Photoprotective Energy Dissipation in Higher Plants Involves Alteration of the Excited State Energy of the Emitting Chlorophyll(s) in the Light Harvesting Antenna II (LHCII)*

Non-photochemical quenching (NPQ), a mechanism of energy dissipation in higher plants protects photosystem II (PSII) reaction centers from damage by excess light. NPQ involves a reduction in the chlorophyll excited state lifetime in the PSII harvesting antenna (LHCII) by a quencher. Yet, little is known about the effect of the quencher on chlorophyll excited state energy and dynamics. Application of picosecond time-resolved fluorescence spectroscopy demonstrated that NPQ involves a red-shift (60 ± 5 cm−1) and slight enhancement of the vibronic satellite of the main PSII lifetime component present in intact chloroplasts. Whereas this fluorescence red-shift was enhanced by the presence of zeaxanthin, it was not dependent upon it. The red-shifted fluorescence of intact chloroplasts in the NPQ state was accompanied by red-shifted chlorophyll a absorption. Nearly identical absorption and fluorescence changes were observed in isolated LHCII complexes quenched in a low detergent media, suggesting that the mechanism of quenching is the same in both systems. In both cases, the extent of the fluorescence red-shift was shown to correlate with the lifetime of a component. The alteration in the energy of the emitting chlorophyll(s) in intact chloroplasts and isolated LHCII was also accompanied by changes in lutein 1 observed in their 77K fluorescence excitation spectra. We suggest that the characteristic red-shifted fluorescence emission reflects an altered environment of the emitting chlorophyll(s) in LHCII brought about by their closer interaction with lutein 1 in the quenching locus.

The photosynthetic apparatus of plants and algae has evolved a process of feedback control of energy input into the photosynthetic reaction centers, revealed in observations of the nonphotochemical quenching of chlorophyll fluorescence (NPQ) 2 (1)(2)(3). Exposure to high light intensity leads to closure of a large fraction of photosystem II (PSII) reaction centers and the buildup of the proton gradient across the photosynthetic mem-brane (⌬pH). The latter prompts the transition of the PSII light harvesting antenna (LHCII) into the photoprotective state (4 -6), where a large proportion of the absorbed energy is rapidly converted into heat, preventing photoinhibitory damage of reaction centers. The molecular mechanism of NPQ is currently under scrupulous multidisciplinary investigation (7)(8)(9)(10)(11)(12)(13)(14).
Many of these studies have centered on the search for pigments associated with light harvesting complexes that may act as energy traps/quenchers. In recent years certain evidence has accumulated that xanthophylls may play a central role in photoprotective energy dissipation. A carotenoid radical cation, suggested to be zeaxanthin, was found to correlate with the level of NPQ and was proposed to act as a direct quencher of chlorophyll excitation (7,10). Other studies indicated the involvement of energy transfer to the S 1 state of LHCII-bound lutein 1 (9,15). Evidence of these putative quenchers occurring in vivo was provided (7,9), yet little is known about how ⌬pH activates them. ⌬pH is believed to protonate luminal residues on the PSII subunit PsbS and LHC complexes causing certain conformational changes to occur, which are necessary for NPQ (5,6). These conformational changes are not well understood but can be traced by changes in LHCII-bound pigments detected by absorption and resonance Raman spectroscopy (9,(15)(16)(17)(18)(19)(20).
Both the proposed zeaxanthin-and lutein-quenching mechanisms assume very close (van der Waals) contact between the quenching xanthophyll and chlorophyll. Such an interaction would, most likely, affect the energy levels of the pigments involved, which may be required to convert the xanthophyll into an energy quencher. Indeed, a subpopulation of zeaxanthin molecules in PSII was reported to undergo an absorption redshift in the NPQ state (18 -20). Recently, lutein absorption was also found to be decreased in the dissipative state (20), interestingly this loss of molar extinction was enhanced by the presence of zeaxanthin.
Whereas many studies have probed the role of xanthophylls, relatively little is known about the state of chlorophylls upon formation of NPQ. Some research suggests that chlorophyll associates may be formed in the NPQ state, which potentially could act as quenchers (21)(22)(23). It has been proposed that such associates have red-shifted absorption and fluorescence spectra. Low temperature (77K) fluorescence spectra of quenched LHCII aggregates possess a 20-nm red-shifted band, F700 (22), which was also detected in leaves and thylakoids frozen in the NPQ state (21,24). The F700 emitting chlorophylls appeared to possess very little or no absorption, indicating they may arise from chlorophyll-chlorophyll excitonic interactions (22). Later, low temperature (77K) time-resolved fluorescence spectra demonstrated F700 was one of several red-shifted bands present in aggregated LHCII, however, because they all possessed much longer lifetimes than the main F680 band it was considered unlikely that they could act as quenchers (25,26). Indeed, the most red-shifted of these bands, F700, was absent in nonaggregated LHCII quenched in a gel media (15). F700 was therefore suggested to reflect intersubunit LHCII chlorophyll interactions rather than being directly related the quencher (15). While the latter study did not exclude the event of LHCII aggregation during NPQ formation, it did demonstrate that an intrinsic change in pigment interactions rather than protein aggregation per se is the cause of quenching. Nonetheless, a small red-shift of both the chlorophyll a absorption and fluorescence emission at room temperature was still observed and found to be in good correlation with the strength of quenching (15).
To investigate the effect of NPQ on chlorophyll excited state energy and dynamics we have undertaken a combined timeresolved and steady state spectroscopic study on isolated intact chloroplasts. Common quenching-induced features were found in the time-resolved and excitation fluorescence spectra in both chloroplasts and LHCII, suggesting a common mechanism of quenching involving changes in the environment of the emitting chlorophyll(s) and lutein 1.

EXPERIMENTAL PROCEDURES
Chloroplast Isolation-Spinach plants were grown for 8 -9 weeks in Sanyo plant growth cabinets with an 8-h photoperiod at a light intensity of 250 mol of photons m Ϫ2 s Ϫ1 and a day/ night temperature of 22/18°C. Intact chloroplasts were prepared as described by Crouchman et al. (27). Chloroplasts devoid of zeaxanthin and antheraxanthin (ϪZea) were prepared from spinach leaves dark adapted for 1 h. Chloroplasts enriched in zeaxanthin (ϩZea) were prepared from leaves pretreated for 30 min at 350 mol of photons m Ϫ2 s Ϫ1 under 98% N 2 , 2% O 2 . Chlorophyll concentration was determined according to the method of Porra et al. (28).
Chlorophyll Fluorescence Induction-Chlorophyll fluorescence was measured with a Dual-PAM-100 chlorophyll fluorescence photosynthesis analyzer (Heinz Walz) using the liquid cell adapter. Intact chloroplasts were measured in a quartz cuvette at a concentration of 12 M chlorophyll under continuous stirring in the presence of 100 M methyl viologen as a terminal electron acceptor. Actinic illumination (350 mol of photons m Ϫ2 s Ϫ1 ) was provided by arrays of 635 nm LEDs. Fo (the fluorescence level with PSII reaction centers open) was measured in the presence of a 10 mol of photons m Ϫ2 s Ϫ1 measuring beam. The maximum fluorescence in the dark-adapted state (Fm), during the course of actinic illumination (FmЈ) and in the subsequent dark relaxation periods was determined using a 0.8-s saturating light pulse (4000 mol of photons m Ϫ2 s Ϫ1 ). The quantum yield of PSII was defined as ((Fm-Fo)/Fm) and NPQ as ((Fm-FmЈ)/FmЈ).
Pigment Analysis-Pigment composition was determined by reversed phase HPLC using a LiChrospher 100 RP-18 column (Merck) and a Dionex Summit chromatography system as previously described (30). Pigments were extracted from LHCII or chloroplasts with and without zeaxanthin/antheraxanthin following induction and subsequent recovery of NPQ as described above. The results of pigment analysis are summarized in supplemental Table S1.
Absorption Spectral Measurements-Absorption changes in the 610 -740-nm region were measured using a SLM DW2000 dual wavelength spectrophotometer. LHCII and ϩZea chloroplasts were measured in a quartz cuvette at a chlorophyll concentration of 4 M under continuous stirring. The instrument slit width was 1 nm, and the scan rate was 1 nm s Ϫ1 .
Steady State Fluorescence Spectroscopy-Room temperature fluorescence emission spectra were recorded on leaves and intact chloroplasts (4 M chlorophyll) between 600 and 800 nm using a Jobin Yvon FluoroMax-3 spectrophotometer. Excitation was defined at 435 nm with a 5-nm spectral bandwidth. The light intensity was 100 mol of photons m Ϫ2 s Ϫ1 . The fluorescence spectral resolution was 1 nm. Low temperature fluorescence excitation spectra of LHCII and intact chloroplasts (4 M chlorophyll) were detected at 680 nm (isolated LHCII) or 685 nm (chloroplasts) with a 4-nm spectral bandwidth. The excitation spectral resolution was 1 nm. Spectra were normalized at 380 nm.
Time-resolved Fluorescence Spectroscopy-Time-correlated single photon counting measurements were performed using a FluoTime 200 ps fluorometer (PicoQuant). Fluorescence lifetime decay kinetics were measured on LHCII and intact chloroplasts (4 M chlorophyll) using excitation provided by a 470-nm laser diode using a 10 MHz repetition rate. These settings were carefully chosen to be far below the onset of singletsinglet exciton annihilation (Ͻ0.1 pJ). Fluorescence was detected at 680 nm (isolated LHCII) or 685 nm (chloroplasts) with a 1-nm slit width. The instrument response function was ϳ50 ps. Time-resolved emission spectra (TRES) were measured in the 655-760-nm detection region with 1-nm steps. The resolution of the time-to-amplitude converter was 4 ps/channel. For lifetime analysis, FluoFit software (PicoQuant) was used. The quality of the fits was judged by the 2 parameter (supplemental Fig. S1A). In addition, the autocorrelation function of the weighted residual data were obtained as a measure of the correlation between residuals in distinct channels separated by various times (supplemental Fig. S1B). Because the residual values should be normally distributed in a good fit (not correlated) the autocorrelation data are randomly distributed around zero and their fluctuations are small (supplemental Fig.  S1B). In addition a support plane analysis was performed (supplemental Fig. S1C) by calculating the 2 as a function of each single parameter to estimate the region of the function below the tolerance level. This analysis is useful for verification of the number of components and possible localization of additional components missed by the fitting procedure. The components of the support plane analysis (functions 2 / 2 min ) were found to be symmetric indicting that the number of components found by the fitting procedure was correct. . 1A shows typical chlorophyll fluorescence induction traces measured on isolated intact chloroplasts possessing violaxanthin as the only xanthophyll cycle carotenoid (ϪZea) (NPQ ϭ 0.9 Ϯ 0.1) and those enriched in zeaxanthin (ϩZea) (NPQ ϭ 2.45 Ϯ 0.2), illuminated for 5 min at 350 mol of photons m Ϫ2 s Ϫ1 . The levels of reversible NPQ in these chloroplasts match those found in leaves (20). Following induction of NPQ, the quenching could be sustained in the dark by the addition of 3% (v/v) of a protein cross-linker, glutaraldehyde, 10 s prior to removing the actinic illumination. A similar technique has previously been used to sustain NPQ present in cyanobacteria (31). The absence of NPQ relaxation was confirmed by application of saturating light pulses (Fig. 1A, trace 3). When administered to chloroplasts prior to the onset of actinic illumination glutaraldehyde greatly slowed the rate of fluorescence quenching and prevented formation of any rapidly relaxing NPQ (Fig. 1A, trace 4). However, in the presence of a lower actinic light intensity (100 mol of photons m Ϫ2 s Ϫ1 ), which was just sufficient to reach the maximum level of fluorescence (Fm), no quenching was observed (Fig. 1A, trace 5). Glutaraldehyde was therefore used to prevent the reversal of NPQ or to sustain the Fm level of fluorescence in the absence of quenching, allowing measurement of steady-state and time-resolved fluorescence spectra in the Fm and NPQ states. The alternative use of an uncoupler, nigericin, to sustain Fm (preventing any NPQ formation) and lowering the pH of the chloroplast incubation medium from 7.6 to 5.2 to prevent relaxation of the NPQ state (32), gave similar results to those obtained with glutaraldehyde (data not shown). The use of 3-(3Ј,4Ј-dichlorophenyl)l,l-dimethylurea (DCMU) to close PSII reaction centers was avoided due to the well known quenching effect of oxidized plastoquinone, which could affect the Fm lifetime. Fig. 1B shows the fluorescence spectra of dilute chloroplasts and those of spinach and Arabidopsis leaves. It is very clear that the chlorophyll concentration in leaves is too high for detection of the correct fluorescence spectrum, which in turn is essential for obtaining not only steady-state but also time-resolved fluorescence spectra (see below). Reabsorption reduces dramatically the PSII fluorescence intensity at 685 nm, causes a red shift of this band but leaves almost unchanged its vibronic satellite band around 740 nm. A small amount of PSI fluorescence, peaking at around 720 nm, also contributes in this region (Fig.  1B, clearly seen as a broad shoulder in Spinach leaf spectrum).

Fig
The absorption spectra in the chlorophyll Q Y region of intact chloroplasts were measured in the Fm and NPQ states. The NPQ-Ϫ-Fm absorption difference spectrum is shown in Fig. 2. It resembles remarkably closely the aggregated-Ϫ-trimeric LHCII difference spectrum (Fig. 2). Both spectra possessed a major positive band in the 684 -687-nm region, a minor band at 660 nm and two negative bands at 675 and 647-652 nm. The emergence of similar absorption red-shift both in vitro and in vivo suggests that quenching in both systems has a selective effect on chlorophyll a. To narrow down the identification of the chlorophyll pigments involved in the shift it was important to perform fluorescence spectral measurements, reflecting the state of only the emitting pigments. Fig. 3A displays reabsorption-free room temperature fluorescence spectra of intact chloroplasts in the Fm and NPQ states. Fluorescence emission in ϩZea chloroplasts in the NPQ state is red-shifted relative to the Fm state by ϳ1.5 nm and there is a slight enhancement of the vibronic satellite in the 710 -760 nm region. Similar, albeit smaller, changes were observed in the ϪZea chloroplasts in the NPQ state (Fig. 3A). Zeaxanthin seems to enhance both the NPQ and the extent of the fluorescence red-shift in chloroplasts. Fig. 3B shows the fluorescence spectra of 10 times quenched LHCII aggregates in comparison to unquenched trimeric LHCII. As in intact chloroplasts, quenching resulted in a clear red-shift (ϳ2.5 nm) of the fluorescence emission and enhancement of the vibronic satellite. These changes are very different from those induced by the denaturation of the complex induced by high temperature or detergent treatment (Ref. 15 and supplemental Fig. S2). Moreover, the observed spectral alterations induced by quenching were completely reversible (15). LHCII is the major contributor to the room temperature fluorescence of intact chloroplasts due to the shallowness of the PSII reaction center trap, its excitation diffusion limited kinetics and the relative size of the peripheral antenna (ϳ250 chlorophylls) compared with the CP43/47 core antenna (ϳ36 chlorophylls) (33). Nonetheless, the enhancement of the vibronic satellite in isolated chloroplasts is likely to be due, at least in part, to an increased relative contribution from photosystem I (PSI) fluorescence particularly at ϳ720 nm, because of quenching of PSII fluorescence.
Time-resolved fluorescence measurements were used to separate different spectral components of the chloroplast emission, find those associated with the NPQ state, and determine whether they possess the red-shifted fluorescence and enhanced vibronic satellite. Fig. 4 displays the fluorescence decay profiles of both intact chloroplasts (A) and isolated LHCII complexes (B). To our knowledge, these are the first time-correlated single photon counting (TCSPC) kinetics ever shown for intact chloroplasts. Indeed, while the published time-resolved fluorescence experiments on thylakoids have shown an NPQ of less than 0.5 (34,35), here we present the decay kinetics corresponding to NPQ levels up to 2.6, i.e. nearly 5 times higher than previously reported. The intensity weighted average fluorescence lifetime in ϪZea chloroplasts was 2.14 Ϯ 0.1 ns. The presence of zeaxanthin caused a distinct reduction of the average lifetime to 1.63 Ϯ 0.1 ns (Fig. 4A). Both the Fm lifetime and the average lifetime reduction as a result of violaxanthin de-epoxidation are consistent with previous reports (20, 36 -38). In the presence of NPQ the average lifetime was reduced to 1.1 Ϯ 0.1 ns in ϪZea chloroplasts and 0.58 Ϯ 0.05 ns in ϩZea chloroplasts (Fig. 4A). The average fluorescence lifetime in isolated trimeric LHCII was 4.1 Ϯ 0.2 ns, which was quenched to 0.4 Ϯ 0.1 ns in aggregated LHCII (Fig. 4B). Fig. 5 shows the results of a lifetime component analysis for the fluorescence decay of intact ϪZea and ϩZea chloroplasts in comparison to the components of isolated trimeric and aggregated LHCII. Three components were resolved in the chloroplast fluorescence decay. An additional longer component of ϳ3.0 -3.5 ns, reported in several studies on thylakoids (34,35,39), was absent in intact chloroplasts, consistent with measurements on leaves (40). It is possible that this longer lifetime is caused by detergent effects/osmotic shock in the isolated thylakoids and PSII particles used in these studies. Another component of 1.26 Ϯ 0.1 ns in our studies had a ϳ30% smaller amplitude than the 2.4 Ϯ 0.2 ns one (Fig. 5A). The two separate

. Relative absorption difference spectra in the chlorophyll Q Y region of aggregated-minus-trimeric LHCII (solid line) and NPQ-minus-Fm ؉Zea chloroplasts (dashed line).
Both Fm and NPQ states in chloroplasts were fixed using 3% (v/v) glutaraldehyde (see Fig. 1). Spectra are averaged from three independent replicates and normalized at 610 nm.  Fig. 1). All spectra were normalized at ϳ683 nm maxima, the 435 nm excitation light intensity was 100 mol of photons m Ϫ2 s Ϫ1 . Spectra are averaged from independent three replicates. AUGUST 28, 2009 • VOLUME 284 • NUMBER 35 PSII lifetime components may arise from some heterogeneity in antenna size and/or state. The shorter 350 -500 ps PSII lifetime component, sometimes reported in the literature (34,35,39), was absent in intact chloroplasts. It appeared only when the exciting light intensity was not high enough to saturate all PSII reaction centers (data not shown), consistent with previous work (33). The shortest lifetime we detected was 95 Ϯ 15 ps in agreement with its assignment to PSI (34,35,40,41). Calculation of the TRES revealed that this component possessed a spectrum distinct from that of the other components. Supplemental Fig. S3 displays the TRES spectra for the lifetime components of ϪZea chloroplasts present in the Fm state. The spectrum of the 95 ps component peaked at ϳ690 nm and possessed a very broad far-red emission with a center at 720 nm. This emission is typical for the room temperature fluorescence spectrum of PSI (42)(43)(44)(45)(46). The other two components have similar spectra with very low amplitudes in the red region accounting only for their corresponding vibronic satellites. These components originate from PSII and are dominated by the LHCII antenna (see above). The component analysis revealed that violaxanthin de-epoxidation shortened both the PSII components to 1.83 Ϯ 0.2 and 0.93 Ϯ 0.1 ns in the ϩZea chloroplasts (Fig. 5A).

Alteration of Excited State Energy of Emitting Chl in LHCII
In the NPQ state both of the PSII lifetimes were shortened significantly. In ϪZea chloroplasts an 800 Ϯ 50 ps component and a 1.7 Ϯ 0.1 ns component were present, while in the ϩZea  chloroplasts a 480 Ϯ 43 ps and 1.0 Ϯ 0.1 ns component were found (Fig. 5B). In both types of chloroplast there was also a slight increase in the amplitude of the shortest lifetime component, the effect was strongest in the ϩZea chloroplasts. The increased amplitude may be due to the presence of an additional short PSII component that we are unable to resolve from PSI. Indeed similar short components were induced by quenching in isolated LHCII, where the main 4.1-ns component in trimers was converted into 210 Ϯ 34-ps and 400 Ϯ 30-ps components in aggregates, similar to that observed previously (25,47) (Fig. 5C). Fig. 6 compares the TRES of the major lifetime components present in ϪZea and ϩZea chloroplasts, in the Fm and NPQ states, and isolated LHCII, in the trimeric and aggregated state. A common feature was observed in all these spectra, a distinct red-shift of the shorter lifetime components. The TRES spectrum of the 480-ps component present in ϩZea chloroplasts in the NPQ state was red-shifted (60 Ϯ 5 cm Ϫ1 ) compared with that of the major Fm component of 1.83 ns (Fig. 6A). Interestingly the vibronic satellite was also slightly enhanced in the NPQ state. Hence, not all of the rise in the vibronic satellite observed in the steady state spectrum (Fig. 3) can be due to increased PSI contribution, consistent with the results obtained in isolated LHCII (Fig. 3 and below). The red-shift and vibronic satellite enhancement of the major 800 ps PSII component in ϪZea chloroplasts in the NPQ state relative to the 2.4 ns Fm component was smaller (25 Ϯ 5 cm Ϫ1 ) than that observed in the ϩZea sample (Fig. 6B). In both types of chloroplasts, the redshift was found to be statistically significant over three independent experiments (Student's t test 99% confidence). In aggregated LHCII, the major 210 ps component spectrum is 80 Ϯ 5 cm Ϫ1 red-shifted with respect to the dominating 4.2-ns lifetime of the LHCII trimer (Fig. 6C). In addition, it possessed a slightly increased vibronic satellite. Therefore, in both chloroplasts and isolated LHCII the effects of NPQ on the TRES spectrum do not strictly require the presence of zeaxanthin. Fig.  7 plots the lifetime of each component and the energy of the red-shift relative to the main Fm (chloroplasts)/trimer (isolated LHCII) component. The plot confirms that the extent of the red-shift increases as the component lifetime decreases, with a very similar pattern observed for LHCII in vitro (here in aggregates and recently in gel media (15)) and for NPQ in vivo (Fig. 7).
Recently measurements of TRES spectra on Arabidopsis leaves have revealed that the major NPQ component possessed a strongly enhanced far-red region (23). To investigate this possibility further, TRES spectra were recorded on Arabidopsis leaves. The major component present in the NPQ state in leaves of 480 Ϯ 25 ps was indeed found to possess a very strong far-red fluorescence band (supplemental Fig. S4). However, because a very similar far-red fluorescence band was also present in the TRES spectrum of the main 2.2 Ϯ 0.2 ns component in the Fm state it seems that the far-red emitting chlorophyll species are not exclusive for NPQ.
To test the state of xanthophylls connected to the emitting chlorophyll molecules in the NPQ state, excitation fluorescence measurements at 77K were performed on chloroplasts and isolated LHCII. Fig. 8A shows F685 fluorescence excitation spectra (PSII core cross-section) in ϪZea and ϩZea intact chlo-  Fig. 1). Fluorescence was detected in the 660 -760-nm region using 470-nm excitation. Data are the average of three independent experiments Ϯ S.E. Spectra are averaged from independent three replicates, all amplitudes are normalized to 1 at ϳ685 nm for comparison. roplasts and the F680 excitation spectra in isolated LHCII in quenched and unquenched states (solid and dashed lines, respectively). All spectra were normalized at 380 nm in order to see if selective spectral changes took place during NPQ. Both NPQ in chloroplasts and quenching in isolated LHCII affected the spectral regions around 440 and 475-500 nm. The calculated difference excitation spectra NPQ-minus-Fm, in chloroplasts, and aggregated-minus-trimeric, in isolated LHCII, are shown in Fig. 8B. All three difference spectra revealed negative bands around 440 nm (chlorophyll a and/or xanthophyll) and 488 nm. The latter is dominated by neoxanthin absorption at 485 nm (48,49). The change around 488 nm band also reveals a shoulder at 495 nm, which is particularly clear in the second derivative spectra, and was assigned to lutein 1 (48,49). This shoulder increases in ϩZea chloroplasts and in particular in aggregated LHCII, paralleling the amount of quenching in each sample. Another feature is present at 455-465 nm in the spectrum of ϩZea chloroplasts, which can be assigned to the vibrational 0 -1 bands of neoxanthin and lutein 1 (49).

DISCUSSION
This work has shown the results of time-resolved and steady state spectroscopic fluorescence studies, performed for the first time using intact chloroplasts possessing the full range of NPQ. This preparation has clear advantages. One of them is possibility to fix PSII in either efficient light harvesting (Fm) or photoprotective (NPQ) states. The other advantage is the absence of the long fluorescence lifetime components reported in thylakoid and PSII particle preparations resulting from detergent effects and osmotic shock. In addition, it was possible to obtain chloroplasts with zero or very high levels of zeaxanthin (and antheraxanthin). Finally, unlike leaves, chloroplasts can be diluted to completely avoid the reabsorption artifacts in the fluorescence spectrum (Fig.  1B). This is particularly important for the interpretation of the time-resolved emission spectra (decay-associated spectra). For example, decay-associated spectra obtained in a recent study on Arabidopsis leaves revealed that the major   [1], ϪZea and trace [2] ϩZea chloroplasts in the Fm state (dashed line) and NPQ states (solid line); trace [3] 77K fluorescence excitation spectra of isolated LHCII (emission 680 nm) in aggregated (solid line) and trimeric states (dashed line). B, NPQ-Ϫ-Fm fluorescence excitation difference spectra and their second derivatives (circles and gray lines) of trace [1] ϪZea, trace [2] ϩZea chloroplasts, and trace [3] in trimeric-Ϫaggregated LHCII. Both Fm and NPQ states in chloroplasts were fixed using 3% (v/v) glutaraldehyde (see Fig. 1). All spectra were normalized at 380 nm, dashed lines represent zero. Spectra are averaged from independent three replicates. lifetime component in the NPQ state possessed strongly enhanced far-red chlorophyll fluorescence forms (23). Comparison of the time-resolved fluorescence spectra of leaves and dilute chloroplasts in this study shows how reabsorption distorts the time-resolved spectra in the same manner as it does the steady-state fluorescence spectra. The reabsorption artifact causes the strong enhancement of the far-red vibronic satellite band (710 -760 nm) relative to the main fluorescence band (see supplemental Fig. S4). Since the observed far-red fluorescence originates from the vibronic satellite of the main fluorescence band it therefore also bares the same lifetime. Therefore, as in chloroplasts, the extent of the enhancement of the vibronic satellite in the NPQ state in leaves is actually rather more modest than that reported by Miloslavina et al. (23).
In this study, it was demonstrated that the main PSII lifetime component in the NPQ state also possesses the characteristic red-shifted fluorescence emission reported in isolated LHCII (15). Furthermore we have demonstrated that the extent of this red-shift is related to the lifetime of the emitting chlorophyll(s) both in vivo and in vitro. The shorter the lifetime the more red-shifted the spectral component. The red-shift cannot be a consequence of formation of intertrimer chlorophyll-chlorophyll interactions because it was also found in LHCII quenched in a gel media in the absence of protein aggregation (15). Neither can it be the result of a zeaxanthin-chlorophyll interaction, since zeaxanthin was absent in the LHCII preparations used in both this and previous studies (supplemental Table S1).
There are therefore two possibilities to explain this phenomenon, either that the identity of the emitting chlorophyll molecules in LHCII change or that their environment changes. The small conservative red-shift in the chlorophyll Qy region of the absorption spectra of both isolated LHCII and chloroplasts suggest that the latter is more likely. Indeed, the hole-burning experiments of Pieper et al. (50) indicated that quenching in LHCII had only a small effect on the excited state structure of the isolated trimer.
The emitting chlorophylls in LHCII: a612, a611, and a610 (51) are in close contact with lutein 1 (lutein 620) ( Fig. 9) (14). Therefore, the changes observed in lutein 1 in the fluorescence excitation spectra of quenched LHCII and chloroplasts provide further evidence of an alteration in the environment of these emitting chlorophylls. Changes in lutein 1 absorption have previously been observed in leaves in the NPQ state (20), the changes in the excitation spectrum likely reflect this drop in oscillator strength. Carotenoid and chlorophyll absorptive properties, such as the oscillator strength and the energy of transition are particularly sensitive to the polarity of the solvent media. In the case of LHCII this solvent is the protein, which binds, orients, and, most importantly, "tunes" pigment energy. This tuning is vital to ensure a broad antenna cross-section as well as directionality of energy transfer, creating a terminal emitter locus, which efficiently collects excitation from all pigments of the complex. The finding that both the transition into the quenched state and transition back to the unquenched state is inhibited by a crosslinker both in vivo and in vitro (15) provides further evidence for the involvement of protein conformation in this alteration of pigment environment.
Because the high resolution crystal structure of LHCII (14) was found to correspond to a partially quenched state, with an average lifetime of 800 -900 ps (8), the interaction of lutein 1 with the emitting chlorophylls a612, a611, and a610 is likely to represent a quenching configuration of this pigment locus (Fig. 9). It is interesting to note the main PSII lifetime in ϪZea chloroplasts in the NPQ state was also ϳ800 ps matching that found in the violaxanthin binding LHCII crystals (8). This is clearly contrary to the recent suggestion that the LHCII crystal structure is that of the unquenched state (52). Although a conformational change is known to be required for lutein 1 to become a quencher (9), the exact molecular details are lacking. A putative disposition of the lutein 1 toward chlorophyll a610 has, however, been proposed (53). Calculations derived from high pressure induced quenching experiments on LHCII indicate that a small change in the volume of the structure (a conformational change) can significantly affect the fluorescence lifetime (54). Indeed, the degree of orbital overlap between lutein and chlorophyll, is likely to be very sensitive to even small alterations in distance and orientation between the molecules involved. Such van der Waals contact between pigments can affect their excited electron energies, resulting in spectral shifts and alterations in electron-vibrational coupling. This contact would be a prerequisite for the Dexter electron exchange mechanism which requires orbital interactions between the energy exchanging molecules. Another possibility is that an enhancement of chlorophyll-lutein 1 interactions promotes an electronic coupling (excitonic interaction) of the pigment S 1 states. 3 Because the quenching-related changes alter the spectra and excited state energy of both of the main PSII lifetime components (Fig. 7), it is feasible to suggest that NPQ is a widespread event within the antenna, involving most, if not all, LHCII complexes. This widespread change is unlikely to involve any selective detachment of LHCII from PSII, as was recently suggested (55), since NPQ did not affect the ratio between chlorophyll a (435 nm) and chlorophyll b (472 nm) in the PSII excitation cross-section spectra (Fig. 8). Such energetic uncoupling would have led to an increase in the amplitude of Chl a to b in the fluorescence excitation spectrum of PSII resulting from the detachment of Chl b-rich LHCII, as is observed in unstacked thylakoids (56).
In conclusion, the data obtained in this study provide novel evidence that the excited state properties of the emitting chlorophyll molecule(s) within LHCII are modified in the NPQ state, and that these alterations are accompanied by changes in the excited state properties of lutein 1. These protein cross-linker-sensitive alterations in pigment interactions within LHCII therefore provide an indication of the subtle changes involved in the formation of the quenching locus.