Origin of the 701 nm fluorescence emission of the Lhca2 subunit of higher plant Photosystem I

Photosystem I of higher plants is characterized by red-shifted spectral forms deriving from chlorophyll chromophores. Each of the four Lhca1 to -4 subunits exhibits a specific fluorescence emission spectrum, peaking at 688, 701, 725, and 733 nm, respectively. Recent analysis revealed the role of chlorophyll-chlorophyll interactions of the red forms in Lhca3 and Lhca4, whereas the basis for the fluorescence emission at 701 nm in Lhca2 is not yet clear. We report a detailed characterization of the Lhca2 subunit using molecular biology, biochemistry, and spectroscopy and show that the 701-nm emission form originates from a broad absorption band at 690 nm. Spectroscopy on recombinant mutant proteins assesses that this band represents the low energy form of an excitonic interaction involving two chlorophyll a molecules bound to sites A5 and B5, the same protein domains previously identified for Lhca3 and Lhca4. The resulting emission is, however, substantially shifted to higher energies. These results are discussed on the basis of the structural information that recently became available from x-ray crystallography (Ben Shem, A., Frolow, F., and Nelson, N. (2003) Nature 426, 630-635). We suggest that, within the Lhca subfamily, spectroscopic properties of chromophores are modulated by the strength of the excitonic coupling between the chromophores A5 and B5, thus yielding fluorescence emission spanning a large wavelength interval. It is concluded that the interchromophore distance rather than the transition energy of the individual chromophores or the orientation of transition vectors represents the critical factor in determining the excitonic coupling in Lhca pigment-protein complexes.


Introduction
Photosystems I and II have a common general organization including a core complex moiety, binding Chl a and β-carotene, and an outer antenna moiety, binding Chl a, Chl b and xanthophylls. Nonetheless, their spectroscopic properties are substantially different. Both compartments of the PSI supercomplex are enriched in red-shifted chlorophyll absorption forms, while those of PSII are not. This property of PSI results in a preferential light harvesting in the near-infrared spectral region (>700 nm) and causes the so-called "Emerson effect" that opened the way to the discovery of the two photosystems (1). In green plants, LHCI, the outer antenna of PSI, is the most red-forms enriched compartment and is composed of four different complexes, the products of the genes Lhca1-4 (2). These complexes are located on one side of the core complex (3;4) and can be isolated as dimers (5). The Lhca proteins belong to the Lhc multigenic family and they share high sequence homology with the antenna complexes of PSII (6). Despite these strong similarities, the spectroscopic properties of Lhca and Lhcb complexes are very different. In particular, the emission spectrum of LHCI peaks at 733 nm, more than 50 nm red-shifted as compared to the emission spectra of the antenna system of PSII (LHCII) (7). In vitro reconstitution has shown that the red shifted Chls are a common characteristic of Lhca complexes, although their energies differ strongly in different subunits. Emission forms at 701-2 nm have been found associated to Lhca1 and Lhca2, while 725 and 733 nm emissions were reported for Lhca3 and Lhca4 respectively (5;8-11). It has been shown that the red emission of Lhca3 and Lhca4 originates from the low energy absorption band of an excitonic interaction involving Chl a in sites A5 and, probably, B5 (12) while the difference in energy is associated with the nature of the Chl A5 ligand, since the substitution of Asn in Lhca3 and Lhca4 by His, the usual ligand for Chl A5 in Lhc complexes, leads to the loss of the red spectral forms (12). In the present work we analyze in detail Lhca2 with the aim of understanding the origin of the 701 nm emission which dominates the fluorescence spectrum at low temperature. It should be noticed that Lhca bears His as a ligand for Chl A5 and yet its fluorescence emission is red shifted by 20 nm with respect to the highly homologous subunit Lhcb6 (CP24), and other Lhcb proteins, thus opening the question whether a different mechanism of modulation for the physico-chemical properties of chromophores is at work in Lhca2. By performing a detailed analysis by biochemical and spectroscopic methods on WT and mutant Lhca recombinant proteins we conclude that the fluorescence emission band at 701 nm derives from excitonic coupling between chromophores A5 and B5, in spite of the presence of a His ligand for Chl A5. These results are discussed on the basis of the structural information that recently became available from X-ray crystallography (1;13). We suggest that in green plant LHCI differences in interchromophore distances rather than in transition energies of the individual chromophores of the interacting pair or in the orientation of their transition vectors represent the critical factor in determining excitonic coupling.

Experimental procedure
DNA constructions and isolation of overexpressed Lhca apoproteins from bacteria. cDNAs of corrected for the wavelength dependence of the detection system. CD spectra were measured at 10°C on a Jasco 600 spectropolarimeter. The wavelength sampling step was 0.5 nm, the scan rate 100 nm/min and spectra were recorded with 8 accumulations. The OD of the samples was 1 at the maximum in the Qy transition for all complexes and the samples were in the same solution as described for the absorption measurements. All spectra presented were normalized to the polypeptide concentration based on the Chl binding stoichiometry.
LD spectra were measured as described in (19) using samples oriented by the polyacrylamide gel squeezing technique.

Analysis of Lhca2-WT complex
Lhca2 WT was obtained by overexpression of the apoprotein in E.coli and reconstitution in vitro with pigments. The complex was purified from excess pigment and unfolded apoprotein by sucrose gradient ultracentrifugation, anion exchange chromatography and glycerol gradient ultracenrifugation, and was obtained in monomeric form, as previously described (11). while in the Chl b region two local maxima can be detected at 644 and 652 nm (Fig. 1). The second derivative analysis allows resolution of several absorption forms: three Chl b at 643, 651 and 653 nm and four Chl a at 661, 672, 680.5 and 688.5 nm. In the blue region, the minimum at 497 nm is associated to the red most peak (S 0 > S 2,0 transition) of the carotenoids, while those at 480 and 471 nm may represent the Soret bands of two Chl b spectral forms. (suggested location figure 1)

Fluorescence measurements
Fluorescence emission spectra of Lhca2 were measured at different temperatures upon excitation in the blue spectral region (Fig. 2). Two emission forms at 688 and 701 nm can be detected. The latter dominates the spectrum at 4K, but the 688 nm form is still present, indicating that the energy transfer between these two forms is not complete, thus suggesting some kind of heterogeneity in the The site-selected fluorescence data allow also determining the peak wavelength of the absorption form responsible for the 701 nm emission. This value corresponds to the excitation wavelength at which the emission peak coincides with that in spectra obtained from non selective excitation (see (20) for more details). By following this procedure the absorption maximum of the red band is calculated to be approximately at 690 nm, in agreement with the second derivative analysis of the absorption spectrum which showed the red-most absorption form at 688.5 nm (see Fig. 1).
Additional information on the characteristics of the red absorption tail can be obtained by fluorescence anisotropy measurements, which gives an indication about the orientation of the absorbing and emitting dipoles. In figure 3B (circles) the values of the fluorescence anisotropy detected at 701 nm are reported as a function of the excitation wavelength. It is clear that the values of the anisotropy also depend on the excitation wavelength. Upon excitation around 680 nm the anisotropy value is low and it rises to a maximum of 0.32 for excitation wavelength above 697 nm, a value very similar to that obtained in the analysis of the 702 nm emission form in a native LHCI preparation (7). This result indicates that above this wavelength no depolarization due to energy transfer takes place, thus suggesting that the same band is responsible for the absorption and the emission. The alternative hypothesis of energy transfer between pigments having the same dipole transition orientation is extremely unlikely.

LD spectra
The LD spectrum of Lhca2 shows strong positive contribution in the Chl a Qy absorption region, with maximum at 679.5 nm and a shoulder around 690 nm (Fig. 4). The signal becomes slightly negative at 667 nm. In the Chl b absorption region, a positive contribution is observed at 653 nm.

Localisation of the low energy absorbing form(s) in to the Lhca2 structure
The above measurements strongly indicate that the red-shifted fluorescence signal at 701 nm is associated with an absorption form peaking at 690 nm. In order to identify which chromophore is responsible for this absorption, mutation analysis at selected Chl-binding residues was performed.
Based on previous analysis with Lhca3 and Lhca4 (12), residues coordinating Chl A5, B5 and B6, involved in the low energy absorption forms, were mutated to non-binding residues (i.e. which can not coordinate the central magnesium of the Chls (see table I

Discussion
The current results show that Lhca2 takes a unique position within the LHCI proteins.
Unlike Lhca3 and Lhca4, it does not show strong absorption above 700 nm, but nevertheless its emission is still 22 nm red shifted as compared to the emission maxima of all Lhcb complexes. In the case of Lhca3 and Lhca4 it was shown that the red absorption is associated with the presence of an Asn as ligand for Chl A5 (12). This is not the case of Lhca2 which has an His in this position, a feature shared with all Lhcb proteins that do not show red-shifted Chl forms. We thus proceeded with the analysis of the chromophore organization in Lhca2 and identified the absorption band responsible for the 701 nm emission band, with the aim of understanding the mechanism which allows modulation of the physico-chemical properties of chlorophyll ligands.
Lhca2 has been shown to have the emission maximum at 701 nm (at 4K). A second emission form was also detected at 688 nm in the 4K spectrum, thus suggesting heterogeneity in the sample.
Several sources of spectroscopic heterogeneity have been reported for Lhc proteins, including phosphorylation-induced spectral changes in CP29 (25), presence of multiple gene products in the LHCII preparations with different spectral features (26), and conformational changes induced by zeaxanthin binding to site L2 (27;28). In the present work, a recombinant Lhca2 preparation was analyzed which was the product of a single gene, did not contain zeaxanthin, was monomeric as detected by sucrose gradient ultracentrifugation and was not phosphorylated. We conclude that the persistence of the two fluorescence emission bands is the product of a previously unrecognized source of molecular heterogeneity which is reflected in the establishment or not of the molecular feature responsible for the 701 nm emission. Two hypotheses can be proposed for explanation of these data: either the Lhca2 protein is present into two conformations in which the two interacting chromophores are set at different distances and/or different orientations with respect to each other, or one of the two sites can be occupied by either Chl a or Chl b. The difference in transition energies between Chl a and Chl b and the difference in the orientation of the transition vectors (29) would make the observable spectral shift of a Chl a/Chl b interaction much smaller with respect to that of a Chl a/Chl a interaction. The likelihood of these hypotheses will be discussed below.
Site-selected fluorescence measurements indicate that the 689 nm emission originates from bulk Chls absorbing around 682 nm, while the 701 nm emission form is associated with a Chl absorption form at 690 nm. This result is confirmed by the mutation analysis which showed that both mutants loosing the 701 nm emission also lack the 690 nm absorption form.

Spectral characteristics of the 690 nm absorption form
In order to get more details on the 690 nm absorption form of Lhca2, the absorption spectrum of the The 690 nm absorption form of Lhca2 is thus characterized by a larger value of both homogeneous and inhomogeneous broadening as compared to the bulk Chl absorption form for which these two values are both around 100 cm -1 (30). In this respect the 690 nm band of Lhca2 shows characteristics very similar to the red absorption forms observed in several organisms (20), although its energy is significantly higher.

The origin of the 701 nm emission
Large values of optical reorganization energy are usually associated with the presence of pigmentpigment interactions. It has already been suggested that the red forms of Lhca3 and Lhca4 are the low energy term of an interaction between two Chl a molecules. Mutational analysis shows that these two chromophores are bound to sites Chl A5 and Chl B5, since mutation at these sites completely abolishes the red emission forms. This conclusion is in agreement with the finding that site A5 binds Chl a in Lhca as well as in any other Lhc protein (22)(23)(24), while site B5 either binds Chl a, or has a mixed Chl a -Chl b occupancy as derived from pigment analysis of WT and mutant protein as well as their spectroscopic analysis. We conclude that the low energy spectral forms originate from the same protein domain in all Lhca complexes and involves Chls A5 and B5.
Nevertheless, the energy level of these red-shifted transitions is modulated to a different extent in each Lhca subunit. (suggested location figure 9) The presence of pigment-pigment interaction inducing the absorption at 688-690 nm of Lhca2 is supported by the comparison of the CD spectra of WT and A5 mutant, which lacks a negative contribution around 687 nm and a positive one at 672 nm (values at RT). Similar components can be detected in the WT minus A5 absorption and LD difference spectra ( fig. 9). In both cases two absorptions, at 688-689 nm and at 672-674 nm, can be detected, representing Chl forms lost by effect of the mutation at site A5 and having positive LD spectra. These data can be interpreted as the loss of an excitonic interaction having the low energy band at 688-690 nm and the high energy band at 672-674 nm, as the effect of A5 mutation. To describe both difference spectra a negative band peaking at 679 nm is required. This band is present in the mutant but not in the WT and shows a positive LD signal (Fig 6). Considering that all spectroscopic techniques consistently indicate that the interaction leading to the 690 nm absorption involves two Chl a molecules, and that in the A5 mutant only one Chl a molecule is lost, the gain in the absorption can be interpreted as the contribution of the non-interacting monomer which is left alone in the A5 mutant. Based on the mutation analysis, which suggested that the two interacting Chls are accommodated in sites A5 and B5, we conclude that the "new" 679 nm form detected in WT minus A5 difference spectra represents the absorption of the monomeric Chl a in site B5. This is in agreement with the analysis of the homologous protein CP29 (22) showing that a non-interacting Chl a absorbing at 679 nm is bound to site B5. Lhca4 has been shown to be 7.9 Å, while in Lhca2 the value is 10.4 Å (3). This is consistent with the proposal that in Lhca4 the presence of the Asn as a ligand for Chl A5 allows a shorter distance between the two interacting Chls as compared to the complexes which have His for Chl A5 ligation (12).