Studying conformational changes of proteins via single-molecule spectroscopy: Cryogenic temperatures versus room temperature

Studying conformational changes of proteins via single-molecule spectroscopy: Cryogenic temperatures versus room temperature

ARTICLE IN PRESS Studying conformational changes of proteins via single-molecule spectroscopy: Cryogenic temperatures versus room temperature € hlera...

NAN Sizes 0 Downloads 0 Views

ARTICLE IN PRESS

Studying conformational changes of proteins via single-molecule spectroscopy: Cryogenic temperatures versus room temperature € hlera,*, Richard J. Cogdellb € rgen Ko Ju a

Spectroscopy of Soft Matter, and Bayreuth Institute of Macromolecular Research (BIMF), University of Bayreuth, Bavarian Polymer Institute, Bayreuth, Germany b Institute of Molecular, Cell & Systems Biology, University of Glasgow, Scotland, United Kingdom *Corresponding author: e-mail address: [email protected]

Contents 1. Introduction 2. Light-harvesting complex 2 (LH2) from purple bacteria 3. Low temperatures: Spectral signatures of protein dynamics 4. Ambient temperatures: Emission intensity fluctuations 5. Conclusion Acknowledgments References

2 5 8 17 25 25 26

Abstract The photophysical properties of chromophores react very sensitively upon changes of their electrostatic environment. This allows conformational fluctuations of a protein to be monitored by optical spectroscopy via the fluctuations of the spectral signatures of chromophores that are embedded in the protein’s matrix. However, to be successful as an approach this requires that the structural fluctuations within the proteins would occur synchronously. This restriction can be overcome by studying the proteins on an individual basis. In this contribution, we illustrate this approach on the example of the peripheral light-harvesting complexes from photosynthetic purple bacteria that contain bacteriochlorophyll a as natural cofactors.

Advances in Botanical Research ISSN 0065-2296 https://doi.org/10.1016/bs.abr.2018.12.002

#

2019 Elsevier Ltd All rights reserved.

1

ARTICLE IN PRESS 2

J€ urgen K€ ohler and Richard J. Cogdell

List of abbreviations ABEL BChl CPA DOPC LH1 LH2 LHC II NPQ PS I PS II PVA RC TLS

anti-Brownian motion electrokinetic trap bacteriochlorophyll change-point analysis dioleoylphosphatidylcholine light-harvesting complex 1 light-harvesting complex 2 light-harvesting complex II non-photochemical quenching photosystem I photosystem II poly(vinyl alcohol) reaction center two-level system

1. Introduction Proteins are heteropolymers that are often required to be rather flexible in order to discharge their biological functions, such as transport of electrons and small molecules, catalysis of biochemical reactions or storage, and transport of energy to fuel metabolic reactions (Stryer, 1995). Although they usually have structures well enough ordered to allow their determination by methods such as X-ray crystallography, the structural fluctuations introduce a microscopic randomness giving rise to rich relaxation dynamics covering many orders of magnitude in time (Frauenfelder, Sligar, & Wolynes, 1991; Frauenfelder, Wolynes, & Austin, 1999; Fritsch, Friedrich, Parak, & Skinner, 1996). In principle the structural heterogeneity of a protein can be mapped out as spectral heterogeneity of a probe molecule that is embedded in the protein matrix. This approach exploits the fact that the electronic energy levels of a chromophore are very sensitive to the interactions with its local surroundings and that allows the conformational dynamics of the protein to be followed using the versatile tools of optical spectroscopy (Lakowicz, 2006). It would be, of course, best to perform these spectroscopic experiments on proteins under ambient conditions with them in their native environment. Unfortunately this is impossible for most of these proteins. Usually the details in the optical spectra from a macroscopic ensemble related to the structural flexibility of the proteins are washed out due to ensemble averaging. These circumstances highlight the need for techniques that are able, at least in part, to diminish these averaging effects. An obvious technique that

ARTICLE IN PRESS Metabolism, structure, and function of chlorophylls

3

avoids ensemble averaging and that can be applied at room temperature to proteins is single-molecule spectroscopy (Bockenhauer, F€ urstenberg, Yao, Kobilka, & Moerner, 2011; Dickson, Cubitt, Tsien, & Moerner, 1997; Lu, 1998; Yang et al., 2003). However, studying optical spectra under ambient conditions suffers from two major drawbacks. First, at room temperature the thermally-induced homogeneous line broadening of the optical transitions of the chromophore may exceed the expected spectral changes induced by the protein matrix by several orders of magnitude. Therefore, many room temperature single-molecule experiments focus on monitoring the fluctuations of the integrated intensity of the emission and/or changes of the fluorescence lifetime rather than on registering spectral fluctuations with very high spectral resolution (Kondo, Chen, & Schlau-Cohen, 2017). Second, due to photobleaching effects of the probe molecule the observation time rarely exceeds a few minutes even when oxygen scavengers are used. Yet, on the proside the protein can be studied under ambient conditions, either in solution or even in its native environment and can be treated with agents for verifying specific influences on reaction pathways (Bockenhauer et al., 2011; De Cremer et al., 2007; Kinosita, Adachi, & Itoh, 2004; Sako & Yanagida, 2003; Xie, Choi, Li, Lee, & Lia, 2008). At cryogenic temperatures, when nuclear motions are frozen out, the optical absorption bands become very narrow. This allows narrow spectral features to be resolved and changes of the protein matrix in the vicinity of the chromophore to be monitored with enhanced sensitivity. Moreover, under cryogenic conditions the samples are extremely photostable and can be studied for long observation times. The readers who consider spectroscopy of proteins at cryogenic temperatures as irrelevant should keep in mind that most of the current threedimensional protein structures, which lay at the heart of our understanding of protein function, have been obtained by X-ray crystallography under cryogenic conditions. In other words these are essentially low-temperature structures. Spectroscopic techniques that are suited to study proteins at low temperatures and that diminish inhomogeneous line-broadening effects either in part or fully can be conducted in the time domain as photon-echo or coherent spectroscopy (Brixner, Hildner, K€ ohler, Lambert, & W€ urthner, 2017; Chenu & Scholes, 2015; Cho, 2008; Ishizaki & Fleming, 2012) or in the frequency (i.e., spectral) domain including fluorescence line narrowing (siteselective spectroscopy), spectral hole burning, combinations thereof, and single-molecule spectroscopy (Berlin, Burin, Friedrich, & K€ ohler, 2006, 2007; Cogdell, Gall, & K€ ohler, 2006; Freiberg, R€atsep, Timpmann,

ARTICLE IN PRESS 4

J€ urgen K€ ohler and Richard J. Cogdell

Trinkunas, & Woodbury, 2003; Friedrich, 1995; Fritsch et al., 1996; Jankowiak, Reppert, Zazubovich, Pieper, & Reinot, 2011; Kondo et al., 2017; Zazubovich, 2014). The latter techniques allow spectral fluctuations of a chromophore inside a protein to be followed either in hole burning as a waiting time-dependent line broadening of the spectral hole, or in singlemolecule spectroscopy directly as a spectral trail ν(t), where ν denotes the optical transition frequency. Another positive feature of using cryogenic conditions is that the dynamics of the system are slowed down allowing intermediate states to be found and identified where they would be hidden in the fast dynamics at room temperature. In particular, a single reporter molecule that undergoes a temporal development between different states is at any time in a distinct, well-defined state and the whole sequence of steps can be studied. Yet, at the same time it is a disadvantage that it is impossible to deduce the time scales of dynamical processes under native conditions. Moreover, working at cryogenic temperatures requires the proteins to be immobilized in a non-natural environment. For obvious reasons, the optical techniques focus mainly on proteins with fluorophore as intrinsic cofactors, because tedious biochemical labeling procedures are avoided. Hence, a class of proteins that were predestined to be studied with optical spectroscopy is the pigment-protein complexes of photosynthesis. Typically, the chromophores in these proteins are the various types of carotenoids and (bacterio-)chlorophylls whose mutual arrangement is determined by the protein scaffold (Mirkovic et al., 2017). Single-molecule techniques for gaining information about the dynamics of such complexes have been extensively applied to the light-harvesting complexes of purple bacteria (Bopp, Sytnik, Howard, Cogdell, & Hochstrasser, 1999; Gall et al., 2015; Hofmann, Aartsma, Michel, & K€ ohler, 2003; Schlau-Cohen, Wang, Southall, Cogdell, & Moerner, 2013; Sch€ orner, Beyer, Southall, Cogdell, & K€ ohler, 2015a, 2015b; Tubasum et al., 2016), and to the antenna systems associated with plant photosystems I and II (PSI and PSII; Brecht, Hussels, Schlodder, & Karapetyan, 2012; Brecht, Radics, Nieder, & Bittl, 2009; Brecht, Radics, Nieder, Studier, & Bittl, 2008; Jelezko, Tietz, Gerken, Wrachtrup, & Bittl, 2000; Kr€ uger, Ilioaia, & van Grondelle, 2011; Kr€ uger, Ilioaia, Valkunas, & van Grondelle, 2011; Kr€ uger, Novoderezhkin, Ilioaia, & van Grondelle, 2010; Kr€ uger, Wientjes, Croce, & van Grondelle, 2011; Skandary, Konrad, Hussels, Meixner, & Brecht, 2015; Tietz et al., 2001; Schlau-Cohen et al., 2015), both at room temperature and under cryogenic conditions, recently excellently reviewed in Kondo et al. (2017).

ARTICLE IN PRESS Metabolism, structure, and function of chlorophylls

5

2. Light-harvesting complex 2 (LH2) from purple bacteria In the following we will illustrate the use of single-molecule techniques for studying the structural fluctuations of proteins using as an example the peripheral light-harvesting complexes from photosynthetic purple bacteria. Briefly, purple bacteria absorb sunlight by a network of antenna pigment proteins and transfer the excitation energy efficiently to the photochemical reaction center (RC), where a charge separation takes place providing the free energy for subsequent chemical reactions. It is known that most of these bacteria typically contain two types of antenna complexes, the core light-harvesting complex 1 (LH1), which is closely associated with the RC, and the more peripheral light-harvesting complex 2 (LH2). The basic building block of LH2 is a protein heterodimer (αβ), which binds three BChl a pigments and one carotenoid molecule (Fig. 1A, left and center). Depending on the bacterial species, most of these LH2 units oligomerize to produce octamers (Rhodospirillum (Rsp.) molischianum; K€ opke, Hu, Muenke, Schulten, & Michel, 1996) or nonamers (Rhodopseudomonas (Rps.) acidophila, Rhodobacter (Rba.) sphaeroides; McDermott et al., 1995; Walz, Jamieson, Bowers, Bullough, & Hunter, 1998) that comprise 24 or 27 BChL a molecules, respectively, that are arranged in two concentric rings (Fig. 1A and B, right). One ring consists of a group of nine (eight for Rps. molischianum) well-separated BChl a molecules absorbing light at about 800 nm, which are referred to as B800 BChl a molecules. The other ring consists of 18 (16 for Rps. molischianum) closely interacting BChl a molecules, which absorb at about 850 nm and that are called B850, accordingly (Fig. 1B). The different arrangement of the BChl a molecules within the two pigment pools results in significant differences in the mutual electronic interactions between the monomers, which is manifested in a distinctively different character of the electronically excited states in the B800 and the B850 manifold, respectively. The electronic energy levels of the BChl a molecules are summarized in the center part of Fig. 2. In the B800 ring the interpigment coupling is relatively weak, and in first approximation, the electronic excitations can be treated as localized on individual pigments. In the B850 ring, in contrast, the interpigment interactions are sufficiently strong to induce electronic couplings leading to delocalized electronic excitations, so-called excitons (van Amerongen, Valkunas, & van Grondelle, 2000).

ARTICLE IN PRESS 6

J€ urgen K€ ohler and Richard J. Cogdell

Fig. 1 (A) General constitution of the peripheral light-harvesting complex, here shown for the B800-B850 complex (light-harvesting complex 2; LH2) from Rps. acidophila. From left to right: One carotenoid (here rhodopin glucoside) and three bacteriochlorophyll a (BChl a) molecules are kept in place by a protein heterodimer referred to as protomer, which oligomerize (here as a 9-mer) into the ring-shaped LH2 complex. (B) UV-vis absorption spectrum of an ensemble of LH2 complexes. Transitions in the UV are associated with transitions of the BChl a molecules into higher excited electronic states. Features around 500 nm are associated with the carotenoids and those around 600 nm with the S0 ! S2 transitions (Qx transition-dipole moments, see part (A) bottom left) of the BChl a molecules. The two strong transitions in the near infrared are associated with the S0 ! S1 transitions (Qy transition-dipole moments, see part (A) bottom left) of the BChl a molecules referred to as B800 and B850, respectively. The spectral difference results from the arrangement of the BChl a molecules in two distinctively different pigment pools as shown next to the spectrum. The figure has been adapted from Kunz, R. (2013). Spectroscopic investigations of light-harvesting 2 complexes from Rps. acidophila. Dissertation.

Given the symmetry of the arrangement only a few of these exciton states carry oscillator strength and can couple to the light field (Alden et al., 1997; Hu, Ritz, Damjanovic, Autenrieth, & Schulten, 2002; L€ ohner, Cogdell, & K€ ohler, 2018; Matsushita et al., 2001; Sauer et al., 1996;

ARTICLE IN PRESS Metabolism, structure, and function of chlorophylls

7

Fig. 2 Summary of the electronic energy landscape of an LH2 complex (center) and the resulting excitation (colored) and emission (gray) spectra for ensembles (left) and individual complexes (right). For details see text. The figure has been adapted from Kunz, R. (2013). Spectroscopic investigations of light-harvesting 2 complexes from Rps. acidophila. Dissertation.

Wu & Small, 1998). However, these distinctive differences between the BChl a molecules in the two assemblies are washed out in an ensemble spectrum due to averaging effects, and the ensemble absorption spectrum consists of two broad featureless bands that appear rather similar in shape (Fig. 2, left). Yet, striking differences between the two absorption bands are observed when ensemble averaging is suppressed (Fig. 2, right). Then in the B800 band several relatively narrow spectral lines can be resolved, whereas the B850 band features a few broad absorption bands. Details are still a matter of debate and have been subject to several studies (Cogdell et al., 2006; Hofmann, Aartsma, & K€ ohler, 2004; Hu et al., 2002; Jang, Silbey, Kunz, Hofmann, & K€ ohler, 2011; Mirkovic et al., 2017).

ARTICLE IN PRESS 8

J€ urgen K€ ohler and Richard J. Cogdell

3. Low temperatures: Spectral signatures of protein dynamics The standard model for the description of the dynamics of materials like (bio-)polymers relies on the assumption that the potential energy hypersurface of a protein corresponds to a space of thousands of dimensions, resulting from the coordinates of the atoms of the protein. This hypersurface features a large number of minima, maxima, and saddle points separated from each other by a broad distribution of barriers, where each minimum in this landscape represents a different conformational state of the protein, i.e., a different arrangement of the atoms. This is commonly referred to as a rugged energy landscape (Frauenfelder, Nienhaus, & Young, 1994; Onuchic, Luthey-Schulten, & Wolynes, 1997; Wolynes, 2005). An extremely simplified two-dimensional sketch of such an energy landscape is shown at the top of Fig. 3. Owing to the relative weak intermolecular interaction between the B800 BChl a molecules in LH2 these are suited to act as reporters to monitor changes of the protein conformation on a local

Fig. 3 At the top, the potential energy landscape of a protein is simplified as a twodimensional surface. The red dot symbolizes the current conformational state of the protein, and a chromophore embedded in the protein is indicated underneath by its electronic states S0 and S1 and its vibrational states in the electronic ground state. Upon electronic excitation of the chromophore from S0 ! S1 the subsequent relaxation of the chromophore releases energy into the protein that might be used for inducing conformational changes, right. This corresponds to a (slight) structural change of the protein which in turn is reflected by a small change of the absorption energy of the embedded chromophore, right (see dashed horizontal line).

ARTICLE IN PRESS Metabolism, structure, and function of chlorophylls

9

scale. Such changes can be induced by optical excitation of the chromophore and subsequent decay of the excited state. Due to the low fluorescence quantum yield of LH2 of about 10%, a large fraction of the average absorbed energy is dissipated by radiationless decay resulting in the excitation of nuclear motions of the protein matrix (Fig. 3, left). Upon relaxation back to thermal equilibrium, there is a finite probability that the system ends up in a different conformational substate, which will be reflected by a change in transition energy of the chromophore (Fig. 3, right). The dissipated photon-energy, about 12,500 cm1 in the case of the B800 chromophores, exceeds by far the thermal energy kT at the temperature of 1.4°K at which the measurements are performed. Therefore the space of conformational substates that is probed by the induced structural fluctuations is not restricted to the part of the energy landscape that is accessible at thermal equilibrium. This approach entails a momentary excitation of nuclear motion and subsequent relaxation back to thermal equilibrium, providing meaningful information about the protein energy landscape. Working at cryogenic temperatures, the various protein substructures are separated by a static distribution of barriers, which effectively trap the biomolecule in distinct physical states. This shifts the timescale for structural fluctuations into a range that is experimentally accessible and offers the opportunity to directly study the detailed organization of the protein’s energy landscape (Hofmann et al., 2003; Schlichter, Friedrich, Herenyi, & Fidy, 2000). Clearly, at low temperatures the observed rates for the conformational changes do not reflect the dynamics of the protein in its native environment. Since structural changes are supposed to only occur in spatially localized regions the complex multidimensional energy landscape of a protein is even more simplified and effectively leads to the concept of double-well potentials, also referred to as two-level systems (TLS; Anderson, Halperin, & Varma, 1972; Phillips, 1972, 1987). The idea is that each of the two minima of a double-well potential represents a specific nuclear configuration, and that changes of this particular configuration, for example the formation or loss of a hydrogen bond, can be represented as a transition between the minima of the TLS. The conformation of the entire protein is then commonly approximated as a distribution of TLS, whose energies and transition rates are randomly distributed. Then the distinct conformation of the protein at time t is represented by the current occupation of the double-well potentials, here abbreviated as Γ(t). Since the current condition of Γ(t) determines the distinct electrostatic environment that is sensed by a chromophore embedded in the protein, conformational fluctuations, i.e., flips of the

ARTICLE IN PRESS 10

J€ urgen K€ ohler and Richard J. Cogdell

TLS, lead to a change of the energies of the electronic states of the chromophore and become observable as a spectral jump (Fig. 4A). Within this framework the conformational fluctuations of the protein are modeled as transitions within the manifold of TLS, and commonly for the theoretical description of the resulting spectral excursions two limiting cases are A

S1

S1 hν2 S0

hν1

S0

Γ(t2)

Γ(t1) B

C

S1

S1

S0

S0

Fig. 4 (A) Left: Distribution of TLS around a chromophore which is represented by the yellowish ellipse. The TLS differ from each other with respect to the depths and steepness of the potential wells, and the distance and orientation from the chromophore. Right: If some of the TLS have been flipped, here indicated by the blue dots, the interaction between the chromophore and its local environment changes giving rise to a shift of the electronic energy levels S0 and S1 which becomes observable as a spectral jump of the absorption frequency, see dotted line. Γ(t) refers to the distinct occupation of the TLS at time t. (B) and (C) Models for describing the interaction of chromophores embedded in a rugged energy landscape. (B) Random distribution of chromophore-TLS distances and (C) chromophore in a cavity model, i.e., there exists a minimum distance between the chromophore and the closest TLS.

ARTICLE IN PRESS Metabolism, structure, and function of chlorophylls

11

distinguished: (i) if the chromophore resides in a binding pocket that is small with respect to its size this leads to a random distribution of the chromophore—TLS distances (Fig. 4B). This corresponds to the standard model used to describe the dynamics of glasses and has been successfully applied to describe the temperature dependence of the specific heat of glasses and glass-like materials ( Jankowiak, Hayes, & Small, 1988; Phillips, 1987). For a dipole-dipole type interaction between the randomly distributed TLSs and the chromophore the model predicts a Lorentzian shape for the first cumulant (Barkai, Silbey, & Zumofen, 2000). In simple words, here the first cumulant corresponds to the distribution of the spectral excursions with respect to the average spectral position of the spectral feature under study. These predictions have been verified by analyzing the spectral diffusion of single molecules in an amorphous polymer (Barkai, Naumov, Vainer, Bauer, & Kador, 2003). (ii) In the other limit the chromophore resides in a cavity that is large with respect to its size, and there exists a minimum distance to the closest TLS interacting with the chromophore (Fig. 4C). Then the profile of the first cumulant is Gaussian irrespective of the type of interaction. This prediction reflects the validity of the central limit theorem, and the physical picture is that large spectral fluctuations are suppressed due to the lack of TLSs in the direct environment of the chromophore. The left-hand side of Fig. 5A displays a stack of 1500 consecutively recorded fluorescence-excitation spectra of the B800 band from an individual LH2 complex from Rba. sphaeroides at 1.4°K (Baier, Richter, Cogdell, Oellerich, & K€ ohler, 2007). The average of all spectra that contribute to the stack is shown at the bottom of the pattern. The single-molecule approach reveals shifts in the spectral position of the absorption and allows the spectral excursions of an individual spectral line to be mapped out as a spectral trail. For illustration the right-hand side of Fig. 5A shows four examples of individual spectra from the boxed region of the top part of Fig. 5A. Such spectral trails can be evaluated with respect to the change in spectral position between two successively recorded spectra, and four examples of such an analysis are presented in Fig. 5B. For a better comparison, the four distributions are displayed as first cumulants, i.e., the changes in spectral position are shown with respect to the spectral mean position of the item under study. The individual distributions are compatible with a sum of Gaussians with either one, two, or three contributions, and none of the obtained histograms could adequately be described with a Lorentzian. Given that for LH2 all the experimentally obtained distributions of the first cumulants were in very good agreement with Gaussian fits, we concluded that the LH2 protein

ARTICLE IN PRESS 12

J€ urgen K€ ohler and Richard J. Cogdell

Fig. 5 (A) Top: Two-dimensional representation of 1500 fluorescence-excitation spectra of the B800 band of an individual LH2 complex from Rb. sphaeroides measured at a temperature of 1.4 K. The horizontal axis corresponds to the laser-excitation energy, the vertical axis to the time and scan number, respectively, and the fluorescence intensity is given by the gray scale. Bottom: Average over all 1500 fluorescence-excitation spectra. (B) Examples for distributions of the change in spectral position between successively recorded spectra. The dashed lines correspond to Gaussian fits, the solid line corresponds to the sum of the fitted Gaussians. The origin of the energy scale displays the spectral mean position of the spectral feature under study. Top, left: The full widths half maximum (FWHM) of the fitted Gaussians are 3.5, 4.5, and 6.1 cm1, respectively. The separations of the Gaussians are 7.7 and 3.6 cm1 (all from left to right). Top, right: The FWHM of the (single) fitted Gaussian is 8.5 cm1. For illustration also a Lorentzian (dotted line) is fitted to the distribution. Bottom, left: The FWHMs of the fitted Gaussians are 8.8 and 7.4 cm1, respectively, the separation is 12.8 cm1. Bottom, right: The FWHMs of the fitted Gaussians are 7.8 and 7.4 cm1, respectively, the separation is 8.9 cm1. (C) Schematic sketch of the connection between the energy landscape (top) and the spectral excursions of a chromophore (bottom). For details see text.

cannot be described by the standard TLS model for glasses, and that the “chromophore in a cavity” limit appears more appropriate here (Baier, Richter, Cogdell, Oellerich, & K€ ohler, 2008). This suggests that the hierarchical organization of the environment of the chromophores in a protein

ARTICLE IN PRESS Metabolism, structure, and function of chlorophylls

13

differs substantially from the more randomly structured insertion sites in a glass or a synthetic polymer. The shape of the distributions presented in Fig. 5B suggests a minimum distance between the chromophores and the TLSs that induce the spectral dynamics, whereas the necessity of superimposing two or three Gaussians indicates the presence of at least two populations of TLS, i.e., those that induce spectral changes of in the order of 10 cm1 and those that induce spectral changes of about 1–3 cm1. The idea is that the center position of the Gaussians in Fig. 5B is determined by a single or at most a few TLS in close vicinity of the chromophore, whereas the widths of the Gaussians are determined by the TLSs that are further away, as illustrated in Fig. 5C. The top part symbolizes a chromophore by its electronic ground state S0 and its first electronically excited state S1, separated by an energy hν1. The local environment of the chromophore is shown as a red circle and accommodates a single TLS (red double-well potential). Upon a conformational change in close vicinity to the chromophore, indicated here as a flip of the close-by TLS, the change of the electrostatic environment of the chromophore affects the energies of its electronic states leading to a change of the spectral position of the electronic transition, here sketched as hν2. Additionally, the exact spectral position of the S0 ! S1 transition is modulated by the distribution of the occupation of the TLSs (green) that are further away from the chromophore. Hence, the intriguing question arises whether it is possible to obtain an estimate for the minimum distance between the chromophore and the TLSs from the data. For a change of the dipole moment of the TLS by, ΔμTLS ¼ ΔR  q where q denotes the electric charge, the change in energy between a chromophore and a TLS at distance R amounts in the dipole-dipole approximation to hΔν ¼

μ  ΔμTLS 1 μ  q ΔR 1 3¼ 4πε0 R 4πε0 R R2

Here μ refers to the dipole moment of the chromophore, which amounts for BChl a to 2 1029 cm corresponding to 6 D (Alden et al., 1997). For a quantitative estimate we assume q  e, and assume that the pressureinduced spectral shift of an absorption of the B800 chromophores in LH2 is 0.1 cm1/MPa (Zazubovich, Jankowiak, & Small, 2002). Using a value of 0.1 GPa1 for the compressibility of a protein (Zazubovich et al., 2002) this translates into relative distance changes ΔR/R  103 for spectral changes in the order of a few wavenumbers. Altogether this yields a minimum distance between the chromophore and the TLS of roughly 1 nm

ARTICLE IN PRESS 14

J€ urgen K€ ohler and Richard J. Cogdell

for the spectral changes that are in the order of a few wavenumbers and a distance of about 0.4 nm for those that are in the order of 10 cm1.Given the fact that the Qy electronic excitations of BChl a are mostly determined by the π-electron system in the bacteriochlorin plane it is reasonable to interpret these distances as the distance from this plane. Then TLSs at a distance of about 1 nm are very likely located outside the protein matrix, whereas TLSs at distances from the chromophore of about 0.4 nm are most probably located inside the protein, leading to the energetic splitting of the Gaussian functions in the bi- or trimodal distributions (Baier et al., 2008). This suggests that these two classes of TLSs reflect “non-slaved” processes from conformational fluctuations that occur within the protein and “slaved” processes induced by the solvent, as suggested in Fenimore, Frauenfelder, Mcmahon, and Parak (2002). Since the multimodal distributions occurred for only about half of the spectral features studied we conclude that these types of TLSs are rare, only on the order of at most 1–2 per protein. The results from hole-burning experiments on a small water soluble protein had already suggested that the standard TLS model for glasses is inappropriate for describing the dynamics of protein (Schlichter et al., 2000). However, these experiments were performed on heme proteins with an altered chromophore to allow for optical spectroscopy, whereas the experiments reported above made use of chromophores that are naturally embedded in a transmembrane protein. These results from two different experimental techniques on distinctively different types of proteins are a clear indication that proteins in general do not behave in a glass-like fashion. In order to link the spectral movements of the electronic transitions to conformational changes in the protein, it is interesting to compare the observed spectral dynamics and the information that is available about structure of the B800 binding pocket for three different species of purple bacteria (Cogdell et al., 2003; He, Sundstr€ om, & Pullerits, 2002; K€ opke et al., 1996; McDermott et al., 1995). In Fig. 6A–D, the distributions of the magnitude of the spectral excursions in the B800 band between two successive scans band are shown for Rb. sphaeroides, Rsp. molischianum, and Rps. acidophila from top to bottom. For these experiments the LH2 complexes were immobilized in a polymer. The data for Fig. 6D have been obtained from LH2 complexes from Rps. acidophila (same species as used for Fig. 6C) that have been reconstituted for control purposes into a dioleoylphosphatidylcholine (DOPC) bilayer to mimic the native membrane environment (Richter, Baier, Cogdell, K€ ohler, & Oellerich, 2007).

ARTICLE IN PRESS Metabolism, structure, and function of chlorophylls

15

Fig. 6 Distributions of the relative change in spectral position between two successively recorded fluorescence-excitation spectra for three species of purple N (A) Rb. sphaeroides (in PVA), (B) Rsp. molischianum (in poly(vinyl alcohol); PVA), (C) Rps. acidophila (in PVA); and (D) Rps. acidophila (in DOPC). For better comparison the scales of the distributions have been normalized to the total number of recorded spectra. (E) Crystallographic B factors given in color code for a spacefill representation of the B800 BChl a molecule in Rsp. molischianum (left) and Rps. acidophila (right). (F) Surface charges in the B800 binding pocket in the vicinity of the bacteriochlorin plane of the BChl a molecule for Rsp. molischianum (left) and Rps. acidophila (right).

ARTICLE IN PRESS 16

J€ urgen K€ ohler and Richard J. Cogdell

Apparently, the spectral dynamics of LH2 complexes embedded either in a polymer or a phospholipid bilayer are very similar. The spectral excursions could be categorized into a group of small spectral changes in the order of 10 cm1, a smaller group of intermediate spectral jumps in the order of some 10 cm1, and finally a third group of relatively rare yet very large spectral excursions of about 250 cm1 in magnitude. Owing to the single-molecule approach such rare events can be made visible, which would be hardly (if at all) observable by (sub) ensemble averaging techniques. Interestingly, the latter fluctuations were observed exclusively for the LH2 complexes from Rsp. molischianum. Another striking difference in the spectral excursions between the three species is that for Rsp. molischianum also the width of the distribution of the small spectral excursions is clearly larger than those for the other two species. As shown in Fig. 6E, the B factors (temperature factor) from Rsp. molischianum are markedly higher across the BChl a molecule in the B800 binding pocket when compared to those for the BChl a molecule in the B800 binding pocket for Rps. acidophila (K€ opke et al., 1996; Papiz, Prince, Howard, Cogdell, & Isaacs, 2003). The B factor describes the displacement of atoms from their mean positions in the structure in the crystal, and can be used as an indirect measure of the mobility of groups of atoms. In particular, very large B factors are observed for the phytol chain of BChl a in Rsp. molischianum suggesting that the protein in that part of the binding pocket allows the phytol chain to be mobile. Despite the nonpolar nature of the phytol chain this results in fluctuations of the local electric fields, which is usually taken into account by a multipole expansion and subsequent summation of the local contributions (Madjet, Abdurahman, & Renger, 2006). Restricting ourselves to the leading term of such an expansion, i.e., the dipole-dipole interaction, and taking into account the estimates made above for the distances of the TLS that are involved in the spectral jumps in the order of 1–10 cm1, i.e., 0.4–1 nm, it is likely that the increase in the minor spectral movements for Rsp. molischianum is associated with the high mobility of the phytol chain for that species. However, spectral changes in the order of 250 cm1, as observed for Rsp. molischianum, are so large that they probably reflect local conformational changes that affect the π-conjugation system of the bacteriochlorin macrocycle, e.g., by affecting the planarity of the ring, through a reorientation of side-groups, or through some rearrangement involving the central-Mg atom and its ligands (Germeroth, Lottspeich, Robert, & Michel, 1993; Gudowska-Nowak, Newton, & Fajer, 1990; He et al., 2002; K€ opke et al., 1996). This is

ARTICLE IN PRESS Metabolism, structure, and function of chlorophylls

17

corroborated when the surface charges of the residues in close proximity of the B800 BChl aπ-conjugated electron system are compared for Rsp. molischianum and Rps. acidophila (Fig. 6F; Petrey & Honig, 2003). In contrast to Rps. acidophila, where the protein environment in the vicinity of the bacteriochlorin ring is fairly neutral, the environment of this ring is negatively charged in the case of Rsp. molischianum giving rise to larger electrostatic interactions. The good agreement of the spectral diffusion statistics for Rps. acidophila and Rb. sphaeroides, i.e., in particular the absence of the very large spectral jumps in the order of 250 cm1, tempted us to conclude that the polarity of the B800 BChl a binding pocket is fairly similar for these two species and to propose a tentative model for the B800 BChl a binding pocket of Rb. sphaeroides (Baier et al., 2009).

4. Ambient temperatures: Emission intensity fluctuations For several reasons, the experimental situation at room temperature is entirely different from the situation at cryogenic temperatures. At room temperature the proteins can be studied under ambient conditions and are much closer to the natural situation. Another advantage is that the influence of agents added to the samples can be studied. However, under ambient atmosphere photochemical reactions in the excited state with oxygen lead to bleaching of the emitter and limit the experimental time to some 10 s or at most to a few minutes if oxygen scavengers are used, whereas bleaching does not play a role at low temperatures (owing to the lack of freely diffusing oxygen). Another issue concerns the spectral width of the absorption bands, which can be very narrow at low temperatures. This makes them ideally suited to act as a sensitive monitor for following minor structural fluctuations as spectral fluctuations (Rebane, 1994). At room temperature, in contrast, thermal broadening effects may cause a substantial overlap of the optical absorption bands, preventing the spectral trail of a distinct absorption line to be followed as a function of time. Therefore an alternative experimental approach that aims at registering the time-dependence of the total emitted fluorescence intensity, sometimes combined with a simultaneous measurement of the emission spectrum and/or the fluorescence lifetime, has been developed (Angeles Izquierdo et al., 2005; Cotlet et al., 2004; Cui, Beyler, Bischof, Wilson, & Bawendi, 2014; Gronheid et al., 2003; Lee et al., 2013; Wang & Moerner, 2013).

ARTICLE IN PRESS 18

J€ urgen K€ ohler and Richard J. Cogdell

In the past often these experiments have been conducted by integrating the number of emitted photons over a short, yet finite, time bin and defining a threshold for discriminating between the different intensity levels, for example between “ON” and “OFF.” For most of the materials studied, including fluorescent by molecules, semiconductor quantum dots, polymers, or proteins (Kr€ uger, Ilioaia, Valkunas et al., 2011; Kr€ uger, Ilioaia, & van Grondelle, 2011; Kr€ uger, Wientjes et al., 2011; Nirmal et al., 1996; Unterkofler, Pflock, Southall, Cogdell, & K€ ohler, 2011; Vanden Bout et al., 1997; Yeow, Melnikov, Bell, de Schryver, & Hofkens, 2006), the probability density distributions P(τ) obtained for finding ON- and OFF-periods of distinct duration, τ, follow an inverse power law P(τ)  tm with an exponential cut-off for the ON-states (Cichos, Vonborczyskowski, & Orrit, 2007). Typically the exponents range from 1 < m < 2 (Cichos et al., 2007; Frantsuzov, Kuno, Janko, & Marcus, 2008). However, binning in combination with thresholding can lead to severe artifacts (Amecke, Heber, & Cichos, 2014; Crouch et al., 2010; Lippitz, Kulzer, & Orrit, 2005; Terentyeva et al., 2012), because variations of the intensity on time scales shorter than the binning time might be interpreted as transitions to (non-existing!) states of intermediate intensity. As has been shown by simulations the exponents of the ON- and OFF-time probability distributions depend crucially on the choice of the threshold preventing reliable statistical data being extracted from these intensity traces (Amecke et al., 2014; Lippitz et al., 2005). Instead of integrating the emitted intensity for a finite dwell time, nowadays the arrival time of each individual photon is registered. This is referred to as time-tagged time-resolved (t3r) detection of single photons. Subsequently the arrival times of the photons are analyzed by a change-point analysis algorithm (CPA) as developed by Watkins (Beausang, Goldman, & Nelson, 2011; Watkins & Yang, 2004). In simple terms, this algorithm analyses the lengths of the time intervals between the arrival of two consecutive photons (Fig. 7A). If the durations of these intervals increase this is equivalent to a decrease of the number of photons emitted, i.e., a change point of the intensity (see Fig. 7B). This procedure avoids ambiguities due to experimental binning and empirical thresholding (Fig. 7C and D). The actual number of different intensity levels that are present in a photon time trace are then determined using a Bayesian information criterion together with expectation-maximization clustering. For illustration Fig. 7E compares the outcome of analyzing an excerpt from a time trace from a single LH2 complex with CPA and binning/ thresholding. More details about the mathematical background of the CPA approach are outlined in Beausang et al. (2011).

ARTICLE IN PRESS Metabolism, structure, and function of chlorophylls

19

Fig. 7 (A) Schematic sequence of photons arriving at times ti separated by Δti. (B) Accumulated number of detected photons as a function of time. The full line indicates how this would look like for a much larger number of photons representing a similar (Continued)

ARTICLE IN PRESS 20

J€ urgen K€ ohler and Richard J. Cogdell

As has been shown in previous work, the pigment-protein complexes from purple bacteria can be immobilized in a polymer film without significant perturbation of the protein (B€ ohm, Kunz, Southall, Cogdell, & K€ ohler, 2013; Richter et al., 2007). In order to avoid the exposure to oxygen the samples were mounted in a sample holder that was flushed with helium gas. This enabled us to follow the fluorescence intermittency of individual LH2 complexes from Rps. acidophila at room temperature for very long time periods. Together with the t3r detection of single photons and the subsequent change-point analysis this allowed for an unambiguous identification of various fluorescence-intensity levels and provided a huge data base for statistical analysis (Sch€ orner et al., 2015b). Examples of fluorescenceintensity traces from three different LH2 complexes embedded in a polymer are shown in Fig. 8A. The traces are displayed as time-binned traces with 100 ms binning time (red lines) that have been reconstructed from the unbinned raw data. Fig. 8A shows on the left-hand side the full dataset recorded for 1 h, and on the right-hand side an expanded view of a short part of the trajectory that is highlighted on the left-hand side (yellow bar). From top to bottom the CPA algorithm identifies changes of the fluorescence intensity between 2, 3, and 4 distinct levels, respectively. For better comparison of these data with those presented in the literature (Schlau-Cohen et al., 2013), the signal levels were referred to in the order of the signals strengths as A-OFF (2 level blinking), A-B-OFF (3 levels), and A-C-B-OFF (4 levels), respectively. For the complexes studied it was found that the observed average intensities within each intensity level reproduce, within experimental accuracy, across the different types of complexes featuring 2-, 3-, or 4-level blinking (Fig. 8B). For each individual complex a histogram of the residence times for each intensity level was obtained. Remarkably, within each blinking Fig. 7—Cont’d relative signal strength. Each change in the slope indicates a change point of the intensity (thick black dashed line). (C) Sketch of the (noiseless) average intensity as a function of time that corresponds to the “real” flow of photons. (D) Intensity levels that would be observed as a function of time for binned detection using the dwell times indicated by the blue lines. (E) Extract from a fluorescenceintensity time trace from a single LH2 complex. The red line refers to a time trace that has been reconstructed from the raw data for a binning time of 100 ms for better display. The blue line corresponds to the assignment of the intensities to the ON or OFF state using the time-binned trace and a threshold (see dotted line) for discrimination. The black lines correspond to the assignment of the intensities to the ON or OFF state applying the CPA to the raw unbinned data.

ARTICLE IN PRESS Metabolism, structure, and function of chlorophylls

21

Fig. 8 (A) Three representative fluorescence-intensity traces from different individual LH2 complexes. For displaying the traces, the red lines show time-binned traces that have been reconstructed from the unbinned data for a 100 ms dwell time. Left: Full dataset recorded for 1 h. Right: Expanded view of a short part of the trajectory, which (Continued)

ARTICLE IN PRESS 22

J€ urgen K€ ohler and Richard J. Cogdell

Table 1 Exponents of the distributions of the mean residence times in the various intensity levels. mA mC mB mOFF

2 levels

1.10  0.07

3 levels

1.06  0.08

4 levels

1.09  0.08

hmi

1.08

1.20  0.12 1.37  0.09

1.22  0.10

1.38  0.14

1.34  0.12

1.17  0.13

1.38

1.35

1.20

category, the distributions from the individual complexes fall all onto the same line. In Fig. 8C, all entries from all complexes with the same blinking characteristics are overlaid and displayed from top to bottom for the 2-level systems, the 3-level systems, and the 4-level systems. These data cover seven orders of magnitude for the probability densities and an unprecedented nine orders of magnitude for the residence times (Sch€ orner et al., 2015b). They can be described by an inverse power law P(τ)  τ -m with the exponents given in Table 1. For the 2-level blinking the residence times in A and OFF (Fig. 8C, top) correspond to the commonly used ON- and OFF-times in fluorescence blinking observed previously for single molecules. The residence time in A (¼ON) is distributed over all experimentally accessible time scales from 106 to 103 s, whereas the OFF-state distribution lacks the residence times shorter than about 100 ms. For the complexes that feature 3-level blinking the residence times in B are between 0.1 and 10 s (Fig. 8C, center). Finally, for 4-level blinking the residence times in C and B strongly overlap covering the range between 0.1 and 10 s (Fig. 8C, bottom). The cut-off for the OFF-time distributions toward time intervals shorter than about 100 ms is observed as well for complexes that feature 3- and 4-level blinking. Form the data (Table 1), it is apparent that the values of the exponents that are associated with the same intensity level agree very closely with each other across the range of the different blinking categories. This demonstrates the strength of the t3r detection of single photons in combination with the CPA for analyzing blinking statistics of data Fig. 8—Cont’d is highlighted on the left-hand side (yellow bar). From top to bottom the number of the different fluorescence-intensity levels that have been identified by the CPA algorithm are 2, 3, and 4. These levels (black lines and dots) are referred to as A, C, B, and OFF. (B) Distributions of the intensities for the LH2 complexes featuring 2, 3, and 4 signal levels (from top to bottom), respectively. (C) Overlay of the probability densities P(τ) of the residence times of all LH2 complexes in the states A (black), C (pink), B (blue), and OFF (red) for 2, 3, and 4 signal levels (from top to bottom), respectively.

ARTICLE IN PRESS 23

Metabolism, structure, and function of chlorophylls

with limited signal-to-noise ratio. In particular the exponents mA and mOFF obtained from the 2-level blinking systems on the one hand and from the 3- and 4-level blinking systems on the other hand would differ strongly from each other if the discrimination of the intensity levels A/C and OFF/B was equivocal. Surprisingly, the values of the exponents provided in Table 1 are significantly smaller than those found in most of the previous studies using binned time traces and empirical thresholding for discriminating the intensity levels. For illustration we determined the exponents for the 2-level blinking systems also for binned time traces that were reconstructed from the raw data for various binning times between 50 ms and 5 s. For all the binning times used we find again an inverse power law P(τ)  τm for the probability density distributions of the residence times τ in the intensity levels A (¼ ON) and OFF. However, these differ significantly in their exponents as a function of the dwell time, as summarized in Table 2. For the ON (OFF) times these exponents vary between 2.6 (2.3) for a dwell time of 50 ms and 1.2 (1.2) for a dwell time of 5 s, revealing the difficulties to extract reliable information from time-binned time traces. Prior to our work the CPA analysis has been applied successfully to investigate the static and dynamic heterogeneity of single LH2 complexes from Rps. acidophila by the Moerner group (Schlau-Cohen et al., 2013). They employed an anti-Brownian motion electrokinetic trap (ABEL) for immobilizing the complexes in solution and recorded simultaneously the fluorescence intensity, emission spectra, and the fluorescence lifetimes. This approach allowed the authors to identify different emissive states of LH2 that were associated with a photoactivated, reversible quenching pathway that most likely involves a conformational change of the protein. In particular it was found that transitions from A ! B in single LH2 complexes were light Table 2 Exponents of the inverse power law as a function of the binning time. Binning time (s) mA 5 mON mOFF

0.05

2.6  0.2

2.3  0.2

0.1

2.1  0.2

2.0  0.2

0.5

1.7  0.2

1.6  0.2

1.0

1.4  0.2

1.3  0.2

5.0

1.2  0.2

1.2  0.2

ARTICLE IN PRESS 24

J€ urgen K€ ohler and Richard J. Cogdell

induced, whereas the back transition was most probably thermally activated. These authors associated the level B with a photoactivated conformational change of the pigment-protein complex, and level C with a photobleached BChl a chromophore in the protein. The rare transitions from C ! B were interpreted as a conformational change of a LH2 with a photobleached BChl a molecule. Hence, if we adopt this interpretation and consider C as synonymous with B plus one photobleached BChl molecule, then this assumption is fully consistent with finding mB  mC for the exponents of the probability density functions, as well as with the observation of very similar residence times for the intensity levels B and C. Having the long time traces of 1 h duration at hand we searched for correlations between successive intensity jumps in the 3- and 4-level systems (Sch€ orner et al., 2015a). From this analysis, we determined the total probability p2(X) that a protein initially fluorescing in intensity level X reappears after two consecutive jumps in this level, i.e., X ! Y ! X for X, Y 2 {A, B, C, OFF}. For example, for the 3-level systems this probability corresponds top2(A) ¼ p(AjB) p1(B) + p(Aj OFF)p1(OFF) where p(AjB) (p(Aj OFF)) refers to the conditional probability that the system is going to switch from B ! A (OFF ! A) given that it switched from A ! B (A ! OFF) in the previous jump, and p1(B) and p1(OFF) refer to the probabilities that the first intensity jump occurred from A ! B, or A ! OFF, respectively. From the analysis we found p2(A)  79% for the 3-level systems and p2(A)  68% for the 4-level systems, exceeding significantly the 50% and 33%, respectively, that would be expected for random transitions between the intensity levels. In other words, we find a memory for the intensity fluctuations, which was strongest if intensity level A was involved. If we relate the measured fluorescence-intensity levels of the intrinsic reporter chromophores with specific conformational substates of the protein, then the observed memory indicates that the protein does not fluctuate randomly between the conformational substates. Also in earlier single-molecule work on the enzyme cholesterol oxidase that catalyzes the oxidation of cholesterol by oxygen, a molecular memory effect was found which was associated with physiological regulation (Lu, 1998). If evolution enabled optimized control of the conformational fluctuations of proteins then such memory effects would be an important precondition, because without memory the randomness makes the possibility of evolving control over conformational states much more unlikely. For example, under strong illumination conditions in oxygenic photosynthetic light harvesting the absorbed energy cannot be processed by the reaction centers leading to highly reactive byproducts that may destroy the light-harvesting machinery. Under these conditions, the plant light-harvesting complexes (LHC) are

ARTICLE IN PRESS Metabolism, structure, and function of chlorophylls

25

converted into a state that dumps the absorbed energy into heat, a process referred to as non-photochemical quenching (NPQ) (Horton, Ruban, & Walters, 1996; M€ uller, Li, & Niyogi, 2001). In recent studies on single LHC II complexes (Kr€ uger, Ilioaia, Valkunas et al., 2011; Kr€ uger, Ilioaia, & van Grondelle, 2011; Kr€ uger, Wientjes et al., 2011; SchlauCohen et al., 2015; Valkunas, Chmeliov, Kr€ uger, Ilioaia, & van Grondelle, 2012), it has been found that both the switching rate between the light-harvesting and the quenched state, as well as the dwell times can be controlled by the environmental conditions, like pH and irradiation intensity. Accordingly, this suggests that nature has learned how to exploit specific protein motions for efficient switching between these two modes. Although purple bacteria perform an oxygenic photosynthesis it is tempting to speculate that conformational memory may well have allowed this control over the efficiency of the light-harvesting system to have evolved in plants.

5. Conclusion We have illustrated how in particular single-molecule spectroscopy can be exploited to study protein dynamics. Experiments can be performed both at room temperature or cryogenic temperatures and provide useful yet complementary information. Under ambient conditions the systems can be studied closer to their natural situation and the influence of for example enzymes or other agents can be investigated and the respective reaction rates can be determined. In contrast, at low temperatures the high sensitivity of the photophysical properties of the electronically excited states of intrinsic chromophores on their local environment can be exploited to map out subtle conformational changes in the vicinity of the binding pockets of the cofactors. As we have pointed out, by taking advantage of the optical excitation energy and its subsequent relaxation, conformational changes can be induced that are not necessarily restricted to the motions of the atoms under thermal equilibrium. By that, also at low temperatures a large part of the space of conformational substates is made accessible for optical spectroscopy.

Acknowledgments We thank our coworkers and collaboration partners who have contributed to this work and whose names are listed in the references. Financial support by the Deutsche Forschungsgemeinschaft (GRK1640, Ko 1359/27-1), the State of Bavaria within the initiatives “Solar Technologies go Hybrid,” and the Elite Study Program “Biological Physics” within the Elite Network of Bavaria (ENB), as well as the BBSRC is gratefully acknowledged.

ARTICLE IN PRESS 26

J€ urgen K€ ohler and Richard J. Cogdell

References Alden, R. G., Johnson, E., Nagarajan, V., Parson, W. W., Law, C. J., & Cogdell, R. J. (1997). Calculations of spectroscopic properties of the LH2 bacteriochlorophyll-protein antenna complex from Rhodopseudomonas acidophila. The Journal of Physical Chemistry B, 101, 4667–4680. Amecke, N., Heber, A., & Cichos, F. (2014). Distortion of power law blinking with binning and thresholding. The Journal of Chemical Physics, 140, 114306 [1-8]. Anderson, P. W., Halperin, B. I., & Varma, C. M. (1972). Anomalous low-temperature thermal properties of glasses and spin glasses. Philosophical Magazine, 25, 1–9. Angeles Izquierdo, M., Bell, T. D. M., Habuchi, S., Fron, E., Pilot, R., Vosch, T., et al. (2005). Switching of the fluorescence emission of single molecules between the locally excited and charge transfer states. Chemical Physics Letters, 401, 503–508. Baier, J., Gabrielsen, M., Oellerich, S., Michel, H., van Heel, M., Cogdell, R. J., et al. (2009). Spectral diffusion and electron-phonon coupling of the B800 BChl a molecules in LH2 complexes from three different species of purple bacteria. Biophysical Journal, 97, 2604–2612. Baier, J., Richter, M. F., Cogdell, R. J., Oellerich, S., & K€ ohler, J. (2007). Do proteins at low temperature behave as glasses? A single-molecule study. The Journal of Physical Chemistry B, 111, 1135–1138. Baier, J., Richter, M. F., Cogdell, R. J., Oellerich, S., & K€ ohler, J. (2008). Determination of the spectral diffusion kernel of a protein by single molecule spectroscopy. Physical Review Letters, 100, 08108. 1–4. Barkai, E., Naumov, A. V., Vainer, Y. G., Bauer, M., & Kador, L. (2003). Levy statistics for random single-molecule line shapes in a glass. Physical Review Letters, 91, 5502. Barkai, E., Silbey, R., & Zumofen, G. (2000). Levy distribution of single molecule line shape cumulants in glasses. Physical Review Letters, 84, 5339–5342. Beausang, J. F., Goldman, Y. E., & Nelson, P. C. (2011). Changepoint analysis for singlemolecule polarized total internal reflection fluorescence microscopy experiments. In M. L. Johnson & L. Brand (Eds.), Methods in enzymology: Computer methods, part C: Academic Press. Berlin, Y. A., Burin, A. L., Friedrich, J., & K€ ohler, J. (2006). Spectroscopy of proteins at low temperature. Part I: Experiments with molecular ensembles. Physics of Life Reviews, 3, 262–292. Berlin, Y. A., Burin, A. L., Friedrich, J., & K€ ohler, J. (2007). Low temperature spectroscopy of proteins. Part II: Experiments with single protein complexes. Physics of Life Reviews, 4, 64–89. Bockenhauer, S., F€ urstenberg, A., Yao, X. J., Kobilka, B. K., & Moerner, W. E. (2011). Conformational dynamics of single G protein-coupled receptors in solution. The Journal of Physical Chemistry B, 115, 13328–13338. B€ ohm, P. S., Kunz, R., Southall, J., Cogdell, R. J., & K€ ohler, J. (2013). Does the reconstitution of RC-LH1 complexes from Rhodopseudomonas acidophila strain 10050 into a phospholipid bilayer yield the optimum environment for optical spectroscopy? The Journal of Physical Chemistry B, 117, 15004–15013. Bopp, M. A., Sytnik, A., Howard, T. D., Cogdell, R. J., & Hochstrasser, R. M. (1999). The dynamics of structural deformations of immobilized single light-harvesting complexes. Proceedings of the National Academy of Sciences of the United States of America, 96, 11271–11276. Brecht, M., Hussels, M., Schlodder, E., & Karapetyan, N. V. (2012). Red antenna states of photosystem I trimers from arthrospira platensis revealed by single-molecule spectroscopy. Biochimica et Biophysica Acta, 1817, 445–452.

ARTICLE IN PRESS Metabolism, structure, and function of chlorophylls

27

Brecht, M., Radics, V., Nieder, J. B., & Bittl, R. (2009). Protein dynamics-induced variation of excitation energy transfer pathways. Proceedings of the National Academy of Sciences of the United States of America, 106, 11857–11861. Brecht, M., Radics, V., Nieder, J. B., Studier, H., & Bittl, R. (2008). Red antenna states of photosystem I from synechocystis PCC 6803. Biochemistry, 47, 5536–5543. Brixner, T., Hildner, R., K€ ohler, J., Lambert, C., & W€ urthner, F. (2017). Exciton transport in molecular aggregates—From natural antennas to synthetic chromophore systems. Advanced Energy Materials, 5, 1700236. Chenu, A., & Scholes, G. D. (2015). Coherence in energy transfer and photosynthesis. Annual Review of Physical Chemistry, 66, 69–96. Cho, M. (2008). Coherent two-dimensional optical spectroscopy: Chemical reviews. Chemical Reviews, 108, 1331–1418. Cichos, F., Vonborczyskowski, C., & Orrit, M. (2007). Power-law intermittency of single emitters. Current Opinion in Colloid & Interface Science, 12, 272–284. Cogdell, R. J., Gall, A., & K€ ohler, J. (2006). The architecture and function of the lightharvesting apparatus of purple bacteria: From single molecules to in vivo membranes. Quarterly Reviews of Biophysics, 39, 227–324. Cogdell, R. J., Isaacs, N. W., Freer, A. A., Howard, T. D., Gardiner, A. T., Prince, S. M., et al. (2003). The structural basis of light-harvesting in purple bacteria. FEBS Letters, 555, 35–39. Cotlet, M., Masuo, S., Luo, G., Hofkens, J., van der Auweraer, M., Verhoeven, J., et al. (2004). Probing conformational dynamics in single donor-acceptor synthetic molecules by means of photoinduced reversible electron transfer. Proceedings of the National Academy of Sciences of the United States of America, 101, 14343–14348. Crouch, C. H., Sauter, O., Wu, X., Purcell, R., Querner, C., Drndic, M., et al. (2010). Facts and artifacts in the blinking statistics of semiconductor nanocrystals. Nano Letters, 10, 1692–1698. Cui, J., Beyler, A. P., Bischof, T. S., Wilson, M. W. B., & Bawendi, M. G. (2014). Deconstructing the photon stream from single nanocrystals: From binning to correlation. Chemical Society Reviews, 43, 1287–1310. De Cremer, G., Roeffaers, M. B. J., Baruah, M., Sliwa, M., Sels, B. F., Hofkens, J., et al. (2007). Dynamic disorder and stepwise deactivation in a chymotrypsin catalyzed hydrolysis reaction. Journal of the American Chemical Society, 129, 15458–15459. Dickson, R. M., Cubitt, A. B., Tsien, R. Y., & Moerner, W. E. (1997). On/off blinking and switching behaviour of single molecules of green fluorescent protein. Nature, 388, 355–358. Fenimore, P. W., Frauenfelder, H., Mcmahon, B. H., & Parak, F. G. (2002). Slaving: Solvent fluctuations dominate protein dynamics and functions. Proceedings of the National Academy of Sciences of the United States of America, 99, 16047–16051. Frantsuzov, P. A., Kuno, M., Janko, B., & Marcus, R. A. (2008). Universal emission intermittency in quantum dots, nanorods and nanowires. Nature Physics, 4, 519–522. Frauenfelder, H., Nienhaus, G. U., & Young, R. D. (1994). Relaxation and disorder in proteins. In R. Richert (Ed.), Disorder effects on relaxational processes (pp. 591–614): Springer Verlag. Frauenfelder, H., Sligar, S., & Wolynes, P. (1991). The energy landscapes and motions of proteins. Science, 254, 1598–1603. Frauenfelder, H., Wolynes, P. G., & Austin, R. H. (1999). Biological physics. Reviews of Modern Physics, 71, S419–S430. Freiberg, A., R€atsep, M., Timpmann, K., Trinkunas, G., & Woodbury, N. W. (2003). Selftrapped excitons in LH2 antenna complexes between 5 K and ambient temperature. The Journal of Physical Chemistry B, 107, 11510–11519.

ARTICLE IN PRESS 28

J€ urgen K€ ohler and Richard J. Cogdell

Friedrich, J. (1995). Hole burning spectroscopy and physics of proteins. Methods in Enzymology, 246, 226–259. Fritsch, K., Friedrich, J., Parak, F., & Skinner, J. L. (1996). Spectral diffusion and the energy landscape of a protein. Proceedings of the National Academy of Sciences of the United States of America, 93, 15141–15145. Gall, A., Ilioaia, C., Kr€ uger, T. P., Novoderezhkin, V. I., Robert, B., & van Grondelle, R. (2015). Conformational switching in a light-harvesting protein as followed by singlemolecule spectroscopy. Biophysical Journal, 108, 2713–2720. Germeroth, L., Lottspeich, F., Robert, B., & Michel, H. (1993). Unexpected similarities of the B800-850 light-harvesting complex from Rhodospirillum molischianum to the B870 light-harvesting complexes from other purple photosynthetic bacteria. Biochemistry, 32, 5615–5621. Gronheid, R., Stefan, A., Cotlet, M., Hofkens, J., Qu, J., Mullen, K., et al. (2003). Reversible intramolecular electron transfer at the single-molecule level. Angewandte Chemie International Edition in English, 42, 4209–4214. Gudowska-Nowak, E., Newton, M. D., & Fajer, J. (1990). Conformational and environmental effects on bacteriochlorophyll optical spectra: Correlations of calculated spectra with structural results. The Journal of Physical Chemistry, 94, 5795–5801. He, Z., Sundstr€ om, V., & Pullerits, T. (2002). Influence of the protein binding site on the excited states of bacteriochlorophyll: DFT calculations of B800 in LH2. The Journal of Physical Chemistry B, 106, 11606–11612. Hofmann, C., Aartsma, T. J., & K€ ohler, J. (2004). Energetic disorder and the B850-exciton states of individual light-harvesting 2 complexes from Rhodopseudomonas acidophila. Chemical Physics Letters, 395, 373–378. Hofmann, C., Aartsma, T. J., Michel, H., & K€ ohler, J. (2003). Direct observation of tiers in the energy landscape of a chromoprotein: A single-molecule study. Proceedings of the National Academy of Sciences of the United States of America, 100, 15534–15538. Horton, P., Ruban, A. V., & Walters, R. G. (1996). Regulation of light harvesting in green plants. Annual Review of Plant Physiology and Plant Molecular Biology, 47, 655–684. Hu, X., Ritz, T., Damjanovic, A., Autenrieth, F., & Schulten, K. (2002). Photosynthetic apparatus of purple bacteria. Quarterly Reviews of Biophysics, 35, 1–62. Ishizaki, A., & Fleming, G. R. (2012). Quantum coherence in photosynthetic light harvesting. Annual Review of Condensed Matter Physics, 3, 333–361. Jang, S., Silbey, R. J., Kunz, R., Hofmann, C., & K€ ohler, J. (2011). Is there elliptic distortion in the light harvesting complex 2 of purple bacteria? The Journal of Physical Chemistry B, 115, 12947–12953. Jankowiak, R., Hayes, J. M., & Small, G. J. (1988). Low-temperature specific heat of glasses: Temperature and time dependence. Physical Review B, 38, 2084. Jankowiak, R., Reppert, M., Zazubovich, V., Pieper, J., & Reinot, T. (2011). Site selective and single complex laser-based spectroscopies: A window on excited state electronic structure, excitation energy transfer, and electron-phonon coupling of selected photosynthetic complexes. Chemical Reviews, 111, 4546–4598. Jelezko, F., Tietz, C., Gerken, U., Wrachtrup, J., & Bittl, R. (2000). Single-molecule spectroscopy on photosystem I pigment  protein complexes. The Journal of Physical Chemistry B, 104, 8093–8096. Kinosita, K., Jr., Adachi, K., & Itoh, H. (2004). Rotation of F1-ATPase: How an ATPdriven molecular machine may work. Annual Review of Biophysics and Biomolecular Structure, 33, 245–268. Kondo, T., Chen, W. J., & Schlau-Cohen, G. S. (2017). Single-molecule fluorescence spectroscopy of photosynthetic systems. Chemical Reviews, 117, 860–898. K€ opke, J., Hu, X., Muenke, C., Schulten, K., & Michel, H. (1996). The crystal structure of the light-harvesting complex II (B800–850) from Rhodospirillum molischianum. Structure, 4, 581–597.

ARTICLE IN PRESS Metabolism, structure, and function of chlorophylls

29

Kr€ uger, T. P. J., Ilioaia, C., Valkunas, L., & van Grondelle, R. (2011). Fluorescence intermittency from the main plant light-harvesting complex: Sensitivity to the local environment. The Journal of Physical Chemistry B, 115, 5083–5095. Kr€ uger, T. P. J., Ilioaia, C., & van Grondelle, R. (2011). Fluorescence intermittency from the main plant light-harvesting complex: Resolving shifts between intensity levels. The Journal of Physical Chemistry B, 115, 5071–5082. Kr€ uger, T. P. J., Novoderezhkin, V. I., Ilioaia, C., & van Grondelle, R. (2010). Fluorescence spectral dynamics of single LHCII trimers. Biophysical Journal, 98, 3093–3101. Kr€ uger, T. P. J., Wientjes, E., Croce, R., & van Grondelle, R. (2011). Conformational switching explains the intrinsic multifunctionality of plant light-harvesting complexes. Proceedings of the National Academy of Sciences of the United States of America, 108, 13516–13521. Lakowicz, J. R. (2006). Principles of fluorescence spectroscopy (3rd ed.). . Springer Verlag. Lee, J.-E., Stepanenko, V., Yang, J., Yoo, H., Schlosser, F., Bellinger, D., et al. (2013). Structure property relationship of perylene bisimide macrocycles probed by atomic force microscopy and single-molecule fluorescence spectroscopy. ACS Nano, 7, 5064–5076. Lippitz, M., Kulzer, F., & Orrit, M. (2005). Statistical evaluation of single nano-object fluorescence. Chemical Physics Chemistry, 6, 770–789. L€ ohner, A., Cogdell, R., & K€ ohler, J. (2018). Contribution of low-temperature singlemolecule techniques to structural issues of pigment–protein complexes from photosynthetic purple bacteria. Journal of the Royal Society Interface, 15, 20170680. Lu, H. P. (1998). Single-molecule enzymatic dynamics. Science, 282, 1877–1882. Madjet, M. E., Abdurahman, A., & Renger, T. (2006). Intermolecular coulomb couplings from ab initio electrostatic potentials: Application to optical transitions of strongly coupled pigments in photosynthetic antennae and reaction centers. The Journal of Physical Chemistry B, 110, 17268–17281. Matsushita, M., Ketelaars, M., van Oijen, A. M., K€ ohler, J., Aartsma, T. J., & Schmidt, J. (2001). Spectroscopy on the B850 band of individual light-harvesting 2 complexes of Rhodopseudomonas acidophila II. Exciton states of an elliptically deformed ring aggregate. Biophysical Journal, 80, 1604–1614. McDermott, G., Prince, S. M., Freer, A. A., Hawthornthwaite-Lawless, A. M., Papiz, M. Z., Cogdell, R. J., et al. (1995). Crystal structure of an integral membrane light-harvesting complex from photosynthetic bacteria. Nature, 374, 517–521. Mirkovic, T., Ostroumov, E. E., Anna, J. M., van Grondelle, R., Govindjee, & Scholes, G. D. (2017). Light absorption and energy transfer in the antenna complexes of photosynthetic organisms. Chemical Reviews, 117, 249–293. M€ uller, P., Li, X. P., & Niyogi, K. K. (2001). Non-photochemical quenching. A response to excess light energy. Plant Physiology, 125, 1558–1566. Nirmal, M., Dabbousi, B. O., Bawendi, M. G., Macklin, J. J., Trautman, J. K., Harris, T. D., et al. (1996). Fluorescence intermittency in single cadmium selenide nanocrystals. Nature, 383, 802–804. Onuchic, J. N., Luthey-Schulten, Z., & Wolynes, P. G. (1997). Theory of protein folding: The energy landscape perspective. Annual Review of Physical Chemistry, 48, 545–600. Papiz, M. Z., Prince, S. M., Howard, T., Cogdell, R. J., & Isaacs, N. W. (2003). The structure and thermal motion of the B800–850 LH2 complex from Rps. acidophila at 2.0A˚ resolution and 100K: New structural features and functionally relevant motions. Journal of Molecular Biology, 326, 1523–1538. Petrey, D., & Honig, B. (2003). GRASP2: Visualization, surface properties, and electrostatics of macromolecular structures and sequences. Methods in Enzymology, 374, 492–509. Phillips, W. A. (1972). Tunneling states in amorphous solids. Journal of Low Temperature Physics, 7, 351–360. Phillips, W. A. (1987). Two-level states in glasses. Reports on Progress in Physics, 50, 1657–1708.

ARTICLE IN PRESS 30

J€ urgen K€ ohler and Richard J. Cogdell

Rebane, K. K. (1994). Zero-phonon line as the foundation stone of high-resolution matrix spectroscopy, persistent spectral hole burning, single impurity molecule spectroscopy. Chemical Physics, 189, 139–148. Richter, M. F., Baier, J., Cogdell, R. J., K€ ohler, J., & Oellerich, S. (2007). Single-molecule spectroscopic characterization of light-harvesting 2 complexes reconstituted into model membranes. Biophysical Journal, 93, 183–191. Sako, Y., & Yanagida, T. (2003). Single-molecule visualization in cell biology. Nature Reviews Molecular Cell Biology, 4(Suppl.), SS1–SS5. https://doi.org/10.1038/nrm1193. Sauer, K., Cogdell, R. J., Prince, S. M., Freer, A., Isaacs, N. W., & Scheer, H. (1996). Structure-based calculations of the optical spectra of the LH2 bacteriochlorophyll-protein complex from Rhodopseudomonas acidophila. Photochemistry and Photobiology, 64, 564–576. Schlau-Cohen, G. S., Wang, Q., Southall, J., Cogdell, R. J., & Moerner, W. E. (2013). Single-molecule spectroscopy reveals photosynthetic LH2 complexes switch between emissive states. Proceedings of the National Academy of Sciences of the United States of America, 110, 10899–10903. Schlau-Cohen, G. S., Yang, H.-Y., Kr€ uger, T. P. J., Xu, P., Gwizdala, M., van Grondelle, R., et al. (2015). Single-molecule identification of quenched and unquenched states of LHCII. The Journal of Physical Chemistry Letters, 6, 860–867. Schlichter, J., Friedrich, J., Herenyi, L., & Fidy, J. (2000). Protein dynamics at low temperatures. The Journal of Chemical Physics, 112, 3045–3050. Sch€ orner, M., Beyer, S. R., Southall, J., Cogdell, R. J., & K€ ohler, J. (2015a). Conformational memory of a protein revealed by single-molecule spectroscopy. The Journal of Physical Chemistry B, 119, 13964–13970. Sch€ orner, M., Beyer, S. R., Southall, J., Cogdell, R. J., & K€ ohler, J. (2015b). Multi-level, multi time-scale fluorescence intermittency of photosynthetic LH2 complexes: A precursor of non-photochemical quenching? The Journal of Physical Chemistry B, 119, 13958–13963. Skandary, S., Konrad, A., Hussels, M., Meixner, A. J., & Brecht, M. (2015). Orientations between red antenna states of photosystem I monomers from Thermosynechococcus elongatus revealed by single-molecule spectroscopy. The Journal of Physical Chemistry B, 119, 13888–13896. Stryer, L. (Ed.), (1995). Biochemistry (3rd ed.). New York: W.H. Freeman and Company. Terentyeva, T. G., Engelkamp, H., Rowan, A. E., Komatsuzaki, T., Hofkens, J., Li, C.-B., et al. (2012). Dynamic disorder in single-enzyme experiments: Facts and artifacts. ACS Nano, 6, 346–354. Tietz, C., Jelezko, F., Gerken, U., Schuler, S., Schubert, A., Rogl, H., et al. (2001). Single molecule spectroscopy on the light-harvesting complex II of higher plants. Biophysical Journal, 81, 556–562. Tubasum, S., Torbj€ ornsson, M., Yadav, D., Camacho, R., S€ oderlind, G., Scheblykin, I. G., et al. (2016). Protein configuration landscape fluctuations revealed by exciton transition polarizations in single light harvesting complexes. The Journal of Physical Chemistry. B, 120, 724–732. Unterkofler, S., Pflock, T., Southall, J., Cogdell, R. J., & K€ ohler, J. (2011). Fluorescence blinking of the RC-LH1 complex from Rhodopseudomonas palustris. Chemical Physics Chemistry, 12, 711–716. Valkunas, L., Chmeliov, J., Kr€ uger, T. P. J., Ilioaia, C., & van Grondelle, R. (2012). How photosynthetic proteins switch: The journal of physical chemistry letters. The Journal of Physical Chemistry Letters, 3, 2779–2784. van Amerongen, H., Valkunas, L., & van Grondelle, R. (2000). Photosynthetic excitons. World Scientific.

ARTICLE IN PRESS Metabolism, structure, and function of chlorophylls

31

Vanden Bout, D. A. V., Yip, W.-T., Hu, D., Fu, D.-K., Swager, T. M., & Barbara, P. F. (1997). Discrete intensity jumps and intramolecular electronic energy transfer in the spectroscopy of single conjugated polymer molecules. Science, 277, 1074–1077. Walz, T., Jamieson, S. J., Bowers, C. M., Bullough, P. A., & Hunter, C. N. (1998). Projection structures of three photosynthetic complexes from Rhodobacter sphaeroides: LH2 at 6 a, LH1 and RC-LH1 at 25 a. Journal of Molecular Biology, 282, 833–845. Wang, Q., & Moerner, W. E. (2013). Lifetime and spectrally resolved characterization of the photodynamics of single fluorophores in solution using the anti-Brownian electrokinetic trap. The Journal of Physical Chemistry B, 117, 4641–4648. Watkins, L. P., & Yang, H. (2004). Detection of intensity change points in time-resolved single-molecule measurements. The Journal of Physical Chemistry B, 109, 617–628. Wolynes, P. G. (2005). Energy landscapes and solved protein-folding problems. Philosophical Transactions. Series A, Mathematical, Physical, and Engineering Sciences, 363, 453–464 [discussion 464-7]. Wu, H.-M., & Small, G. J. (1998). Symmetry-based analysis of the effects of random energy disorder on the excitonic level structure of cyclic arrays: Application to photosynthetic antenna complexes. The Journal of Physical Chemistry B, 102, 888–898. Xie, X. S., Choi, P. J., Li, G.-W., Lee, N. K., & Lia, G. (2008). Single-molecule approach to molecular biology in living bacterial cells. Annual Review of Biophysics, 37, 417–444. Yang, H., Luo, G., Karnchanaphanurach, P., Louie, T.-M., Rech, I., Cova, S., et al. (2003). Protein conformational dynamics probed by single-molecule electron transfer. Science, 302, 262–266. Yeow, E. K. L., Melnikov, S. M., Bell, T. D. M., de Schryver, F. C., & Hofkens, J. (2006). Characterizing the fluorescence intermittency and photobleaching kinetics of dye molecules immobilized on a glass surface. The Journal of Physical Chemistry A, 110, 1726–1734. Zazubovich, V. (2014). Fluorescence line narrowing and Δ-FLN spectra in the presence of excitation energy transfer between weakly coupled chromophores. The Journal of Physical Chemistry B, 118, 13535–13543. Zazubovich, V., Jankowiak, R., & Small, G. J. (2002). A high-pressure spectral hole burning study of correlation between energy disorder and excitonic couplings in the LH 2 complex from Rhodopseudomonas acidophila. The Journal of Physical Chemistry B, 106, 6802–6814.