Photosynth Res (2012) 111:53–62 DOI 10.1007/s11120-011-9696-3

REGULAR PAPER

Reconstituted CP29: multicomponent fluorescence decay from an optically homogeneous sample Erica Belgio • Giorgio Tumino • Stefano Santabarbara Giuseppe Zucchelli • Robert Jennings

•

Received: 9 February 2011 / Accepted: 28 September 2011 / Published online: 16 October 2011 Ó Springer Science+Business Media B.V. 2011

Abstract The multiexponential fluorescence decay of the CP29 complex in which the apoprotein and pigments were reconstituted in vitro was examined. Of the three decay components observed only the two dominant ones, with about 3 and 5 ns lifetimes, were studied. The main question addressed was whether the multicomponent decay was associated with sample optical heterogeneity. To this end, we examined the optical absorption and fluorescence of the CP29 sample by means of two different and independent experimental strategies. This approach was used as the wavelength positions of the absorption/fluorescence spectral forms has recently been shown to be a sensitive indicator of the binding site-induced porphyrin ring deformation (Zucchelli et al. Biophys J 93:2240–2254, 2007) and hence of apoprotein conformational changes. The data indicate that this CP29 sample is optically homogeneous. It is hypothesised that the different lifetimes are explained in terms of multiple detergent/CP29 interactions leading to different quenching states, a suggestion that allows for optical homogeneity. Erica Belgio and Giorgio Tumino have contributed equally to this study.

Electronic supplementary material The online version of this article (doi:10.1007/s11120-011-9696-3) contains supplementary material, which is available to authorized users. E. Belgio  G. Tumino  S. Santabarbara  G. Zucchelli  R. Jennings CNR-Istituto di Biofisica, Sede di Milano, Via G. Celoria 26, 20133 Milan, Italy E. Belgio  G. Tumino  S. Santabarbara  G. Zucchelli  R. Jennings (&) Dipartimento di Biologia, Universita` degli Studi di Milano, Via G. Celoria 26, 20133 Milan, Italy e-mail: [email protected]

Keywords Antenna complexes  CP29  Multicomponent fluorescence decay  Optical homogeneity  Quenched/ unquenched conformers

Introduction Over the years, many fluorescence decay studies have been published on the isolated antenna complexes of plant photosystems (e.g. Gruszecki et al.1997; Moya et al. 2001; Crimi et al. 2001; Huyer et al. 2004; Avenson et al. 2008; Passarini et al. 2010). An intriguing feature of all these studies is that the decay is multiexponential. At first sight, this is an unexpected result as the spectral equilibration time of most antenna complexes has been demonstrated to be just a few picoseconds (Bittner et al. 1995; Visser et al. 1996; Connelly et al. 1997; Gradinaru et al.1998; Cinque et al. 2000), which is about three orders of magnitude faster than the decay time itself. Thus, assuming homogeneity of the antenna sample, one expects a monoexponential decay which, as stated above, is not observed. It would therefore seem reasonable to conclude that the multiexponential decay is associated with some kind of sample heterogeneity, with each decay component representing a different antenna population. In fact, suggestions of this kind have been frequently put forward in the literature for a variety of isolated antenna complexes. The most common suggestion, based on the initial studies by Moya et al. (2001) using a number of photosystem I and II external antenna complexes, invokes different protein conformational states associated with the presence or absence of the xanthophyll zeaxanthin (e.g. Crimi et al. 2001; Pascal et al. 2005; Avenson et al. 2008; Johnson et al. 2009; Passarini et al. 2010). This suggestions is however not applicable to the

123

54

multicomponent decay of CP47 (Huyer et al. 2004) as this complex does not bind zeaxanthin (Alfonso et al. 1994). This point has in fact already been made by Huyer et al. (2004), who suggested that a dynamic equilibrium may exist between chlorophyll–protein complex conformers in which at least one chlorophyll has a modified fluorescence yield, due to differences in non-radiative decay. In the present study, we further analyse this problem using the minor photosystem II outer antenna complex CP29, in which the recombinant apoprotein was reconstituted with pigments in vitro (e.g. Giuffra et al. 1996; Belgio et al. 2010). This complex has a relatively simple decay, characterised by just two dominant decay processes of around 3 and 5 ns and a minor fast decay. We have addressed the question of whether this sample exhibits optical heterogeneity which may correlate with the multicomponent decay. It has recently been demonstrated by Zucchelli et al. (2007) that the chlorophyll absorption (fluorescence) properties are modulated by protein-induced distortions of the porphyrin rings and are therefore sensitive indicators of protein structural differences (see also ‘Discussion’ section). The high sensitivity of absorption shifts to even small porphyrin structure deformations is indicated by the fact that the small site structural fluctuations, associated with each single chlorophyll binding site, generate wavelength shifts in the Qy electronic transitions which are commonly of up to 5–6 nm for the same chlorophyll binding site (Zucchelli et al. 2007). The present experimental strategy involved two independent approaches: (i) analysis of the relation between the absorption and fluorescence spectra according to the Stepanov equation; (ii) the use of the fluorescence quencher dibromobenzothymoquinone (DBMIB), shown to selectively quench one of the main decay components, followed by measurement of the fluorescence emission spectrum. Both approaches suggest that the CP29 sample is optically homogeneous, which in the light of the Zucchelli et al. (2007) analysis suggests that the sample does not contain different protein conformers. Furthermore, as the DBMIB quenching is shown to be static and not diffusional, its selective quenching of one decay component renders the idea that the multicomponent decay is due to protein dynamics unlikely. The hypothesis is suggested that the heterogeneous decay in recombinant CP29 could be caused by different interactions between the solubilising detergent and the complex itself, rather than intrinsically different CP29 conformers.

Materials and methods Recombinant and reconstituted CP29 was obtained as previously described (Belgio et al. 2010). Steady-state absorption and fluorescence spectra of recombinant complexes were measured with a liquid

123

Photosynth Res (2012) 111:53–62

nitrogen-cooled CCD array detector (Princeton Applied Research) through a spectrograph (Spectra Pro-300i, Acton Research Co. USA) equipped with a 300 grooves/mm grating. The wavelength scale of the instrument was calibrated with a neon spacing calibration source (Cathodeon). The spectral resolution of this setup is 0.28 nm. The excitation beam (440 nm) was provided by a water-cooled xenon lamp and monochromator of a J-500 spectropolarimeter (Jasco). Fluorescence spectra were corrected for the instrument detection wavelength sensitivity. Samples were diluted to 0.4 OD/cm at the Qy absorption maximum for absorption measurements and to 0.1 OD/cm for fluorescence, in a buffer containing 10 mM Hepes pH 7.6, 0.045% DM. Steady-state fluorescence spectra quenched by DBMIB were measured in 10 mM Hepes pH 7.6, 0.045% DM, kept at 8°C, using an EG&G OMA III (model 1460) multichannel spectrometer with a cooled (-33°C) diode array (model 1420). Excitation light (440 nm) was provided by a xenon lamp and filtered by a Heath monochromator plus two Corning 4–96 filters. Fluorescence emission was filtered across a Schott OG 530 filter. Quenched emission spectra were determined after addition of 10 lM DBMIB dissolved in ethanol (1% of final volume). Spectra were accumulated for several samples in order to have about 3 9 104 counts in the peak channel. Time-resolved fluorescence decay measurements of the CP29 were performed in 10 mM Hepes pH 7.6, 0.045% DM by time correlated single photon counting technique (TCSPC). The samples were diluted to 0.1 OD/cm at 677 nm and kept at 4°C. The excitation source was a Pulsed Diode Laser (PicoQuant GmbH, Berlin Germany), peaking at 632 nm, controlled by a controller PicoQuant PDL 800-B, and operating at a repetition rate of 10 MHz. The excitation light was collimated by two prisms and was directed onto the sample through an optical fibre. The emission decays were collected through a monochromator (Jasco CT-10, Japan) at 10 nm intervals in the wavelength range between 670 and 710 nm and detected by a cooled microchannel plate photomultiplier tube (Hamamatsu R3809U-51). Quenched fluorescence decays were measured at 680 nm by means of a titration with the quinone quencher DBMIB dissolved in ethanol (up to 2.5% final volume of the sample). The photomultiplier tube was kept at -10°C by a Peltier cooler, in order to improve the signal/noise ratio. Over the longest measurement time (about 20 min) about 10 dark counts were recorded per channel, with respect to the 104 signal counts accumulated in the peak channel. The channel spacing was 7.6 ps, resulting in a 30 ns time window. The decay parameters, time decay constants and amplitudes, were obtained by an iterative reconvolution between a sum of weighted exponential decay functions and a response function (IRF). Moreover, the background (dark) intensity and the offset of the time scale are

Photosynth Res (2012) 111:53–62

55

parameters in the decay function definition. The resulting curve is fitted to the experimental data by minimising the v2 on varying the set of parameters. The IRF was obtained measuring the fluorescence decay of a fluorochrome having a single known lifetime (e.g. Zuker et al. 1985). The fluorochrome was DCI’ (Exciton, Inc., Dayton,Ohio) dissolved in ethanol, having a characteristic lifetime around 35 ps that was directly estimated by the fluorescence decay measurements and the value used in the fitting process. The advantages of this method were previously discussed (Tumino et al. 2008). The minimisation procedure has been performed with a modified version of the algorithm previously developed in the laboratory to use the measured decay of a known fluorochrome as the source of IRF. The software is based on the Minuit library from CERN, distributed with a C-fortran interface as ‘C-minuit’. Both the ‘C-minuit’ and our Minuit-based algorithm are freely available as opensource software [https://sourceforge.net/projects/multiexpfit; https://sourceforge.net/projects/c-minuit]. The algorithm allows not only the minimisation, but also error analysis and determination of v2. After a number of tests, we consider that the effective time resolution under our experimental conditions is close to 10 ps.

Results CP29 fluorescence decay In Fig. 1a, the fluorescence decay of the recombinant and reconstituted CP29 is presented together with the residuals and the reduced v2 for both the linear sum of two and three exponentials. In both cases, prominent decays are found close to 3.5 and 5 ns (see figure legend for details). In the two-component analyses the fit seems inadequate in the first few channels, a problem which is largely overcome by a third exponential component with an approximately 110 ps lifetime, of minor amplitude. An example of the decay associated spectra, obtained by a global fitting analysis of data points collected between 670 and 710 nm is presented in Fig. 1b. Though the DAS have a similar overall band shape, we do not wish to interpret this in terms of optical homogeneity due the small number of points and the inherently high noise levels of this type of measurement. Detailed analysis of sample optical homogeneity/ heterogeneity requires the spectral resolution of steadystate measurements. These time resolved data are, in general terms, similar to previously reported measurements of this kind on CP29 (Moya et al. 2001; Crimi et al. 2001; Huyer et al. 2004) which also show two dominant long lifetime decays as well as a minor fast decaying component. This minor, short lifetime, decay is unexpected in an isolated antenna complex and, to date, its physical origins

Fig. 1 a Typical fluorescence decay measurement of the reconstituted CP29 chlorophyll/protein complex and the errors analysis for the 2 and 3 exponential fits. b Decay associated spectra of the three exponential descriptions: circles s = 3.5 ns; squares s = 5 ns; triangles s = 110 ps

are not properly understood. It may represent a small fraction of aggregated complexes, known to have a very low fluorescence yield and short lifetime and/or complexes which are in the so-called ‘non-photochemical quenching’ conformation (van Oort et al. 2007). van Oort et al. (2007) induced the formation of a short lifetime component by application of high pressure to the sample. In the present

123

56

Photosynth Res (2012) 111:53–62

study, as in most of the earlier ones in the literature, we concentrate on the two long lifetime decays. CP29 fluorescence decay: effect of quinone fluorescence quencher In the present study, we have used DBMIB in an attempt to gain information on the multiexponential fluorescence decay of CP29. This quencher has been extensively used in studies on intact photosystems and isolated chlorophyll/protein complexes (e.g. Hodges and Barber 1984; Zucchelli et al. 1995; Bukhov et al. 2003; Rajagopal et al. 2003; Santabarbara and Jennings 2005). It is well known that quinones are efficient quenchers of the chlorophyll singlet excited state due to the formation of a charge transfer state complex with chlorophyll (e.g. Karukstis et al. 1988, Santabarbara et al. 1999, 2001). Artificial quinones and their analogues are known to bind to natural quinone protein sites. DBMIB acts as a plastoquinone analogue in the Cyt b6f complex. As quinone binding sites are not present in antenna proteins, there is no reason to suspect that DBMIB would bind to specific sites of CP29 and, at the same time, interact with chlorophyll(s) to form a charge transfer complex. The extremely lipophylic character of DBMIB suggests that its access to the chlorophyll sites will not be affected by protein conformation changes which do not modify the lipophylic index of a protein. In general terms, chlorophyll fluorescence quenching occurs by either a dynamic (diffusional) mechanism, which is the well-known Stern–Volmer quenching (Stern and Volmer 1919), or by a static mechanism. The diffusional mechanism depends on the probability of collision between the quencher and the excited state of the fluorophore. This is the classical Stern–Volmer quenching which may be described according to Eq. 1 for an exponential decay:   F0 F ¼ s 0 s ¼ 1 þ k q s 0 ½ Q  ð1Þ where F0 and F are the fluorescence yield in the absence and presence, respectively, of quencher, and are determined in the usual way from fluorescence lifetime data P ðF ¼ i Ai si Þ. s0 and s are the fluorescence lifetimes in the absence and presence of quencher, kq is the bimolecular quenching rate constant, [Q] is the quencher concentration and Ai are the fit decay amplitudes (counts). On the other hand, the static quenching mechanism is based on the formation of a stable complex between the fluorophore and the quencher. This complex is often without measurable fluorescence, i.e. the fluorophore is completely quenched and, for the case of an exponential decay, may be described by Eq. 2 (Lakowicz 1999): F0 =F ¼ A0 =A ¼ 1 þ Ks ½Q

ð2Þ

In this case, Ks is the association constant (equilibrium constant) of the quencher with the fluorophore.

123

In both these cases, the F0 =F plot is linear with the quencher concentration though, as indicated above, the physical meaning of the slope is different. Quenching, in the static case (Eq. 2), is due to a decrease in the concentration of effectively fluorescing fluorophores and hence is characterised by changes only in the amplitude of the fluorescence decay. This, of course, does not occur for diffusional quenching (Eq. 1). In Fig. 2, data are presented for the experiments in which the reconstituted CP29 complex was titrated with DBMIB and the fluorescence decay measured. Figure 2a shows the linear F0 =F plot, thus indicating a non-anomalous fluorescence quenching of recombinant CP29 with DBMIB. It will also be noticed that the s0/s ratio, where the

Fig. 2 Titration with DBMIB in reconstituted CP29 preparations. The decay data were fitted to a three exponential function. Parameters F0 =F are presented in a where F0 and F are the fluorescence yield in the absence and presence, respectively, of quencher, and are determined in the usual way from fluorescence lifetime data (F = RAisi). b The parameters A0/A (filled circle) together with s0/s (open circle) are presented. A0 and A are the amplitude sums of the 3 and 5 ns components in the absence and presence of the quencher, respectively. The s values are the mean lifetime values and were determine as RAis2i /RAisi. c The F0 =F plots for the 5 ns (open square) and 3 ns (filled square) decay components. Three different titration samples are shown

Photosynth Res (2012) 111:53–62

57

s values indicate the mean lifetimes (see legend to Fig. 2), are almost constant. From Fig. 2b, it can be seen that the amplitude plot has a pronounced slope, similar to the F0 =F plot, indicating that the fluorescence quenching is largely due to amplitude changes. Thus, we conclude that quenching is essentially of the static kind. The conventional interpretation of static quenching, as indicated above, is that the quencher forms a stable complex with the chromophore, which in this case would be one of the six chlorophyll a molecules bound to CP29. Rapid, subpicosecond, energy transfer would then ensure that the entire complex become effectively non-fluorescent. We have examined the effect of this quinone quencher on the fluorescence yield of the two main decay components of CP29 in an attempt to gain information on the nature of the sample heterogeneity which they seem to represent. The data (Fig. 2c) show that, while the 3 ns component is strongly quenched, with a linear F0 =F versus [DBMIB] plot, the 5 ns decay displays an approximately sevenfold reduced sensitivity. This means that the association constant between DBMIB and the two CP29 states differs by a factor of about seven. This interesting observation will be taken up later.

CP29 sample: absorption and fluorescence homogeneity In this section, we analyse the recombinant CP29 sample for spectroscopic heterogeneity. As indicated in greater detail below, the chlorophyll absorption/fluorescence bands are sensitive to distortions of the tetrapyrrole ring which are brought about by the conformation of the binding sites themselves (Zucchelli et al. 2007). Two independent approaches have been used. We initially present data on the recombinant CP29 sample using the thermodynamic absorption/fluorescence relation elaborated by Stepanov (1957; Eq. 3): FðmÞ=AðmÞ ¼ DðTÞm2 e

khmT B

which are not spectroscopically identical. This was demonstrated quantitatively by van Metter and Knox (1979). The important question in this case is the spectral resolution of the approach, the details of which can be found in the supplementary material. It is of paramount importance for this type of analysis, in which the emission spectrum is calculated from the absorption spectrum, that the absorption baseline must be unambiguously determined. Our measurements were performed with a high resolution CCD array. The absorption spectrum for a typical reconstituted CP29 sample is presented in Fig. 3 where the continuous line represents the spectrum with an uncorrected instrumental baseline, which is slightly sloping. The dashed curve represents the spectrum from which this slightly sloping instrumental base line was subtracted. The difference between these two curves near 760 nm is of the order 0.0005 of the absorption maximum, near 678 nm (Fig. 3). When the zero point is taken in the range 750–800 nm, using either baseline, the calculated emission spectra are indistinguishable between 720 and 685 nm and, in particular, in the wavelength region near the maxima (±5 nm; data not presented). For all subsequent calculations, it is the corrected baseline spectrum which is used. In Fig. 4a, the calculated emission spectrum, using the measured absorption spectrum and Eq. 3 is compared with the measured emission spectrum. It should be underlined that these spectra were measured with the same CCD instrument, so between-instrument wavelength differences do not exist.

ð3Þ

As is well known, this equation is based on the assumption of complete thermalisation of the excited state. It was developed for a pigment solution but may be used also for coupled pigment systems, such as chlorophyll–protein complexes (e.g. van Metter and Knox 1979; Zucchelli et al. 1995; Jennings et al. 2000, 2003; Dau 1996; Tumino et al. 2008). It has been also applied, with success, to both native and recombinant CP29 suspensions (Giuffra et al. 1997; Belgio et al. 2010). It is known to provide accurate data substantially in the absorption/fluorescence overlap interval which, in the present case of CP29, is between about 660 and 690 nm. As one might intuitively imagine, this relation is violated in the case of an inhomogeneous sample of complexes; for example, in a suspension of CP29 complexes

Fig. 3 Absorption spectra, measured at room temperature, of the Qy absorption region of the reconstituted CP29 preparation. The continuous line represents the spectrum with an uncorrected instrumental baseline, which is slightly sloping. The dashed curve represents the spectrum from which this slightly sloping instrumental baseline was subtracted

123

58

Visual comparison indicates, as previously reported (Giuffra et al. 1997; Belgio et al. 2010), that close agreement exists between the Stepanov calculated emission and that measured. In the wavelength region around the maxima (±5 nm) the spectra are indistinguishable (Fig. 4b). There are, however, differences outside this wavelength range. A small deviation is observed in the short wavelength region and this is clearly seen in the difference spectrum ‘measured minus calculated’ (Fig. 4c). This short wavelength difference spectrum structure peaks near 670 nm and is, therefore, presumably due to a small amount of uncoupled chlorophyll present in the sample. Uncoupled chlorophylls, the emission of which is detected in the measured spectrum, do not equilibrate with the pigment reticulum and their signal is therefore absent in the calculated spectrum. As can be seen this signal is very small, indicative of a tiny fraction of uncoupled chlorophylls, which may be ignored from the point of view of two main CP29 fluorescence decay populations. It will also be noticed that the measured and calculated spectra deviate above 688 nm, and this is clearly seen in the difference spectrum (Fig. 4c). This deviation is associated with distortions in the calculated fluorescence spectrum outside the area of good absorption/fluorescence overlap. It is normally encountered in analyses of this kind (e.g. Giuffra et al. 1997; Zucchelli et al. 1995). The data presented in Fig. 4 therefore suggest that, with the exception of the tiny fraction of uncoupled chlorophylls, the sample seems to be spectroscopically homogeneous. This interpretation is largely based on the almost perfect correspondence between the calculated and measured spectra in the 10 nm range around the wavelength maximum. It should be mentioned that this conclusion concerns only the chla molecules bound to the reconstituted CP29, as the chlb molecules have a very low effective fluorescence yield and are thus largely excluded from this kind of analysis. We have examined the effective resolution of this approach in detecting absorption/fluorescence heterogeneity in this CP29 sample for the situation described in the fluorescence decay measurements in which the relative amplitudes of the two principle decay components are 0.6 and 0.4 and the relative fluorescence yields of these two components are both 0.5. Both parameters contributed to the simulation of spectroscopic heterogeneity. This analysis (supplementary material) leads us to conclude that the Stepanov approach which we have used has quite a high resolution and can probably detect sample heterogeneity differences down to about 0.5 nm in terms of the wavelength position of CP29 absorption, with or without spectral broadening or narrowing. The experimental data do not indicate such differences and so we conclude that if sample absorption heterogeneity does exist it must be no greater than 0.5 nm for the recombinant CP29 sample.

123

Photosynth Res (2012) 111:53–62

Fig. 4 Absorption/fluorescence Stepanov analysis of the reconstituted CP29 preparation. a The measured fluorescence emission spectrum (dashed line) and the emission spectrum calculated from the absorption spectrum (continuous line). b A ‘blow up’ of the spectrum presented in (a) for the wavelength interval 676–686 nm. c The difference fluorescence spectrum ‘measured minus calculated’. Please note the scale

Photosynth Res (2012) 111:53–62

In order to further examine this aspect of sample heterogeneity more directly we have used another approach, which takes advantage of the observation (Fig. 2) that the fluorescence quencher DBMIB is about seven times more effective in quenching fluorescence associated with the 3 ns component than that of 5 ns component. Thus, DBMIB may be used to as a quasi selective quencher of the 3 ns component. In Fig. 5, data are presented for an experiment in which the steady-state emission spectra of the CP29 sample was measured both before and after the addition of a concentration of DBMIB which quenched fluorescence by about 50%. It is apparent that the two spectra, normalised at their maxima, are very similar. The minor broadening on the short wavelength side, which has its difference spectrum maximum near 670 nm (data not shown), is due to a small amount of uncoupled chlorophylls, which was also observed in the Stepanov analysis. At other wavelengths in the 676–690 nm range the minor oscillations do not exceed 0.002, and are about four times lower than the expected statistical fluctuations. Thus, the two spectra in this wavelength interval may be considered to be identical. According to the quasi selectivity of DBMIB for the 3 ns decay, the spectrum of the quenched sample should represent the 5 ns component to about 90% and the 3 ns component to about 10%. It is therefore concluded that the two fluorescence decays are generated by sample populations which are spectroscopically indistinguishable, according to this absorption/fluorescence analysis.

Fig. 5 Steady-state fluorescence emission spectra of the reconstituted CP29 preparation in the presence (dashed line) and absence (continuous line) of the quinone quencher DBMIB. To enable comparison spectra have been normalised at the peak wavelength. The quenching induced by DBMIB was approximately 50%

59

Discussion The data presented in the present paper pertain to the question of whether the multicomponent fluorescence decay in reconstituted CP29, and in particular to the two dominant decay components (3.5 and 5 ns for our samples), are associated with the presence of an optically heterogeneous sample (see ‘Introduction’ section). Recently, on the basis of single molecular fluorescence spectroscopy Kru¨ger et al. (2010) have shown that marked optical fluctuations occur in isolated LHCII trimers. These fluctuations, some of which seem to be long lived (s) are suggested to be in dynamic equilibrium and could give rise to a multicomponent fluorescence decay. In our approach, using CP29, two independent strategies were used: (i) the Stepanov absorption/fluorescence relation (Eq. 3); (ii) the selective fluorescence quenching of the 3 ns decay component by the quinone quencher DBMIB. On the basis of these experiments, we conclude that the sample displays no detectable absorption/fluorescence spectral heterogeneity. Thus, it is concluded that in reconstituted CP29 the multicomponent decay is not generated by optically different complex subpopulations. This conclusion has relevant implications concerning the chlorophyll binding sites. As previously described (see ‘Introduction’ section), the chlorophyll binding sites induce distortions in the tetrapyrrole ring leading to sizeable shifts on the pigment absorption energies (Zucchelli et al. 2007) and may therefore be used as markers for the chlorophyll binding site configuration. The absence of detectable optical heterogeneity in the samples used in this study, therefore, suggests that the relative chlorophyll a binding sites in all, or almost all, complexes have the same structure. In the present study, we have used the lipophylic fluorescence quencher DBMIB to study the two long lifetime decay components of CP29. The data demonstrate that the quenching is substantially static and thus indicate that stable quenching complexes are formed between the quencher and the chlorophyll(s) of the CP29 complexes. Therefore, if two conformers were present, in dynamic equilibrium, one would expect that both states be quenched to a somewhat similar extent by the stable chlorophyll/ DBMIB interaction. This is not the case as the 3 ns decay is approximately sevenfold more quenched than the 5 ns one. We therefore are lead to conclude that the multiple fluorescence decay in reconstituted CP29 is not generated by conformers associated with protein dynamics. As mentioned above, the existence of dynamic spectral heterogeneity of LHCII trimers has, in fact, been recently demonstrated by single molecule fluorescence spectroscopy (Kru¨ger et al. 2010) and it is certainly conceivable that this might be the basis of the multiexponential decay. This is certainly an interesting point, however, at least for

123

60

the time being, the static quenching by DBMIB of just one decay component is not in favour of this suggestion for the CP29 decay. While the present study on reconstituted CP29 is not in agreement with some kind of intrinsic conformational protein heterogeneity, it is difficult to escape the conclusion that some kind of sample heterogeneity must be involved. This heterogeneity, according to our analysis, does not involve the absorption/fluorescence spectra, but is characterised by different fluorescence lifetimes and thus by different excited state/ground state relaxation kinetics. As already mentioned in the Introduction, it has been often suggested that the observation of multiexponential fluorescence decay in isolated antenna complexes of PS II could be due to heterogeneity in the carotenoid binding in specific protein binding sites and, in particular, relating to the relative abundance of xanthophyll cycle pigments (e.g. Gruszecki et al. 1997; Moya et al. 2001; Crimi et al. 2001; Avenson et al. 2008). This would in turn lead to the presence of different subpopulations of complexes. Even though there is evidence that the presence of zeaxanthin in place of violaxanthin does lead to quenching of the excited state lifetime (e.g. Ahn et al. 2008; Avenson et al. 2008), it is worth noticing that the fluorescence decay remains multiexponential even in complexes reconstituted either in the presence of only violaxanthin or zeaxanthin (Avenson et al. 2008). Also, as pointed out in the Introduction, the fluorescence decay is multiexponential also in the antenna core complex CP47 (Huyer et al. 2004), that does not bind xanthophylls. Thus, it is difficult to think that differences in the carotenoid content, per se, could account significantly for the heterogeneity observed in the fluorescence lifetime analysis of this study and several others in the literature (e.g. Gruszecki et al. 1997; Moya et al. 2001; Crimi et al. 2001; Huyer et al. 2004; Avenson et al. 2008; van Oort et al. 2007). In addition, we point out that the selective quenching of one decay component by the lipophylic quencher DBMIB is difficult to envisage in terms of different xanthophyll contents, as these are not expected to significantly alter the lipophylic index of the chlorophyll– protein complex, and hence should not modify DBMIB binding to chlorophyll. Thus, we are led to suggest that another source of heterogeneity should be considered. We propose, as an hypothesis, that it may in fact be the interaction with the solubilising agent, be it detergent or liposome, that generates this lifetime heterogeneity. This possibility was considered by Huyer et al. (2004), who rejected it on the bases of the apparent improbability that the different solubilising conditions would lead to just ‘two distinct micelle structures with similar distribution probabilities’. The comment on ‘two micelle structures’ derives from their fluorescence decay analysis on three isolated antenna complexes,

123

Photosynth Res (2012) 111:53–62

LHCII, CP29 and CP47, in which just two decay components were observed in the 2–3 and 4–5 ns range. However in contrast to this, the literature contains a large number of reports in which the fluorescence decay shows greater complexity than that reported by Huyer et al. (2004) (e.g. Gruszecki et al.1997; Moya et al. 2001; Crimi et al. 2001; Avenson et al. 2008; Passarini et al. 2010, the present study). It is therefore clear that the solubilised antenna complexes display complex decay kinetics and there seems to be no reason to exclude the possibility that this decay heterogeneity may be generated by a solubilisation-induced heterogeneity, i.e. detergent/protein interactions, on this basis. We note that this point of heterogeneity of the detergent/protein interactions, at the detergent concentrations conventionally used, is implicit in the non-linear polarisation data for LHCII measured over a wide range of detergent concentrations (Voigt et al. 2008). Furthermore, the selectivity in DBMIB-induced quenching of the 3 and 5 ns components reported here is also readily understood in terms of different quinone partition into multiple detergent/ protein environments. Micelle packing heterogeneity, for example, which is a well known phenomenon (Garavito and Ferguson-Miller 2001), can be expected to modify DBMIB partition between the detergent phase and the solubilised CP29 complex. As detectable spectroscopic heterogeneity is absent in our CP29 sample though lifetime heterogeneity is present, we suggest that the different decay components are generated by different non-radiative decay rates, i.e. internal conversion or intersystem conversion. While changes in the electronic energy levels, and hence spectroscopic changes, are known to modify the rate of the non-radiative decay processes (e.g. Englman and Jortner 1979) the opposite does not necessarily apply. To cite one example, a modified rate of collisional quenching (internal conversion) is not expected to have detectable absorption/fluorescence spectroscopic effects but will modify the fluorescence lifetime. This detergent-induced multicomponent fluorescence decay hypothesis is compatible with optical homogeneity. Acknowledgement (2007WBRA42).

This research was financed by PRIN 2007

References Ahn TK, Avenson TJ, Ballottari M, Cheng YC, Niyogi KK, Bassi R, Fleming GR (2008) Architecture of a charge-transfer state regulating light harvesting in a plant antenna protein. Science 320:794–797 Alfonso M, Montoya G, Cases R, Rodriguez R, Picorel R (1994) Core antenna complexes, CP43 and CP47, of higher plant photosystem II. Spectral properties, pigment stoichiometry, and amino acid composition. Biochemistry 33:10494–10500

Photosynth Res (2012) 111:53–62 Avenson TJ, Ahn TK, Zigmantas D, Niyogi KK, Li Z, Ballottari M, Bassi R, Fleming GR (2008) Zeaxanthin radical cation formation in minor light-harvesting complexes of higher plant antenna. J Biol Chem 283:3550–3558 Belgio E, Casazza AC, Zucchelli G, Garlaschi FM, Jennings RC (2010) Band shape heterogeneity of the low-energy chlorophylls of CP29: absence of mixed binding sites and excitonic interactions. Biochemistry 49:882–892 Bittner T, Wiederrecht GP, Irrgang K-D, Renger G, Wasielewski M (1995) Femtosecond transient absorption spectroscopy on the light-harvesting Chl a/b protein complex of Photosystem II at room temperature and 12 K. Chem Phys 194:311–322 Bukhov NG, Sridharana G, Egorova EA, Carpentier R (2003) Interaction of exogenous quinones with membranes of higher plant chloroplasts: modulation of quinone capacities as photochemical and non-photochemical quenchers of energy in Photosystem II during light–dark transitions. Biochim Biophys Acta 1604:115–123 Cinque G, Croce R, Holzwarth A, Bassi R (2000) Energy transfer among CP29 chlorophylls: calculated Fo¨rster rates and experimental transient absorption at room temperature. Biophys J 79:1706–1717 Connelly JP, Mu¨ller MG, Hucke M, Gatzen G, Mullineaux CW, Ruban AV, Horton P, Holzwarth AR (1997) Ultrafast spectroscopy of trimeric light-harvesting complex II from higher plants. J Phys Chem B 101:1902–1909 Crimi M, Dorra D, Bosinger CS, Giuffra E, Holzwarth AR, Bassi R (2001) Time-resolved fluorescence analysis of the recombinant photosystem II antenna complex CP29. Effects of zeaxanthin, pH and phosphorylation. Eur J Biochem 268:260–267 Dau H (1996) On the relation between absorption and fluorescence emission spectra of photosystems—derivation of a Stepanov relation for pigment clusters. Photosynth Res 48:139–145 Englman R, Jortner J (1979) The energy gap law for radiationless transitions in large molecules. Mol Phys 18:145–165 Garavito RM, Ferguson-Miller S (2001) Detergents as tools in membrane biochemistry. J Biol Chem 276:32403–32406 Giuffra E, Cugini D, Croce R, Bassi R (1996) Reconstitution and pigment-binding properties of recombinant CP29. Eur J Biochem 238:112–120 Giuffra E, Zucchelli G, Sandona D, Croce R, Cugini D, Garlaschi FM, Bassi R, Jennings RC (1997) Analysis of some optical properties of a native and reconstituted photosystem II antenna complex, CP29: pigment binding sites can be occupied by chl a or chl b and determine spectral forms. Biochemistry 36:12984–12993 ¨ zdemir S, Gu¨len D, van Stokkum IHM, van Gradinaru CC, O Grondelle R, van Amerongen H (1998) The flow of excitation in LHCII monomers. Implications for the structural model of the major plant antenna. Biophys J 75:3064–3077 Gruszecki WI, Matuła M, Mys´liwa-Kurdziel B, Kernen P, Krupa Z, Strzałka K (1997) Effect of xanthophyll pigments on fluorescence of chlorophyll a in LHCII embedded to liposomes. J Photochem Photobiol B 37:84–90 Hodges M, Barber J (1984) Analysis of chlorophyll fluorescence quenching by DBMIB as a means of investigating the consequences of thylakoid membrane phosphorylation. Biochim Biophys Acta 767:102–107 Huyer J, Eckert HJ, Irrgang KD, Miao J, Eichler HJ, Renger G (2004) Fluorescence decay kinetics of solubilized pigment protein complexes from the distal, proximal, and core antenna of Photosystem II in the range of 10–277 K and absence or presence of sucrose. J Phys Chem B 108:3326–3334 Jennings RC, Elli G, Garlaschi FM, Santabarbara S, Zucchelli G (2000) Selective quenching of the fluorescence of core chlorophyll-protein complexes by photochemistry indicates that

61 Photosystem II is partly diffusion limited. Photosynth Res 66:225–233 Jennings RC, Zucchelli G, Croce R, Garlaschi FM (2003) The photochemical trapping rate from red spectral states in PSILHCI is determined by thermal activation of energy transfer to bulk chlorophylls. Biochim Biophys Acta 1557:91–98 Johnson MP, Pe´rez-Bueno ML, Zia A, Horton P, Ruban AV (2009) The zeaxanthin-independent and zeaxanthin-dependent qE components of nonphotochemical quenching involve common conformational changes within the Photosystem II antenna in Arabidopsis. Plant Physiol 149:1061–1075 Karukstis KK, Gruber SM, Fruetel JA, Boegeman SC (1988) Quenching of chlorophyll fluorescence by substituted anthraquinones. Biochim Biophys Acta 932:84–90 Kru¨ger TP, Novoderezhkin VI, Ilioaia C, van Grondelle R (2010) Fluorescence spectral dynamics of single LHCII trimers. Biophys J 98:3093–3101 Lakowicz JR (1999) Principles of fluorescence spectroscopy, 2nd edn. Plenum Press, New York Moya I, Silvestri M, Vallon O, Cinque G, Bassi R (2001) Timeresolved fluorescence analysis of the Photosystem II antenna proteins in detergent micelles and liposomes. Biochemistry 40:12552–12561 Pascal AA, Liu Z, Broess K, van Oort B, van Amerongen H, Wang C, Horton P, Robert B, Chang W, Ruban AV (2005) Molecular basis of photoprotection and control of photosynthetic lightharvesting. Nature 436:134–137 Passarini F, Wientjes E, van Amerongen H, Croce R (2010) Photosystem I light-harvesting complex Lhca4 adopts multiple conformations: Red forms and excited-state quenching are mutually exclusive. Biochim Biophys Acta 1797:501–508 Rajagopal S, Egorova EA, Bukhov NG, Carpentier R (2003) Quenching of excited states of chlorophyll molecules in submembrane fractions of Photosystem I by exogenous quinones. Biochim Biophys Acta 1606:147–152 Santabarbara S, Jennings RC (2005) The size of the population of weakly coupled chlorophyll pigments involved in thylakoid photoinhibition determined by steady-state fluorescence spectroscopy. Biochim Biophys Acta 1709:138–149 Santabarbara S, Garlaschi FM, Zucchelli G, Jennings RC (1999) The effect of excited state population in Photosystem II on the photoinhibition-induced changes in chlorophyll fluorescence parameters. Biochim Biophys Acta 1409:165–170 Santabarbara S, Neverov KV, Garlaschi FM, Zucchelli G, Jennings RC (2001) Involvement of uncoupled antenna chlorophylls in photoinhibition in thylakoids. FEBS Lett 491:109–113 Stepanov BI (1957) A universal relation between the absorption and luminescence spectra of complex molecules. Dokl Akad Nauk SSSR 112:839–843 (Dokl Akad Nauk SSSR Phys Sect (Engl Transl) 2:81–84) Stern O, Volmer M (1919) On the quenching time of fluorescence. Phys Z 20:183–188 Tumino G, Casazza AP, Engelmann E, Garlaschi FM, Zucchelli G, Jennings RC (2008) Fluorescence lifetime spectrum of the plant Photosystem II core complex: photochemistry does not induce specific reaction center quenching. Biochemistry 47:10449– 10457 van Metter RL, Knox RS (1979) Fluorescence of light-harvesting chlorophyll a/b-protein complexes: implications for the photosynthetic unit. In: Wolstenholme G, Fitzsimons DW (eds) Chlorophyll organisation and energy transfer in photosynthesis. Excerpta Medica, Amsterdam, pp 177–186 van Oort B, van Hoek A, Ruban AV, van Amerongen H (2007) Equilibrium between quenched and nonquenched conformations of the major plant light-harvesting complex studied with high-

123

62 pressure time-resolved fluorescence. J Phys Chem B 111:7631–7637 Visser HM, Kleima FJ, van Stokkum IHM, van Grondelle R, van Amerongen H (1996) Probing the many energy-transfer processes in the photosynthetic light-harvesting complex II at 77 K using energy selective sub-picosecond transient absorption spectroscopy. Chem Phys 210:297–312 Voigt B, Krikunova M, Lokstein H (2008) Influence of detergent concentration on aggregation and spectroscopic properties of light-harvesting complex II. Photosynth Res 95(2–3):317– 325

123

Photosynth Res (2012) 111:53–62 Zucchelli G, Garlaschi FM, Croce R, Bassi R, Jennings RC (1995) A Stepanov relation analysis of steady-state absorption and fluorescence spectra in the isolated D1/D2/cytochrome b-559 complex. Biochim Biophys Acta 1229:59–63 Zucchelli G, Brogioli D, Casazza AP, Garlaschi FM, Jennings RC (2007) Chlorophyll ring deformation modulates Qy electronic energy in chlorophyll-protein complexes and generates spectral forms. Biophys J 93:2240–2254 Zuker M, Szabo AG, Bramall L, Krajcarski DT, Selinger B (1985) Delta function convolution method (DFCM) for fluorescence decay experiments. Rev Sci Instrum 56:14–22