Abstract

Results

To isolate the OCP, C-terminal His-tagged OCP mutants were constructed in Synechocystis PCC 6803, using kanamycin or spectinomycin resistance for mutant selection [supporting information (SI) Figs. S1 and S2]. In the WT and the kanamycin resistant mutant, blue-green light induced similar fluorescence quenching, indicating that the presence of the His-tag did not affect the activity of the OCP. In the spectinomycin resistant mutant, the fluorescence quenching was slightly decreased relative to the WT (Fig. 1A); this correlates with a lower concentration of the OCP in this mutant (probably because of a destabilization of the mRNA) (Fig. 1B). In addition, a mutant was constructed in which the slr1963 gene containing a C-terminal His-tag was expressed using the psbA2 promoter region as the promoter. In this strain, large quantities of the OCP were present (see Fig. 1B) and a very large fluorescence quenching was observed (see Fig. 1A), confirming the relationship between OCP concentration and excitation energy quenching. The OCP was isolated using affinity Ni-chromatography followed by an ion-exchange column (Fig. S3). The absorption spectrum of the His-tagged OCP isolated from Synechocystis 6803 is essentially identical to that of A. maxima (compare Fig. 2B dark form with the spectrum in Fig. 2 in ref. 20). As in the latter strain, it mainly binds hECN (Fig. S4).

Open in a separate window

Fig. 1.

Relationship between blue-green light induced NPQ and the OCP in whole cells. (A) Decrease of maximal fluorescence (Fm') during exposure of WT (red squares), over-expressing OCP mutant (blue circles), His-tagged OCP/K_m resistant mutant (green romboids) and His-tagged OCP/Sp resistant mutant (black triangles) cells to 740-μmol photons m⁻² s⁻¹ of blue-green light (400–550 nm). The graph is the average of four independent experiments and the error bars show the maximum and the minimum fluorescence value for each point. The cells were diluted to 3-μg chlorophyll/ml. (B) Coomassie blue-stained gel electrophoresis and immunoblot detection (bottom panel) in membrane-phycobilisome fractions from His-tagged OCP/Sp resistant mutant (1), His-tagged OCP/K_m resistant mutant (2), WT (3), and over-expressing OCP mutant (4). The arrow indicates the OCP. Each lane contained 1.5-μg chlorophyll. (C) OCP^r is present in whole cells under NPQ-inducing conditions. Difference light-minus-dark absorbance spectrum of the isolated OCP (red) compared with the double difference absorbance spectrum: light-minus-dark overexpressing OCP cells spectrum minus light-minus-dark ΔOCP cells [strain without OCP (9)] (average of five independent experiments). The illumination of cells and spectra recording were realized at 11°C.

Open in a separate window

Fig. 2.

Isolated OCP is responsive to blue-green light : Photoconversion and dark recovery. (A) Photograph of isolated OCP° and OCP^r. To obtain OCP^r the isolated protein was illuminated with a blue-green light at 740-μmol photons m⁻² s⁻¹ at 12°C for 2 min. (B) Absorbance spectra of the dark orange form (OCP°; black) and the light red form (OCP^r; red). (C) Darkness red OCP^r to orange OCP° conversion (decrease of the absorbance at 580 nm) and (D) photoconversion from the OCP° to the OCP^r form using a 350-μmol photon m⁻² s⁻¹ blue-green light intensity (increase of the absorbance at 580 nm) at different temperatures: 32°C (violet), 28°C (rose), 24°C (blue), 19°C (green), 15°C (red), and 11°C (black). (E) OCP^r accumulation at 11°C and different light intensities: 20 (black), 50 (red), 120 (green), 210 (blue), 350 (rose), 740 (orange), and 1,200 (violet)-μmol photons m⁻² s⁻¹ of blue-green (400–550 nm). The protein concentration was OD 0.2 at 495 nm.

In darkness or dim light the OCP appears orange (OCP°) and its absorption spectrum presents a typical carotenoid shape, reflecting the strongly allowed S₀-S₂ transition with three distinct vibrational bands; the 0–0 vibrational peak is located at 496 nm (see Fig. 2 A and B). The position and shape of this transition corresponds to that of hECN locked into an all-trans conformation by the surrounding protein (20), in agreement with the crystallographic structure (19). Strikingly, upon illumination with blue-green light (400–550 nm) at 10°C, the OCP° is completely photoconverted to a red form, OCP^r. The red-shifted spectrum of OCP^r with a maximum at 500 nm loses the resolution of the vibrational bands (see Fig. 2 A and B). To elucidate whether the conversion of the OCP° to OCP^r occurs in whole cells under NPQ-inducing conditions, absorbance spectra of whole cells of the overexpressing OCP strain and the ΔOCP strain, lacking the OCP (9), were measured before and after illumination with high intensities of blue light at 11°C and the light-minus-dark difference spectra calculated. Our results strongly suggested that OCP^r is accumulated in vivo under the conditions inducing energy dissipation and fluorescence quenching (Fig. 1C).

In darkness, OCP^r spontaneously reverts to the orange form (Fig. 2C). This step is not accelerated by illumination (Fig. S5). The back-reaction kinetics (red to orange) shows a large temperature dependence, the half-time of recovery varying from 30 sec at 32°C to ≈45 min at 11°C. In vivo, the recovery NPQ kinetics also show a large temperature dependence; however, the kinetics are slower (8, 9, 14) than the OCP^r to OCP° dark conversion, suggesting that the OCP^r form is more stable in vivo than in vitro or that the fluorescence quenching remains longer than the OCP^r form. In contrast, the initial rate of the light photoconversion was temperature-independent (Fig. 2D). Thus, at a fixed light intensity, the steady-state concentration of OCP^r depended on the temperature. This is reflected in vivo, where a larger NPQ was observed at 7°C than at 33°C (Fig. S6). The kinetics of OCP^r formation strongly depended on light intensity (Fig. 2E). This correlates with the observed dependency on light intensity of the blue-light induced NPQ in whole cells (9).

Collectively, these results strongly suggested that OCP^r is the active form of the OCP and that pigment-protein interactions are critical to this activity. To confirm this, a C-terminal His-tagged W110S-OCP mutant and an overexpressing C-terminal His-tagged W110S OCP mutant were constructed and characterized. High intensities of blue-green light induced no fluorescence quenching in the mutant (Fig. 3B), indicating that Trp-110 is critical to the OCPs photoprotective function. This was substantiated by construction of an overexpressing W110S-OCP mutant in which very little quenching was induced, although large quantities of OCP were present (see Fig. 3 B and C). The absorbance spectrum of the isolated mutated OCP was similar to that of the WT OCP (Fig. 3D). Moreover, each isolated mutant OCP contains a carotenoid molecule as judged by comparison to WT OCP (similar A₄₆₇/A₂₈₀ ratio, data not shown). High intensities of blue-green light converted only a very small portion of the OCP° to OCP^r, even after 15 min of illumination at 10°C (see Fig. 3D). Thus, there is a direct correlation between the accumulation of OCP^r and the induction of the fluorescence quenching.

Resonance Raman spectroscopy was used to observe light-induced structural changes of the hECN. The Raman spectra of the OCP° from A. maxima and Synechocystis PCC 6803 are nearly identical, indicating that the carotenoid-peptide interactions in these two proteins are remarkably homologous (Fig. 4A). Thus, the X-ray structure from the former strain may be used as a model for the OCP of the latter. This has been confirmed by the determination of the 1.65 Å structure of the Synechocystis PCC6803 OCP (C.A.K., unpublished results). Before illumination, the resonance Raman spectra of OCP-bound hECN shows that this carotenoid is in an all-trans conformation, in good agreement with the X-ray structure (19). The relatively high intensity of the bands in the ν₄ region, however, indicates that this conformation is not planar, but highly distorted around the C-C bonds (21). Upon illumination the position of C = C stretching mode shifts from 1,528 cm⁻¹ to 1,521 cm⁻¹, indicating that the apparent conjugation length of hECN increased by about one conjugated bond (22), in agreement with the observed red-shift (see Fig. 4A). Trans-cis isomerisation of hECN should shift down the position of this band, and should result in the appearance of new bands in the spectra because of the change in the molecular symmetry. As this is not observed, we can conclude that hECN is still all-trans in its red form. Of course, we cannot formally discard presence of end chain isomerisation (such as 7-cis), whose effect on the Raman spectra (and on the electronic absorption spectra) is limited. The decrease of intensity of the 950 cm⁻¹ indicates that upon photoconversion, hECN reaches a less distorted, more planar structure (22). Similar differences were observed using the OCP from A. maxima (data not shown). Further studies will be necessary to elucidate the specific nature of the changes in the carotenoid.

Open in a separate window

Fig. 4.

Changes in the carotenoid and the protein upon illumination. (A) Resonance Raman spectra of Synechocystis PCC 6803 OCP^r and OCP° and Arthrospira maxima OCP° in the 1,740 to 820 cm⁻¹ range and (B) light-minus-dark FTIR spectra of OCP in H₂O (black) and in D₂O ≈5 (red line) and 50 (blue line) seconds after illumination with a blue LED in the carbonyl region between 1,730 and 1,600 cm − ¹ (Amide I region). The OCP^r minus OCP° Amide I differential pattern is assigned to H-bond loosening (upshift) of the loops and α-helices carbonyl modes together with a H-bond strengthening (downshift of carbonyl modes) of the β-sheet.

Light-induced Fourier transform infrared (FTIR) difference spectroscopy was applied to probe the light-induced structural changes in the protein of the OCP. Upon illumination, large changes in the OCP FTIR spectra can be observed ≈1,650 and 1550 cm⁻¹, where the Amide I and Amide II protein modes are expected to contribute. In the light-minus-dark FTIR spectra, these protein contributions, with specific frequencies for each type of protein secondary structure (loops, α-helix, and β-sheet) (23), may be formally attributed to the protein modes on the basis of their sensitivity to D₂O exchange (hECN molecules contain only one, nonconjugated, exchangeable proton) (Fig. 4B). The differential signal at 1,677(−)/1,697(+) cm⁻¹ is assigned to portions of the protein without secondary structure: that is, the loops. The α-helix and β-sheet carbonyl region shows an intense and broad negative band centered at 1,643 cm⁻¹, which has a corresponding induced absorption at 1,663(+) and 1,618(+)cm⁻¹ assigned to α-helix and β-sheet, respectively. Light-induced Amide I changes in the OCP are large, suggesting a major rearrangement of the protein backbone. An up-shift of the Amide I vibrations of α-helix corresponds to a weakening of the NH − C = O hydrogen bonds that connect the turns, indicating a less rigid helical structure in a significant part of OCP upon formation of the red form. A down-shift of the Amide I vibrations of β-sheet corresponds to a shortening (ie, a strengthening) of the NH − C = O H-bonds between the β strands, resulting in a compaction of the β-sheet domain upon formation of the red form. The Amide I frequency-shift pattern observed here for OCP is strikingly similar to that observed in the LOV2 domain of phototropin, a flavin-binding blue-light photoreceptor (24).

Ultrafast spectroscopy was applied to examine the primary photochemistry of the OCP and the efficiency of product formation. Fig. 5A shows the result of a global analysis of the time-resolved data. Four kinetic components are required for an adequate fit. The 100-fs component corresponds to internal conversion of the initially excited, optically allowed, S₂ state of the carotenoid to the low-lying, optically forbidden S₁ state coupled to an intramolecular charge-transfer (ICT) state. The 1-ps and 4.5-ps components correspond to decay of the S₁ and ICT states, respectively (20). After decay of the hECN excited states on a ps timescale, a photoproduct remains at very low amplitude (Fig. 5B). The spectral signature of this product species is similar to that of the light-minus-dark spectrum of the OCP (see Fig. 5A). However, the primary product is slightly red-shifted with respect to that of the light-minus-dark, indicating that these states may be similar but not identical and that relaxation takes place on timescales longer than nanoseconds. The amplitude of the product state at 565 nm is only 1% of that of the original carotenoid bleach near 475 nm immediately after excitation (see Fig. 5A), consistent with the assumption that the OCP is a low quantum-yield photoactive protein.

Open in a separate window

Fig. 5.

Primary photochemistry of OCP revealed by femtosecond spectroscopy. OCP was excited with 50-fs flashes at 475 nm and the absorbance changes were monitored between 400 and 675 nm at time delays up to 3 ns. (A) Global analysis of time resolved data in the form of Evolution-Associated Difference Spectra (EADS). Four kinetic components with time constant of 100 fs (black), 1 ps (red), 4.5 ps (blue), and a long-lived component (green, expanded 30×). (B) Kinetic trace recorded at 565 nm (symbols) and the result of the global analysis (solid line). After decay of carotenoid excited-state signals on the picosecond timescale, a nondecaying photoproduct is formed at low yield. Note that the time axis is linear up to 10 ps and logarithmic thereafter. The vertical axis is split into two linear axes to show the long-lived product.

Discussion

Previously, we demonstrated that the OCP is specifically involved in a photoprotective phycobilisome-associated NPQ mechanism (9). Here, we show that the OCP is a photoactive protein sensing light intensity. Light induces the conversion between a dark stable orange form (OCP°) and a metastable stable red form (OCP^r). The accumulation of the OCP^r is essential for the induction of the photoprotective mechanism. The OCP^r is present in whole cells under illuminations inducing the NPQ mechanism. Moreover, high intensities of blue-green light induced almost no fluorescence quenching in an OCP mutant in which the OCP photoconvertion was nearly abolished. The OCP° to OCP^r conversion occurs with a very low quantum yield (≈0.03); this is likely to be because the OCP protein is involved in a photoprotective process that must be induced only under high light conditions. Moreover, the ratio of the two forms under illumination depends on light intensity and temperature: more red active OCP^r at high light intensities and low temperatures. This provides a mechanism for the protein to sense photoinhibitory conditions: high light and light plus cold. Indeed, in vivo, greater fluorescence quenching is induced by increasing the light intensity and by lowering the temperature.

Retinal, a carotenoid derivative, is the chromophore of rhodopsins, a family of photoactive membrane-embedded proteins that are used throughout the animal and microbial kingdom as light receptors (25). Because retinal and hECN share structural similarities, it is worth considering whether OCP photoactivation could be similar to that of rhodopsin. However, our results suggest a completely different mechanism. Retinal is covalently bound to the protein by a protonated Schiff-base linkage to a Lys residue. The hECN in the OCP is close to an absolutely conserved Arg, but there is no covalent bond (19). Light induces an isomerisation of the retinal initiating the photocycle involving proton transfer between the Schiff-base and a carboxylate residue of the rhodopsin. In the OCP, no isomerisation was detected. Thus, although the specific changes in the carotenoid of OCP remain to be elucidated, they are clearly different from those occurring in retinal.

Light-driven structural changes in the carotenoid correlate with the protein conformational changes observed by FTIR; the structural rearrangement in the OCP may be a precondition for interaction with the phycobilisome or other proteins. Collectively, our data suggest a mechanism analogous to that of other blue-light responsive proteins, such as PYP and LOV domain-containing proteins. In each of these, blue light causes changes in the chromophore that induces disruption of hydrogen bonding networks and large conformational changes in the protein required for the light-responsive function (26, 27). By analogy, we posit that the absorption of blue light by the OCP alters the strength of one or both of the hydrogen bonds between the carotenoid and the protein. The resulting conformational changes in the protein could render the carotenoid more accessible and form a signal propagation pathway from the carotenoid to the surface of the OCP.

The C-terminal domain of the OCP (between residues 196 to 296) is structurally similar to the 7.8 kDa core linker protein (Noam Adir, personal communication; Fig. 6A), although lacking significant sequence homology. The linker protein consisting of a three-stranded β-sheet and one α-helix was cocrystallized with allophycocyanin (APC) trimers that constitute part of the core of the phycobilisome (28). The structure revealed that the linker protein is bound within the central aperture of the APC trimer and directly interacts with the β84 chromophores of two APC subunits. Therefore, we hypothesize that the C-terminal OCP domain, by interacting with the center of an APC trimer, may bring the carotenoid into proximity of the APC chromophores. In addition, the Trp-290 and Tyr-203 residues, which interact with the carbonyl of the carotenoid, belong to the central strand of the β-sheet and to the α-helix of the C-terminal domain, respectively (Fig. 6B). Modifications in this domain caused by light-induced carotenoid changes could then possibly regulate the interaction between the carotenoid and the APC chromophores

Open in a separate window

Fig. 6.

Structural similarity between the OCP C-terminal domain and the core linker protein APL_C ^7.8. (A) Superposition of the three strand β-sheet and the α-helix of the linker protein (blue) and of the C-terminal domain of the OCP (red). (B) The OCP C-terminal domain (red) showing the carotenoid and the Y203 and W290 residues (green stick representation) superimposed onto the linker protein structure (blue).

Carotenoid molecules are able to quench excitation energy from tetrapyrrole molecules in vitro through energy transfer to their optically forbidden S₁ state, coupled to an intramolecular charge-transfer state (29). It was recently proposed that this mechanism is at the origin of the NPQ process in higher plants (7). Hydroxyechinenone, which contains a conjugated carbonyl group, clearly exhibits such an intramolecular charge transfer state (20). Therefore, we hypothesize that the OCP is not only the sensor of light but also the fluorescence quencher. Absorption of light by the hECN induces conformational changes of the carotenoid, shifting the absorption spectra toward the red region of the visible spectrum. The red shift is likely necessary to tune the optically forbidden S₁ state of hECN to a position to allow the energy transfer from the phycobilisome to the OCP. This mechanism is reminiscent of the gear-shift mechanism proposed to explain the role of zeaxanthin in higher plant NPQ (30). Possibly, an increase of charge-transfer character of the hECN excited states upon formation of OCP^r plays a role as well.

Although the molecular mechanism of the orange to red conversion and the exact role of the OCP in the cyanobacterial photoprotective mechanism await elucidation, our results suggest several mechanistic hypotheses that open an unexplored frontier in research on the functions of carotenoids: this is a unique report of a carotenoid protein functioning as a photoactive protein sensing light intensity.

Materials and Methods

Supplementary Material

Acknowledgments.

Footnotes

References