PCCP View Article Online
Published on 27 October 2014. Downloaded by University of Cambridge on 18/11/2014 05:27:25.
PAPER
Cite this: Phys. Chem. Chem. Phys., 2014, 16, 26059
View Journal | View Issue
Ultrafast excited state processes in Roseobacter denitrificans antennae: comparison of isolated complexes and native membranes† Marco Ferretti,*a Katia Duquesne,b James N. Sturgisc and Rienk van Grondelle*a Roseobacter (Rsb.) denitrificans is a marine aerobic anoxygenic photosynthetic purple bacterium with an unusually high-800 nm absorption band. Ultrafast excited state processes have been intensively studied in the past in order to understand why the energy transfer efficiency between photosynthetic antennae approaches unity and recently it has been proved that the organization of the antennae proteins within the membranes plays an important role. Thanks to the development of genetic manipulation and to the capability of Rsb. denitrificans to grow anaerobically as well, it is possible to construct several mutants in order to compare the ultrafast dynamics between isolated complexes and complexes embedded in membrane environments. Time resolved fluorescence and transient absorption have been applied to isolate LH2, genetically modified membranes with LH2-only and wild type membranes with both LH2
Received 8th July 2014, Accepted 27th October 2014
and LH1 antennae of Rsb. denitrificans, in order to understand the effect of the membrane environment
DOI: 10.1039/c4cp02986k
states of LH2 and LH1, and although there is shortening of the relaxation lifetime of the LH2-only membranes with respect to the isolated LH2, we find an energy transfer efficiency from LH2 to LH1 of
www.rsc.org/pccp
95%, which still approaches unity.
on the energy transfer efficiency. A global analysis is applied to calculate the lifetime of the excited
1 Introduction In natural photosynthesis the energy transfer between pigment–protein complexes is considered one of the most efficient
a
VU University, Amsterdam, The Netherlands. E-mail: [email protected] Aix-Marseille Universite´, Centrale Marseille, CNRS, ISM2 UMR 7313, 13397 Marseille, France c Aix-Marseille University, CNRS, LISM, Marseille, France † Electronic supplementary information (ESI) available: Membrane samples such as the Dpuf or the wild type membranes used in this experiment may in general be affected by sample turbidity, different from the isolated complexes. The sample turbidity scatters the light and too much scattering may affect the time resolved fluorescence dynamics when the measurement is performed using a streak camera. However the sample turbidity does not affect the dynamics measured via transient absorption, where the signal originates only from the section of the sample where there is a strong overlap between the pump and probe, which corresponds to the non-scattered part of the laser beam. The fact that the transient absorption results of global analysis reproduce the results of time resolved fluorescence is a proof that the turbidity of the sample does not affect the results shown in this paper. However to show that the turbidity does not affect the steady state absorption, an absorption spectrum starting at 220 nm is shown in the following figure. It is clear from the baseline that the effect of the scattering is minimum even in the UV region where this effect is maximum. The peak appearing at 260/280 nm in the membranes samples (Dpuf and wild type) is due to the presence of some residual ribosomes and it is not relevant either for photosynthesis or for the analysis applied in this paper, which is the focus of the near-IR dynamics. See DOI: 10.1039/c4cp02986k b
This journal is © the Owner Societies 2014
ultra-fast processes. For example in the purple photosynthetic bacteria the incident light energy is collected by the antennae, which are able to transfer the excitation energy to the reaction centers or to other antennae with an efficiency approaching unity. The organization of the pigments within the proteins and of the proteins within the membrane is fundamental for tuning the spectral and energy transfer properties of the complexes in order to ensure high efficiency.1–5 The light harvesting (LH) complexes of purple bacteria can be divided into two groups: the core light-harvesting complex (LH1–RC) which is composed of the LH1 antenna surrounding the reaction center (RC) and the peripheral antenna, LH2, which transfers the absorbed energy to LH1. The LH1 pigment–protein complex is typically formed by an apparently closed ring of 16 subunits,1,2,6,7 each subunit being composed of a bacteriochlorophyll (BChl) dimer non-covalently bound to the protein scaffold of an a–b-polypeptide8 pair and a carotenoid molecule. The LH2 peripheral antenna consists of a ring of 89 or 910 subunits, each containing 3 BChl and one or two carotenoids. The BChls form two rings: one of quasi-monomeric, weakly coupled, pigments absorbing near 800 nm (B800) and one of more strongly coupled dimeric pigments absorbing typically at 850 nm (B850), though the position of the absorption maximum varies between species and with growth conditions.11 A few LH2 complexes show a high-800 nm absorption, with the 850 nm band blue-shifted
Phys. Chem. Chem. Phys., 2014, 16, 26059--26066 | 26059
View Article Online
Published on 27 October 2014. Downloaded by University of Cambridge on 18/11/2014 05:27:25.
Paper
and super-imposed on the 800 nm band. Examples are the LH2s of Rhodopseudomonas (Rps.) palustris grown at very low light intensities12 and Roseobacter (Rbs.) denitrificans.13 However it is still debated which structural modification may cause this shift13 and whether there is any advantage for the light-harvesting efficiency of the LH2 antennae. Recently, due to the development of atomic force microscopy, the importance of the organization of the proteins within the membrane for the energy transfer functionality of purple bacteria has been shown,1,14,15 therefore it is extremely interesting to compare excitation transfer dynamics between isolated complexes and complexes embedded in different membrane environments, in order to understand the origins and the consequences of any observable differences. Because of the ultra-fast time scale of energy transfer (a few to several hundreds of picoseconds), ultra-fast spectroscopic techniques such as transient absorption and time resolved fluorescence are instrumental for studying the dynamics of the energy transfer processes. These two techniques have been intensively used to study these dynamics in photosynthetic antennae in bacteria and plants; however most of the past studies include comparison between the dynamics of isolated antennae vs. wild type membranes. Nowadays, thanks to the development of genetic manipulation, it is possible to construct mutants in several species7 in order to study the properties of the LH2 complexes within the membrane environment, the results of which should better reflect the in vivo conditions than those obtained for a detergent-isolated complex. In this work we apply transient absorption and time resolved fluorescence to isolated LH2, genetically modified membranes with only LH2 complexes (LH2-only membranes) and wild type membranes (with both LH2 and LH1–RC) of Rbs. denitrificans. In this way we can compare the dynamics of isolated complexes and LH2-only membranes in order to investigate the role of the protein organization within the membranes for the energy transfer efficiency. Global analysis is applied to transient absorption and time resolved fluorescence traces in order to calculate the energy transfer time from LH2 to LH1 and the intrinsic relaxation time of the LH2. The main difference between isolated and in membrane LH2 is shortening of the relaxation lifetime and the appearance of a fast decay component in the latter case. Therefore the study of the dynamics as a function of the photon excitation density is done in order to understand the nature of the fast decay component, which can arise from energy reorganization, energy annihilation or energy quenching.
2 Results Steady state absorption The steady state absorption spectra of isolated LH2, LH2-only and wild type membranes are shown in Fig. 1. The isolated LH2 (in blue) absorption spectrum shows one peak in the near infrared at 800 nm which is the convolution between the monomeric B800 and the oligomeric B850 bands, differently
26060 | Phys. Chem. Chem. Phys., 2014, 16, 26059--26066
PCCP
Fig. 1 Steady state absorption of isolated LH2 (blue), LH2-only membranes or Dpuf membranes (black) and wild type membranes (red) at room temperature.
from LH2 from other species. In fact the 850 band of Rsb. denitrificans is blue-shifted with respect to the one of Rps. acidophila and at room temperature is superimposed on the B800 band13 as it is shown in Fig. 1. The LH2-only membrane absorption (Dpuf mutant, in black) shows one major near-IR absorption peak at 800 nm, as in the isolated LH2; however there is also a minor peak near 760 nm. This is due to the presence of bacteriopheophytin a and bacteriochlorophyll a in the membranes that are not associated with photosynthetic complexes. The 760 nm peak shown in Fig. 1 is more pronounced than that shown in Duquesne et al.,8 indeed the size of this peak, which is always present to some degree in LH2-only membranes, varies somewhat between preparations. The presence of this peak could be explained by the presence of damaged complexes within the membranes, or more probably by accumulation of Bchl a in the absence of the LH1–RC to which it is normally bound. In the visible region the peaks at 520–550 nm and 410 nm are due to the presence of relatively large amounts of cytochrome in the membranes, probably also a consequence of the inability to synthesize the LH1–RC complex. The wild type membrane absorption (in red) shows two major peaks in the near infrared, one at 800 nm arising essentially from the LH2 absorption bands and one at 870 nm due to the LH1–RC absorption band.
Time resolved spectroscopy The isolated LH2, the LH2-only and the wild type membranes are excited at 790 nm, in the blue wing of the 800 nm absorption bands of the LH2, in order to study the energy transfer from LH2 to LH1–RC. The transient absorption signal is detected in the near-infrared region, with an instrument response function (IRF) of about 100 fs. In Fig. 2, the left hand frames show a series of
This journal is © the Owner Societies 2014
View Article Online
Published on 27 October 2014. Downloaded by University of Cambridge on 18/11/2014 05:27:25.
PCCP
Paper
Fig. 2 Transient absorption (left frame) and time resolved fluorescence (right frame) spectra for isolated LH2 (top), LH2-only membranes (center) and wild type membranes (bottom). The points correspond to the data and the lines to the result of the global analysis.
spectra, recorded at different time delays between the pump and probe. The displayed signal is measured at room temperature with an energy per pump pulse of 15 nJ (which corresponds to a photon density below 1013 cm2), this is well below the excitation density where annihilation processes are typically observed in isolated LH2 from other purple bacteria species16,17 The transient absorption spectrum of LH2-only membranes shows a negative peak at 818 nm, which arises from the convolution between ground state bleaching (GSB) and stimulated emission (SE) of the LH2 antenna. From these kinetics it is possible to measure the relaxation lifetime of the LH2 in membranes where no energy transfer to the LH1–RC is possible. In the bottom-left frame the transient absorption of the wild type membranes is shown. The main difference with respect to the LH2-only membranes is that after tens of picoseconds the GSB and the SE of the LH2 vanish while a negative peak appears at 890 nm, which is due to GSB and SE of LH1, and a positive peak at 850 nm, which corresponds to the excited state absorption (ESA) of the LH1 antenna.18 From these kinetics it is possible to measure the energy transfer time from LH2 to LH1–RC and, knowing the excitation lifetime of the isolated LH2 or of the LH2-only membranes, the quantum efficiency of this process can be evaluated.19 In this way we can compare the differences between isolated LH2 and LH2 in membranes, focusing in particular on the effects of the membrane environment on the energy transfer efficiency. The fluorescence kinetics are recorded using a streak camera, after excitation with sub-picosecond pulses with an energy per pulse of 14 nJ. The IRF is about 5–6 ps and is limited by the detector response. In Fig. 2 right frame, some relevant spectra are shown. From the fluorescence kinetics it is clear that the lifetime of the isolated LH2 is much longer than the lifetime of
This journal is © the Owner Societies 2014
LH2 in membranes, in agreement with the transient absorption results discussed above. In fact the former shows a signal after more than 1 ns, whereas in the latter the fluorescence is almost zero after 200 ps. The wild type membrane kinetics shows energy transfer from LH2 to LH1–RC, in fact right after time zero the LH2 is emitting and the LH1 fluorescence is almost zero, whereas already after only 30 ps the fluorescence signal is dominated by emission from LH1–RC. A global analysis was performed on both transient absorption and time resolved fluorescence kinetics in order to evaluate the relaxation lifetime of the LH2 and the energy transfer time from LH2 to LH1 (for an explanation of the global analysis model refer to the section Experimental). In Fig. 3 the evolution associated spectra (EAS) corresponding to time-resolved fluorescence (right frame) and the evolution associated difference spectra (EADS) corresponding to transient absorption (left frame) are shown. The top frame is related to isolated LH2 and the kinetics can be well described by a mono-exponential decay (only one component), in agreement with previous results on different species of purple bacteria.8,20,21 The mono-exponential decay supports the assumption of the absence of exciton annihilation in the isolated LH2. The corresponding fluorescence-lifetime is 670 5 ps, in agreement with the LH2 lifetime of Rps. acidophila17,22 and Rhodobacter (Rb.) sphaeroides,8,23,24 despite the difference in spectral characteristics. Notice that the error arises essentially from the IRF which is obtained as a result of global analysis. In the LH2-only membranes the experimental fluorescence decay can be best described by two components (bi-exponential decay), with time constants of 11 5 and 162 5 ps. The transient absorption (Fig. 3 central left frame) can also be described by two components with lifetimes of 9.6 0.1 and
Phys. Chem. Chem. Phys., 2014, 16, 26059--26066 | 26061
View Article Online
Published on 27 October 2014. Downloaded by University of Cambridge on 18/11/2014 05:27:25.
Paper
PCCP
Fig. 3 Result from global analysis: evolution associated spectra (EAS) for time resolved fluorescence and evolution associated difference spectra (EADS) for transient absorption. A model with one component is used for isolated LH2 (top-right) whereas a model with 2 components is used for LH2-only membranes (center) and wild type membranes (bottom). In the transient absorption of wild type membranes (bottom-left) the negative areas (GSB + SE) correspond to: A1 = 2333 cm1 mOD and A2 = 902 cm1 mOD for the EADS1 and A3 = 2618 cm1 mOD for the EADS2. There is a good agreement between transient absorption and time resolved fluorescence lifetimes. In the scheme (top-left) the lifetimes of LH2-only and wild type membranes are summarized. The quenchers in the LH2-only membranes are probably the BChl a which are accumulated in the absence of the LH1–RC to which they are normally bound or large aggregates of LH2.
160.0 0.1 ps, these are in close agreement with the lifetimes obtained from fluorescence measurements. Notice that although the IRF and therefore the time resolution of the transient absorption experiment is about 100 fs, we cannot resolve the B800 to ‘‘B850’’ energy transfer. This may be caused by the spectral overlap between the monomeric and excitonic Bchl a absorption bands which causes the excitation of both bands even at the shortest time delays (as is shown in Fig. 2 bottom-left frame, in the red spectrum), or by ultra-rapid equilibration. In fact because of the spectral overlap, most likely the B800 localized states are strongly mixed (or more mixed than in other purple bacteria without the high-800 nm absorption band) with the delocalized ‘‘B850’’ states, making the ultra-rapid relaxation possible. Although we cannot resolve the ultra-rapid relaxation, we can estimate that the relaxation time is smaller than 100 fs, which corresponds to the IRF of the transient absorption experiment. The wild type membrane signal is also well modeled by a two exponential component fit. In the fluorescence signal (Fig. 3 bottom right frame) we find a fast component of 10 4 ps and a slow component of 130 4 ps. The fast decay is a bit slower than the energy transfer time from LH2 to LH1–RC observed in wild type membranes of Rb. sphaeroides,25 whereas the lifetime of 130 ps is in agreement with the one of closed26 LH1–RC of
26062 | Phys. Chem. Chem. Phys., 2014, 16, 26059--26066
other species. We assume that the reaction centers are closed (unavailable for charge separation) because of the high repetition rate of the laser in both experiments (50 kHz for fluorescence and 40 kHz for transient absorption). The transient absorption lifetimes of 8.9 0.2 and 134 0.2 ps are in very good agreement with the fluorescence results. Notice that in the fastest EADS (evolution associated difference spectra) of transient absorption there is a negative shoulder at 808 nm, which shows the bleaching of the B800 band of LH2. However we can see that the bleaching and the stimulated emission of the ‘‘B850’’ band (around 818 nm) is present already from time zero and this confirms that due to the overlap between the two absorption bands (B800 and blue-shifted B850) the pump at 790 nm excites both bands. In order to investigate the origin of the appearance of a fast component in the LH2-only membranes, a study of the fluorescence lifetimes and EAS amplitudes, as a function of the photon excitation density, has been performed, which is shown in Fig. 4. We repeated the time-resolved fluorescence experiment with several different pulse energies (1, 5 and 10 nJ) for both the LH2-only and the wild type membranes. In Fig. 4 the amplitude of the fast and slow components as a function of the excitation density is shown. It is clear from the linear relationship that there is no annihilation occurring at these photon excitation densities.
This journal is © the Owner Societies 2014
View Article Online
Published on 27 October 2014. Downloaded by University of Cambridge on 18/11/2014 05:27:25.
PCCP
Fig. 4 Amplitude of the EAS of the time resolved fluorescence kinetics as a function of the excitation energy (proportional to the density of photons). Blue and green: fast and slow components of LH2-only membranes, red and black. The linear behaviour suggests the absence of energy annihilation.
3 Discussion The excited state lifetime of the isolated LH2 complex of Rsb. denitrificans is in good agreement with that observed for the LH2 of other species such as Rps. acidophila and Rb. sphaeroides.8,17,22–24 The LH2-only membrane kinetics are very different with respect to the isolated LH2 ones, with much shorter lifetimes as compared to the isolated complexes and exhibiting a bi-exponential decay instead of a mono-exponential decay. This was already reported for membrane-embedded LH2: in the case of reconstitution of the LH2 complexes of Rps. acidophila and Rb. sphaeroides into phospholipid vesicles,21 and in the case of an LH2-only mutant of Rb. sphaeroides.27 In the work of Hunter et al.27 the fast component was interpreted as an energy transfer from the main B850 band to a red-shifted sub-population band, now known as the inhomogeneously distributed lowest exciton component of LH2. More recently the bi-exponential decay has been interpreted as annihilation caused by LH2–LH2 energy transfer in the membranes due to clustering of the LH complexes.21 Support to the idea of clustering comes from AFM measurements on native membranes28 and membrane-reconstituted LH complexes.29 The appearance of a fast component has also been observed in chromatophores of other LH2 aggregates.30–33 However the energy dependence of the decay amplitudes in Fig. 4 shows a linear relationship between EAS amplitudes and the photon excitation density. This makes the annihilation interpretation rather unlikely for our samples. The bi-exponential decay is expected for LH1–RC membranes, in fact in this case we expect two components because of two sequential reactions: the LH2 to LH1 energy transfer and the LH1 to RC energy transfer. The case of LH2-only membranes is
This journal is © the Owner Societies 2014
Paper
different, in fact the presence of two components could be interpreted as a re-equilibration within the LH2 antennae, but in this case a red-shift of the slower component, with respect to the faster, is expected due to equilibration and as you can see in Fig. 3 the two EAS corresponding to LH2-only membranes are almost identical. Therefore we interpret the presence of two time constants as being due to sample heterogeneity caused by clustering. In fact some quencher, such as oxidized bacteriopheophytin or free BChl a which are accumulated in the absence of the LH1–RC to which they are normally bound, may quench part of the energy in the LH2-only membranes. Notice that the energy quenching is in agreement with the linear behavior of the EAS amplitudes in Fig. 4. In this picture the wild type membranes and the LH2-only membranes have very similar dynamics, with the fast lifetime which correspond to energy transfer from LH2 to LH1–RC in the wild type membranes and from LH2 to quenchers in the LH2-only membranes, where the quenchers are present because of the lack of LH1 (see scheme in Fig. 3 top left panel). The nature of the fast component is very important for calculating the quantum efficiency of the energy transfer from LH2 to LH1–RC. In fact we can generally define the quantum efficiency of the energy transfer process (F) as: tET1 F ¼ P 1 ti where tET is the lifetime of the energy transfer from LH2 to LH1–RC and ti are all the possible channels of decay (the different decay components) of the excited LH2. It is clear that the calculation of quantum efficiency upon laser excitation with the same excitation density of this experiment has to take into account both lifetimes (fast and slow) of the LH2-only membranes. This leads to a quantum efficiency of 50 1%. Notice that the quantum efficiency of the in vivo complexes depend on the nature of the fast component in the LH2-only membrane. In fact if the fast component is caused by energy annihilation, the sun light photon density (1.4 kW m2) may be below the annihilation threshold. Therefore in this case only the slow component has to be taken into account, leading to a quantum efficiency of 95 1%. In the case of energy quenching, the quenchers in the LH2-only membranes may be present only in the mutant because of the lack of LH1, or the quenchers may be related to pools of LH2 aggregates,34 the size of which depends on the growth light conditions34,35 of the membranes and the presence of large LH2 aggregates may be a consequence of the lack of LH1–RC. In both cases the fast lifetime of LH2-only membranes is not physiological and it has not to be taken into account and the quantum efficiency is 95 1%. Support for this idea comes from the comparison between the GSB plus SE of the LH1 and the ones of the LH2, in the wild type membrane (Fig. 3). In fact it is possible to evaluate the quantum efficiency (or at least a lower limit) also from the ratio between the two negative areas of fast and slow EADS of wild type membranes, which in this experiment correspond to about 80% (see Fig. 3). The main issue of this method is that the GSB and SE are convoluted with the ESA and the deconvolution of the negative
Phys. Chem. Chem. Phys., 2014, 16, 26059--26066 | 26063
View Article Online
Published on 27 October 2014. Downloaded by University of Cambridge on 18/11/2014 05:27:25.
Paper
area has not a trivial solution, but depends on the relative dipole moments.36 Therefore, considering that after deconvolution the GSB plus SE area increases with respect to the convoluted area,36,37 the 80 1% represent only a lower limit. Anyway this lower limit is higher than the quantum efficiency calculated taking into account both lifetimes of LH2-only membranes. Recently it has been proposed that depending on the growth light conditions,34 large aggregates of unconnected LH2 may be present, and because they are poorly connected to the LH1–RC, these aggregates may play a role in the photoprotection mechanisms. In fact the overall quantum efficiency of the energy transfer from LH2 to LH1–RC depends on the average size of the clusters of LH2. This is in agreement with the present work, where extremely large LH2 aggregates may be present in the Dpuf membranes due to the adaptation to the lack of LH1– RC. The results presented in this paper can help in the understanding of the role of large aggregates of LH2 in purple bacteria membranes, which is still an open question. Notice that in the work of Driscoll et al., in the power study of the time resolved fluorescence, the amplitude of the slower component is not in a linear relationship with the excitation power like in this work. They explain this behavior with energy annihilation, differently from our results. The presence of annihilation in the Driscoll et al. work may be related to singlet–triplet annihilation due to the very high repetition rate of their experiment. In any case the quantum efficiency calculated using only the slow lifetime of the LH2-only membranes (95 1%) results in a lower number than the one calculated using the lifetime of the isolated complex, which is 98 1% and represent always an upper limit of the real quantum efficiency of the energy transfer. Finally notice that the LH2 of Rsb. denitrificans represents an interesting complex to be studied by 2 Dimension electronic spectroscopy (2DES) because of the strong exciton mixing between B800 and B850 bands, and this article can help future studies for the interpretation of the 2DES spectra.
4 Experimental Samples The growth of wild type and mutant cells, the preparation of membranes and the purification of LH2 complexes are described in detail elsewhere.13,38 Briefly the Rsb. denitrificans strains were grown aerobically in Marine Broth 2216 (Difco) medium at 28 1C with agitation at 200 rpm. Cells were harvested in the late log phase. For the preparation of the intracytoplasmic membranes (wild type or only-LH2), cells were resuspended in 10 mM Tris pH 7.5, 5% sucrose, 0.6 mg mL1 lysozyme, 100 mg mL1 DNAase, 20 mg mL1 RNAase, complete antiprotease cocktail (Roche) and were broken by twice passages through a French pressure cell at 900 kPa. The resulting lysate is layered directly onto 15–70% sucrose density gradients and centrifuged 90 min in a SW27 swinging bucket rotor at 25 000 rpm for 4 hours. The colored band is collected, diluted with Tris 20 mM pH 7.5, and pelleted by centrifugation for 90 min at 40 000 rpm in a Ti45 rotor. The LH2-only and the
26064 | Phys. Chem. Chem. Phys., 2014, 16, 26059--26066
PCCP
wild type membranes are resuspended in buffer containing 20 mM Tris buffer, pH 7.5. For the preparation of LH2 complexes, purified membranes described above were solubilized with dodecyl maltoside (5% final concentration in the same buffer) for 2 h overnight. Unsolubilized material was removed by centrifugation for 30 min at 60 000 rpm, and the supernatant was loaded directly onto a Resource Q column, equilibrated with Tris 20 mM pH 7.5, 20 mM NaCl, and 0.05% dodecyl maltoside. LH2 is eluted with a linear gradient from 20 mM NaCl to 1.0 M NaCl. The fractions with the highest ratio of absorbance at 800 to 280 nm are pooled and concentrated using centrifugal concentrators (30 K cutoff). The concentrated fraction is loaded onto a superose 6 gel-filtration column and eluted with Tris 20 mM pH 7.5, 20 mM NaCl, and 0.05% dodecyl maltoside. The fractions with the highest ratio of absorbance at 800 to 280 nm are again pooled and concentrated using centrifugal concentrators (30 K cutoff). The isolated LH2 are in a buffer solution containing Tris 20 mM pH 7.5 and 0.05% of b-DM. The LH2-only and the wild type membranes are in a buffer solution containing Tris 20 mM pH 7.5. Spectroscopy For the time resolved fluorescence experiments the samples were excited at 795 nm using a pulsed Titanium:Sapphire (Ti:Sa) laser system, consisting of a mode-locked oscillator (Vitesse, Coherent) and a regenerative amplifier (ReGa, Coherent), operated at a repetition rate of 50 kHz, a pulse width of about 200 fs and an energy per pulse of 1, 5, 10 and 14 nJ. Only the fluorescence signal, polarized at the magic angle, was detected at 901 with respect to the excitation beam by a streak camera device. The instrument response function (IRF) was about 4–5 ps, which is completely limited by the detector response. For the transient absorption the samples were excited at 795 nm using a pulsed Ti:Sa laser system made of a modelocked oscillator (Mira, Coherent) and a regenerative amplifier (ReGa, Coherent), operated at a repetition rate of 40 kHz, a pulse width of about 90–100 fs and an energy per pulse of 14 nJ. The absorption was measured by means of a super-continuum light beam obtained by focusing the output of the laser system on a sapphire crystal. The transmission of the super-continuum was detected by a spectrograph coupled to a photo-diode array of 76 diodes. Global analysis The main assumption of global analysis is that the time and wavelength properties are separable, which reflects the properties of spectra of being independent of time and the dynamics of spectra to be independent of the wavelength. Therefore the lifetimes are global parameters which are wavelength independent. For this analysis a sequential model was used and the assumption of this model is that each state (or component which is a function of the wavelength) evolves into the next one with an increasing lifetime (or a decreasing k rate).39–42 Notice that because of the sequential model, each component does not
This journal is © the Owner Societies 2014
View Article Online
Published on 27 October 2014. Downloaded by University of Cambridge on 18/11/2014 05:27:25.
PCCP
Paper
necessarily represent one real state but may be a mix of different states (e.g. in the LH2 the mix between B800 and B850 bands). A brief description of the mathematical model is as follows: According to linear system theory, when the impulse response of the system is an exponential decay, it has to be convolved with the instrument response function (IRF). Therefore a contribution of an exponentially decaying component can be described as: ð1 P Cm ðtÞ ¼ IRFðt; m; DÞekm ðttÞ dt 1
where km = 1/tm is the decay rate and IRF(t) is a gaussian function with parameters for its width (D) and position (m). In general, according to the superposition principle, the measured data C(t,l) can be expressed as a superposition of spectral properties el(l) of the components weighted by their concentration cl(t): Cðt; lÞ ¼
ncomponents X
cl ðtÞeðlÞ
l¼1
The case of a sequential model, where the lifetimes are increasing (the rates decreasing), gives rise to evolution associated spectra (EAS) for the time resolved fluorescence, where each concentration is a linear combination of exponential decays: Cðt; lÞ ¼
ncomponents X
cSl EASl ðlÞ;
l¼1
cSl ¼
l X
bjl cSl ðkl Þ
lifetime of the LH1–RC, which are respectively 8.9 and 134 ps. The shortening of the lifetimes and the appearance of a fast component in the LH2-only membranes is in good agreement with previous results on reconstituted membranes of Rps. acidophila and on mutants of Rb. sphaeroides.27 Furthermore in order to study the nature of the fast component appearing in the membrane samples, this work includes a study of the fluorescence kinetics as a function of the excitation density of photons. The linear relationship between excitation density of photons and fluorescence amplitudes suggests that the fast component in the LH2-only membranes arises from energy quenching instead of energy annihilation. A comparison between LH2 and LH1–RC ground state bleaching and stimulated emission of the wild type membranes shows the absence of the quenchers in the latter membranes, which makes the fast component of the LH2-only membranes non-physiological for the energy transfer efficiency.
Acknowledgements M. F. and R. v. G. were supported by the VU University Amsterdam, TOP grant (700.58.305) from the Foundation of Chemical Sciences part of NWO, by the advanced investigator grant (267333, PHOTPROT) from the European Research Council and by the EU FP7 project PAPETS (GA 323901). R. v. G. gratefully acknowledges his Academy Professor grant from the Royal Netherlands Academy of Arts and Sciences (KNAW). M.F. acknowledges Ivo H. M. van Stokkum for the global analysis section. K. D. acknowledges her support from the Laserlab-Europe and for her EMBO fellowship. The global analysis was performed using Glotaran software, available at www.glotaran.org.
j¼1
where the superscript S stands for sequential, respectively and the amplitudes bjl are: , l 1 l Y Y bjl ¼ km kn kj ; for j l: m¼1
n¼1; naj
The case of transient difference absorption spectra can be described with the same formalism, with evolution associated difference spectra (EADS) instead of EAS: Cðt; lÞ ¼
ncomponents X
cSl EADSl ðlÞ
l¼1
5 Conclusions The transient absorption and the time resolved fluorescence are measured in the near infrared for isolated LH2, LH2-only membranes and wild type membranes of Rsb. denitrificans. A mono-exponential decay with a lifetime of 670 ps is found for the isolated LH2, whereas a bi-exponential decay with lifetimes of 9.6 and 160 ps is found for the LH2-only membranes. From the wild type membranes it is possible to estimate the energy transfer time from LH2 to LH1–RC and the relaxation
This journal is © the Owner Societies 2014
Notes and references 1 F. Fassioli, A. Olaya-Castro, S. Scheuring, J. N. Sturgis and N. F. Johnson, Biophys. J., 2009, 97, 2464–2473. 2 S. Scheuring and J. N. Sturgis, Science, 2005, 309, 484–487. 3 R. W. Visschers, W. Crielaard, G. J. S. Fowler, C. N. Hunter and R. van Grondelle, Biochim. Biophys. Acta, Bioenerg., 1994, 1183, 483–490. 4 R. Van Grondelle and V. I. Novoderezhkin, Phys. Chem. Chem. Phys., 2006, 8, 793–807. 5 V. I. Novoderezhkin and R. van Grondelle, Phys. Chem. Chem. Phys., 2010, 12, 7352–7365. 6 S. J. Jamieson, P. Wang, P. Qian, J. Y. Kirkland, M. J. Conroy, C. N. Hunter and P. a. Bullough, EMBO J., 2002, 21, 3927–3935. 7 P. Qian, M. Z. Papiz, P. J. Jackson, A. a. Brindley, I. W. Ng, J. D. Olsen, M. J. Dickman, P. a. Bullough and C. N. Hunter, Biochemistry, 2013, 52, 7575–7585. 8 R. Monshouwer, M. Abrahamsson, F. Van Mourik and R. Van Grondelle, J. Phys. Chem. B, 1997, 101, 7241–7248. 9 J. Koepke, X. Hu, C. Muenke, K. Schulten and H. Michel, Structure, 1996, 4, 581–597. 10 G. McDermott, S. M. Prince, A. A. Freer, A. M. HawthornthwaiteLawless, M. Z. Papiz, R. J. Cogdell and N. W. Isaacs, Nature, 1995, 374, 517–521.
Phys. Chem. Chem. Phys., 2014, 16, 26059--26066 | 26065
View Article Online
Published on 27 October 2014. Downloaded by University of Cambridge on 18/11/2014 05:27:25.
Paper
11 J. N. Sturgis, V. Jirsakova, F. Reiss-Husson, R. J. Cogdell and B. Robert, Biochemistry, 1995, 34, 517–523. 12 K. McLuskey, S. M. Prince, R. J. Cogdell and N. W. Isaacs, Biochemistry, 2001, 40, 8783–8789. 13 K. Duquesne, C. Blanchard and J. N. Sturgis, Biochemistry, 2011, 50, 6723–6729. 14 S. Bahatyrova, R. N. Frese, C. A. Siebert, J. D. Olsen, K. O. Van Der Werf, R. Van Grondelle, R. A. Niederman, P. A. Bullough, C. Otto and C. N. Hunter, Nature, 2004, 430, 1058–1062. 15 C. Mascle-Allemand, K. Duquesne, R. Lebrun, S. Scheuring and J. N. Sturgis, Proc. Natl. Acad. Sci. U. S. A., 2010, 107, 5357–5362. 16 T. Pullerits, F. van Mourik, R. Monshouwer, R. W. Visschers and R. van Grondelle, J. Lumin., 1994, 58, 168–171. 17 H. Stiel, D. Leupold, K. Teuchner, F. Nowak, H. Scheer and R. J. Cogdell, Chem. Phys. Lett., 1997, 276, 62–69. 18 H. M. Visser, O. J. Somsen, F. van Mourik, S. Lin, I. H. van Stokkum and R. van Grondelle, Biophys. J., 1995, 69, 1083–1099. 19 M. Maiuri, J. J. Snellenburg, I. H. M. Van Stokkum, S. Pillai, K. Wongcarter, D. Gust, T. A. Moore, A. L. Moore, R. Van Grondelle, G. Cerullo and D. Polli, J. Phys. Chem. B, 2013, 117, 14183–14190. 20 R. Monshouwer, I. O. de Zarate, D. Zarate, F. Van Mourik and R. van Grondelle, Chem. Phys. Lett., 1995, 246, 341–346. ¨hler and 21 T. Pflock, M. Dezi, G. Venturoli, R. J. Cogdell, J. Ko S. Oellerich, Photosynth. Res., 2008, 95, 291–298. 22 M. A. Bopp, Y. Jia, L. Li, R. J. Cogdell and R. M. Hochstrasser, Proc. Natl. Acad. Sci. U. S. A., 1997, 94, 10630–10635. 23 X.-H. Chen, L. Zhang, Y.-X. Weng, L.-C. Du, M.-P. Ye, G.-Z. Yang, R. Fujii, F. S. Rondonuwu, Y. Koyama, Y.-S. Wu and J. P. Zhang, Biophys. J., 2005, 88, 4262–4273. 24 P. Sebban, G. Jolchine and I. Moya, Photochem. Photobiol., 1984, 39, 247–253. 25 S. Hess, M. Chachisvilis, K. Timpmann, M. R. Jones, ¨m, Proc. Natl. G. J. Fowler, C. N. Hunter and V. Sundstro Acad. Sci. U. S. A., 1995, 92, 12333–12337. 26 K. Timpmann, F. Geng, A. Freiberg and V. Sundstrm, Biochim. Biophys. Acta, 1993, 1183, 185–193.
26066 | Phys. Chem. Chem. Phys., 2014, 16, 26059--26066
PCCP
¨m, R. van Grondelle and 27 C. N. Hunter, H. Bergstro ¨m, Biochemistry, 1990, 29, 3203–3207. V. Sundstro 28 S. Scheuring, R. P. Gonçalves, V. Prima and J. N. Sturgis, J. Mol. Biol., 2006, 358, 83–96. ´vy, J. Seguin and 29 R. P. Gonçalves, J. Busselez, D. Le S. Scheuring, J. Struct. Biol., 2005, 149, 79–86. 30 A. Schubert, A. Stenstam, W. J. D. Beenken, J. L. Herek, ¨m, Biophys. J., 2004, R. Cogdell, T. Pullerits and V. Sundstro 86, 2363–2373. 31 M. Vos, R. van Grondelle, F. W. van der Kooij, D. van de Poll, J. Amesz and L. N. M. Duysens, Biochim. Biophys. Acta, Bioenerg., 1986, 850, 501–512. 32 R. van Grondelle, C. N. Hunter, J. G. C. Bakker and H. J. M. Kramer, Biochim. Biophys. Acta, 1983, 723, 30–36. 33 W. H. J. Westerhuis, M. Vos, R. van Grondelle, J. Amesz and R. A. Niederman, Biochim. Biophys. Acta, 1998, 1366, 317–329. 34 B. Driscoll, C. Lunceford, S. Lin, K. Woronowicz, R. A. Niederman and N. W. Woodbury, Phys. Chem. Chem. Phys., 2014, 16, 17133–17141. 35 K. Timpmann, M. Chenchiliyan, E. Jalviste, J. A. Timney, C. N. Hunter and A. Freiberg, Biochim. Biophys. Acta, 2014, 1837, 1835–1846. 36 V. Novoderezhkin, R. Monshouwer and R. van Grondelle, Biophys. J., 1999, 77, 666–681. 37 V. I. Novoderezhkin, R. Monshouwer and R. van Grondelle, J. Phys. Chem. B, 1999, 103, 10540–10548. 38 K. Duquesne and J. N. Sturgis, Methods Mol. Biol., 2010, 601, 205–217. 39 I. H. M. van Stokkum, D. S. Larsen and R. van Grondelle, Biochim. Biophys. Acta, 2004, 1657, 82–104. 40 I. H. M. van Stokkum, D. S. Larsen and R. van Grondelle, Biochim. Biophys. Acta, Bioenerg., 2004, 1658, 262. 41 I. H. M. van Stokkum, B. van Oort, F. van Mourik, B. Gobets and H. van Amerongen, Biophysical Techniques in Photosynthesis, Volume II, Series Advances in Photosynthesis and Respiration, 2008. 42 J. J. Snellenburg, P. L. Seger, R. Serger, K. M. Mullen and I. H. M. van Stokkum, J. Stat. Softw., 2012, 49, 1–22.
This journal is © the Owner Societies 2014
Published on 27 October 2014. Downloaded by University of Cambridge on 18/11/2014 05:27:25.
PAPER
Cite this: Phys. Chem. Chem. Phys., 2014, 16, 26059
View Journal | View Issue
Ultrafast excited state processes in Roseobacter denitrificans antennae: comparison of isolated complexes and native membranes† Marco Ferretti,*a Katia Duquesne,b James N. Sturgisc and Rienk van Grondelle*a Roseobacter (Rsb.) denitrificans is a marine aerobic anoxygenic photosynthetic purple bacterium with an unusually high-800 nm absorption band. Ultrafast excited state processes have been intensively studied in the past in order to understand why the energy transfer efficiency between photosynthetic antennae approaches unity and recently it has been proved that the organization of the antennae proteins within the membranes plays an important role. Thanks to the development of genetic manipulation and to the capability of Rsb. denitrificans to grow anaerobically as well, it is possible to construct several mutants in order to compare the ultrafast dynamics between isolated complexes and complexes embedded in membrane environments. Time resolved fluorescence and transient absorption have been applied to isolate LH2, genetically modified membranes with LH2-only and wild type membranes with both LH2
Received 8th July 2014, Accepted 27th October 2014
and LH1 antennae of Rsb. denitrificans, in order to understand the effect of the membrane environment
DOI: 10.1039/c4cp02986k
states of LH2 and LH1, and although there is shortening of the relaxation lifetime of the LH2-only membranes with respect to the isolated LH2, we find an energy transfer efficiency from LH2 to LH1 of
www.rsc.org/pccp
95%, which still approaches unity.
on the energy transfer efficiency. A global analysis is applied to calculate the lifetime of the excited
1 Introduction In natural photosynthesis the energy transfer between pigment–protein complexes is considered one of the most efficient
a
VU University, Amsterdam, The Netherlands. E-mail: [email protected] Aix-Marseille Universite´, Centrale Marseille, CNRS, ISM2 UMR 7313, 13397 Marseille, France c Aix-Marseille University, CNRS, LISM, Marseille, France † Electronic supplementary information (ESI) available: Membrane samples such as the Dpuf or the wild type membranes used in this experiment may in general be affected by sample turbidity, different from the isolated complexes. The sample turbidity scatters the light and too much scattering may affect the time resolved fluorescence dynamics when the measurement is performed using a streak camera. However the sample turbidity does not affect the dynamics measured via transient absorption, where the signal originates only from the section of the sample where there is a strong overlap between the pump and probe, which corresponds to the non-scattered part of the laser beam. The fact that the transient absorption results of global analysis reproduce the results of time resolved fluorescence is a proof that the turbidity of the sample does not affect the results shown in this paper. However to show that the turbidity does not affect the steady state absorption, an absorption spectrum starting at 220 nm is shown in the following figure. It is clear from the baseline that the effect of the scattering is minimum even in the UV region where this effect is maximum. The peak appearing at 260/280 nm in the membranes samples (Dpuf and wild type) is due to the presence of some residual ribosomes and it is not relevant either for photosynthesis or for the analysis applied in this paper, which is the focus of the near-IR dynamics. See DOI: 10.1039/c4cp02986k b
This journal is © the Owner Societies 2014
ultra-fast processes. For example in the purple photosynthetic bacteria the incident light energy is collected by the antennae, which are able to transfer the excitation energy to the reaction centers or to other antennae with an efficiency approaching unity. The organization of the pigments within the proteins and of the proteins within the membrane is fundamental for tuning the spectral and energy transfer properties of the complexes in order to ensure high efficiency.1–5 The light harvesting (LH) complexes of purple bacteria can be divided into two groups: the core light-harvesting complex (LH1–RC) which is composed of the LH1 antenna surrounding the reaction center (RC) and the peripheral antenna, LH2, which transfers the absorbed energy to LH1. The LH1 pigment–protein complex is typically formed by an apparently closed ring of 16 subunits,1,2,6,7 each subunit being composed of a bacteriochlorophyll (BChl) dimer non-covalently bound to the protein scaffold of an a–b-polypeptide8 pair and a carotenoid molecule. The LH2 peripheral antenna consists of a ring of 89 or 910 subunits, each containing 3 BChl and one or two carotenoids. The BChls form two rings: one of quasi-monomeric, weakly coupled, pigments absorbing near 800 nm (B800) and one of more strongly coupled dimeric pigments absorbing typically at 850 nm (B850), though the position of the absorption maximum varies between species and with growth conditions.11 A few LH2 complexes show a high-800 nm absorption, with the 850 nm band blue-shifted
Phys. Chem. Chem. Phys., 2014, 16, 26059--26066 | 26059
View Article Online
Published on 27 October 2014. Downloaded by University of Cambridge on 18/11/2014 05:27:25.
Paper
and super-imposed on the 800 nm band. Examples are the LH2s of Rhodopseudomonas (Rps.) palustris grown at very low light intensities12 and Roseobacter (Rbs.) denitrificans.13 However it is still debated which structural modification may cause this shift13 and whether there is any advantage for the light-harvesting efficiency of the LH2 antennae. Recently, due to the development of atomic force microscopy, the importance of the organization of the proteins within the membrane for the energy transfer functionality of purple bacteria has been shown,1,14,15 therefore it is extremely interesting to compare excitation transfer dynamics between isolated complexes and complexes embedded in different membrane environments, in order to understand the origins and the consequences of any observable differences. Because of the ultra-fast time scale of energy transfer (a few to several hundreds of picoseconds), ultra-fast spectroscopic techniques such as transient absorption and time resolved fluorescence are instrumental for studying the dynamics of the energy transfer processes. These two techniques have been intensively used to study these dynamics in photosynthetic antennae in bacteria and plants; however most of the past studies include comparison between the dynamics of isolated antennae vs. wild type membranes. Nowadays, thanks to the development of genetic manipulation, it is possible to construct mutants in several species7 in order to study the properties of the LH2 complexes within the membrane environment, the results of which should better reflect the in vivo conditions than those obtained for a detergent-isolated complex. In this work we apply transient absorption and time resolved fluorescence to isolated LH2, genetically modified membranes with only LH2 complexes (LH2-only membranes) and wild type membranes (with both LH2 and LH1–RC) of Rbs. denitrificans. In this way we can compare the dynamics of isolated complexes and LH2-only membranes in order to investigate the role of the protein organization within the membranes for the energy transfer efficiency. Global analysis is applied to transient absorption and time resolved fluorescence traces in order to calculate the energy transfer time from LH2 to LH1 and the intrinsic relaxation time of the LH2. The main difference between isolated and in membrane LH2 is shortening of the relaxation lifetime and the appearance of a fast decay component in the latter case. Therefore the study of the dynamics as a function of the photon excitation density is done in order to understand the nature of the fast decay component, which can arise from energy reorganization, energy annihilation or energy quenching.
2 Results Steady state absorption The steady state absorption spectra of isolated LH2, LH2-only and wild type membranes are shown in Fig. 1. The isolated LH2 (in blue) absorption spectrum shows one peak in the near infrared at 800 nm which is the convolution between the monomeric B800 and the oligomeric B850 bands, differently
26060 | Phys. Chem. Chem. Phys., 2014, 16, 26059--26066
PCCP
Fig. 1 Steady state absorption of isolated LH2 (blue), LH2-only membranes or Dpuf membranes (black) and wild type membranes (red) at room temperature.
from LH2 from other species. In fact the 850 band of Rsb. denitrificans is blue-shifted with respect to the one of Rps. acidophila and at room temperature is superimposed on the B800 band13 as it is shown in Fig. 1. The LH2-only membrane absorption (Dpuf mutant, in black) shows one major near-IR absorption peak at 800 nm, as in the isolated LH2; however there is also a minor peak near 760 nm. This is due to the presence of bacteriopheophytin a and bacteriochlorophyll a in the membranes that are not associated with photosynthetic complexes. The 760 nm peak shown in Fig. 1 is more pronounced than that shown in Duquesne et al.,8 indeed the size of this peak, which is always present to some degree in LH2-only membranes, varies somewhat between preparations. The presence of this peak could be explained by the presence of damaged complexes within the membranes, or more probably by accumulation of Bchl a in the absence of the LH1–RC to which it is normally bound. In the visible region the peaks at 520–550 nm and 410 nm are due to the presence of relatively large amounts of cytochrome in the membranes, probably also a consequence of the inability to synthesize the LH1–RC complex. The wild type membrane absorption (in red) shows two major peaks in the near infrared, one at 800 nm arising essentially from the LH2 absorption bands and one at 870 nm due to the LH1–RC absorption band.
Time resolved spectroscopy The isolated LH2, the LH2-only and the wild type membranes are excited at 790 nm, in the blue wing of the 800 nm absorption bands of the LH2, in order to study the energy transfer from LH2 to LH1–RC. The transient absorption signal is detected in the near-infrared region, with an instrument response function (IRF) of about 100 fs. In Fig. 2, the left hand frames show a series of
This journal is © the Owner Societies 2014
View Article Online
Published on 27 October 2014. Downloaded by University of Cambridge on 18/11/2014 05:27:25.
PCCP
Paper
Fig. 2 Transient absorption (left frame) and time resolved fluorescence (right frame) spectra for isolated LH2 (top), LH2-only membranes (center) and wild type membranes (bottom). The points correspond to the data and the lines to the result of the global analysis.
spectra, recorded at different time delays between the pump and probe. The displayed signal is measured at room temperature with an energy per pump pulse of 15 nJ (which corresponds to a photon density below 1013 cm2), this is well below the excitation density where annihilation processes are typically observed in isolated LH2 from other purple bacteria species16,17 The transient absorption spectrum of LH2-only membranes shows a negative peak at 818 nm, which arises from the convolution between ground state bleaching (GSB) and stimulated emission (SE) of the LH2 antenna. From these kinetics it is possible to measure the relaxation lifetime of the LH2 in membranes where no energy transfer to the LH1–RC is possible. In the bottom-left frame the transient absorption of the wild type membranes is shown. The main difference with respect to the LH2-only membranes is that after tens of picoseconds the GSB and the SE of the LH2 vanish while a negative peak appears at 890 nm, which is due to GSB and SE of LH1, and a positive peak at 850 nm, which corresponds to the excited state absorption (ESA) of the LH1 antenna.18 From these kinetics it is possible to measure the energy transfer time from LH2 to LH1–RC and, knowing the excitation lifetime of the isolated LH2 or of the LH2-only membranes, the quantum efficiency of this process can be evaluated.19 In this way we can compare the differences between isolated LH2 and LH2 in membranes, focusing in particular on the effects of the membrane environment on the energy transfer efficiency. The fluorescence kinetics are recorded using a streak camera, after excitation with sub-picosecond pulses with an energy per pulse of 14 nJ. The IRF is about 5–6 ps and is limited by the detector response. In Fig. 2 right frame, some relevant spectra are shown. From the fluorescence kinetics it is clear that the lifetime of the isolated LH2 is much longer than the lifetime of
This journal is © the Owner Societies 2014
LH2 in membranes, in agreement with the transient absorption results discussed above. In fact the former shows a signal after more than 1 ns, whereas in the latter the fluorescence is almost zero after 200 ps. The wild type membrane kinetics shows energy transfer from LH2 to LH1–RC, in fact right after time zero the LH2 is emitting and the LH1 fluorescence is almost zero, whereas already after only 30 ps the fluorescence signal is dominated by emission from LH1–RC. A global analysis was performed on both transient absorption and time resolved fluorescence kinetics in order to evaluate the relaxation lifetime of the LH2 and the energy transfer time from LH2 to LH1 (for an explanation of the global analysis model refer to the section Experimental). In Fig. 3 the evolution associated spectra (EAS) corresponding to time-resolved fluorescence (right frame) and the evolution associated difference spectra (EADS) corresponding to transient absorption (left frame) are shown. The top frame is related to isolated LH2 and the kinetics can be well described by a mono-exponential decay (only one component), in agreement with previous results on different species of purple bacteria.8,20,21 The mono-exponential decay supports the assumption of the absence of exciton annihilation in the isolated LH2. The corresponding fluorescence-lifetime is 670 5 ps, in agreement with the LH2 lifetime of Rps. acidophila17,22 and Rhodobacter (Rb.) sphaeroides,8,23,24 despite the difference in spectral characteristics. Notice that the error arises essentially from the IRF which is obtained as a result of global analysis. In the LH2-only membranes the experimental fluorescence decay can be best described by two components (bi-exponential decay), with time constants of 11 5 and 162 5 ps. The transient absorption (Fig. 3 central left frame) can also be described by two components with lifetimes of 9.6 0.1 and
Phys. Chem. Chem. Phys., 2014, 16, 26059--26066 | 26061
View Article Online
Published on 27 October 2014. Downloaded by University of Cambridge on 18/11/2014 05:27:25.
Paper
PCCP
Fig. 3 Result from global analysis: evolution associated spectra (EAS) for time resolved fluorescence and evolution associated difference spectra (EADS) for transient absorption. A model with one component is used for isolated LH2 (top-right) whereas a model with 2 components is used for LH2-only membranes (center) and wild type membranes (bottom). In the transient absorption of wild type membranes (bottom-left) the negative areas (GSB + SE) correspond to: A1 = 2333 cm1 mOD and A2 = 902 cm1 mOD for the EADS1 and A3 = 2618 cm1 mOD for the EADS2. There is a good agreement between transient absorption and time resolved fluorescence lifetimes. In the scheme (top-left) the lifetimes of LH2-only and wild type membranes are summarized. The quenchers in the LH2-only membranes are probably the BChl a which are accumulated in the absence of the LH1–RC to which they are normally bound or large aggregates of LH2.
160.0 0.1 ps, these are in close agreement with the lifetimes obtained from fluorescence measurements. Notice that although the IRF and therefore the time resolution of the transient absorption experiment is about 100 fs, we cannot resolve the B800 to ‘‘B850’’ energy transfer. This may be caused by the spectral overlap between the monomeric and excitonic Bchl a absorption bands which causes the excitation of both bands even at the shortest time delays (as is shown in Fig. 2 bottom-left frame, in the red spectrum), or by ultra-rapid equilibration. In fact because of the spectral overlap, most likely the B800 localized states are strongly mixed (or more mixed than in other purple bacteria without the high-800 nm absorption band) with the delocalized ‘‘B850’’ states, making the ultra-rapid relaxation possible. Although we cannot resolve the ultra-rapid relaxation, we can estimate that the relaxation time is smaller than 100 fs, which corresponds to the IRF of the transient absorption experiment. The wild type membrane signal is also well modeled by a two exponential component fit. In the fluorescence signal (Fig. 3 bottom right frame) we find a fast component of 10 4 ps and a slow component of 130 4 ps. The fast decay is a bit slower than the energy transfer time from LH2 to LH1–RC observed in wild type membranes of Rb. sphaeroides,25 whereas the lifetime of 130 ps is in agreement with the one of closed26 LH1–RC of
26062 | Phys. Chem. Chem. Phys., 2014, 16, 26059--26066
other species. We assume that the reaction centers are closed (unavailable for charge separation) because of the high repetition rate of the laser in both experiments (50 kHz for fluorescence and 40 kHz for transient absorption). The transient absorption lifetimes of 8.9 0.2 and 134 0.2 ps are in very good agreement with the fluorescence results. Notice that in the fastest EADS (evolution associated difference spectra) of transient absorption there is a negative shoulder at 808 nm, which shows the bleaching of the B800 band of LH2. However we can see that the bleaching and the stimulated emission of the ‘‘B850’’ band (around 818 nm) is present already from time zero and this confirms that due to the overlap between the two absorption bands (B800 and blue-shifted B850) the pump at 790 nm excites both bands. In order to investigate the origin of the appearance of a fast component in the LH2-only membranes, a study of the fluorescence lifetimes and EAS amplitudes, as a function of the photon excitation density, has been performed, which is shown in Fig. 4. We repeated the time-resolved fluorescence experiment with several different pulse energies (1, 5 and 10 nJ) for both the LH2-only and the wild type membranes. In Fig. 4 the amplitude of the fast and slow components as a function of the excitation density is shown. It is clear from the linear relationship that there is no annihilation occurring at these photon excitation densities.
This journal is © the Owner Societies 2014
View Article Online
Published on 27 October 2014. Downloaded by University of Cambridge on 18/11/2014 05:27:25.
PCCP
Fig. 4 Amplitude of the EAS of the time resolved fluorescence kinetics as a function of the excitation energy (proportional to the density of photons). Blue and green: fast and slow components of LH2-only membranes, red and black. The linear behaviour suggests the absence of energy annihilation.
3 Discussion The excited state lifetime of the isolated LH2 complex of Rsb. denitrificans is in good agreement with that observed for the LH2 of other species such as Rps. acidophila and Rb. sphaeroides.8,17,22–24 The LH2-only membrane kinetics are very different with respect to the isolated LH2 ones, with much shorter lifetimes as compared to the isolated complexes and exhibiting a bi-exponential decay instead of a mono-exponential decay. This was already reported for membrane-embedded LH2: in the case of reconstitution of the LH2 complexes of Rps. acidophila and Rb. sphaeroides into phospholipid vesicles,21 and in the case of an LH2-only mutant of Rb. sphaeroides.27 In the work of Hunter et al.27 the fast component was interpreted as an energy transfer from the main B850 band to a red-shifted sub-population band, now known as the inhomogeneously distributed lowest exciton component of LH2. More recently the bi-exponential decay has been interpreted as annihilation caused by LH2–LH2 energy transfer in the membranes due to clustering of the LH complexes.21 Support to the idea of clustering comes from AFM measurements on native membranes28 and membrane-reconstituted LH complexes.29 The appearance of a fast component has also been observed in chromatophores of other LH2 aggregates.30–33 However the energy dependence of the decay amplitudes in Fig. 4 shows a linear relationship between EAS amplitudes and the photon excitation density. This makes the annihilation interpretation rather unlikely for our samples. The bi-exponential decay is expected for LH1–RC membranes, in fact in this case we expect two components because of two sequential reactions: the LH2 to LH1 energy transfer and the LH1 to RC energy transfer. The case of LH2-only membranes is
This journal is © the Owner Societies 2014
Paper
different, in fact the presence of two components could be interpreted as a re-equilibration within the LH2 antennae, but in this case a red-shift of the slower component, with respect to the faster, is expected due to equilibration and as you can see in Fig. 3 the two EAS corresponding to LH2-only membranes are almost identical. Therefore we interpret the presence of two time constants as being due to sample heterogeneity caused by clustering. In fact some quencher, such as oxidized bacteriopheophytin or free BChl a which are accumulated in the absence of the LH1–RC to which they are normally bound, may quench part of the energy in the LH2-only membranes. Notice that the energy quenching is in agreement with the linear behavior of the EAS amplitudes in Fig. 4. In this picture the wild type membranes and the LH2-only membranes have very similar dynamics, with the fast lifetime which correspond to energy transfer from LH2 to LH1–RC in the wild type membranes and from LH2 to quenchers in the LH2-only membranes, where the quenchers are present because of the lack of LH1 (see scheme in Fig. 3 top left panel). The nature of the fast component is very important for calculating the quantum efficiency of the energy transfer from LH2 to LH1–RC. In fact we can generally define the quantum efficiency of the energy transfer process (F) as: tET1 F ¼ P 1 ti where tET is the lifetime of the energy transfer from LH2 to LH1–RC and ti are all the possible channels of decay (the different decay components) of the excited LH2. It is clear that the calculation of quantum efficiency upon laser excitation with the same excitation density of this experiment has to take into account both lifetimes (fast and slow) of the LH2-only membranes. This leads to a quantum efficiency of 50 1%. Notice that the quantum efficiency of the in vivo complexes depend on the nature of the fast component in the LH2-only membrane. In fact if the fast component is caused by energy annihilation, the sun light photon density (1.4 kW m2) may be below the annihilation threshold. Therefore in this case only the slow component has to be taken into account, leading to a quantum efficiency of 95 1%. In the case of energy quenching, the quenchers in the LH2-only membranes may be present only in the mutant because of the lack of LH1, or the quenchers may be related to pools of LH2 aggregates,34 the size of which depends on the growth light conditions34,35 of the membranes and the presence of large LH2 aggregates may be a consequence of the lack of LH1–RC. In both cases the fast lifetime of LH2-only membranes is not physiological and it has not to be taken into account and the quantum efficiency is 95 1%. Support for this idea comes from the comparison between the GSB plus SE of the LH1 and the ones of the LH2, in the wild type membrane (Fig. 3). In fact it is possible to evaluate the quantum efficiency (or at least a lower limit) also from the ratio between the two negative areas of fast and slow EADS of wild type membranes, which in this experiment correspond to about 80% (see Fig. 3). The main issue of this method is that the GSB and SE are convoluted with the ESA and the deconvolution of the negative
Phys. Chem. Chem. Phys., 2014, 16, 26059--26066 | 26063
View Article Online
Published on 27 October 2014. Downloaded by University of Cambridge on 18/11/2014 05:27:25.
Paper
area has not a trivial solution, but depends on the relative dipole moments.36 Therefore, considering that after deconvolution the GSB plus SE area increases with respect to the convoluted area,36,37 the 80 1% represent only a lower limit. Anyway this lower limit is higher than the quantum efficiency calculated taking into account both lifetimes of LH2-only membranes. Recently it has been proposed that depending on the growth light conditions,34 large aggregates of unconnected LH2 may be present, and because they are poorly connected to the LH1–RC, these aggregates may play a role in the photoprotection mechanisms. In fact the overall quantum efficiency of the energy transfer from LH2 to LH1–RC depends on the average size of the clusters of LH2. This is in agreement with the present work, where extremely large LH2 aggregates may be present in the Dpuf membranes due to the adaptation to the lack of LH1– RC. The results presented in this paper can help in the understanding of the role of large aggregates of LH2 in purple bacteria membranes, which is still an open question. Notice that in the work of Driscoll et al., in the power study of the time resolved fluorescence, the amplitude of the slower component is not in a linear relationship with the excitation power like in this work. They explain this behavior with energy annihilation, differently from our results. The presence of annihilation in the Driscoll et al. work may be related to singlet–triplet annihilation due to the very high repetition rate of their experiment. In any case the quantum efficiency calculated using only the slow lifetime of the LH2-only membranes (95 1%) results in a lower number than the one calculated using the lifetime of the isolated complex, which is 98 1% and represent always an upper limit of the real quantum efficiency of the energy transfer. Finally notice that the LH2 of Rsb. denitrificans represents an interesting complex to be studied by 2 Dimension electronic spectroscopy (2DES) because of the strong exciton mixing between B800 and B850 bands, and this article can help future studies for the interpretation of the 2DES spectra.
4 Experimental Samples The growth of wild type and mutant cells, the preparation of membranes and the purification of LH2 complexes are described in detail elsewhere.13,38 Briefly the Rsb. denitrificans strains were grown aerobically in Marine Broth 2216 (Difco) medium at 28 1C with agitation at 200 rpm. Cells were harvested in the late log phase. For the preparation of the intracytoplasmic membranes (wild type or only-LH2), cells were resuspended in 10 mM Tris pH 7.5, 5% sucrose, 0.6 mg mL1 lysozyme, 100 mg mL1 DNAase, 20 mg mL1 RNAase, complete antiprotease cocktail (Roche) and were broken by twice passages through a French pressure cell at 900 kPa. The resulting lysate is layered directly onto 15–70% sucrose density gradients and centrifuged 90 min in a SW27 swinging bucket rotor at 25 000 rpm for 4 hours. The colored band is collected, diluted with Tris 20 mM pH 7.5, and pelleted by centrifugation for 90 min at 40 000 rpm in a Ti45 rotor. The LH2-only and the
26064 | Phys. Chem. Chem. Phys., 2014, 16, 26059--26066
PCCP
wild type membranes are resuspended in buffer containing 20 mM Tris buffer, pH 7.5. For the preparation of LH2 complexes, purified membranes described above were solubilized with dodecyl maltoside (5% final concentration in the same buffer) for 2 h overnight. Unsolubilized material was removed by centrifugation for 30 min at 60 000 rpm, and the supernatant was loaded directly onto a Resource Q column, equilibrated with Tris 20 mM pH 7.5, 20 mM NaCl, and 0.05% dodecyl maltoside. LH2 is eluted with a linear gradient from 20 mM NaCl to 1.0 M NaCl. The fractions with the highest ratio of absorbance at 800 to 280 nm are pooled and concentrated using centrifugal concentrators (30 K cutoff). The concentrated fraction is loaded onto a superose 6 gel-filtration column and eluted with Tris 20 mM pH 7.5, 20 mM NaCl, and 0.05% dodecyl maltoside. The fractions with the highest ratio of absorbance at 800 to 280 nm are again pooled and concentrated using centrifugal concentrators (30 K cutoff). The isolated LH2 are in a buffer solution containing Tris 20 mM pH 7.5 and 0.05% of b-DM. The LH2-only and the wild type membranes are in a buffer solution containing Tris 20 mM pH 7.5. Spectroscopy For the time resolved fluorescence experiments the samples were excited at 795 nm using a pulsed Titanium:Sapphire (Ti:Sa) laser system, consisting of a mode-locked oscillator (Vitesse, Coherent) and a regenerative amplifier (ReGa, Coherent), operated at a repetition rate of 50 kHz, a pulse width of about 200 fs and an energy per pulse of 1, 5, 10 and 14 nJ. Only the fluorescence signal, polarized at the magic angle, was detected at 901 with respect to the excitation beam by a streak camera device. The instrument response function (IRF) was about 4–5 ps, which is completely limited by the detector response. For the transient absorption the samples were excited at 795 nm using a pulsed Ti:Sa laser system made of a modelocked oscillator (Mira, Coherent) and a regenerative amplifier (ReGa, Coherent), operated at a repetition rate of 40 kHz, a pulse width of about 90–100 fs and an energy per pulse of 14 nJ. The absorption was measured by means of a super-continuum light beam obtained by focusing the output of the laser system on a sapphire crystal. The transmission of the super-continuum was detected by a spectrograph coupled to a photo-diode array of 76 diodes. Global analysis The main assumption of global analysis is that the time and wavelength properties are separable, which reflects the properties of spectra of being independent of time and the dynamics of spectra to be independent of the wavelength. Therefore the lifetimes are global parameters which are wavelength independent. For this analysis a sequential model was used and the assumption of this model is that each state (or component which is a function of the wavelength) evolves into the next one with an increasing lifetime (or a decreasing k rate).39–42 Notice that because of the sequential model, each component does not
This journal is © the Owner Societies 2014
View Article Online
Published on 27 October 2014. Downloaded by University of Cambridge on 18/11/2014 05:27:25.
PCCP
Paper
necessarily represent one real state but may be a mix of different states (e.g. in the LH2 the mix between B800 and B850 bands). A brief description of the mathematical model is as follows: According to linear system theory, when the impulse response of the system is an exponential decay, it has to be convolved with the instrument response function (IRF). Therefore a contribution of an exponentially decaying component can be described as: ð1 P Cm ðtÞ ¼ IRFðt; m; DÞekm ðttÞ dt 1
where km = 1/tm is the decay rate and IRF(t) is a gaussian function with parameters for its width (D) and position (m). In general, according to the superposition principle, the measured data C(t,l) can be expressed as a superposition of spectral properties el(l) of the components weighted by their concentration cl(t): Cðt; lÞ ¼
ncomponents X
cl ðtÞeðlÞ
l¼1
The case of a sequential model, where the lifetimes are increasing (the rates decreasing), gives rise to evolution associated spectra (EAS) for the time resolved fluorescence, where each concentration is a linear combination of exponential decays: Cðt; lÞ ¼
ncomponents X
cSl EASl ðlÞ;
l¼1
cSl ¼
l X
bjl cSl ðkl Þ
lifetime of the LH1–RC, which are respectively 8.9 and 134 ps. The shortening of the lifetimes and the appearance of a fast component in the LH2-only membranes is in good agreement with previous results on reconstituted membranes of Rps. acidophila and on mutants of Rb. sphaeroides.27 Furthermore in order to study the nature of the fast component appearing in the membrane samples, this work includes a study of the fluorescence kinetics as a function of the excitation density of photons. The linear relationship between excitation density of photons and fluorescence amplitudes suggests that the fast component in the LH2-only membranes arises from energy quenching instead of energy annihilation. A comparison between LH2 and LH1–RC ground state bleaching and stimulated emission of the wild type membranes shows the absence of the quenchers in the latter membranes, which makes the fast component of the LH2-only membranes non-physiological for the energy transfer efficiency.
Acknowledgements M. F. and R. v. G. were supported by the VU University Amsterdam, TOP grant (700.58.305) from the Foundation of Chemical Sciences part of NWO, by the advanced investigator grant (267333, PHOTPROT) from the European Research Council and by the EU FP7 project PAPETS (GA 323901). R. v. G. gratefully acknowledges his Academy Professor grant from the Royal Netherlands Academy of Arts and Sciences (KNAW). M.F. acknowledges Ivo H. M. van Stokkum for the global analysis section. K. D. acknowledges her support from the Laserlab-Europe and for her EMBO fellowship. The global analysis was performed using Glotaran software, available at www.glotaran.org.
j¼1
where the superscript S stands for sequential, respectively and the amplitudes bjl are: , l 1 l Y Y bjl ¼ km kn kj ; for j l: m¼1
n¼1; naj
The case of transient difference absorption spectra can be described with the same formalism, with evolution associated difference spectra (EADS) instead of EAS: Cðt; lÞ ¼
ncomponents X
cSl EADSl ðlÞ
l¼1
5 Conclusions The transient absorption and the time resolved fluorescence are measured in the near infrared for isolated LH2, LH2-only membranes and wild type membranes of Rsb. denitrificans. A mono-exponential decay with a lifetime of 670 ps is found for the isolated LH2, whereas a bi-exponential decay with lifetimes of 9.6 and 160 ps is found for the LH2-only membranes. From the wild type membranes it is possible to estimate the energy transfer time from LH2 to LH1–RC and the relaxation
This journal is © the Owner Societies 2014
Notes and references 1 F. Fassioli, A. Olaya-Castro, S. Scheuring, J. N. Sturgis and N. F. Johnson, Biophys. J., 2009, 97, 2464–2473. 2 S. Scheuring and J. N. Sturgis, Science, 2005, 309, 484–487. 3 R. W. Visschers, W. Crielaard, G. J. S. Fowler, C. N. Hunter and R. van Grondelle, Biochim. Biophys. Acta, Bioenerg., 1994, 1183, 483–490. 4 R. Van Grondelle and V. I. Novoderezhkin, Phys. Chem. Chem. Phys., 2006, 8, 793–807. 5 V. I. Novoderezhkin and R. van Grondelle, Phys. Chem. Chem. Phys., 2010, 12, 7352–7365. 6 S. J. Jamieson, P. Wang, P. Qian, J. Y. Kirkland, M. J. Conroy, C. N. Hunter and P. a. Bullough, EMBO J., 2002, 21, 3927–3935. 7 P. Qian, M. Z. Papiz, P. J. Jackson, A. a. Brindley, I. W. Ng, J. D. Olsen, M. J. Dickman, P. a. Bullough and C. N. Hunter, Biochemistry, 2013, 52, 7575–7585. 8 R. Monshouwer, M. Abrahamsson, F. Van Mourik and R. Van Grondelle, J. Phys. Chem. B, 1997, 101, 7241–7248. 9 J. Koepke, X. Hu, C. Muenke, K. Schulten and H. Michel, Structure, 1996, 4, 581–597. 10 G. McDermott, S. M. Prince, A. A. Freer, A. M. HawthornthwaiteLawless, M. Z. Papiz, R. J. Cogdell and N. W. Isaacs, Nature, 1995, 374, 517–521.
Phys. Chem. Chem. Phys., 2014, 16, 26059--26066 | 26065
View Article Online
Published on 27 October 2014. Downloaded by University of Cambridge on 18/11/2014 05:27:25.
Paper
11 J. N. Sturgis, V. Jirsakova, F. Reiss-Husson, R. J. Cogdell and B. Robert, Biochemistry, 1995, 34, 517–523. 12 K. McLuskey, S. M. Prince, R. J. Cogdell and N. W. Isaacs, Biochemistry, 2001, 40, 8783–8789. 13 K. Duquesne, C. Blanchard and J. N. Sturgis, Biochemistry, 2011, 50, 6723–6729. 14 S. Bahatyrova, R. N. Frese, C. A. Siebert, J. D. Olsen, K. O. Van Der Werf, R. Van Grondelle, R. A. Niederman, P. A. Bullough, C. Otto and C. N. Hunter, Nature, 2004, 430, 1058–1062. 15 C. Mascle-Allemand, K. Duquesne, R. Lebrun, S. Scheuring and J. N. Sturgis, Proc. Natl. Acad. Sci. U. S. A., 2010, 107, 5357–5362. 16 T. Pullerits, F. van Mourik, R. Monshouwer, R. W. Visschers and R. van Grondelle, J. Lumin., 1994, 58, 168–171. 17 H. Stiel, D. Leupold, K. Teuchner, F. Nowak, H. Scheer and R. J. Cogdell, Chem. Phys. Lett., 1997, 276, 62–69. 18 H. M. Visser, O. J. Somsen, F. van Mourik, S. Lin, I. H. van Stokkum and R. van Grondelle, Biophys. J., 1995, 69, 1083–1099. 19 M. Maiuri, J. J. Snellenburg, I. H. M. Van Stokkum, S. Pillai, K. Wongcarter, D. Gust, T. A. Moore, A. L. Moore, R. Van Grondelle, G. Cerullo and D. Polli, J. Phys. Chem. B, 2013, 117, 14183–14190. 20 R. Monshouwer, I. O. de Zarate, D. Zarate, F. Van Mourik and R. van Grondelle, Chem. Phys. Lett., 1995, 246, 341–346. ¨hler and 21 T. Pflock, M. Dezi, G. Venturoli, R. J. Cogdell, J. Ko S. Oellerich, Photosynth. Res., 2008, 95, 291–298. 22 M. A. Bopp, Y. Jia, L. Li, R. J. Cogdell and R. M. Hochstrasser, Proc. Natl. Acad. Sci. U. S. A., 1997, 94, 10630–10635. 23 X.-H. Chen, L. Zhang, Y.-X. Weng, L.-C. Du, M.-P. Ye, G.-Z. Yang, R. Fujii, F. S. Rondonuwu, Y. Koyama, Y.-S. Wu and J. P. Zhang, Biophys. J., 2005, 88, 4262–4273. 24 P. Sebban, G. Jolchine and I. Moya, Photochem. Photobiol., 1984, 39, 247–253. 25 S. Hess, M. Chachisvilis, K. Timpmann, M. R. Jones, ¨m, Proc. Natl. G. J. Fowler, C. N. Hunter and V. Sundstro Acad. Sci. U. S. A., 1995, 92, 12333–12337. 26 K. Timpmann, F. Geng, A. Freiberg and V. Sundstrm, Biochim. Biophys. Acta, 1993, 1183, 185–193.
26066 | Phys. Chem. Chem. Phys., 2014, 16, 26059--26066
PCCP
¨m, R. van Grondelle and 27 C. N. Hunter, H. Bergstro ¨m, Biochemistry, 1990, 29, 3203–3207. V. Sundstro 28 S. Scheuring, R. P. Gonçalves, V. Prima and J. N. Sturgis, J. Mol. Biol., 2006, 358, 83–96. ´vy, J. Seguin and 29 R. P. Gonçalves, J. Busselez, D. Le S. Scheuring, J. Struct. Biol., 2005, 149, 79–86. 30 A. Schubert, A. Stenstam, W. J. D. Beenken, J. L. Herek, ¨m, Biophys. J., 2004, R. Cogdell, T. Pullerits and V. Sundstro 86, 2363–2373. 31 M. Vos, R. van Grondelle, F. W. van der Kooij, D. van de Poll, J. Amesz and L. N. M. Duysens, Biochim. Biophys. Acta, Bioenerg., 1986, 850, 501–512. 32 R. van Grondelle, C. N. Hunter, J. G. C. Bakker and H. J. M. Kramer, Biochim. Biophys. Acta, 1983, 723, 30–36. 33 W. H. J. Westerhuis, M. Vos, R. van Grondelle, J. Amesz and R. A. Niederman, Biochim. Biophys. Acta, 1998, 1366, 317–329. 34 B. Driscoll, C. Lunceford, S. Lin, K. Woronowicz, R. A. Niederman and N. W. Woodbury, Phys. Chem. Chem. Phys., 2014, 16, 17133–17141. 35 K. Timpmann, M. Chenchiliyan, E. Jalviste, J. A. Timney, C. N. Hunter and A. Freiberg, Biochim. Biophys. Acta, 2014, 1837, 1835–1846. 36 V. Novoderezhkin, R. Monshouwer and R. van Grondelle, Biophys. J., 1999, 77, 666–681. 37 V. I. Novoderezhkin, R. Monshouwer and R. van Grondelle, J. Phys. Chem. B, 1999, 103, 10540–10548. 38 K. Duquesne and J. N. Sturgis, Methods Mol. Biol., 2010, 601, 205–217. 39 I. H. M. van Stokkum, D. S. Larsen and R. van Grondelle, Biochim. Biophys. Acta, 2004, 1657, 82–104. 40 I. H. M. van Stokkum, D. S. Larsen and R. van Grondelle, Biochim. Biophys. Acta, Bioenerg., 2004, 1658, 262. 41 I. H. M. van Stokkum, B. van Oort, F. van Mourik, B. Gobets and H. van Amerongen, Biophysical Techniques in Photosynthesis, Volume II, Series Advances in Photosynthesis and Respiration, 2008. 42 J. J. Snellenburg, P. L. Seger, R. Serger, K. M. Mullen and I. H. M. van Stokkum, J. Stat. Softw., 2012, 49, 1–22.
This journal is © the Owner Societies 2014
