INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL RESULTS DISCUSSION SUMMARY REFERENCES

Green sulfur bacteria, such as Chlorobium (Cb.) tepidum and Cb. limicola, and green nonsulfur bacteria, such as Chloroflexus (Cf.) aurantiacus, are separated by a large evolutionary distance according to 16S rRNA analysis (Woese, 1987; Olsen et al., 1994). Their distinctiveness is reflected in the different types of reaction centers and membrane-bound antenna complexes found in the two families, as well as their entirely distinct metabolic and ecological characteristics (see Blankenship et al., 1995; Olson, 1998, and references therein). Interestingly, they both contain chlorosomes; an antenna complex of flattened ellipsoidal shapes of ~100 × 30 × 10 nm³ for the green nonsulfur bacteria and ~200 × 70 × 12 nm³ for the green sulfur bacteria. Chlorosomes consist of up to ~10,000 aggregated bacteriochlorophyll (BChl) c, d, or e molecules surrounded by a monolayer lipid envelope and a relatively small amount of protein. BChl a molecules are associated with the baseplate of the chlorosomes (Blankenship et al., 1995; Olson, 1998), which serves as an intermediate in the excitation energy transfer from the chlorosomes to BChl a-containing antenna complexes in both types of microorganisms: the B808-866 complex for the green nonsulfur bacteria and the Fenna-Matthews-Olson (FMO) complex for the green sulfur bacteria. Fig. 1 is a schematic of the arrangement of photosynthetic complexes in green sulfur bacteria. The molar ratio of BChl c (d or e) to BChl a in the chlorosomes varies with species and growth conditions. Molar ratios of 20:1 for chlorosomes from Cf. aurantiacus (Schmidt, 1980) and 90:1 for those from Cb. limicola (Gerola and Olson, 1986) have been reported.

View larger version (22K): [in this window] [in a new window]   FIGURE 1   Schematic model for the arrangement of the chlorosome, the FMO complex and the reaction center in green sulfur bacteria. Chlorosomes consist of BChl c aggregates in the form of rods surrounded by a lipid monolayer with the BChl a-containing baseplate attached. The FMO complex and reaction center (P840) bind BChl a molecules, and the reaction center contains an iron-sulfur cluster.

Compared with other antenna complexes, chlorosomes have relatively low protein to BChl c, d, or e ratios, ranging from ~0.5 to ~2, depending on the species and growth conditions (Cruden and Stanier, 1970; Schmidt, 1980; Feick and Fuller, 1984; Chung et al., 1994; Blankenship et al., 1995). It is generally accepted that pigment-pigment interactions within chlorosomes are mainly responsible for the pigment organization, rather than pigment-protein interactions, as is commonly found in other antenna complexes. BChl c, d, or e can readily form aggregates in nonpolar organic solvents (Bystrova et al., 1979; Smith et al., 1983; Brune et al., 1987; Olson and Pedersen, 1990) or aqueous lipid/detergent micelles (Hirota et al., 1992; Miller et al., 1993; van Noort et al., 1997) in vitro with spectroscopic properties similar to those observed in chlorosomes. That is, the Q_y band of the BChl c monomers at ~670 nm is red-shifted to the ~720-760-nm region because of the formation of aggregates. Based on resonance Raman, Fourier transform infrared spectroscopy, and absorption studies, the BChl c molecules are essentially pentacoordinated, and the 3¹ hydroxyl and 13¹ keto groups are involved in the aggregation (Brune et al., 1988; Lutz and van Brakel, 1988; Nozawa et al., 1990; Hildebrandt et al., 1994). It is believed that the aggregated BChl c, d, or e molecules form rods 5-10 nm in diameter within chlorosomes, as was seen in freeze-fracture electron micrographs of Cb. limicola and Cf. aurantiacus cells (Staehelin et al., 1978, 1980; Golecki and Oelze, 1987). Various pigment aggregation models have been proposed for the chlorosome aggregates that are based on antiparallel chains (Brune et al., 1988; Nozawa et al., 1994; Minuro et al., 1995), parallel stepped chains (Brune et al., 1988; Mizoguchi et al., 1998), and cylindrical arrangements of BChl molecules (Somsen et al., 1996). Larger scale structural models for the rod elements, aided by quantum chemical calculations or molecular modeling, have also been proposed (Holzwarth and Schaffner, 1994; Buck and Struve, 1996; Fetisova et al., 1996).

Over the past 20 years nonphotochemical hole-burning (NPHB) spectroscopy has been applied to chromophores embedded in disordered host media whose optical spectra suffer from significant inhomogeneous broadening. Photosynthetic protein complexes with an inhomogeneous broadening contribution of ~100 cm¹ to the absorption bandwidths are good candidates for NPHB spectroscopy, which can improve spectral resolution by two to three orders of magnitude. Information that is obtainable includes the inhomogeneity of the protein complex, excited-state lifetimes, exciton-level structures, the electron-phonon coupling strength of optical transitions and frequencies, and Franck-Condon factors of chlorophyll (Chl) modes. For general reviews covering NPHB and its applications, see Moerner (1988) and Jankowiak et al. (1993). For reviews pertaining to NPHB studies of photosynthetic complexes, see Johnson et al. (1991), Reddy et al. (1992a), and Jankowiak and Small (1993).

Recently, NPHB, in combination with high pressure and electric (Stark) fields, has been applied to several photosynthetic protein complexes. The results obtained led to new insights into excitation energy transfer dynamics and the Q_y electronic structures of Chl molecules. Of particular relevance to this study are the high-pressure- and Stark-NPHB experiments of Reddy et al. (1996), Wu et al. (1996, 1997a, 1998), and Rätsep et al. (1998a,b) on light-harvesting (LH) complexes from purple bacteria. Not only are the results of those studies consistent with the structure of LH2 as determined by x-ray crystallography; they also provide benchmarks for electronic structure calculations by establishing that electron exchange coupling between nearest-neighbor pigments and energy disorder associated with the cyclic LH2 and LH1 rings cannot be ignored. This is important because a firm understanding of the Q_y states of Chl molecules bound to the proteins of LH complexes is essential for understanding their excitation energy transfer/relaxation dynamics. The usefulness of the external field/NPHB combination (or external field/optical spectroscopy combination in a broader sense) in the study of dynamics, structures, interactions, and functions is not limited to photosynthetic complexes. In principle, any system with a probe is an ideal candidate for such a study. An example of a probe intrinsic to the system is the heme group in hemoglobin, but artificial labeling is also possible, such as the labeling of human serum albumin with hypericinate ions (Falk and Meyer, 1994). An example of an external field/optical spectroscopy combination study that is nonbiologically oriented is the photon echo investigation of optical dephasing in pentacene-doped polymethyl methacrylate under high pressure (Berg and Chronister, 1997). The reader interested in the effects of high pressure and Stark fields on the spectroscopic properties of pigments bound to proteins is referred to the review articles of Kohler et al. (1998) and Fidy et al. (1998).

The chlorosome Q_y bands of Cf. aurantiacus and Cb. limicola, which have absorption maxima at 742 and 751 nm, respectively, have been investigated by NPHB in the fluorescence excitation mode at low temperature (Fetisova and Mauring, 1992, 1993). The results were consistent with the chlorosome pigments being excitonically coupled. The lowest exciton level of the Q_y band was characterized by zero-phonon hole (ZPH) action spectroscopy and found to carry an inhomogeneous full width at half-maximum (FWHM) of ~100 cm¹ with a band center at ~752 nm for Cf. aurantiacus and ~774 nm for Cb. limicola. The ZPH action spectrum is generated by burning a series of ZPHs across the absorption band with constant laser fluence and is a powerful method for obtaining the width, center, and intensity of an inhomogeneously broadened distribution mixed with other absorption contributions. An example of such a study on photosynthetic antenna complexes is the identification of the lowest exciton level (B870) of LH2 complex from Rhodopseudomonas acidophila (Wu et al., 1997c). Broad nonresonant holes resulting from fast interexciton-level relaxation from the higher exciton components, which carry most of the oscillator strength, were also observed. Recently, ZPH action spectroscopy was applied in a study of the lowest exciton level of chlorosomes of Cb. tepidum (Psencik et al., 1998). The lowest exciton level was found to lie in the range of 760-800 nm and to be centered at ~780 nm. The fractional hole depth of the most intense hole was 0.04 in the action spectrum. Different redox conditions were also employed to examine their effects on hole widths, as the excited-state lifetime of BChl c aggregates and energy transfer to the baseplate were known to be regulated by the redox potential for green sulfur bacteria. The ZPH width was 1.8 cm¹ under anaerobic conditions and 4.0 cm¹ under aerobic conditions and was independent of wavelength and temperature up to 25 K. Psencik et al. (1998) proposed that the ZPH width is due to the BChl c-to-BChl c energy transfer. A width of 1.8 cm¹ corresponds to a BChl c-to-BChl c energy transfer time (T₁) of 5.8 ps, whereas a width of 4.0 cm¹ corresponds to a transfer time of 2.7 ps. The shortening to 2.7 ps may be due to an additional quenching mechanism, involving one or both of the quinone molecules chlorobiumquinone and menaquinone-7 under oxidizing conditions (Frigaard et al., 1997).

Currently the exact structural arrangement of the pigments within the chlorosomes is unknown, but models generally include J-aggregate-like long chains arranged circularly to form the rod elements observed by electron microscopy. The homogeneous line shift and broadening of pseudoisocyanine (PIC) J-aggregates under pressure (<6.5 MPa) have been investigated by hole-burning spectroscopy at 4.2 K (Hirschmann and Friedrich, 1992). The holes within the J-band were found to shift with a linear rate of about 0.3 cm¹/MPa, which depends on the burn frequency within the band. The hole broadening induced by pressure was found to be surprisingly small. The largest linear broadening rate observed was ~0.025 cm¹/MPa; a dependence on the burn frequency (color effect) was observed. Hirschmann and Friedrich argued that the low hole-broadening rate is the result of motional narrowing of structural heterogeneity due to large exciton coherence lengths.

Before this study, Stark hole burning was applied to the J-band of aggregated pseudoisocyanine iodide (PIC-I) at 2 K (Wendt and Friedrich, 1996). The aggregation of PIC-I reduced the linear Stark effect by more than two orders of magnitude compared to the monomer. At a field strength of 300 kV/cm, the quadratic Stark effect was significant for the J-aggregate (comparable to the linear contribution). The large difference in the polarizability between the ground and excited states, which is responsible for the quadratic Stark effect, was argued to be due to the aggregation of PIC-I molecules on the basis of exciton theory. Wendt and Friedrich argued that certain structural constraints such as a helical pigment organization could explain the large reduction in the linear Stark effect of the aggregates.

Recently, conventional Stark modulation spectroscopy at 77 K was used in a study of chlorosomes from Cf. aurantiacus (in both wild-type and carotenoid-deficient cells) (Frese et al., 1997). The Stark spectrum of the Q_y band resembles a first-derivative-type lineshape, consistent with there being a large difference in polarizability between the ground and excited states (Tr() = 1650 ± 100 Å³/f², where Tr() denotes the trace of the polarizability difference tensor). The second-derivative contribution to the Stark spectrum was found to be negligible, which indicates that the dipole moment change between the ground and excited states is very small. Frese et al. suggested that the aggregated BChl c in chlorosomes adopt an antiparallel structure, as opposed to various parallel models proposed by others (Brune et al., 1988; Alden et al., 1992; Mizoguchi et al., 1998).

We present here the results of high-pressure and Stark hole-burning experiments on isolated chlorosomes from BChl c containing Cb. tepidum, as well as Stark hole-burning results for BChl c monomers in a polymer film. The motivations for the experiments were to gain additional insights into the arrangement of BChl c molecules in chlorosomes and the exciton level structure of the aggregates (with emphasis on the lowest Q_y exciton level) and to provide benchmarks for electronic structure calculations based on structural models for the BChl c aggregates.

	EXPERIMENTAL

TOP ABSTRACT INTRODUCTION EXPERIMENTAL RESULTS DISCUSSION SUMMARY REFERENCES

Cb. tepidum cells were grown at ~40°C in a 90-liter clear plastic vessel in a modified version of the CH1 medium reported by Olson et al. (1973) and illuminated with four banks of fluorescent tube lights ( 60 µEinsteins m² s¹) over 3-4 days with gentle stirring. Cells were harvested using a continuous flow centrifuge and used in the preparation of chlorosomes.

Chlorosomes were isolated following the method of Gerola and Olson (1986) with minor revisions. Four grams of cell paste was mixed with 16 ml of chlorosome buffer (10 mM sodium phosphate (pH 7.4), 10 mM sodium ascorbate, and 2 M sodium thiocyanate) until an even suspension was achieved. Cells were broken by sonication in the presence of DNase I, and the sonicate was centrifuged to remove cell debris. The dark green supernatant was layered on top of a 20-50% continuous sucrose density gradient. The sucrose gradient was spun at 45,000 rpm for 18 h at 4°C, and the chlorosome fraction was removed. Chlorosomes (FMO depleted) equilibrate between 25% and 30% sucrose.

BChl c pigments used in the Stark hole-burning experiment were extracted from whole cells of Cb. tepidum according to the procedure described by van Noort et al. (1997). BChl c was dissolved in methanol, which prevents the formation of aggregates, and was added to poly(vinyl butyral-co-vinyl alcohol-co-vinyl acetate) (Aldrich, Milwaukee, WI)/dichloromethane solution (0.5 g of the polymer with an average molecular weight of 90,000-120,000 in 15 ml dichloromethane). The solution was dried overnight in a flat Petri dish covered by a glass plate (to reduce the drying rate). The dried film was kept in an oven under vacuum at 40°C for ~2 h to remove the residual solvent. Typically the film separated from the Petri dish after it was placed in a darkened closed box with a wet towel for ~2 days. The resulting OD and FWHM of the BChl c absorption band at 667 nm at room temperature were 0.07 and  400 cm¹, respectively.

Hole-burned spectra of chlorosomes were detected in the absorption mode (Lyle et al., 1993). A Ti:sapphire ring laser (model 899-21, linewidth ~0.07 cm¹; Coherent, Santa Clara, CA) pumped by a Coherent 15-W argon ion laser was used to burn holes in the Q_y absorption band. The spectra before and after burning were recorded with a Bruker IFS 120 HR Fourier transform spectrometer (Bruker, Billerica, MA). Burn fluence and spectral reading resolutions are given in the figure captions.

Because of the smaller ZPH width, spectral hole burning of BChl c monomers was detected in the fluorescence excitation mode with an apparatus described by Reinot et al. (1996) and Kim et al. (1995). The burn and read laser was a Coherent 699-21 ring dye laser (linewidth <30 MHz) pumped by a 6-W Coherent Innova 90 Ar ion laser stabilized by a LS 100 power stabilizer (Cambridge Research and Instrumentation, Cambridge, MA). Fluorescence was detected by a GaAs photomultiplier tube (RCA C31034) and a photon counter (SR-400; Stanford Research Systems, Sunnyvale, CA). Scattered laser light was eliminated with cutoff filters at 750 nm. Laser intensities used for hole burning were in the range of ~10 µW/cm², and hole depths were typically ~25%. For hole reading, the duration of which was ~300 s, the laser was attenuated by a factor of ~60.

High pressures of up to 1.5 GPa were generated by compressing helium gas with a Unipress U11 three-stage compressor (Warsaw, Poland). The sample was contained in a gelatin capsule and placed in a high-pressure cell (rated for pressures less than 800 MPa) with two sapphire windows installed. A custom-made liquid helium cryostat (Janis, Wilmington, MA) was used to achieve low temperatures for the high-pressure hole-burning experiments. For further details, see Reddy et al. (1995).

A Trek model 610 C dc high-voltage power supply (0 to ± 10 kV; Trek, Medina, NY) was used to generate the Stark field. By changing the polarity of the power supply, we were able to achieve a maximum Stark field difference of 100 kV/cm for two copper electrodes separated by 2 mm. For the quadratic Stark effect, however, the highest field achievable was 50 kV/cm, because the positive and negative polarities of the power supply would result in the same effect. The gelatin capsule containing the sample was allowed to soften at room temperature for 5 min before being squeezed into the space between the electrodes (optical path length ~6 mm). To avoid dielectric breakdown, all measurements were performed at 1.8 K in a Janis 10 DT liquid helium cryostat. This setup allowed for study of the Stark effect with varying laser polarization relative to the external field. The interested reader is referred to Rätsep et al. (1998a) for further details.

The Stark cell used in the fluorescence excitation mode for the BChl c monomer study consisted of two vertical Teflon walls and two copper electrodes perpendicular to the walls (Milanovich et al., 1998). A separation distance of 5 mm between electrodes was maintained by placing them in the grooves on the inside of the Teflon walls. A slit was made in one of the Teflon walls to allow the laser access to the sample. Fluorescence was collected at a 90° angle relative to the incident laser light through an opening on the other Teflon wall in the Stark cell. A polarizer placed in front of the Stark cell allowed for control of the laser polarization relative to the applied field direction.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL RESULTS DISCUSSION SUMMARY REFERENCES

Pressure-dependent studies of the Q_y band and its lowest exciton level

Fig. 2 shows the near-IR absorption spectrum of chlorosomes at 15 and 752 MPa (1 MPa  10 atm) at 100 K. A red shift of 320 cm¹ (from 13,250 ± 5 to 12,930 ± 5 cm¹) and a band broadening of 95 cm¹ (from 550 ± 5 to 645 ± 5 cm¹) with increasing pressure are observed. Fig. 3 plots the absorption band positions and widths versus pressure. Within the pressure range employed, the red shift and broadening are reversible and linear with pressure (see the regression lines in Fig. 3). The linear pressure shift and broadening rates were 0.44 and 0.12 cm¹/MPa, respectively.

View larger version (23K): [in this window] [in a new window]   FIGURE 2   Absorption spectra (100 K) of the Qy band of isolated chlorosomes from Cb. tepidum at 15 and 752 MPa. As pressure increases the Qy band shifts from 13,250 cm1 (755 nm) to 12,930 cm1 (773 nm), and the bandwidth increases from 550 to 645 cm1. See Fig. 3 for the pressure shift and broadening rates. The inset shows the zero-phonon hole burned at 12,861.8 cm1 at 15 MPa and 12 K and its shift and broadening when the pressure was increased to 34 MPa. The hole detection was made with a Fourier transform infrared spectrometer in the absorption mode. The burn fluence and read resolution are 100 J/cm2 and 2 cm1, respectively.

View larger version (16K): [in this window] [in a new window]   FIGURE 3   Pressure shift () and broadening () of the Qy absorption band of isolated chlorosomes from Cb. tepidum at 100 K. As indicated by the two regression lines, the pressure dependence of the band shifts and widths is linear within the pressure range employed. The linear shift and broadening rates are 0.44 ± 0.01 cm1/MPa and 0.12 ± 0.01 cm1/MPa, respectively. No irreversible pressure-induced effects were observed.

As in the studies of LH2 and LH1 antenna complexes of purple bacteria (Reddy et al., 1992b, 1993; Wu et al., 1997c, 1998), the lowest exciton level of the chlorosome Q_y band was studied by means of ZPHs burned at the red edge of the band. This level was determined by ZPH action spectroscopy at ambient pressure to have an inhomogeneous distribution width (FWHM) of ~100 ± 10 cm¹ and was centered at ~12,900 ± 10 cm¹ (775 nm), consistent with the findings of Psencik et al. (1998) for Cb. tepidum. The pressure shifting of the lowest exciton level of the Q_y band at 12 K is shown in Fig. 4. Five ZPHs were burned at 15 MPa in this spectral region, which shift to the red with increasing pressure. The inset of Fig. 2 shows the ZPHs burned at 12,862 cm¹ at 12 K and 15 MPa. The shifting and broadening of that hole resulting from an increase in pressure to 34 MPa are also shown. Because of the solidification of the pressure-transmitting medium, helium gas, at ~75 MPa at 12 K, as well as rapid hole filling and broadening, no data were obtained beyond 57 MPa. As with the absorption band, the shift of ZPHs is linearly dependent on the pressure (see the regression lines and their slopes in Fig. 4) with an averaged linear pressure shift rate of 0.54 cm¹/MPa. Unlike the lowest exciton level (B870) of B850 molecules of LH2 from Rps. acidophila and Rb. sphaeroides (Wu et al., 1997a), which exhibits an increase in shift rate in going from the high-energy to low-energy sides of the band, the shift rate of the lowest exciton level of the chlorosome Q_y band is essentially constant.

View larger version (19K): [in this window] [in a new window]   FIGURE 4   Linear pressure shifting of zero-phonon holes (holes were read in the absorption mode with a read resolution = 2 cm1) burned on the low-energy side of the Qy absorption band of chlorosomes from Cb. tepidum. Initially, five holes were burned at 12,910.9, 12,894.3, 12,877.9, 12,861.8, and 12,844.9 cm1, at 15 MPa and 12 K. One of the ZPH profiles is shown in the inset of Fig. 2. A burn fluence of 100 J/cm2 was used. All holes shifted linearly (see the solid lines) to the red with increasing pressure with rates (in cm1/MPa) indicated in the figure. The uncertainties of the shift rates are ±0.03, ±0.02, ±0.01, ±0.01, and ±0.01 cm1/MPa from top to bottom. No irreversible pressure-induced effects were observed.

Stark-hole burning studies

The Stark shift of the optical transition frequency of the absorber is given by

(1)

 is the local field correction factor and is taken to be a scalar. µ₀ is the molecular dipole moment change, and ₀ is the molecular polarizability difference tensor. E_int is the matrix field experienced by the absorber and the induced dipole moment µ_ind = E_int. The first and second terms in the square brackets depend linearly and quadratically on the Stark field E_s, respectively. As reviewed in Rätsep et al. (1998b), E_int in molecular systems is large,  10⁶ V/cm (see also Kador et al., 1990; Meixner et al., 1992; Middendorf et al., 1993), which is an order of magnitude larger than the maximum Stark fields used in Stark hole-burning experiments. As a result, only the linear Stark effect was observed for the B800 and B870 bands of the LH2 complex, the B896 band of the LH1 complex, the 825-nm band of the FMO complex, and all isolated chromophores in polymer, glass, and protein matrices (see Rätsep et al., 1998a, and references therein). (The B870 and B896 bands correspond to the lowest exciton level of the LH2 and LH1 BChl a rings.) However, a large quadratic Stark effect was observed in the J-aggregate, in which was enhanced because of aggregation (Wendt and Friedrich, 1996).

Considering further the case of isolated molecules in glassy matrices with the domination of the linear Stark effect, let  be the angle between the molecular dipole moment difference vector µ_o and the transition dipole vector d. Linearly polarized light preferentially burns out those molecules with d parallel to the polarization vector eof the light. The experimentally observed dipole moment change can be written as µ = µ_o + µ_ind, where µ_ind is the matrix-induced dipole moment change equal to E_int. When µ_o is dominant, photoselection enables one to probe molecules for which the angle between µ_o and e is well defined (shallow hole limit). As discussed by Meixner et al. (1986) and Gafert et al. (1995), Stark splitting of the hole can be observed for an angle between e and E_s, which depends on the value of . For example, for  = 0 or , Stark splitting occurs for parallel polarization, while symmetrical broadening occurs for perpendicular polarization. The situation is reversed for  = ±/2. However, when µ_ind is dominant, only Stark broadening is expected for both polarizations. This is because the orientation of µ_ind relative to d or e is random for a glassy matrix, i.e., the matrix field varies significantly from site to site.

The assumption of random orientations for µ_ind for Chl molecules in photosynthetic complexes is questionable because the structure of the protein around these chromophores is well defined, although the Q_y absorption bands do suffer from significant inhomogeneous broadening. For example, Gafert et al. (1995) observed Stark hole splitting for two of the three sites of mesoporphyrin substituted in horseradish peroxidase. For the same molecule in a glass, only Stark broadening was observed for both laser polarizations. Gafert et al. introduced the notion of random and nonrandom protein contributions to µ_ind, i.e., µ_ind(random) and µ_ind(nonrandom). Recently, Rätsep et al. observed Stark splitting and broadening with parallel and perpendicular laser polarizations, respectively, for the lowest exciton level of the FMO complex (Rätsep et al., 1998a). However, for the B800 and B870 bands of the LH2 complex and B896 of the LH1 complex, Stark broadening of the ZPH was observed for both parallel and perpendicular laser polarizations (Rätsep et al., 1998a,b).

Stark hole-burning studies of the Q_y band of BChl c monomers

1

1

1

View larger version (25K): [in this window] [in a new window]   FIGURE 5   Stark effect for a zero-phonon hole burned at 670.2 nm in the Qy band of BChl c monomer in a poly(vinyl butyral) copolymer film at 1.8 K. The burn fluence used was 1.3 mJ/cm2. Holes were read in the fluorescence excitation mode. The laser polarization is perpendicular to the Stark field. The horizontal axis is offset to center the hole at zero. The dependence of the hole width on the electric field is shown in Fig. 6.

View larger version (12K): [in this window] [in a new window]   FIGURE 6   Stark broadening of a zero-phonon hole burned at 670.2 nm in the Qy band of the BChl c monomer in a poly(vinylbutyral) copolymer film, T = 1.8 K. Holes were read in the fluorescence excitation mode. The solid curve is the theoretical fit obtained using the theory of Kador et al. (1987), with fµ = 0.35 D.

Stark hole-burning studies of chlorosomes

1

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL RESULTS DISCUSSION SUMMARY REFERENCES

Absorption spectra

Compared with the 4.2 K absorption bands of other photosynthetic antenna complexes such as B800 (FWHM = 125 cm¹) and B850 of LH2 (FWHM = 200 cm¹) of purple bacteria and the three partially resolved Q_y bands (FWHM  100 cm¹) of the FMO complex, the chlorosome Q_y band is considerably broader (FWHM  550 cm¹) and clearly asymmetrical (Fig. 2). Given that the inhomogeneous broadening of the lowest exciton level of the chlorosome is only ~100 cm¹, a significant fraction of the 550 cm¹ width is most likely due to higher energy exciton levels. These levels are inhomogeneously and homogeneously broadened because of structural heterogeneity and downward interexciton level relaxation, as proposed in the earlier hole-burning papers discussed in the Introduction.

Before discussing the spectral heterogeneity of chlorosomes, it is useful to discuss the exciton level structure of a J-aggregate. For a linear chain of N identical molecules, the Hamiltonian in the absence of energy disorder is (Fidder et al., 1991; Alden et al., 1992)

(2)

where | label the states in which molecule  ( = 1, 2, ... N) is excited and all others are in the ground state. e is the excitation energy of the chromophore, and V is the coupling energy between molecules  and . Under the nearest-neighbor coupling approximation V_,+1 ± V_,1 = V, the Hamiltonian can be diagonalized exactly to give the eigenfunctions and eigenenergies as

(3)

and

(4)

where j = 1, 2, ... , N. For simplicity, e of Eq. 2 has been set equal to zero. From Eq. 3 the transition dipole strength of |j is found to be (Fidder et al., 1991)

(5)

(6)

µ_mon denotes the transition dipole moment of the chromophore. It follows from Eq. 5 that the j = 1 state carries the largest transition dipole strength. It also follows that f_j decreases with increasing j and that the larger the value of N, the smaller is the percentage of the total transition dipole strength carried by the j = 1 state. In the limit of large N, 81% of the total transition dipole strength is carried by the j = 1 state.

For J-aggregates, the coupling energy V between the nearest neighbors must be negative. As a result, the lowest exciton level (j = 1) carries most of the transition dipole strength in the absence of energy disorder. However, the J-aggregate picture is inconsistent with NPHB studies of chlorosomes which establish that the lowest exciton level is only weakly allowed (see Results and Fetisova and Mauring, 1992, 1993, and Psencik et al., 1998). Buck and Struve (1996) proposed that this issue can be resolved by introducing a tubular arrangement of J-aggregates and diagonal energy disorder to explain the exciton level structure of chlorosomes. Their tubular arrangement leads to the lowest exciton level being forbidden in the absence of energy disorder. Energy disorder endows this level with some intensity stolen from the allowed levels. This situation is similar to those of B870 of LH2 and B896 of LH1 of purple bacteria (Reddy et al., 1992b, 1993; Wu et al., 1997b,c; Wu and Small, 1997, 1998). The picture that emerges is that the broad and asymmetrical Q_y band of chlorosomes is a superposition of exciton levels carrying different absorption intensities and, possibly, bandwidths. However, the tailing on the blue side of the Q_y absorption band seen in Fig. 2 may be contributed to by weak intramolecular vibronic transitions of the BChl c molecules (Cherepy et al., 1996).

Spectral equilibration among different components of chlorosome BChl c molecules of Cb. tepidum was observed in the femtosecond pump-probe experiments of Savikhin et al. (1995). A red shift in the absorption difference spectrum was observed upon 720-nm excitation. The one-color and two-color isotropic data also confirmed the existence of spectral heterogeneity within the chlorosome absorption band. Similar results have been obtained for chlorosomes from BChl e-containing Cb. phaeovibrioides and BChl c- and d-containing Cb. vibrioforme by picosecond pump-probe laser spectroscopy (van Noort et al., 1994). Interestingly, chlorosomes from Cf. aurantiacus, which contain BChl c and trace quantities of BChl d molecules, do not show such downhill energy transfer within BChl aggregates. At room temperature, the FWHM of the chlorosome Q_y band of Cb. tepidum and Cf. aurantiacus is 800 and 530 cm¹, respectively, suggesting greater heterogeneity for Cb. tepidum. As mentioned in the Introduction, the chlorosomes from green sulfur bacteria such as Cb. tepidum are larger in size and have higher BChl c/BChl a ratios than those of green nonsulfur bacteria such as Cf. aurantiacus. This, together with the spectral equilibration data obtained by Savikhin et al. (1995), suggests that more than one pool of BChl aggregates exists in green sulfur bacteria such as Cb. tepidum to compensate for the lower energy transfer efficiency to the baseplate due to larger chlorosome sizes and lower amounts of BChl a molecules. The different pools of BChl aggregates may be due to variations in the size of the rod elements and/or the relative numbers of BChl c, d, and e molecules and their homologs.

High-pressure studies

Pressure shifting of the Q_y band

1

1

1

(7)

1

1

1

1

1

1

1

1

1

1

1

(8)

1

1

1

1

1

Pressure broadening of the Q_y band

1

1

1

1

1

Stark hole-burning studies

Stark hole burning of BChl c monomers in polymer films

(9)

Stark hole burning of isolated chlorosomes from Cb. tepidum

1

1

1

1

1

1

(10)

	SUMMARY

TOP ABSTRACT INTRODUCTION EXPERIMENTAL RESULTS DISCUSSION SUMMARY REFERENCES

The strongly coupled nature of BChl c molecules in chlorosomes was confirmed by the large linear pressure shift rate of the Q_y band (0.44 cm¹/MPa) and the lowest exciton level (0.54 cm¹/MPa) at low temperature. According to the analysis presented, about half of the contribution to the shift rate is from electron exchange interactions, which require a rather compact arrangement of pigments. This important finding serves as a guideline for electronic structure calculations of the chlorosome in the absence of knowing its exact structural arrangement of pigments.

The dipole moment change (fµ) and polarizability change () for the Q_y transition of BChl c monomers in a polymer film were determined to be 0.35 D and 19 Å³, respectively. µ is dominated by the matrix-induced contribution. The absence of any Stark effect in the lowest exciton level of chlorosomes led us to conclude that pigments in chlorosomes possibly adopt certain structural constraints to attain small dipole moments and polarizability changes. The basic tubular models put forth by Buck and Struve (1996) and Somsen et al. (1996), as well as the helical structure proposed by Wendt and Friedrich (1996), are all in agreement with the Stark hole-burning results, as long as the orientations of the pigments result in the reduction of µ₀ and _ind.