Pigment structure in the violaxanthin–chlorophyll-a-binding protein VCP
Resonance Raman spectroscopy was used to evaluate pigment-binding site properties in the violaxanthin–chlorophyll-a-binding protein (VCP) from Nannochloropsis oceanica. The pigments bound to this antenna protein are chlorophyll-a, violaxanthin, and vaucheriaxanthin. The molecular structures of bound Chl-a molecules are discussed with respect to those of the plant antenna proteins LHCII and CP29, the crystal structures of which are known. We show that three populations of carotenoid molecules are bound by VCP, each of which is in an all-trans configuration. We assign the lower-energy absorption transition of each of these as follows. One violaxanthin population absorbs at 485 nm, while the second population is red-shifted and absorbs at 503 nm. The vaucheriaxanthin population absorbs at 525 nm, a position red-shifted by 2138 cm−1 as compared to isolated vaucheriaxanthin in n-hexane. The red-shifted violaxanthin is slightly less planar than the blue-absorbing one, as observed for the two central luteins in LHCII, and we suggest that these violaxanthins occupy the two equivalent binding sites in VCP at the centre of the cross-brace. The presence of a highly red-shifted vaucheriaxanthin in VCP is reminiscent of the situation of FCP, in which (even more) highly red-shifted populations of fucoxanthin are present. Tuning carotenoids to absorb in the green-yellow region of the visible spectrum appears to be a common evolutionary response to competition with other photosynthetic species in the aquatic environment.
KeywordsLight-harvesting complex VCP Resonance Raman Nannochloropsis oceanica Carotenoids
During the photosynthetic process, solar photons are first absorbed by specialized light-harvesting pigments bound to antenna proteins, and the resulting excitation energy is transferred to the reaction centers where it is transformed into electrochemical energy. The overall structure of reaction centers is conserved in all photosynthetic organisms, while antenna complexes are far more diversified, responding to the light constrictions of the biotope. The antenna proteins of most oxygen-evolving eukaryotes are members of a multigenic family of proteins with a common evolutionary origin, called light-harvesting complexes or LHC proteins (Engelken et al. 2010; Koziol et al. 2007; Neilson and Durnford 2010). LHCs share a preserved structural organization, characterized by three membrane-spanning α-helixes connected by stromal and lumen-exposed loops. Two of these helices are homologous and present a “generic LHC motif”, consisting of a highly hydrophobic sequence containing glutamic acids involved in chlorophyll (Chl) binding and in the stabilization of the protein scaffold through salt bridges with arginines in the other helix (Liu et al. 2004). From a common origin, which has been proposed to be the HLiD photoprotective proteins from cyanobacteria (Dolganov et al. 1995; Neilson and Durnford 2010), LHCs have diversified in different groups of photosynthetic eukaryotes, such as the Chl-a/b-binding proteins found in Viridiplantae (plants and green algae), fucoxanthin Chl-a/c-binding proteins (FCP, or LHCF) in diatoms and brown algae, peridinin Chl-a/c-binding proteins (acpPC) in dinoflagellates, and LHCR in red algae and diatoms (Engelken et al. 2010; Neilson and Durnford 2010; Busch and Hippler 2011). The pigment-binding properties of LHCs diversified to adapt to the amount and quality of light in specific habitats. Chl-a, Chl-b, and Chl-c may be bound to these proteins, as well as a large diversity of carotenoid molecules. For instance, plant LHCII binds two lutein molecules, one neoxanthin, and one violaxanthin per monomer (Liu et al. 2004), while carbonyl carotenoids fucoxanthin and peridinin are typically found in brown algae, diatoms, and dinoflagellates (Apt et al. 1995; Hofmann et al. 1996) .
Nannochloropsis is a genus of eukaryotic algae belonging to the Eustigmatophyceae, a group of organisms related to diatoms. The photosynthetic apparatus of Nannochloropsis oceanica is not fully characterized, but interest in this organism is increasing owing to its high rate of lipid production and accumulation (Rodolfi et al. 2009; Radakovits et al. 2012; Sforza et al. 2012; Simionato et al. 2011). N. oceanica antenna proteins bind Chl-a, violaxanthin, and vaucheriaxanthin (Fig. 1), the latter in the form of 19′ deca/octanoate esters (Britton et al. 2008), along with traces of zeaxanthin and antheraxanthin (<5% of total carotenoid content) (Basso et al. 2014; Litvín et al. 2016). The violaxanthin/vaucheriaxanthin stoichiometric ratio and Chl-a/carotenoid ratio were estimated to be ~1.35 and ~2.2, respectively (Keşan et al. 2016). The major apoprotein of these violaxanthin–chlorophyll-a-binding protein (VCP) complexes is a 22-kDa polypeptide which shows similarity to FCP (Basso et al. 2014). The allenic carotenoid vaucheriaxanthin, with its bulky esterified chain, shows much less structural similarity to lutein than violaxanthin. This and the pigment stoichiometry have been used to argue that the two symmetry-related LHCII luteins around the conserved A and B helices (Liu et al. 2004; Paulsen et al. 1990) are replaced by violaxanthin molecules in VCP (Basso et al. 2014). Sequence analysis further suggests that VCP exhibits the conserved nucleus of pigments observed in the LHC superfamily, containing these two central carotenoids (Liu et al. 2004) and seven bound chlorophyll molecules (Carbonera et al. 2014).
In this manuscript, we present a resonance Raman study of VCP from N. oceanica in order to characterize pigment structure in this antenna protein at the molecular level. Resonance Raman spectroscopy provides selective vibrational information for each type of pigment bound to the protein (Chl-a and carotenoid molecules), according to the excitation wavelength used to produce the resonance effect (Ruban et al. 1995). As a vibrational method, resonance Raman yields information on the molecular configuration and conformation of these chromophores (Gall et al. 2015).
Materials and methods
Nannochloropsis oceanica cultivation and VCP purification have been described elsewhere (Keşan et al. 2016).
Room-temperature absorption spectra were measured using a UV-300 (Spectronic Unicam, Cambridge, UK) spectrophotometer with a 1.0-cm-pathlength cuvette. Measurement at 77 K was performed in an Optistat DN2 cryostat (Oxford Instruments, UK), on VCP mixed with 57% (v/v) glycerol. The spectrum was recorded using a Shimadzu UV-1800 spectrophotometer (Japan) with a 1.0-cm-pathlength cuvette.
Resonance Raman spectra were recorded at 77 K using an LN2 flow cryostat (Air Liquide, France). Laser excitations at 406.7 and 413.1 nm, and 476.5, 488.0, 501.7, 514.5, and 528.7 nm, were obtained with Coherent Kr+ (Innova 90) and Ar+ (Sabre) lasers, respectively. Output laser powers of 10–100 mW were attenuated to <5 mW at the sample. Scattered light was focused into a Jobin-Yvon U1000 double-grating spectrometer (1800 grooves/mm gratings) equipped with a red-sensitive, back-illuminated, LN2-cooled CCD camera. Sample stability and integrity were assessed based on the similarity between the first and last Raman spectra. Immersion in LN2 prevents thermal degradation, and the N2 stretch at 2326.6 cm−1 was used for calibration.
Results and discussion
The absorption spectra of the VCP complex at 77 K, and of violaxanthin and vaucheriaxanthin in n-hexane at room temperature, are displayed in Fig. 1. Violaxanthin presents a typical carotenoid absorption spectrum with three distinct bands, corresponding to the vibrational sublevels of the S0 → S2 transition (the position of the lower-energy 0–0 sublevel is usually used to describe the position of this transition). This transition energy depends on the effective conjugation length of the carotenoid, as well as on the polarizability of the local environment (Mendes-Pinto et al. 2013b). For violaxanthin (conjugation length 9 C=C), the S0 → S2 transition is at 470 nm in n-hexane, as previously reported (Llansola-Portoles et al. 2016). Vaucheriaxanthin has eight conjugated C=C double bonds, but it appears that the allene group present at one end of the conjugated chain affects its bond order somewhat, as its electronic absorption transition peaks at 472 nm in n-hexane, 2 nm redder than violaxanthin with 9 C=Cs. The absorption spectrum of VCP exhibits peaks corresponding to the Chl-a Soret band at 437 nm, and several bands corresponding to the S0 → S2 transitions of bound carotenoids—two distinct bands at 485 and 503 nm and a weaker, broad transition around 520–525 nm (more obvious in the second derivative; Fig. 1 inset). These absorption peaks have not yet been attributed to specific carotenoid species (violaxanthin or vaucheriaxanthin). In the red region, the intense Qy (S0 → S1) of Chl-a can be observed, with a major peak at 672 nm and a shoulder at 682 nm. Weaker vibronic satellites of this transition can be observed at 585 and 625 nm (Reimers et al. 2013).
Resonance Raman spectra of Chl-a molecules
Resonance Raman spectra contain bands which arise from those vibrational modes coupled to the electronic transition used to produce the resonance effect—in the case of Chl-a, this is usually the highest energy Soret band, to avoid interference from the intrinsic fluorescence (Lutz 1977). These modes often arise from vibrations delocalized on the Chl-a macrocycle, which are sensitive to its conformation. The modes in the low and middle frequency ranges (around 300 and 800–1500 cm−1, respectively) do not present shifts large enough so that they can be used for conclusive analysis of complexes containing as many chlorophylls as LHCs. In the higher-frequency region, a band around 1550 cm−1, attributed to vibrational modes of the chlorin ring (Lutz and Mäntele 1991), together with the methine bridge stretching mode around 1600 cm−1, is sensitive to the macrocycle core size and conformation (Näveke et al. 1997; Fujiwara and Tasumi 1986). The latter methine bridge mode has been widely used to assess the coordination of the central magnesium of Chls. It is located at 1600 cm−1 when the magnesium is six-coordinated (i.e. two external ligands), and at 1610–1615 cm−1 for five-coordinated magnesium (one external ligand). Stretching modes of conjugated C=O groups (keto carbonyls for Chl-a) are observed above 1640 cm−1 (Feiler et al. 1994). The frequency of this mode is at ca 1700 cm−1 when the C=O is free from interactions in non-polar environments (Lapouge et al. 1998). Intermolecular interactions downshift this mode by up to 40 cm−1, with the extent of the downshift being dependent on the strength of the interaction, while an increase in polarity of the C=O environment causes a smaller downshift (up to ~10 cm−1).
Figure 2 presents the high-frequency region of resonance Raman spectra at two Chl-a excitations, 413.1 and 406.7 nm, for VCP and the higher plant antenna proteins LHCII and CP29 (Pascal et al. 2000). The different C=O frequencies deduced from these spectra are shown in Table 1. In LHCII and CP29, modes sensitive to Chl conformation are localized around 1612 and 1550 cm−1, respectively, indicating that the central magnesium atoms of most if not all Chl-a molecules are five-coordinated in each case. These results were confirmed in the resolved crystal structure of each complex (Liu et al. 2004; Pan et al. 2011). In VCP, this region shows similar components at 1554 and 1612 cm−1. We may thus conclude that the central magnesium of VCP-bound Chl-a are all five-coordinated, and that the macrocycles of these molecules do not experience specific distortions as compared to relaxed, five-coordinated Chl-a. The envelope of bands in the 1650–1700 cm−1 range for VCP indicates a number of non-equivalent keto carbonyl modes. Two bands are observed corresponding to strongly H-bonded C=O’s at both excitation wavelengths, one at 1658 cm−1 which must be assigned to at least one Chl-a, and a more intense mode at 1664 cm−1—suggesting the contribution of three Chl-a at this latter frequency. In the central C=O region, a broad contribution is seen at 1680–1685 cm−1—this arises from at least two Chl-a, with C=O’s involved in medium/weak H bonds. Finally, the weak band vibrating at 1701 cm−1 indicates a single Chl-a, the carbonyl group of which is either weakly bonded or free from interaction in a non-polar environment. When comparing the two Chl-a excitations, resonance of the broad central region (corresponding to medium/weak interactions) is enhanced at 406.7 nm relative to that of the strongly H-bonded carbonyls. Hence, the distribution of frequencies and their relative intensities suggest that the presence of a minimum of 7–8 Chl-a in four different environments (see Table 1) is necessary to account for the Raman spectra. This is consistent with the presence in VCP of 7–8 Chl-binding amino acids in LHCII (Carbonera et al. 2014). The resonance Raman spectra obtained for LHCII, CP29, and VCP samples show quite a different distribution of frequencies, reflecting differences in environment of the bound Chl-a molecules, and more precisely in the H-bonding state of their conjugated keto carbonyls. As the H bonding to these groups was recently shown to have no influence on the absorption properties of the Chl-a molecules (Kish et al. 2016), and as no distortion of these molecules could be observed anywhere in the Raman spectra, we must conclude that the splitting in the absorption of the Qy of Chl-a in VCP must arise from excitonic interactions rather than from differences in pigment site energies.
Frequencies of Chl-a keto carbonyl stretching modes in CP29, LHCII, and VCP (in cm−1), measured by resonance Raman spectroscopy
Resonance Raman spectra of carotenoid molecules
Carotenoid resonance Raman spectra display four main groups of bands, denoted ν1–ν4, which can be satisfactorily modelled by DFT calculations (Pendon et al. 2005; Dokter et al. 2002; Macernis et al. 2015, 2014). The highest frequency ν1 band above 1500 cm−1 arises from stretching vibrations of C=C double bonds. Its position depends on the length of the π-electron conjugated chain and on the molecular configuration of the carotenoid (Rimai et al. 1973; Mendes-Pinto et al. 2013b). The ν2 band at 1160 cm−1 contains contributions from stretching vibrations of C–C single bonds coupled with C–H in-plane bending modes, and this region is a fingerprint for the assignment of carotenoid isomerization states (Koyama et al. 1988, 1983, 1982). The ν3 band at 1000 cm−1 arises from in-plane rocking vibrations of the methyl groups attached to the conjugated chain, coupled with in-plane bending modes of the adjacent C–H’s (Saito and Tasumi 1983). It was reported to be a fingerprint of the conjugated end-cycle configuration (Mendes-Pinto et al. 2013a; Llansola-Portoles et al. 2017), a hypothesis which is also supported by theoretical modelling (Macernis et al. 2014). Finally, the ν4 band around 960 cm−1 arises from C–H out-of-plane wagging motions coupled with C=C torsional modes (out-of-plane twists of the carbon backbone) (Rimai et al. 1973). When the carotenoid conjugated system is planar, these out-of-plane modes are not coupled with the electronic transition (which is oriented along the plane), and so they exhibit little resonance enhancement. However, distortions around C–C single bonds increase the coupling of (some of) these modes with the electronic transition, resulting in an increase in their intensity (Rimai et al. 1973; Lutz et al. 1987).
Figure 3a shows the 77 K resonance Raman spectra in the ν1 and ν2 regions for isolated violaxanthin and vaucheriaxanthin in n-hexane, excited close to the lower energy maximum of their S0 → S2 transition (476.5 nm). The ν1 frequency observed for violaxanthin is 1533.4 cm−1—taking into account that this band upshifts by 4–5 cm−1, when decreasing from room temperature to 77 K (Andreeva et al. 2011), this frequency corresponds very well to that expected for a 9 C=C carotenoid (Mendes-Pinto et al. 2013b). For vaucheriaxanthin, the ν1 peaks at 1533.0 cm−1—suggesting that this carotenoid has a slightly longer effective conjugation length, consistent with the small red shift in absorption (Fig. 1). In the ν2 region, the main peak does not occur at the same frequency for the two carotenoids—1163.4 cm−1 for vaucheriaxanthin compared to 1160.1 cm−1 for violaxanthin—and in addition there is a clear difference in shape and position for the satellite bands appearing at higher frequency. Figure 3b displays resonance Raman spectra in the same region for VCP at different excitations, from 488.0 to 528.7 nm. The ν1 frequency appears to be excitation dependent, appearing at 1535.0 cm−1 at 488.0 nm, and at 1531 cm−1 for excitations above 501.7 nm. This is an indication that different carotenoid populations contribute in each case. At 488.0 nm excitation, the ν1 band is narrow (FWHM ∼ 11 cm−1), indicating that only one population of carotenoid contributes to the spectrum. At 496.5 nm excitation, this band broadens (FWHM ∼ 15 cm−1) and is clearly composed of two different contributions. At all higher excitation wavelengths, this band is at the same frequency (1535.0 cm−1), suggesting that the same carotenoid population contributes over this whole excitation range. These results are in agreement with the two absorption peaks determined in the carotenoid region in VCP absorption spectra (Fig. 1), with one species of carotenoid absorbing at 485 nm and the other at 503 nm. These observations additionally show that the carotenoid absorbing at 503 nm has either a longer effective conjugation length than that absorbing at 488 nm, or is present in an environment with higher polarizability. In the ν2 region of VCP (Fig. 3b), the contributions of both VCP-bound carotenoids can be distinguished clearly. Comparing with spectra of the isolated carotenoids, the shoulder at 1198 cm−1 can be assigned to violaxanthin, while the 1183 cm−1 satellite can be assigned to vaucheriaxanthin. Note that, even though the bands are shifted in VCP relative to the corresponding bands of the isolated carotenoids, they are sufficiently separated to make this assignment. The 1198 cm−1 violaxanthin mode is only present at 488.0, 496.5, and 501.7 nm excitations, while the vaucheriaxanthin satellite at 1183 cm−1 is present for all excitations. As no new satellite bands are observed in this region for any excitation wavelength, we conclude that all VCP-bound carotenoids are in the all-trans configuration.
Figure 4a shows the 77 K resonance Raman spectra for violaxanthin and vaucheriaxanthin in the ν3 and ν4 regions (n-hexane, excitation 476.5 nm). For violaxanthin, ν3 consists of a single mode peaking at 1007.3 cm−1 with a shoulder at 1003.6 cm−1, whereas for vaucheriaxanthin it is a doublet (1002.7 and 1008.9 cm−1). The ν3 band mainly arises from in-plane rocking vibrations of the CH3 substituents along the conjugated chain (positions 9, 9′, 13, and 13′ for violaxanthin; see the molecular structure in Fig. 1). These CH3 groups are nearly equivalent in linear carotenoids, so that ν3 is a single peak as seen here for isolated violaxanthin. Splitting of this band has already been reported for allenic carotenoids (Llansola-Portoles et al. 2016; Premvardhan et al. 2009; Pascal et al. 1998). It was concluded that the presence of the non-equivalent CH3 at position nine, in close proximity to the allene group (see Fig. 1), lifts the degeneracy of this vibrational mode so that it appears as a doublet (Premvardhan et al. 2009). Figure 4b depicts VCP resonance Raman spectra in this region for different excitation energies. At 528.7 nm excitation, a clear doublet peak is observed with components at 1005.4 and 1011.3 cm−1, and their relative intensities are similar to those observed for isolated vaucheriaxanthin. At 514.5 nm excitation, the relative intensity of the higher-frequency peak increases, indicating a minor contribution from violaxanthin molecules. At 501.7 and 496.5 nm, this 1008.4 cm−1 band is even more intense, while at 488.0 nm only a single peak is observed at 1008 cm−1, indicating that violaxanthin contributions dominate. From these results, the following assignment in the absorption spectrum of VCP can be made: the red-most absorption around 525 nm corresponds to a vaucheriaxanthin pool, while the 485 and 503 nm peaks arise from two different violaxanthin pools. Using the relationship between the ν1 frequency and the position of the 0–0 electronic transition, we can conclude that the difference in absorption between these two violaxanthins populations arises from the different polarizability of their binding pockets (Mendes-Pinto et al. 2013b; Llansola-Portoles et al. 2016). The polarizability of the environment for each of these violaxanthins pools can be estimated using their ν1 frequencies (Llansola-Portoles et al. 2016; Mendes-Pinto et al. 2013a). The blue-shifted violaxanthin (485 nm; ν1 at room temperature estimated at 1530.5 cm−1) must be in a binding pocket with a polarizability of ~0.28 (close to the polarizability of toluene, ~0.29), whereas the red-shifted violaxanthin (503 nm; ν1 ~ 1527.5 cm−1) is lying in an environment with a polarizability of ~0.32. In the ν4 region, a resonance enhancement of the vibrational modes peaking at 957 and 968 cm−1 is observed when exciting at 528.7 nm. Since we have determined, from analysis of the ν1 region, that vaucheriaxanthin contributions dominate at 528.7 nm, it is safe to conclude that these enhanced modes correspond to this allenic carotenoid. These modes are also observed, to a decreasing extent, at 514.5 and 501.7 nm. At the latter excitation (501.7 nm), an additional, dominant peak at 966 cm−1 (also visible at 514.5 nm) can be assigned to the reddest violaxanthin. At 496.5 nm, the 966 cm−1 mode still dominates the ν4 envelope, while at 488.0 nm excitation, where the major contribution must come from the blue-absorbing violaxanthin, only two broad signals are visible, corresponding in the main to post-resonance contributions from the other two carotenoids (vaucheriaxanthin and redder violaxanthin). This suggests that the blue-absorbing violaxanthin is in a relaxed configuration, with only minor contributions in ν4. Thus, the vaucheriaxanthin molecule and the red-absorbing violaxanthin both display specific distortions, while the blue-absorbing violaxanthin is relatively undistorted.
From the analysis of the Chl-a conjugated C=O frequencies, we estimate the total number of Chl-a molecules bound to VCP is at least seven, each of which exhibits a relaxed macrocycle containing a five-coordinated central Mg atom. We further conclude that VCP binds two populations of all-trans violaxanthin molecules and one population of all-trans vaucheriaxanthin. Following this model, the simplest conclusion is that VCP binds two molecules of violaxanthin, displaying different absorption properties and one population of vaucheriaxanthin with very red-shifted absorption transition. That ratio is however quite different from that estimated in the literature, ~1.35 (Keşan et al. 2016), which rather suggests a 3:2 violaxanthin:vaucheriaxanthin stoichiometry (some of the violaxanthin being more loosely bound, and partially lost during the purification process). In this case, as we observe only one population of vaucheriaxanthin, both of these molecules would absorb at 525 nm, possibly because of carotenoid–carotenoid excitonic interactions, as already proposed for other LHC from algae containing allene carotenoids (Llansola-Portoles et al. 2016). The third violaxanthin, necessarily loosely bound as partially lost during purification, would absorb at 485 nm, the strongest transition of the complex. However, this 1.5 carotenoid ratio leads us to a total number of 11 Chl-a, a stoichiometry closer to that observed in LHCII and CP29, but with four chlorophyll proteins coordinated with undetermined amino acids (Carbonera et al. 2014), as well as with a large number of degenerate bands in the carbonyl stretching region of the resonance Raman spectra. In any case, we conclude that two populations of violaxanthins are strongly bound to VCP, with red-shifted absorption transitions (485 and 503 nm) with respect to their absorption in n-hexane, and the most red-absorbing violaxanthin displays a distorted conformation. This situation is reminiscent of the two central luteins in plant LHCII, which also display different absorption and conformation (Ruban et al. 2000). However, the two luteins in LHCII lie in the same polarizability environment of 0.32, and the absorption shift between these two luteins is ensured by a change in the configuration of their conjugated end cycle (Mendes-Pinto et al. 2013a). Here in VCP, the two violaxanthin populations also differ in absorption but the analysis of their C=C stretching mode frequency indicates that this is due to different local polarizability (Mendes-Pinto et al. 2013b). It is striking that two different molecular mechanisms result in these two complexes in a similar difference in the energy transitions of the two main carotenoids. It is also worth noting, in this respect, that violaxanthin does not possess conjugated β-rings (see Fig. 1), and so its absorption transition could not be red-shifted by the same mechanism as for lutein in LHCII. Finally, the absorption of VCP-bound vaucheriaxanthin is significantly red-shifted (525 nm, compared to 472 nm in n-hexane, i.e. more than 2000 cm−1). This situation is also reminiscent of another protein from the LHC family, FCP, in which a population of fucoxanthin (another allenic carotenoid) displays an absorption shifted far into the red. It is probable that N. oceanica, like diatoms and brown algae, evolved to absorb light in the green-yellow region of the visible spectrum because of competition from other photosynthetic species in the aquatic environment.
This work was supported by the ERC funding agency (PHOTPROT project) and The French Infrastructure for Integrated Structural Biology (FRISBI). The research in the Czech Republic was supported by the Czech Science Foundation Grant 14-01377P and by institutional funding RVO:60077344.
- Basso S, Simionato D, Gerotto C, Segalla A, Giacometti GM, Morosinotto T (2014) Characterization of the photosynthetic apparatus of the Eustigmatophycean Nannochloropsis gaditana: evidence of convergent evolution in the supramolecular organization of photosystem I. Biochim Biophys Acta 1837(2):306–314CrossRefPubMedGoogle Scholar
- Koyama Y, Kito M, Takii T, Saiki K, Tsukida K, Yamashita J (1982) Configuration of the carotenoid in the reaction centers of photosynthetic bacteria. Comparison of the resonance Raman spectrum of the reaction center of Rhodopseudomonas sphaeroides G1C with those of cis–trans isomers of β-carotene. Biochim Biophys Acta 680(2):109–118CrossRefGoogle Scholar
- Koyama Y, Takii T, Saiki K, Tsukida K (1983) Configuration of the carotenoid in the reaction centers of photosynthetic bacteria. 2. Comparison of the resonance Raman lines of the reaction centers with those of the 14 different cis–trans isomers of β-carotene. Photobiochem Photobiophys 5:139–150Google Scholar
- Koyama Y, Takatsuka I, Nakata M, Tasumi M (1988) Raman and infrared spectra of the all-trans, 7-cis, 9-cis, 13-cis and 15-cis isomers of β-carotene: key bands distinguishing stretched or terminal-bent configurations form central-bent configurations. J Raman Spectrosc 19(1):37–49CrossRefGoogle Scholar
- Lutz M, Mäntele W (1991). In: Scheer H (ed) The chlorophylls. CRC Press Inc., Boca Raton, pp 855–902Google Scholar
- Lutz M, Szponarski W, Berger G, Robert B, Neumann J-M (1987) The stereoisomerization of bacterial, reaction-center-bound carotenoids revisited: an electronic absorption, resonance Raman and 1H-NMR study. Biochem Biophys Acta 894:423–433Google Scholar