will first discuss delta-FLN spectra obtained within the 663 nm absorption band, which are shown in the upper four frames of Figure 7. The first two burn/ excitation wavelengths of 662.1 and 664.2 nm reflect the case of burning at the high-energy wing of the main fluorescence origin band of Chl b-WSCP peaking at 666.1 nm. As discussed above for Chl a-WSCP, the conventional preburn FLN spectra have not attained maximum selectivity in this case. This is especially apparent for the excitation wavelength of 662.1 nm, where the fluorescence peak of ∼666 nm is almost the same as that of the nonline-narrowed fluorescence spectrum. In contrast, the deltaFLN spectrum always reveals an asymmetric PSB with a peak at ∼23 cm-1. Higher selectivity in FLN is reached for the burn/excitation wavelengths of 666.4 and 669.4 nm, which are located to the red of the main fluorescence origin band of Chl b-WSCP. In this case, both FLN and delta-FLN spectra exhibit highly structured PSBs with a pronounced peak at ∼23 cm-1, a shoulder at ∼46 cm-1, and a further peak at ∼82 cm-1.

Figure 6. 4.5 K absorption (black dotted line) and nonline-narrowed fluorescence spectra (red line) of Chl b-WSCP labeled with their peak wavelengths. The blue line is a delta-FLN spectrum excited/burned at 666.4 nm and recorded with a fluence of 5.6 mJ/cm2, while the fluence applied for hole burning was 170 mJ/cm2. The inset shows a magnification of the delta-FLN spectrum plotted on a wavenumber scale. The vertical numbers label selected vibrational frequencies in wavenumbers. See Table 1 for a complete list of all discernible vibrational ZPL. The frequencies labeled in blue have not been observed for Chl in solution.

Furthermore, these spectra exhibit higher intensity around ∼80 cm-1 compared with the delta FLN spectra obtained at the burn/excitation wavelengths of 662.1 and 664.2 nm. This observation indicates that the latter delta-FLN spectra may be contaminated with a nonselective contribution which could arise, e.g., from EET or exciton relaxation from higher-energy electronic or exciton states and may thus lead to an overestimation of the S-factor. Similar conclusions can be drawn from an inspection of deltaFLN spectra obtained within the 676 nm absorption band, which are shown in the lower four frames of Figure 6. Again, the first burn/excitation wavelength of 671.4 nm is located at the highenergy wing of the 676 nm absorption band so that the conventional FLN spectra have not reached maximum selectivity. This leads to an additional fluorescence peak clearly visible at ∼676 nm in the FLN spectrum, which can be mistakenly interpreted as a higher intensity of the ∼80 cm-1 mode. Full 4048

dx.doi.org/10.1021/jp111455g |J. Phys. Chem. B 2011, 115, 4042–4052

The Journal of Physical Chemistry B

ARTICLE

Figure 8. Simulation (red line) of the 4.5 K delta-FLN spectrum of Chl b-WSCP excited at 666.4 nm (black curve) calculated using a HuangRhys factor S of 0.85 and three one-phonon profiles with peak phonon frequencies of 23, 46, and 82 cm-1, respectively. See Table 2 for details.

Table 3. Parameters of the One-Phonon Profiles Used to Fit the 4.2 K Delta-FLN Spectra of Chl b-WSCP Shown in Figure 8 Chl b

Figure 7. Typical 4.5 K delta-FLN spectra (lower red curves) of Chl bWSCP excited/burned in the wavelength range between ∼660 and ∼685 nm. The actual excitation/burn wavelength is given in each frame, and the ZPLs are cut off at ∼14% of their full intensity. The black and blue curves show the corresponding pre- and postburn FLN spectra, respectively, recorded with a fluence of 5.6 mJ/cm2. The fluence applied for hole burning was 170 mJ/cm2.

selectivity in FLN is reached for the burn/excitation wavelengths of 680.1 and 683.2 nm located at the low-energy side of the 676 nm absorption band. In this case, both FLN and delta-FLN spectra exhibit quite similar PSB shapes as observed before for the 666 nm absorption band (see discussion above). The low-frequency region, i.e., the phonon sideband, of the delta-FLN spectrum of Chl b-WSCP excited at 666.4 nm is shown as a black curve in Figure 8. Its shape is similarly structured as that of Chl a-WSCP with a pronounced peak at ∼23 cm-1, a shoulder at ∼46 cm-1, and a further peak at ∼82 cm-1. A fit of this spectrum is shown as a red line in Figure 8 and again requires the use of eq 3 with the assumption of three individual onephonon profiles composed of Gaussian and Lorentzian wings at the high- and low-energy sides (see Table 3 for parameters). This analysis yields an integral Huang-Rhys factor S of 0.85 at 666.4 nm. Similar fits have been obtained for all delta-FLN spectra shown in Figure 7. The wavelength dependence of the Huang-Rhys factors S is shown in Figure 5 (red triangles) along with the Gaussian IDFs of the lower (fluorescent) exciton levels of Chl b-WSCP determined in the accompanying paper (see part II, DOI 10.1021/jp111457t). The data reveal a slight decrease of S with increasing excitation wavelength converging toward

profile I

profile II

profile III

Huang-Rhys factor

Stotal = 0.85

peak phonon frequency

ωm [cm-1]

23

46

82

fwhm of Gaussian wing

Γg [cm-1]

25

20

24

fwhm of Lorentzian wing

ΓL [cm-1]

25

50

50

fwhm, one-phonon profile

Γ [cm-1]

25

35

37

0.42

0.26

0.17

S = 0.84 at the red side of the IDF. A similar behavior is observed for the 676 nm absorption band with S-values converging toward a slightly higher value of ∼0.9. This indicates that the nonselectively excited contribution to the delta-FLN spectra at the blue side of the IDF is much smaller than observed above for Chl a-WSCP, which is consistent with the lower burn efficiency of the lower exciton level of Chl b-WSCP (see part II, DOI 10.1021/jp111457t). In summary, the wavelength dependence of the Huang-Rhys factors S reported here for the lower exciton level of WSCP is essentially different from that observed for the widely localized lowest Qy-state of the antenna complex CP2926 and for isolated Chl in polymer matrices,41 where an increase of S with excitation wavelength of ∼10% (CP29) or higher (isolated Chl) has been found.

4. DISCUSSION Delta-FLN spectroscopy at 4.5 K has been used to investigate electron-phonon and electron-vibrational coupling strengths of the lower exciton level of the water-soluble chlorophyllbinding protein (WSCP) from cauliflower reconstituted with Chl a or Chl b, respectively. The results presented characterize electron-phonon and electron-vibrational coupling of a strongly coupled Chl-homodimer with interpigment interaction energies in the order of 100 and 85 cm-1 for Chl a- or Chl b-WSCP, respectively (see accompanying paper II, DOI 10.1021/ jp111457t). WSCP Electron-Vibrational Coupling. A rich structure of vibrational frequencies corresponding to the S1fS0 electronic transition was resolved in the wavenumber range between 180 and 1665 cm-1 for both Chl a- and Chl b-WSCP. All discernible 4049

dx.doi.org/10.1021/jp111455g |J. Phys. Chem. B 2011, 115, 4042–4052

The Journal of Physical Chemistry B vibrational frequencies and their corresponding individual Huang-Rhys factors falling in the range between 0.0011 and 0.05 are listed in Table 1. To the best of our knowledge, this is the first report of S-factors for vibrational modes of Chl b. A comparison with data obtained for Chl in solution32 or for the widely localized lowest-energy state of CP2926 reveals that the vibronic structure observed for WSCP is mainly that of the bound pigment molecule. Furthermore, both the relative and integral Huang-Rhys factors of Chl a-WSCP are highly conserved in comparison with those of Chl a in the PSI-200 complex,42 LHC II,43 and CP2926 (see Table 1). This indicates that excitonic coupling in the range of ∼100 cm-1 has little effect on electron-vibrational coupling strengths; i.e., the S-values remain virtually unaffected by the strong excitonic coupling within the Chl-homodimers of WSCP. Although the vibrational properties of Chl a and Chl b appear to be widely conserved in WSCP, the presence of two additional modes at 228 and 327 cm-1 for Chl a-WSCP is most remarkable. Similar vibrational modes are observed at 221 and 325 cm-1 for Chl b-WSCP. In addition, Chl b-WSCP exhibits a further additional mode at 283 cm-1 which is also observed in Chl a-WSCP and Chl a in solution but not in Chl b in solution. A direct assignment of these modes is quite difficult because vibrations in the 200-600 cm-1 range are typically associated with a variety of collective vibrational motions of the tetrapyrrole backbone of Chl.44 In this case, however, one would expect that the Chl binding to WSCP affects frequency and intensity of existing Chl vibrations. This could lead to frequency shifts or suppression of certain backbone motions by formation of additional bonds to the protein environment. Alternatively, vibrational frequencies in the 200-400 cm-1 range are also expected to arise from H-bonds,45 which are formed between the carbonyl group of ring B of Chl and a water molecule as illustrated in the inset of Figure 1. In addition, the structure of Horigome et al.10 suggests that the formyl group of Chl b can form an additional H-bond with the Leu 91 side chain of WSCP. This H-bond is not formed in Chl a-WSCP since the formyl group is replaced by a methyl group in Chl a. Thus, it appears reasonable to identify the two additional modes present in both Chl a (228 and 327 cm-1) and Chl b-WSCP (221 and 325 cm-1) as vibrations of the H-bonds of the water molecule bound close to the Chl molecule. The third additional 283 cm-1 mode observed only for Chl b-WSCP could be assigned to the H-bond formed by its formyl group. This idea is even consistent with the presence of a mode in this spectral region for Chl a because molecular dynamics studies assigned 269 and 285 cm-1 modes to methyl group motions.46,47 Thus, a motion of the methyl group of ring B of Chl a found at 283 cm-1 may be replaced by that of a H-bond between the corresponding formyl group of Chl b and the protein backbone. WSCP Electron-Phonon Coupling. As for electron-phonon coupling, the one-phonon profile of Chl a-WSCP is found to reveal an interesting substructure with a pronounced peak at ∼24 cm-1, a shoulder at ∼48 cm-1, and a further peak at ∼88 cm-1. Chl b-WSCP exhibits a similarly structured onephonon profile with slightly shifted peak frequencies and altered relative intensities. The integral Huang-Rhys factors determined at the red edge of the fluorescence origin bands were ∼0.8 and 0.85 for Chl a and Chl b-WSCP, respectively. The unusual dependence of the S-values on excitation wavelength appears to originate mainly from nonresonant excitation, so that the S-factors found at the red edge of the fluorescence origin

ARTICLE

bands can be viewed as the true electron-phonon coupling strengths of the lower exciton level of WSCP. Nevertheless, there is a possibility that a small contribution of electronic states of weakly coupled Chl-dimers, as suggested in refs 9 and 21, is located at the blue edge of the fluorescence origin bands. This interpretation will be discussed in more detail in the accompanying paper (see part II, DOI 10.1021/jp111457t). Broad and asymmetric one-phonon profiles are generally observed for photosynthetic pigment-protein complexes,22,26,35-39 while a similar structure within a one-phonon profile has so far only been reported for LH2.40 A most remarkable feature is the finding that the higher-frequency peaks are close to higher harmonics of the main ∼24 cm-1 peak, while for the given integral S (see above) they cannot be expressed as multiple phonon transitions (not shown). Because of the excitonic nature of the fluorescing state of WSCP, the distinct peaks within the one-phonon profile could be a coherent contribution equivalent to quantum beats observed in time-resolved experiments.48 Unfortunately, low-temperature time-resolved experiments with femtosecond resolution are not yet available for Chl a- and Chl b-WSCP, so the latter possibility cannot be verified at the moment. One may also argue that in the case of Chl b-WSCP the shape of the one-phonon profile is quite similar for both the 666 and the 676 nm absorption bands, although the latter band belongs most likely to a heterodimer consisting of a Chl b and a Chl-derivative. This heterodimer is expected to be characterized by much smaller excitonic delocalization (cf. ref 16 for consideration of a Chl a-, Chl b heterodimer of similar diagonal disorder). The latter observation suggests that the shape of the one-phonon profile is determined by the properties of the protein environment, i.e., by the protein density of states (see below). Therefore, it is attractive to speculate about an alternative interpretation in terms of the highly ordered structure of the WSCP tetramer being composed of four structurally equivalent subunits. Briefly, the one-phonon profile l1 (ω) represents the product of the density of states (DOS) of the phonon modes and of an electron-phonon coupling term. Both terms cannot be determined separately by site-selective optical spectroscopic techniques like SHB and FLN. However, the density of phonon states can be obtained separately using the complementary method of inelastic neutron scattering (INS).49 Although according to INS data the density of phonon states appears to be relatively similar for many different proteins (see refs 49-52), its spectral form was shown to vary with protein size (see, e.g., refs 50 and 52) and is due to interaction with the solvent (see, e.g., ref 51). The shift of the phonon peak frequency to lower values with delocalization over larger protein subunits may also explain the present results for WSCP. Accordingly, one may assume that the low-frequency peak at ∼88 cm-1 corresponds to the protein vibrations of a monomeric WSCP subunit and that the harmonic peaks at ∼24 and ∼48 cm-1 can be identified with vibrations delocalized over WSCP tetramers and dimers, respectively. A similar possibility has been discussed for LHC II;49 however, the larger structural heterogeneity of the latter protein with several structurally inequivalent helices may broaden the corresponding features so that no distinct peaks in the one-phonon profile can be observed. Nevertheless, further studies, e.g., molecular dynamics simulations, have to be carried out to verify this interpretation. Finally, it is remarkable to note that the integral S-factors determined here for the 666 nm absorption band of Chl b-WSCP are about 10% lower than those observed for the 676 nm 4050

dx.doi.org/10.1021/jp111455g |J. Phys. Chem. B 2011, 115, 4042–4052

The Journal of Physical Chemistry B absorption band, although the protein environment can be assumed to be highly similar except for special binding motifs of Chl b vs a Chl b derivative. As discussed above, the 676 nm absorption band may belong to a heterodimer with smaller excitonic delocalization. Thus, the difference of the S-factors observed for both bands may be interpreted as a suppression of S with increasing excitonic delocalization.53,54

5. CONCLUDING REMARKS The 4.5 K difference fluorescence line-narrowing experiments presented above provide detailed quantitative insight into electron-phonon and electron-vibrational coupling of excitonically coupled Chl homodimers in recombinant class-IIa WSCP from cauliflower. These parameters are essential for a precise simulation of the excitonic states’ lineshapes and thus an important prerequisite for a proper understanding of pigment-protein interactions in photosynthetic antenna and reaction center complexes in general. The electron-phonon coupling strength characterized by integral Huang-Rhys factors S in the order of 0.81-0.85 appears to be quite similar to that typically reported for pigment-protein complexes. However, the one-phonon profile associated with the lower exciton state of Chl homodimers of WSCP reveals an unusually structured shape with peak phonon frequencies (ωm) of ∼24, 48, and 88 cm-1, respectively. It is attractive to speculate that the latter structure may be due to different delocalization domains of the protein phonons within the WSCP tetramer. Furthermore, a rich structure of S1fS0 vibrational frequencies in the wavenumber range between 180 and 1665 cm-1 and their individual Huang-Rhys factors are assigned. Two distinct modes at 228 and 327 cm-1 have not been reported for chlorophyll in solution before, so that they are most likely attributed to specific structural motifs in the pigment binding pocket of WSCP, i.e., to H-bond formation in the vicinity of the chlorophyll molecules. The wealth of information on electron-vibrational coupling gathered by difference fluorescence line-narrowing underlines the great potential of this type of spectroscopy. ’ AUTHOR INFORMATION Corresponding Author

*Technical University, PC14, Strasse des 17. Juni 135, 10623 Berlin, Germany. Phone: þ49-30-31421067. Fax: þ49-30-31421122. E-mail: [email protected]. Present Addresses #

Institute of Physics, University of Tartu, Riia 142, 51014 Tartu, Estonia. Phone: þ(372) 737 4627. Fax: þ(372) 738 3033. E-mail: [email protected]. Author Contributions ^

Authors contributed equally.

’ ACKNOWLEDGMENT Estonian Science Foundation (Grant No. 8674) and Estonian Ministry of Education and Science (Grant No. SF0180055s07) have supported this work. J.P. and G.R. gratefully acknowledge support from Deutsche Forschungsgemeinschaft (SFB 429, TP A1). We are also grateful to S. Kussin and M. Wess (TU Berlin) for their help in sample preparation.

ARTICLE

’ REFERENCES (1) Renger, G.; Holzwarth, A. R. Primary electron transfer. In Photosystem II: The water/plastoquinone oxido-reductase in photosynthesis; Wydrzynski, T., Satoh, K., Eds.; Kluwer Academic Publishers: Dordrecht, The Netherlands, 2005; pp 139-175. (2) van Amerongen, H.; Croce, R. In Primary processes of photosynthesis: Basic principles and apparatus; Renger, G., Ed.; RSC Publ.: Cambridge, U.K., 2008; Vol. I, p 329. (3) Liu, Z.; Yan, H.; Wang, K.; Kuang, T.; Zhang, J.; Gui, L.; An, X.; Chang, W. Nature 2004, 428, 287. (4) Standfuss, J.; Lamborghini, M.; K€uhlbrandt, W.; van Scheltinga, A. C. T. EMBO J. 2005, 24, 919. (5) Loll, B.; Kern, J.; Saenger, W.; Zouni, A.; Biesiadka, J. Nature (London, U. K.) 2005, 438, 1040. (6) Guskov, A.; Kern, J.; Gabdulkhakov, A.; Broser, M.; Zouni, A.; Saenger, W. Nat. Struct. Mol. Biol. 2009, 16, 334. (7) Satoh, H.; Uchida, A.; Nakayama, K.; Okada, M. Plant Cell Physiol. 2001, 42, 906. (8) Schmidt, K.; Fufezan, C.; Krieger-Liszkay, A.; Satoh, H.; Paulsen, H. Biochemistry 2003, 42, 7427. (9) Theiss, C.; Trostmann, I.; Andree, S.; Schmitt, F. J.; Renger, T.; Eichler, H. J.; Paulsen, H.; Renger, G. J. Phys. Chem. B 2007, 111 (46), 13325. (10) Horigome, D.; Satoh, H.; Itoh, N.; Mitsunaga, K.; Oonishi, I.; Nakagawa, A.; Uchida, A. J. Biol. Chem. 2007, 282 (9), 6525. (11) Satoh, H.; Uchida, A.; Nakayama, K.; Okada, M. Plant Cell Physiol. 2001, 42, 906. (12) Satoh, H.; Nakayama, K.; Okada, M. J. Biol. Chem. 1998, 46, 30568. (13) Takamiya, K.; Tsuchiya, T.; Ohta, H. Trends Plant Sci. 2000, 5, 426. (14) H€ortensteiner, S. Cell. Mol. Life Sci. 1999, 56, 330. (15) Matile, P.; Schellenberg, M.; Vicentini, F. Planta 1997, 201, 96. (16) Renger, T.; Trostmann, I.; Theiss, C.; Madjet, M. E.; Richter, M.; Paulsen, H.; Eichler, H. J.; Knorr, A.; Renger, G. J. Phys. Chem. B 2007, 111 (35), 10487. (17) Renger, T.; Madjet, E.; M€uh, F.; Trostmann, I.; Schmitt, F. J.; Theiss, C.; Paulsen, H.; Eichler, H. J.; Knorr, A.; Renger, G. J. Phys. Chem. B 2009, 113, 9948. (18) Hughes, J. L.; Razeghifard, R.; Logue, M.; A. Oakley Wydrzynski, T.; Krausz, E. J. Am. Chem. Soc. 2006, 128, 3649. (19) Parson, W. In Primary processes of photosynthesis: Basic principles and apparatus; Renger, G., Ed.; RSC Publ.: Cambridge, 2008; Vol. II, p 57. (20) Renger, G. In Primary processes of photosynthesis: Basic principles and apparatus; Renger, G., Ed.; RSC Publ.: Cambridge, 2008; Vol. II, p 237. (21) Schmitt, F.-J.; Trostmann, I.; Theiss, C.; Pieper, J.; Renger, T.; Fuesers, J.; Hubrich, H.; Paulsen, H.; Eichler, H. J.; Renger, G. J. Phys. Chem. B 2008, 112, 13951. (22) Jaaniso, R. Proc. Acad. Sci. Est. SSR. Phys. Math. 1985, 34, 277. (23) F€unfschilling, J.; Glatz, D.; Zschokke-Gr€anacher, I. J. Lumin. 1986, 36, 85. (24) R€atsep, M.; Freiberg, A. Chem. Phys. Lett. 2003, 377, 371. (25) R€atsep, M.; Freiberg, A. J. Lumin. 2007, 127, 251. (26) R€atsep, M.; Pieper, J.; Irrgang, K.-D.; Freiberg, A. J. Phys. Chem. B 2008, 112, 110. (27) Hobe, S.; Fey, H.; Rogl, H.; Paulsen, H. J. Biol. Chem. 2003, 278 (8), 5912–5919. (28) Pieper, J.; Voigt, J.; Renger, G.; Small, G. J. Chem. Phys. Lett. 1999, 310, 296. (29) Reppert, M.; Naibo, V.; Jankowiak, R. J. Chem. Phys. 2010, 133, 9. (30) Hayes, J. M.; Lyle, P. A.; Small, G. J. J. Phys, Chem. 1994, 98, 7337. (31) Rebane, K. K. Impurity Spectra of Solids; Plenum: New York, 1970. 4051

dx.doi.org/10.1021/jp111455g |J. Phys. Chem. B 2011, 115, 4042–4052

The Journal of Physical Chemistry B

ARTICLE

(32) Avarmaa, R. A.; Rebane, K. K. Spectrochim. Acta 1985, 41, 1365. (33) R€atsep, M.; Linnanto, J.; Freiberg, A. J. Chem. Phys. 2009, 130, 194501. (34) Hughes, J. L.; Conlon, B.; Wydrzynski, T.; Krausz, E. Phys. Procedia 2010 3, 1591–1599. (35) Peterman, E. J. G.; Pullerits, T.; van Grondelle, R.; van Amerongen, H. J. Phys. Chem. B 1997, 101, 4448. (36) Pieper, J.; Sch€odel, R.; Irrgang, K.-D.; Voigt, J.; Renger, G. J. Phys. Chem. B 2001, 105, 7115. (37) Timpman, K.; R€atsep, M.; Hunter, C. N.; Freiberg, A. J. Phys. Chem. B 2004, 108, 10581. (38) Hofmann, C.; Michel, H.; van Heel, M.; K€ohler, J. Phys. Rev. Lett. 2005, 94, 195501. (39) Brecht, M.; Studier, H.; Radics, V.; Nieder, J. B.; Bittl, R. J. Am. Chem. Soc. 2008, 130 (51), 17487. (40) Freiberg, A.; R€atsep, M.; Timpman, K.; Trinkunas, G. Chem. Phys. 2009, 357, 102. (41) R€atsep, M.; Pajusalu, M.; Freiberg, A. Chem. Phys. Lett. 2009, 479, 140. (42) Gillie, J. K.; Small, G. J.; Golbeck, J. H. J. Phys. Chem. 1989, 93, 1620. (43) Pieper, J.; R€atsep, M.; Jankowiak, R.; Irrgang, K.-D.; Voigt, J.; Renger, G.; Small, G. J. J. Phys. Chem. A 1999, 103, 2412. (44) Lutz, M. J. Raman Spectrosc. 1974, 2, 497. (45) Steiner, T. Angew. Chem., Int. Ed. 2002, 41, 48. (46) Goupil-Lamy, A. V.; Smith, J. C.; Yunoki, J.; Parker, S. F.; Kataoka, M. J. Am. Chem. Soc. 1997, 119 (39), 9268. (47) Holderna-Natkaniec, K.; Kazimierz, J.; Natkaniec, I.; Nowak, D.; Szyczewski, A. Chem. Phys. 2005, 317, 178. (48) Toutounji, M.; Small, G. J.; Mukamel, S. J. Chem. Phys. 1998, 109, 7949. (49) Pieper, J.; Irrgang, K.-D.; Renger, G.; Lechner, R. E. J. Phys. Chem. B. 2004, 108, 10556. (50) Orecchini, A.; Paciaroni, A.; Bizzarri, A. R.; Cannistraro, S. J. Phys. Chem. B 2002, 106, 7348. (51) Paciaroni, A.; Orecchini, A.; Cinelli, S.; Onori, G.; Lechner, R. E.; Pieper, J. Chem. Phys. 2003, 292, 397. (52) Kurkal-Siebert, V.; Smith, J. C. J. Am. Chem. Soc. 2006, 128, 2356. (53) Hochstrasser, R. M.; Prasad, P. N. J. Chem. Phys. 1972, 56, 2814. (54) Merrifield, R. E. J. Chem. Phys. 1964, 40, 445.

4052

dx.doi.org/10.1021/jp111455g |J. Phys. Chem. B 2011, 115, 4042–4052