Temperature Dependence of Chlorophyll Triplet Quenching in Two

Article Cite This: J. Phys. Chem. B XXXX, XXX, XXX−XXX

pubs.acs.org/JPCB

Temperature Dependence of Chlorophyll Triplet Quenching in Two Photosynthetic Light-Harvesting Complexes from Higher Plants and Dinoflagellates Ivo S. Vinklárek,† Till L. V. Bornemann,‡,§ Heiko Lokstein,† Eckhard Hofmann,‡ Jan Alster,† and Jakub Pšenčík*,† J. Phys. Chem. B Downloaded from pubs.acs.org by UNIV OF LOUISIANA AT LAFAYETTE on 09/24/18. For personal use only.

†

Department of Chemical Physics and Optics, Faculty of Mathematics and Physics, Charles University, Ke Karlovu 3, 121 16 Prague 2, Czech Republic ‡ Protein Crystallography, Faculty of Biology and Biotechnology, Ruhr-University Bochum, D-44780 Bochum, Germany S Supporting Information *

ABSTRACT: Chlorophyll (Chl) triplet states generated in photosynthetic light-harvesting complexes (LHCs) can be quenched by carotenoids to prevent the formation of reactive singlet oxygen. Although this quenching occurs with an efficiency close to 100% at physiological temperatures, the Chl triplets are often observed at low temperatures. This might be due to the intrinsic temperature dependence of the Dexter mechanism of excitation energy transfer, which governs triplet quenching, or by temperature-induced conformational changes. Here, we report about the temperature dependence of Chl triplet quenching in two LHCs. We show that both the effects contribute significantly. In LHC II of higher plants, the core Chls are quenched with a high efficiency independent of temperature. A different subpopulation of Chls, which increases with lowering temperature, is not quenched at all. This is probably caused by the conformational changes which detach these Chls from the energy-transfer chain. In a membrane-intrinsic LHC of dinoflagellates, similarly two subpopulations of Chls were observed. In addition, another part of Chl triplets is quenched by carotenoids with a rate which decreases with temperature. This allowed us to study the temperature dependence of Dexter energy transfer. Finally, a part of Chls was quenched by triplet− triplet annihilation, a phenomenon which was not observed for LHCs before.

■

INTRODUCTION In photosynthesis, photons are absorbed by pigments in lightharvesting complexes (LHCs). In this way formed excitations are transferred between the pigments toward the reaction centers (RCs), where the excitation energy is used for charge separation. The LHCs enlarge the absorption cross section and also the rate with which excitations are delivered to the RCs. However, if this rate exceeds the turnover rate of the RCs, excitations may accumulate in the LHCs, which increase the probability of harmful reactions. Chlorophylls (Chls) or bacteriochlorophylls in LHCs may undergo intersystem crossing from the lowest excited singlet state to a relatively long-lived triplet state. The triplet states can be quenched by molecular oxygen and generate a singlet excited state of oxygen(1O2), which is highly reactive and can cause damage to the photosynthetic apparatus. An efficient protection against 1 O2 is provided by carotenoids. These pigments can quench the triplet states of Chls or 1O2.1,2 Most of the excitation energy transfer (EET) toward the RCs occurs between the singlet states of the pigments and can be often described by the Förster mechanism or (modified) Redfield theory, depending on the extent of involvement of © XXXX American Chemical Society

excitonic interactions between the pigments. These mechanisms exhibit only indirect dependence on temperature through a Boltzmann weighting of transfer rates and spectral overlaps. By contrast, the protective quenching of the Chl triplet states occurs via triplet energy transfer, which can be described by the Dexter mechanism of EET (DET) and which is intrinsically temperature dependent because of an involvement of a thermal barrier.3 The DET can be described as a double-electron exchange, and its rate constant is therefore also strongly dependent on the distance between the donor and the acceptor. From the Marcus theory, it follows that the rate constant can be written as kDET =

|Hif |2 ijj π 2 yzz A i −ΔG yz iA y zz = 1 expjjj 2 zzz jj zz expjjj ℏ k λRT { T k RT { kT {

(1)

where T is the absolute temperature, H is the Hamiltonian of the electron coupling interaction between the donor and the Received: July 14, 2018 Revised: August 29, 2018 Published: September 4, 2018 A

DOI: 10.1021/acs.jpcb.8b06751 J. Phys. Chem. B XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry B acceptor, R is the molar gas constant, and λ and ΔG are the reorganization energy and free activation energy, respectively.3 The parameters A1 and A2 are used later in the text for the fitting of the measured temperature dependence. The equation implies that the rate constant of DET decreases as the temperature decreases. In this study, we focus on the experimental study of DET efficiency between 77 K and room temperature in two LHCs: LHC II of higher plants, which is the most abundant membrane protein on the earth, and the Chl a−Chl c2− peridinin−protein complex (acpPC) from the dinoflagellate alga Amphidinium carterae. Both LHC II and acpPC belong to the so-called LHC II superfamily. Sequence identities between the members are between 25 and 80%.4,5 Despite the protein sequence similarities, the pigment composition of both complexes is significantly different. One monomer of LHC II contains 14 Chls (8 Chls a and 6 Chls b) and 4 xanthophyll carotenoids (2 luteins, 1 neoxanthin, 1 violaxanthin, or its de-epoxidation products due to xanthophylls cycle activity).6 The two luteins are positioned in the center of the complex close to the red-most Chls and therefore are assumed to have a crucial role in Chl triplet quenching.7,8 In contrast to LHC II, the structure and composition of acpPC has not been determined yet, and therefore, the available information is rather limited. Purified acpPC consists of 18−19 kDa polypeptide unit most likely formed by three transmembrane helices, as in LHC II.9−11 It was proposed that there are five conserved Chl binding sites, and some Chls a are replaced by carotenoids in acpPC.5,11,12 The two binding sites occupied by luteins in LHC II are supposedly conserved in acpPC and occupied by peridinins.5 Niedzwiedzki et al.13 showed that the 77 K absorption spectrum of acpPC from another dinoflagellate alga, Symbiodinium, can be deconvoluted into contributions from 4 Chls a, 6 Chls c2, 6 peridinins, and 1 diadinoxanthin. The larger content of carotenoids as compared to LHC II is apparently a result of adaptation of dinoflagellates to the conditions below the surface of the ocean, where the available light is enriched in the blue-green part of the spectrum.14,15 Previously, it was observed that the Chl triplets in acpPC were rapidly quenched by carotenoids at room temperature (295 K), with an efficiency close to 100%.16 However, relatively long-lived Chl triplet states were observed in acpPC at 77 K.13 This seems to be in line with the results reported for many other LHCs in which triplet states of Chls were detected at low temperatures, although they were efficiently quenched at physiological conditions. For instance, no Chl triplets were observed in LHC II at room temperature, but the efficiency of quenching gradually decreased to 94% at 77 K and 82% at 4 K.17 However, for acpPC, only 7% of Chl triplets was estimated to be quenched by carotenoids at 77 K.13 One of the goals of this work was to understand why these two related LHCs exhibit such a different efficiency of Chl triplet quenching at 77 K. Another aim was to elucidate whether the changes in the quenching efficiency are primarily caused by the intrinsic temperature dependence of DET, or by conformational changes caused by the freezing of the complexes. Our results provide evidence that both effects contribute. Most likely, the outer regions of the LHCs are affected more by the conformational changes, whereas their central parts remain highly efficient in Chl triplet quenching even at low temperatures. In acpPC, we have observed triplet−triplet annihilation, which was not observed in any other LHC so far.

This phenomenon explains the previously observed differences in quenching efficiencies at low temperature.

■

MATERIALS AND METHODS Sample Preparation. LHC II was isolated from fresh spinach leaves as described previously18 with slight modifications. The buffer for measurements at 295 K consisted of 20 mM (pH 7.8) tricine and 0.06% n-dodecyl β-D-maltoside (DDM, Glycon). The buffer for low-temperature samples was prepared from 20 mM tricine and 0.12% DDM and mixed with glycerol (Sigma-Aldrich) in a ratio of 1:2. The slightly different composition of the buffer was required to prevent sample aggregation in the presence of glycerol. The acpPC complex was isolated as described in ref 19. For measurements at 295 K, the sample was dissolved in a buffer (pH 7.5) containing 20 mM tricine (Sigma-Aldrich), 20 mM KCl (Sigma-Aldrich), and 0.16% n-decyl β-D-maltoside (DM, Glycon). The composition of the buffer was modified for lowtemperature measurements to contain 70 mM tricine, 70 mM KCl, and 0.56% DM. The buffer was mixed with glycerol in a ratio of 1:2. The absorbance of the sample was set to 0.4−0.6 in a 1 cm path cuvette at the Qy band for all samples. All measurements were performed under aerobic conditions. Optical Spectroscopy. The steady-state absorption spectra were measured using a SPECORD 250 spectrophotometer (Analytik Jena) before and after every transient absorption experiment to monitor sample intactness. The lowtemperature absorption spectra were measured with the sample placed in an optical cryostat (OptistatDN2, Oxford Instruments). Transient spectra were measured as described previously.16 In brief, a sample was excited by ∼3 ns (full width at half-maximum) pulses from an optical parametric oscillator (PG122, EKSPLA) pumped by a Q-switched Nd:YAG laser (NL303G/TH, EKSPLA). The energy of the excitation pulses was adjusted to values between 300 and 500 μJ by a set of neutral density filters. A xenon flash lamp (LS1130-1 Flashpack with a FX-1161 flashtube, PerkinElmer) was used as a source of probe and reference pulses. The probe light transmitted through the sample together with the reference light were detected using an intensified CCD camera (PI-MAX 512RB, Roper Scientific) coupled to an imaging spectrometer (iHR 320, Horiba Jobin Yvon). Transient absorption spectra were measured between ∼380 and 745 nm at defined delays after excitation. A 2 ns gate width was used. For lowtemperature measurements, the sample was placed in an OptistatDN2 cryostat controlled by a Mercury iTC unit (Oxford Instruments). The sample was placed in plastic cuvettes with 1 cm optical path. The transient absorption spectra were processed by global analysis as described in ref 16 except for the spectral and temporal regions containing the fluorescence signal. Spontaneous emission cannot be fitted simultaneously with the data recalculated to transient absorption.16,20 The transient spectra were measured at delays up to 3.6 ms, and therefore, a long-lived millisecond component could not be determined with a high accuracy. For the measurement of the delayed fluorescence (DF), a 50 ns gate scanned from 125 ns to 5 μs and a higher gain were used. Prompt fluorescence of acpPC was measured at 77 K using the same setup with a gate of 50 ns at a zero time, without scanning and thus effectively integrating all the prompt fluorescence. B


Article

The Journal of Physical Chemistry B

Figure 1. Steady-state absorption spectra of LHC II (a) and acpPC (b) at 295 K (black lines) and 77 K (red lines). The arrows indicate the excitation wavelengths used in the transient absorption experiments.

■

RESULTS Steady-State Absorption Spectra of LHC II and acpPC. The absorption spectra of LHC II at 295 and 77 K are shown in Figure 1a. The blue-green spectral region is dominated by the Soret bands of Chl a (maximum at 436 nm) and Chl b (471 nm) overlapping with the absorption of carotenoids. The Qy bands of Chl a and Chl b have maxima at 672 and 653 nm, respectively. The bands in the spectral region between 630 and 650 nm are ascribed to the vibronic structure of both the Q-bands. 17,21 The spectral resolution is significantly improved at low temperature. The spectrum at 77 K shows several peaks centered at 662, 672, and 676 nm, which fit subpopulations of Chl a as reported earlier.21 The Qy band of Chl b is also better resolved as well as the Soret band region, where several bands at 461, 486, and 494 nm can be discerned. The absorption spectra of acpPC are shown in Figure 1b. The absorption between 350 and 570 nm, together with the sidebands at ∼500 and ∼540 nm, is mainly due to carotenoids. Two maxima at 441 and 461 nm are assigned to the Soret bands of Chl a and Chl c2, respectively. A peak with a maximum at 672 nm is due to the Qy transition of Chl a. Its blue-shifted sidebands at 636 and 650 nm are due to its vibrational structure, and the Qx band overlaps with the Qbands of Chl c2 at 632 and 649 nm, which were identified by circular dichroism and low-temperature absorption spectroscopy.1,10 Transient Absorption and Decay-Associated Spectra of LHC II at 295 K. The transient absorption spectra of LHC II in the buffer without glycerol for excitation at 675 nm (the Qy band of Chl a) and selected delays at 295 K are shown in Figure 2a. The spectrum at 1 ns is dominated by the negative Soret band ground-state bleaching (GSB) of Chl a with a minimum at around 440 nm and a strong fluorescence signal at around 680 nm, which overlaps with the GSB of the Qy band of Chl a. Fluorescence vanishes within 50 ns. No Soret band GSB of Chl b was observed, not even in the case of its direct excitation at 470 nm (Figure S1a), proving the intactness of the complex and that singlet EET between Chl b and Chl a occurs fast and efficiently.22 GSB of Chl a decays within a few tens of nanoseconds. It is replaced by a broad GSB of

carotenoids, with a partly resolved vibrational structure and the main minimum at ≈425 nm,17 and two positive bands in the spectral region between 460 and 570 nm. This positive signal is assigned to excited-state absorption (ESA) from the triplet states (triplet ESA) of carotenoids.8,17 At later times (>50 ns), only carotenoid triplets contribute to the transient absorption spectra and their signal decays within ∼10 μs. These results agree well with the previously published data with a lower time resolution.17 The maximum of the main ESA band shifts from 510 to 515 nm during the decay because of the presence of two carotenoid triplets with different lifetimes (see below). The ESA band with a maximum at 475 nm remains at the same position. The same spectral features were observed when LHC II was measured in a buffer−glycerol mixture, which was used for low-temperature experiments (Figure S1). The only difference between the experiments in both buffers was a slower decay of the formed carotenoid triplet states, which was caused by limited diffusion of oxygen in a glycerol-containing buffer (see the Discussion section). The transient absorption spectra were processed by global analysis. At room temperature, two components were sufficient to provide satisfactory fits, with the lifetimes of ∼5.5 ns and ∼2.6 μs, respectively (Figure 2b). When the data were fitted with three components, the microsecond component splits into two, with the spectral shape and lifetimes very similar to what was resolved at the aerobic conditions previously8,17 (Figure S2). The microsecond component(s) are ascribed to the decay of the carotenoid triplet states, most probably located on the two central luteins with a different oxygen accessibility.8,17 The results for LHC II in a buffer-containing glycerol were the same, except for a prolongation of the carotenoid triplet-state lifetime from ∼2.6 to ∼5 μs. Only a single microsecond component could be resolved (Figure S3). It shows that the limited diffusion of oxygen in the glycerolcontaining buffer affects both luteins in a similar way. In addition to the features related to the decay of carotenoid triplet states, the ∼microsecond component also exhibits a socalled interaction peak at ∼675 nm23,24 (see the Discussion section). The ∼5.5 ns component was not affected by the change of buffer. It describes the decay of the excited states of Chl a and the population of the carotenoid triplets. The C


Article


Figure 2. Transient absorption spectra of LHC II in a buffer at 295 K after excitation at 675 nm (a) and DAS obtained after fitting with two decay components (b). The DAS after fitting with three components is shown in Figure S2. The inverted steady-state absorption spectra at 295 K (the black dotted lines) are shown for comparison.

Figure 3. Transient absorption spectra of acpPC in a buffer at 295 K after excitation at 670 nm (a) and DAS (b). The inverted steady-state absorption spectra at 295 K (the black dotted lines) are shown for comparison.

spectral features originating from Chls are typical for their singlet states. The carotenoid triplet states are formed with the same kinetics as observed for Chl singlet state decay. Because carotenoid triplet states can only be generated by triplet energy transfer from Chls on this time scale, it means that the triplet states of Chl a are generated and quenched by carotenoids with a transfer time faster than our resolution, as discussed in previous studies.16,20 Therefore, the component does not exhibit any Chl a triplet feature. Transient Absorption and DAS of acpPC at 295 K. Figure 3a presents the transient absorption spectra of acpPC in buffer at 295 K at selected delays. As in the case of LHC II, the sample was excited in the Qy band of Chl a, which has a maximum at 670 nm for acpPC. The transient spectrum at 1 ns is again dominated by a GSB of the Chl a Soret band and its

negative band with a minimum at 440 nm and a shoulder at 420 nm is due to the Soret band GSB of Chl a. The singlet ESA of Chl a at 450−500 nm is overlapping with a carotenoid triplet ESA with minima at 479 and 510 nm. The triplet ESA bands of carotenoids are negative because the ∼5.5 ns component reflects their rise. The rising (positive) GSB of carotenoids (∼380−475 nm) is overlapping with the decaying (negative) Soret band of Chl a. This, together with a higher extinction coefficient of carotenoids,25,26 leads to a relatively small contribution of Chl a to this component. The ∼5.5 ns component of the decay-associated spectra (DAS) corresponds to the decay of the singlet states of Chl a. The lifetime of this component 5.5 (±0.5) ns is similar to the previously reported fluorescence lifetime of Chl a in LHC II at 295 K,27 and of acpPC, as measured in this work (see below). In addition, the D


Article


Figure 4. Transient absorption spectra of LHC II at 77 K after excitation at 675 nm (a) and DAS (b). The inverted steady-state absorption spectra at 77 K (black dotted lines) are shown for comparison. Figure S6 shows the millisecond component of DAS alone.

Figure 5. Transient absorption spectra of acpPC at 77 K after excitation at 670 nm (a) and DAS (b). The inverted steady-state absorption spectra at 77 K (the black dotted lines) are shown for comparison.

strong fluorescence.16 At longer times (>50 ns), the features belonging to Chl a vanish (including ESA around 475 and 610 nm) and are replaced by the features of carotenoid triplets characterized by a shallow negative band with a maximum at 440 nm, assigned to GSB and positive ESA bands at 472, 514, and 555 nm. The temporal evolution estimated from the DAS spectra (Figure 3b) is rather similar to the one observed for LHC II. The excited singlet states of Chls disappear with an ∼5.6 ns lifetime, and the carotenoid triplets are formed with the same kinetics. Later, the carotenoid triplet states decay with an ∼2.4 μs lifetime. The measured transient absorption spectra and the obtained DAS are very similar to the previously reported ones,16 proving a good reproducibility of both sample preparation and spectroscopic measurements. The main difference with respect to LHC II is a broad ESA band centered at 555 nm, which is due to the triplet state of

peridinin in a polar environment.16 The results obtained in a buffer-containing glycerol were the same except for a prolongation of the carotenoid triplet decay (Figure S4). The spectrum shows a negligible signal at wavelengths above 600 nm, where the interaction peak would be expected. Transient Absorption and DAS of LHC II at 77 K. The transient absorption spectra of LHC II at different temperatures were measured with excitation at 675 nm (the Qy band of Chl a, Figure 4) and at 470 nm (the Soret band of Chl b and carotenoids, Figure S5). No significant differences between both the excitation wavelengths were observed. The transient spectra measured after excitation at 675 nm are shown in Figure 4a. The time evolution and the shapes of the transient spectra at different delays are rather similar, as at 295 K. The only apparent difference is a slower decay of the carotenoid triplet states. E


Article


energy to the Chl a at low temperatures. In summary, the main features of the transient absorption spectra at 295 K are preserved at 77 K. However, some differences were discernible, which are discussed with the help of a global analysis in the next paragraphs. The transient spectra excited at 670 nm require four components to obtain a satisfactory fit (Figure 5b). The fastest DAS component has a similar lifetime of 5.9 (±0.7) ns and a spectral shape as the fast component resolved at 295 K. It is assigned to the decay of Chl a singlet states (GSB of the Soret band and a positive singlet ESA band in the spectral regions between 460−530 and 595−630 nm). Sharp minima at 475 and 520 nm are ascribed to the increasing signal of the triplet ESA of carotenoids. The DAS component thus shows that at least a fraction of Chl a triplet states is quenched by carotenoids with a high efficiency as was observed at 295 K. The main difference compared to the fast DAS resolved at 295 K is the absence of a rise of the triplet ESA of peridinin in a polar environment, which would manifest itself as a broad band with a minimum at ∼560 nm, see Figures 3b and S4. These peridinins are apparently not able to quench triplets of Chl a with the same rate as at 295 K. Such a rise of peridinin ESA is, however, discernible in the second DAS component with a lifetime of 133 (±15) ns, which was not resolved at room temperature. It further consists of a negative signal corresponding to the GSB of the Soret band of Chl a, with a minimum at the same wavelength (∼439 nm) as the ∼6 ns component; however, its lifetime is more than 20 times longer than the lifetime of Chl a singlet states, suggesting that these Chls a are in triplet states. This is supported by the blue shift of the ESA in the spectral region between 446 and 530 nm as compared to that resolved in the ∼6 ns component, which is typical for the triplet ESA of Chl a.13,16 Two sharp minima at 468 and 515 nm are due to the rise of triplet ESA of carotenoids other than peridinin in a polar environment and are blue-shifted as compared to the minima at 475 and 520 nm of the ∼6 ns component. This observation indicates that a different subpopulation of carotenoids is involved in the formation of the ∼130 ns component. The part of the ∼130 ns DAS component between 350 and 650 nm thus corresponds to the quenching of Chl triplets by a subpopulation of carotenoids, which is much slower than the quenching at 295 K and does not prevent accumulation of Chl triplet states. The strong negative band with a minimum at 680 nm, however, does not fit into this interpretation and must have a different origin. The band is approximately three times stronger than the GSB in the region of the Soret band, and therefore, it cannot be explained either by the GSB of the Qy band or as an interaction peak. The most likely explanation is that the band has its origin in emission (spontaneous and/or stimulated) from the S1 state of Chl a. However, emission at around 680 nm should not be observable at delays longer than the excited-state lifetime of the Chl a singlet states. Such a signal can only be explained by one of the mechanisms leading to the occurrence of DF. In the Discussion section, we will provide evidence that the mechanism is likely to be triplet− triplet annihilation occurring between the two Chl molecules in the triplet states. The DAS component with a lifetime of 15.0 (±0.4) μs shares most of the features with the slow component resolved in the data measured at 295 K. It describes the decay of the carotenoid triplets populated by triplet energy transfer from Chls. Their rise was resolved in both the ∼6 and ∼130 ns DAS

The global analysis confirmed these observations and revealed spectrally similar DAS as at 295 K (Figure 4b). The fast component has only a slightly longer lifetime (∼6.5 ns) as compared to 295 K, whereas the decay of the carotenoid triplet was prolonged more substantially, from ∼5.5 to ∼12.3 μs. In addition, however, the global analysis revealed one more, even slower component, which was too weak to be apparent in the transient absorption spectra, and was not resolved at 295 K. The prolongation of the nanosecond component is most likely due to an intrinsic dependence of the S1-state lifetime on temperature, which was observed before.28 No signs of Chl triplets could be resolved in the nanosecond component, meaning that the quenching of Chl triplets by the two luteins is efficient even at a low temperature. The carotenoid triplets decay with a lifetime of ∼12 μs, which is in line with the previously reported results.17 The most significant difference as compared to 295 K is that an additional weak DAS component was resolved, which decays with a millisecond lifetime. For excitation at 675 nm, this component exhibits a spectrum typical for Chl a in the triplet state, while the spectrum exhibits features of both Chl a and Chl b triplets in the case of excitation at 470 nm (Figure S6). The population of Chl triplets is less than 7%. A similar result was reported previously.17 It was estimated that the efficiency of Chl triplet-state quenching was 94(±2)% at 77 K. From the difference between the emission spectra excited at 675 nm (Figure 4a) and 470 nm (Figure S5a), it is obvious that Chl b contributes to the emission at 77 K, but not at 295 K (compare Figures 2 and S1). Our results thus indicate that there are two subpopulations of Chl molecules within LHC II: a major part of the Chl triplet states is quenched with a high efficiency as at room temperature, and a minor fraction is not quenched at all. Transient Absorption and DAS of acpPC at 77 K. The transient absorption spectra of acpPC at low temperatures were measured upon excitation at 670 nm (the Qy band of Chl a), 540 nm (peridinin), and 460 nm (the Soret band of Chl c2). Figure 5a presents the transient absorption spectra of acpPC at 77 K after excitation at 670 nm. The GSB of the Soret band remains discernible after an initial fast decay even at 100 ns after excitation. The fluorescence decay is biphasic, and the fluorescence is observable until ∼250 ns. This is rather unusual because all the singlet states of Chl a decay only slightly slower (with ∼6 ns, see below) than at 295 K. The broad singlet ESA of Chl a between 460 and 525 nm is rapidly replaced by two triplet ESA bands of carotenoids with maxima at 472 and 514 nm and a broad triplet ESA of peridinin in a polar environment at 555 nm. The decay of the carotenoid triplet states in the spectral region between 450 and 620 nm is significantly slower as compared to 295 K, similarly as observed for LHC II. Figure S7a shows the transient absorption spectra of acpPC excited at 540 nm, where mainly peridinin absorbs. However, a certain amount of Chl c2 and diadinoxanthin is also excited. The differences between the excitation at 540 and 670 nm in the spectral region between 375 and 630 nm are negligible. The situation is different above 630 nm. This region is dominated by a strong fluorescence signal from Chl a; however, the fluorescence of Chl c2 is discernible at ∼650 nm after excitation at 540 nm. This is surprising because no fluorescence from Chl c2 was observed at 295 K (all the excitation from Chl c2 is transferred to Chl a within 1 ns).1,13 Apparently, some of the Chl c2 molecules cannot transfer F


Article


Temperature Dependence. The temperature dependences of the lifetimes of the DAS components of LHC II and acpPC were estimated for different temperatures between 77 and 295 K. Figure 7 shows the temperature dependence of the

components. The spectral region between 460 and 600 nm is a mirror image of the combination of both nanosecond components of the DAS. The last DAS component resembles the millisecond component resolved for LHC II at low temperatures. It has a lifetime of 2.3 (±0.7) ms and a spectrum typical for Chl a in the triplet state. The DAS for excitation at 540 nm, where mainly peridinin absorbs, is shown in Figure S7b. Except for a weak GSB signal from the Soret band of Chl c2 in the triplet state at 465 nm in the millisecond component, the DAS are very similar to those resolved after excitation at 670 nm. Apparently, some Chl c2 molecules cannot transfer their excitation energy to Chl a at low temperatures. This leads to the observation of the abovementioned fluorescence and population of the triplet state. Because its lifetime is in a millisecond range, we may speculate that no carotenoid is located in their close-enough neighborhood, at least at 77 K. The population of the longlived triplets of Chl c2 is less than 7% as estimated from the analysis of the transient absorption spectra. The signal from Chl c2 in the triplet state is much stronger after direct excitation at 460 nm and dominates the millisecond DAS component (Figure 6), while the faster three components (∼6

Figure 7. Temperature dependence of the lifetimes of the 5−6 ns and 2−12 μs DAS components resolved for LHC II. The lifetimes were averaged from the data obtained after excitation at 675 and 470 nm.

5.5−6.5 ns and 5−12 μs DAS components resolved for LHC II. The temperature dependence of the millisecond lifetimes ascribed to unquenched Chl triplets was not analyzed because of their weak amplitude. This led to a large error of the estimated values. The results for excitation at 675 and 470 nm were analyzed together because the differences between the lifetimes are less than the experimental error, and no significant differences were observed for the DAS and TA spectra. The temperature dependence of the lifetimes is fitted by the Arrhenius equation kdecay = B1 exp( − B2 /kT )

(2)

according to ref 29 where B1 is the pre-exponential (or frequency) factor and B2 is the activation energy. In the case of acpPC, the transient spectra excited at 670 nm were evaluated (Figure 8). The ∼5 ns component shows only a weak temperature dependence, with changes less than the experimental error. Therefore, the data were fitted by a constant. The lifetime of the slower nanosecond component (∼10−130 ns) is fitted by eq 1. This component describes a temperature-dependent quenching of Chl triplet states by carotenoids and, thus, presumably the direct dependence of the DET rate on temperature. It should be noted that this dependence could not be satisfactorily fitted by the Arrhenius law (2). The lifetime of this component and the amplitude decrease with temperature and become indistinguishable from the faster nanosecond component above 250 K. The shape of the fit is in good agreement with experimental results, and the determined parameters from eq 1 are A1 = 3531 K·ns−1 and A2 = 1327 K [ΔG = −A2·R = ∼11 kJ mol−1 (∼0.1 eV)]. The temperature dependencies of the nanosecond (between 175 and 77 K) and millisecond component are fitted by the

Figure 6. DAS of acpPC at 77 K after excitation at 460 nm. The inverted steady-state absorption spectrum at 77 K (the black dotted lines) is shown for comparison.

ns, ∼130 ns, and 14 μs) are similar to those resolved after excitation at 670 and 540 nm. The spectrum of Chl c2 in the triplet state very well fits the spectrum of the third triplet state resolved previously in acpPC at 295 K after excitation at 460 nm, which was tentatively attributed to a third carotenoid triplet.16 The weak signal observed at 295 K probably derives from a small number of pigments released from the protein during the isolation and copurified with the protein in DM micelles (or, alternatively, pigments still bound to the apoprotein but detached from the EET chain). On the other hand, the strong signal resolved at 77 K apparently results from pigments that cannot transfer their excitation energy, probably because of some temperature-dependent conformational change. G


Article


Figure 8. Temperature dependence of lifetimes of the DAS components resolved for acpPC between 77 and 295 K. The lifetimes were determined from the data obtained after excitation at 670 nm.

Figure 9. DF of acpPC at 77 K after excitation at 460 nm. (a) DF spectra at different delays. (b) DF kinetics at 680 nm fitted by a biexponential function (see the text for more details).

DF at 77 K. The lifetime of the prompt fluorescence was determined to be 5.6 (±0.2) ns, which is in accordance with previously reported results, both at room temperature and 77 K.13,25 In the second step, we have measured the emission at delays between 125 ns and 5 μs following the excitation pulse, when the prompt fluorescence cannot contribute. If a signal is observed, it must be due to DF. The sample was excited at 460 nm to prevent any confusion of the DF with laser reflections after multiple passes through optical fibers or lenses. Despite the fact that Chl c2 is predominantly excited at this wavelength,

Arrhenius law because they presumably reflect the temperature dependence of the triplet-state relaxation rates to the ground states by ISC.29 The complex temperature dependence of the nanosecond component exhibits pronounced changes at around 175 and 270 K, which are partly caused by a contribution of dissolved oxygen to triplet-state quenching (see the Discussion section). Delayed Fluorescence. In order to obtain experimental evidence to explain the strong negative band around 680 nm in the ∼130 ns component, we have measured the prompt and H


Article

The Journal of Physical Chemistry B only emission from Chl a could be detected at these delays (Figure 9). This further proves that the observed signal has a different origin than the prompt fluorescence, which contained a contribution from Chl c2. The DF can be detected even at 1 μs, when the prompt fluorescence intensity is proportional to ∼exp(−1000/5), which is definitely under the signal-to-noise ratio of the used setup. The decay of DF could be satisfactorily fitted by two components with lifetimes of ∼60 ns (relative amplitude ≈ 95%) and ∼540 ns (∼5%). The measurement proves that the singlet states of Chl a are populated in acpPC at the times when they would be fully relaxed under standard conditions.

dependent. The lifetime increases from 5.5 ns at 295 K to 6.5 ns at 225 K and remains constant below this temperature (Figure 7). Moreover, carotenoid triplets are formed with the same kinetics, as follows from the fastest DAS component. Because carotenoid triplet states can only be generated by triplet transfer from another molecule on this time scale, this indicates that the Chl triplet states are formed, but they are efficiently quenched by carotenoids with a rate constant much faster than 5 ns at both room and low temperatures. Although the dynamics of the quenching process can be assumed to be temperature dependent, the rate of the process is fast enough between 77 and 295 K to prevent any observable accumulation of triplet states on these Chl molecules. This renders the determination of the Chl-to-carotenoid triplet transfer time difficult, even in an experiment with a higher temporal resolution. The issue was discussed in more detail in our previous publications,16,20 in which it was estimated that the transfer time of triplets from Chls to carotenoids must be shorter than 0.1 ns,16 in agreement with the previous estimates.13,31 It has been argued that structural fluctuations are responsible for reducing the triplet quenching time into the sub-nanosecond time scale.32 These efficiently protected Chls are probably the red-most Chls located close to the two central lutein molecules, which are the dominant triplet quenchers.7,8,20 The structure of this main quenching site thus seems to have evolved in a way that preserves its high efficiency of Chl triplet quenching at a wide range of temperatures. However, at low temperatures, a small portion of Chl molecules, both Chls a and b, was observed to form long-lived (>1 ms) triplet states. The percentage of such Chls was found to increase with a decreasing temperature, from 0% at 295 K to

Temperature Dependence of Chlorophyll Triplet Quenching in Two

Recommend Documents