Protoporphyrin IX

The pigment binding behaviour of water-soluble chlorophyll protein (WSCP)

Philipp Girr, Jessica Kilper, Anne-Christin Pohland and Harald Paulsen *

Abstract

Water-soluble chlorophyll proteins (WSCPs) are homotetrameric proteins that bind four chlorophyll (Chl) molecules in identical binding sites, which makes WSCPs a good model to study protein–pigment interactions. In a previous study, we described preferential binding of Chl a or Chl b in various WSCP versions. Chl b binding is preferred when a hydrogen bond can be formed between the C7 formyl of the chlorin macrocycle and the protein, whereas Chl a is preferred when Chl b binding is sterically unfavorable. Here, we determined the binding affinities and kinetics of various WSCP versions not only for Chl a/b, but also for chlorophyllide (Chlide) a/b and pheophytin (Pheo) a/b. Altered KD values are responsible for the Chl a/b selectivity in WSCP whereas differences in the reaction kinetics are neglectable in explaining different Chl a/b preferences. WSCP binds both Chlide and Pheo with a lower affinity than Chl, which indicates the importance of the phytol chain and the central Mg2+ ion as interaction sites between WSCP and pigment. Pheophorbide (Pheoide), lacking both the phytol chain and the central Mg2+ ion, can only be bound as Pheoide b to a WSCP that has a higher affinity for Chl b than Chl a, which underlines the impact of the C7 formyl-protein interaction. Moreover, WSCP was able to bind protochlorophyllide and Mg-protoporphyrin IX, which suggests that neither the size of the π electron system of the macrocycle nor the presence of a fifth ring at the macrocycle notably affect the binding to WSCP. WSCP also binds heme to form a tetrameric complex, suggesting that heme is bound in the Chl-binding site.

Introduction

Tetrapyrroles are the pigments of life.1,2 As one of the most ancient prosthetic groups, they are involved in several crucial tasks in organisms like light-harvesting in photosynthesis, electron transport in photosynthesis and respiration, transport of gasses, and catalysis of enzymatic reactions.2 All tetrapyrroles consist of four pyrrole rings that are linked by unsaturated methine groups. These can form macrocycles that can then be reopened secondarily to form linear tetrapyrroles (bilins).5 The chemical and physical properties of tetrapyrroles are tuned by the number of conjugated double bonds of the macrocycle (size of the π electron system), by the diversity in the side chains of the macrocycle and by the chelation of different metal ions by the macrocycle.4 To fulfill their various functions, tetrapyrroles are bound to proteins, which then fine-tune their properties.6 Besides their beneficial chemical properties, which are essential for life, tetrapyrroles can be dangerous by acting as photosensitizers and thus producing reactive oxygen species (ROS) upon illumination, which damages different cellular components and can ultimately lead to cell death.7 To prevent ROS production and to ensure a correct function, tetrapyrroles are usually tightly bound to proteins.
Plants contain four types of tetrapyrroles: chlorophylls (Chl), hemes, sirohemes, and phytochromobilins, which are all synthesized in the plastid.3 However, tetrapyrroles, especially hemes, as essential molecules are present in all cellular compartments. Yet, the most abundant tetrapyrroles Chls, whose macrocycle chelate a Mg2+ ion, have an additional fifth ring and have a phytol chain as typical side chain,8,9 are only present in chloroplasts. To optimize functioning in photosynthesis, and to prevent ROS production, all Chl molecules are non-covalently bound to carotenoid containing Chlbinding proteins, which are inserted in the thylakoid membrane.8,10 These Chl-binding proteins bind a large number of Chl molecules (Chl a and Chl b) in non-identical binding sites as wells as carotenoids, which prevent the formation of ROS.11 The variability of (bacterio)Chl binding has been tested with many proteins, for instance by the exchange of Mg complexes with Zn or other metal complexes, or with synthetic proteins (see ref. 9 and references reviewed within). However, systematic studies are lacking of how structural changes in the tetrapyrrole pigments change their binding to proteins, although the specific binding of Chls or other tetrapyrroles to Chl-binding proteins is important for their biogenesis and function (for details see ref. 12) as well as for potential applications in synthetic biology.13 Pigment specificity is difficult to address in most Chl-binding proteins because as outlined before all photosynthetic Chl-binding proteins bind several Chl molecules in non-identical binding sites. Thus, water-soluble chlorophyll proteins (WSCPs) from Brassicaceae have been introduced as a model system to study tetrapyrrole specificity in Chl-binding proteins.12,14
WSCPs strikingly differ from other Chl-binding proteins in higher plants. WSCPs are water-soluble, bind four Chl molecules in identical binding sites and no carotenoids at all, and are not involved in photosynthesis.15 Furthermore, all WSCPs are remarkably stable against denaturation, they even withstand boiling in aqueous solution and harsh pH conditions unharmed.16,17 The biological function of WSCP remains enigmatic, even though functions as Chl (or Chl metabolite) carrier in Chl metabolism or during the rearrangement of the photosynthetic apparatus under stress,18 as a Chl scavenger after cell disruption,19,20 as a protease inhibitor,21–26 and as ROS source in ROS signaling27 have been proposed.
WSCP is perfectly suited as a model system to study tetrapyrrole specificity because WSCP binds a broad range of tetrapyrroles,14 recombinant expression is established and thus mutational studies are possible,14,18–20,28 and structural information is available.12,29,30 In addition, two WSCP subclasses exist that differ in their Chl a/b ratio. Class IIA, which was isolated from Arabidopsis, Brassica, and Raphanus, is characterized by Chl a/b ratios of >6, whereas class IIB, which has so far only been isolated from Lepidium virginicum, Arabis alpina and Brassica rapa subsp. pekinensis,31 shows Chl a/b ratios of <3.5.15 It has been demonstrated by using recombinant WSCP that class IIA and class IIB indeed exhibit different preferences towards Chl a and b, with class IIA having no clear preference for either Chl a or b and class IIB showing a preference towards Chl b.12,19,20,28 Structural information on both subclasses is available with the crystal structures of the WSCP from Brassica oleracea var. botrytis (BobWSCP; class IIA)30 and of the WSCP of Lepidium virginicum (LvWSCP, class IIB).29 Both BobWSCP and LvWSCP form homotetrameric complexes with one Chl molecule bound per subunit.29,30 The central Mg2+ ion of Chl is ligated from the unfavorable32 β-side by the backbone carbonyl of a conserved proline residue. The Chl molecules are arranged in dimers in a so-called open-sandwich conformation in the center of the protein, which shields the Chl molecules from the surrounding solvent. In the open-sandwich conformation, the two Chl molecules face one another with their chlorin macrocycles, which leads to excitonic coupling of the molecules within a dimer.33,34 The phytol chains of the Chl molecules of a dimer are intertwined and interact with the phytol chains of the second Chl dimer in the heart of the tetramer, which significantly stabilizes the tetramer.35 The protein itself has a typical β-trefoil fold with six antiparallel β-sheets and interconnecting unfolded loops. Although BobWSCP and LvWSCP show a remarkably conserved overall structure and a conserved Chl ligation, there are distinctive differences in the protein environment of the bound Chl molecules between BobWSCP and LvWSCP. In a previous study,12 we investigated the Chl a/b specificity in LvWSCP and BobWSCP. A loop was identified that is responsible for the Chl a/b selectivity in WSCP. This loop consists of four amino acids in LvWSCP (LCPS) and of six amino acids in BobWSCP (PVCNEL) in an otherwise conserved region of the proteins. Consequently, we created WSCP variants, where the corresponding loops were interchanged. Thereby, a BobWSCP version (BobWSCP LCPS) was created, which has a notably higher preference for Chl b than the WT, and a LvWSCP version (LvWSCP PVCNEL), which has a strongly reduced Chl b preference compared to the WT. More strikingly, by exchanging a single amino acid a LvWSCP version was discovered that has a strong Chl a preference. LvWSCP PCPS, where we exchanged Leu91 of the LCPS loop to Pro, shows a relative Chl a/b binding affinity altered by a factor of 40 compared to the WT. Structural analysis revealed that the LCPS/PVCNEL loop is in close contact with the side chain of the C7 atom of the chlorin macrocycle, where Chl a has a methyl and Chl b a formyl group. In LvWSCP, the backbone nitrogen of L91 is close enough (3.0 Å) to the formyl group of Chl b to form a hydrogen bond, whereas in BobWSCP and LvWSCP PCPS the hydrogen bond donor closest to the formyl group of Chl b is 4.5 Å and 4.9 Å away, respectively.12,36 The close proximity of a hydrogen bond donor explains the higher Chl b affinity of LvWSCP. In contrast, BobWSCP has no hydrogen bond donor in close proximity of the formyl group of Chl b and thus shows neither a preference for Chl b nor Chl a, which implies that BobWSCP binds both Chls with the same affinity. However, we cannot explain the Chl a preference of LvWSCP PCPS with the distance of the next hydrogen bond donor to the formyl group of Chl b, which is similar in LvWSCP PCPS and BobWSCP. A possible explanation for the preference of Chl a by LvWSCP PCPS comes from comparing its structure with Chl a bound to its structure with Chl b bound. With Chl b bound, Pro91 adopts its energetically unfavorable exo-conformation, which is most likely caused by steric reasons due to the bulkier formyl group. Thus, Chl b binding is less favorable in LvWSCP PCPS compared to LvWSCP, explaining the Chl a preference of LvWSCP PCPS. In all these studies, only relative Chl a/b binding affinities in a competition situation between Chl a and b were obtained, where the apparent relative Chl affinities may be either thermodynamically or kinetically controlled. In this study, we investigated the tetrapyrrole binding specificity of LvWSCP, BobWSCP and LvWSCP PCPS. A circular dichroism (CD) spectroscopy-based method was used to determine KD values for the binding of different tetrapyrroles to WSCP and additionally to study the kinetics of those reactions. In particular, we investigated in detail the roles of the C7 side chain, of the phytol chain and of the central Mg2+ ion for the binding of a tetrapyrrole to WSCP and analyzed the impact of the type of central ion, of the size of the π electron system of the macrocycle and of the fifth ring of the macrocycle for the binding of a tetrapyrrole to WSCP. Materials and methods Pigments Chl a and b were isolated from pea plants and purified as described previously.37 Pheophytin a and b (Pheo) were prepared from the corresponding Chl according to the method of Hyninnen.38 Chlorophyllide (Chlide) and pheophorbide (Pheoide) were prepared from Chl and Pheo, respectively, as described previously with the enzyme chlorophyllase, and subsequently purified by reversed-phase chromatography.27 A sample of protochlorophyllide (PChlide) prepared as described by Kruk and Myśliwa-Kurdziel39 was generously gifted by Jerzy Kruk, Jagiellonian University Krakow. Mg-Protoporphyrin IX (Mg-Proto IX) and protoporphyrin IX (Proto IX) were purchased from Frontier Scientific. Heme b was prepared from hemin (Sigma-Aldrich) by incubation with sodium dithionite (for details see below). Proteins In this study, WSCPs from Lepidium virginicum (LvWSCP)20 and from Brassica oleracea var. botrytis (BobWSCP)40 with a C-terminal hexahistidine-tag were used. Furthermore, a variant of LvWSCP (LvWSCP PCPS) with a single amino acid exchange in a motif involved in Chl binding was used. The amino acid exchange from LCPS in the WT to PCPS in the mutant strongly alters the relative Chl a/b affinity (for details see ref. 12). All WSCPs were expressed recombinantly and purified as described previously with a few alterations.40 In short, E. coli BL21 (DE3) were cultivated at 37 °C in 800 ml lysogeny broth (LB) medium with 50 µg ml−1 kanamycin, until an OD600 nm of 0.6 was reached. Then the protein expression was induced by the addition of IPTG (1 mM final). The induced culture was incubated overnight at 37 °C and then harvested by centrifugation (5 min, 8000g). The cells were lysed by sonication with a tip sonicator (Vibra cell, Sonics & Materials) for 5 min in 20 mM sodium phosphate pH 7.8, 300 mM NaCl, 15 mM imidazole. After centrifugation of the lysate (30 000g, 15 min, 4 °C), soluble WSCP apoprotein in the supernatant was further purified by Ni2+-affinity chromatography. The supernatant was applied to a Ni2+ loaded Chelating Sepharose Fast Flow column (GE Healthcare, column volume (CV) 5 ml), equilibrated with 2 CV sodium phosphate buffer pH 7.8. After two subsequent washing steps with 25 and 50 mM imidazole, in 20 mM sodium phosphate pH 7.8 (2 CV each), the bound WSCP was eluted from the column with 1.5 CV 300 mM imidazole in 20 mM sodium phosphate pH 7.8. The eluate was desalted with a Zeba spin column (Thermo Fisher Scientific) according to the manufacturer’s instructions. Protein concentrations were determined by absorption spectroscopy using extinction coefficients that were calculated by the protein sequences. Reconstitution of WSCP with pigments WSCPs were reconstituted with Chl, Chlide, Pheo, and Pheoide by using Triton X-114 (TX-114) as described previously.12 For the reconstitution of LvWSCP with PChlide, PChlide was dissolved in diethyl ether and added to a protein solution in 20 mM sodium phosphate pH 7.8 to final concentrations of 100 µM LvWSCP, 300 µM PChlide and 20% (v/v) diethyl ether. After overnight incubation at 4 °C at 30 rpm in an overhead shaker (neolab), the WSCP-containing water phase was separated by centrifugation (5 min, 10 000g, 4 °C) from the ether phase. Mg-Proto IX was dissolved in 20 mM sodium phosphate pH 7.8 and then added to LvWSCP at final concentrations of 100 µM LvWSCP and 300 µM Mg-Proto IX. After incubation at 4 °C overnight, unbound Mg-Proto IX was removed with a home-packed 5 ml desalting column (Sephadex G-25 fine, GE Healthcare) equilibrated with 2 CV 20 mM sodium phosphate pH 7.8. 1 ml of the reconstitution mixture was loaded onto the column, which was subsequently washed with one CV 20 mM sodium phosphate pH 7.8. WSCP eluted immediately from the column and was collected. For the reconstitution of LvWSCP with heme, 2 mg hemin were dissolved in 1 ml methanol. After centrifugation (10 000g, 5 min), the supernatant was mixed with 1 ml 2 mM sodium dithionite in 20 mM sodium phosphate pH 7.8 to reduce hemin to heme. After incubation of 30 min at RT, the heme concentration was determined spectroscopically using the extinction coefficient of heme in DMSO (170 000 M−1 cm−1 at 404 nm).41 Heme solution was added to LvWSCP to 100 µM LvWSCP and 300 µM heme (final, <20% methanol and <0.8 mM dithionite). After incubation at 4 °C overnight, unbound heme was removed with a home-packed 5 ml desalting column (for details see above). The reconstitution products were analyzed and purified by size-exclusion chromatography (SEC). The samples were loaded onto a Superose 12 10/300 GL prepacked column (GE Healthcare) operated by an NGC chromatography system (Biorad). After SEC purification, the pigmented fractions were pooled and analyzed spectroscopically. Prior to SEC, the reconstitution products were additionally analyzed by native polyacrylamide gel electrophoresis (native PAGE) with 15% polyacrylamide gels. Spectroscopic measurements UV–Vis absorption spectra of WSCP were recorded at RT with a V-550 UV/Vis spectrophotometer (Jasco) between 450 and 250 nm (apoprotein) or 750 and 250 nm (pigmented tetramer) in a 10 mm quartz cuvette (scan speed 200 nm min−1; bandwidth 2 nm). Circular dichroism (CD) was measured in a J-810 spectropolarimeter (Jasco). CD spectra were recorded at RT between 750 and 350 nm in 2, 5 or 10 mm OS cuvettes (1 nm data pitch, 100 nm min−1 scan speed, 4 s response time, and 1× or 4× accumulation). Intrinsic tryptophan (Trp) fluorescence of WSCP was determined with a FluoroMax-2 spectrometer (Horiba Scientific). Fluorescence emission was recorded between 300 and 500 nm after excitation at 280 nm (2 nm slits, 1 nm increment, 1 s integration time, 3× accumulation, 5 × 5 mm quartz cuvette). The WSCP concentration was adjusted to 5 µM. Kinetic measurements of the pigment binding to WSCP Kinetics of the pigment binding to WSCP were analyzed by timeresolved CD measurements. 10 mm OS cuvettes were preloaded with pigment (20 µM final) in Triton X-100 (TX-100) micelles (0.1% final). After the addition of WSCP (40 µM final) the CD signal was tracked at the Qy maximum (positive extremum) for Chl a and Chlide a, and the Soret minimum (negative extremum) for Chl b, Chlide b, Pheo a/b for 30 min. For each reaction at least two independent kinetic measurements were performed. To obtain amplitudes (A) and reaction rates (k), kinetic traces were fitted. Different fits were tested. Fits assuming two consecutive first-order reactions were the simplest ones that gave reasonable results with the following equation: Time constants (τ) are the reciprocals of the calculated reaction rates. Determination of KD values for the pigment binding to WSCP To determine KD values for the pigment binding to WSCP, we titrated pigment solutions against WSCP. WSCP was added to a constant pigment concentration (20 µM) to mixtures with increasing WSCP concentrations (0–100 µM). After overnight incubation at 4 °C, CD spectra of the samples were measured. To obtain binding curves, the CD signal of the Qy maximum for Chl a and Chlide a, and the Soret minimum for Chl b, Chlide b, and Pheo a of three independent measurements were averaged and plotted against the WSCP concentration. For KD determination, the binding curves were fitted with the following equation:42 whereby c is the pigment concentration and k is a CD coefficient to convert concentration into a CD signal. Results WSCP has a high affinity for Chl The KD values for the binding of Chl a and b to LvWSCP, BobWSCP, and LvWSCP PVCNEL were determined by incubating solutions with a constant pigment concentration with increasing WSCP concentrations. After overnight incubation CD spectra of the samples were measured. Unfortunately, all titrations were stoichiometric titrations and only upper estimates for the KD values (Fig. 1; Table S1†) could be determined. All determined values are in the nM range, varying between 500 nM and 30 nM. Thus, WSCP has a high affinity for both Chl a and b. However, due to the stoichiometric titration and the rough estimation of the KD values, it is impossible to determine differences in the binding of Chl a and b to the WSCPs. Unfortunately, the detection limit of the method (around 1 µM) does not allow us to reduce the concentrations of Chl and WSCP to determine proper KD values. To obtain Chl a/b binding kinetics, CD spectroscopy was used to track the formation of the WSCP Chl complex. Binding kinetics for LvWSCP, BobWSCP and LvWSCP PCPS were measured with both Chl a and b, respectively (Fig. 2, for a focus on the earlier time points see Fig. S1†). For better comparison, the kinetic traces were fitted assuming two consequent first-order reactions. The resulting fits revealed that all reactions have a substantially faster phase, which has a significantly larger amplitude than the slower phase (for details see Table S2†). Due to the experimental dead time and the fast first phase of the reactions, the fits are of lower quality for the earlier time points (see Fig. S1†). The time constants of Chl b binding to WSCP are comparable between LvWSCP and BobWSCP with ∼6 s for the faster and ∼250 s for the slower phase. In contrast, the time constants of Chl a binding differ between the two WSCPs. BobWSCP binds Chl a with smaller time constants (∼3 s and ∼90 s) than LvWSCP (∼9 s and ∼140 s). The time constants of both Chl a and b binding to LvWSCP PCPS are altered in comparison to LvWSCP. Chl b binds more slowly to LvWSCP PCPS (time constants around 8 s and 300 s) than to LvWSCP. LvWSCP PCPS binds Chl a with a faster fast phase (∼3 s) and a slower slow phase than LvWSCP (∼200 s). Thus, differences in the kinetics between Chl a and b binding might correlate with the strong preference for Chl a observed with LvWSCP PCPS since Chl a is considerably faster bound than Chl b by this protein. However, BobWSCP binds Chl a faster as well, but we do not observe a Chl a preference in BobWSCP. Moreover, LvWSCP has a Chl b preference but does not bind Chl b faster than Chl a. Thus, the relative Chl a/b binding affinities are at least not predominantly kinetically controlled but are rather determined by different affinities for the two Chl. The phytol chain is important for the pigment-WSCP interaction To evaluate the impact of the phytol chain, KD values for the binding of Chlide a and b were determined, which in contrast to Chl have no phytol chain, but even so bind to WSCP.14 In addition, we chose to use Chlide a/b affinities as a proxy for Chl a/b affinities, since we could not determine KD values for the binding of Chl to WSCP. In the structure of WSCP tetramers,29,30 the phytol chains of the Chl molecules seem to stabilize the complex by interacting with each other. Consistently, WSCP-Chlide complexes have a reduced stability in comparison to WSCP-Chl complexes,35 which suggests a lower affinity for Chlide than for Chl. The titrations of WSCP with both Chlide a and b revealed binding curves (Fig. 3), which allowed to determine proper KD values (for details see Table S1†). These determined values are all in the µM range and thus WSCP has – as expected – a lower affinity for Chlide than for Chl. With KD values of 1.6 µM and 1.3 µM BobWSCP has a similar affinity for Chlide a and b, which is in good agreement with the previously determined relative Chl a/b binding affinity. In contrast, the affinities for Chlide a and b binding to LvWSCP do not support a Chlide b preference, which was indicated by the relative Chl a/b binding affinity. The KD values of 2.6 µM and 3.7 µM for Chlide a and b, respectively, point to a higher affinity for Chlide a than Chlide b. However, this may be due to the limited accuracy of the method considering the large error margins for both KD values (1.2 and 2.3 µM for Chlide a and b binding, respectively). Nevertheless, the binding curves for Chlide binding to LvWSCP PCPS reveal a remarkably reduced affinity for Chlide b (KD = 24.7 µM) compared to LvWSCP, which is in accordance with the relative Chl a/b binding affinities that show a Chl a preference. In addition, the Chlide a binding affinity is also altered in LvWSCP PCPS compared to LvWSCP. LvWSCP PCPS shows a higher affinity for Chlide a than LvWSCP (KD = 0.5 µM). In addition to the determination of the KD values, the kinetics of Chlide binding to WSCP were also investigated with time-resolved CD spectroscopy (Fig. 4, for a focus on the earlier time points see Fig. S2†). Fitting of the time traces leads (for details see Table S2†) to similar results as observed for Chl-binding: all reactions have a substantially faster phase, which has also a significantly larger amplitude than the slower phase. In general, Chlide is faster bound to WSCP than Chl, especially the slower phase of the Chlide-binding is notably faster compared to the slower phase of Chl-binding (with the exception of Chlide b binding to LvWSCP). The fast Chlide binding again shows some differences between the experimental data and the fits (see Fig. S2†). The time constants of the faster phases are also altered between Chl and Chlide binding. Only the time constants of the binding of Chlide a to BobWSCP and LvWSCP PCPS are higher than those of the binding of Chl a to these WSCPs, all other time constants of the faster phase are lower compared to the analogous reaction with Chl. Comparing the Chlide a and b binding to the three investigated WSCPs, all bind Chlide a faster than Chlide b, in analogy to the Chl binding kinetics. The time constants of Chlide a binding are comparable between all three WSCPs (4–5 s and ∼60 s). In contrast, the time constant of Chlide b binding differs between the proteins. The time constants of the faster phases of Chlide b binding to LvWSCP and LvWSCP PCPS are with ∼4.5 s equal, but BobWSCP shows a time constant of this reaction of 1.6 s. The time constants of the slower phases are comparable between BobWSCP and LvWSCP PCPS with around 160 s, but different from that of LvWSCP (∼329 s). The central Mg2+ ion is important for the pigment-WSCP interaction In contrast to previous studies,14 recent results (unpublished data) showed that WSCP is also able to bind pheophytin (Pheo; Chl without Mg2+ central ion). This opened the possibility to investigate the influence of the central ion on the binding to WSCP. The affinity for Pheo a was determined by titration. Unfortunately, we could not titrate with Pheo b because Pheo b aggregates rapidly in TX-100. The aggregates are in fact spectroscopically distinguishable from WSCPbound Pheo b, but due to aggregation the free Pheo b concentration is lowered and thus an apparent binding saturation is already reached at low WSCP concentrations (stoichiometric titration; data not shown). The KD values for Pheo a binding were determined (Fig. 5; for details see Table S1†). Compared to Chl a binding the affinity of WSCP for Pheo a is lower. The KD values for Pheo a binding differ strongly between LvWSCP, BobWSCP and LvWSCP PVCNEL. Whereas BobWSCP (KD of 2.7 µM) binds Pheo a with approximately the same affinity as Chlide a, LvWSCP (KD of 72.9 µM) shows a lower affinity for Pheo a than for Chlide a. The affinity of LvWSCP PCPS for Pheo a is remarkably reduced. The affinity is so low that a proper titration curve with LvWSCP PCPS could not be recorded (Fig. S4A†). Thus, the KD of Pheo a binding to LvWSCP PCPS is higher than 100 µM. Nevertheless, LvWSCP PCPS can bind Pheo a as shown with SEC, CD spectroscopy and native PAGE (Fig. S4B and C†). In contrast, we could not observe any sign of Pheo b binding to LvWSCP PCPS. Yet, BobWSCP and LvWSCP bind Pheo b. In addition to the affinities, the kinetics of Pheo binding to WSCP (Fig. 6, for a focus on the earlier time points see Fig. S3†) were investigated by recording time-resolved CD measurements of Pheo a/b binding to LvWSCP and BobWSCP. However, Pheo a binding to LvWSCP PCPS could not be tracked over time since the binding affinity and with that the tetramer yield were very low. Again, fitting of the time traces revealed that one phase is remarkably faster than the other phase (for details see Table S2†). However, the faster phase of Pheo binding has not necessarily the larger amplitude. For Pheo a binding to LvWSCP the slower phase has a larger amplitude and for Pheo b binding to BobWSCP both phases have similar amplitudes. Interestingly, Pheo b binding to LvWSCP (time constants of ∼7 s and ∼120 s) is even faster than Chl b binding with the time constant of the slower phase being lower for Pheo b binding compared to Chl b. In contrast, LvWSCP binds Pheo a with slightly smaller time constants (∼12 s and ∼160 s) than Chl a. Pheo binding to BobWSCP is also slower than Chl binding. Especially, the binding of Pheo a to BobWSCP (∼6 s and ∼315 s) is notably slowed down compared to Chl a binding, the time constants of both phases are affected. Furthermore, it was tested whether LvWSCP and BobWSCP are able to bind pheophorbide (Pheoide; Pheo without phytol). BobWSCP bound neither Pheoide a nor b, and LvWSCP was not able to bind Pheoide a, but did bind Pheoide b (Fig. 7). In the SEC elution diagram (Fig. 7A) LvWSCP Pheoide b elutes at approximately 12.4 ml a little earlier than LvWSCP Chl a at ∼12.8 ml, indicating a bigger size of the LvWSCP Pheoide b complexes. The native PAGE analysis (Fig. 7C) shows that this peak consists of LvWSCP Pheoide b tetramers because we observe only one fluorescent band that runs a little lower than LvWSCP Pheo b tetramers. This indicates that the shift in the SEC elution diagram is due to column performance issues or a slightly different shape of the tetrameric complex. The shoulder around 13.4 ml in the LvWSCP Pheoide b elution shows that there is still some WSCP apoprotein left, which suggests a lower reconstitution yield with Pheoide b than with Chl a. Finally, the recorded CD spectrum of purified LvWSCP Pheoide b tetramers (Fig. 7B), with a strong signal in the Soret region with a minimum at 458 nm, indicates excitonic coupling of the bound Pheoide b molecules within the complex, implying a similar arrangement of the Pheoide b molecules in dimers as that of WSCP-bound Chl molecules. WSCP binds Chl precursors In addition to the investigation of the influence of the C7 side chain, of the phytol and of the Mg2+ central ion for the pigment binding to WSCP, the impact of other structural features of the tetrapyrrole for the binding to WSCP was explored. In particular, the influence of the size of the π electron system and the fifth ring of the macrocycle were investigated. Therefore, the interaction of WSCP with the Chl precursors protochlorophyllide (PChlide), Mg-protoporphyrin IX (MgProto IX) and protoporphyrin IX (Proto IX) was analyzed. All three precursors are porphyrins and thus have an electron system of 22 π electrons in the macrocycle, whereas chlorins like Chl have 20 π electrons. The only additional difference between PChlide a and Chlide a is the C8 side chain of the macrocycle (vinyl and ethyl groups, respectively). In contrast to PChlide, Mg-Proto IX lacks the fifth ring of the macrocycle and Proto IX has the same structure as Mg-Proto IX, but no Mg2+ central ion. LvWSCP was reconstituted with the pigments, and the reconstitution mixture was analyzed by SEC and native PAGE (Fig. 8). In the SEC elution diagram of the PChlide reconstitution (Fig. 8A and Fig. S5A†) LvWSCP PChlide elutes at around 13 ml like LvWSCP Chl a tetramers do. In the native PAGE (Fig. 8C), tetramer formation with PChlide is confirmed. However, the reconstitution yield of WSCP with PChlide is lower than that with Chl a, since the LvWSCP apoprotein peak in the SEC elution diagram is bigger than the tetramer peak. In addition, the coomassie stained native PAGE revealed a notable amount of unpigmented apoprotein. Furthermore, in the native PAGE a second pigmented band of the PChlide reconstitution is visible by fluorescence detection. This band migrates further than tetramers in the gel but not as far as apoprotein, which may indicate the formation of LvWSCP PChlide dimers. In the reconstitution of LvWSCP with MgProto IX only this dimer band is visible in the fluorescence image of the gel. In this case, SEC supports the formation of pigmented dimers (Fig. 8B and Fig. S5B†). Besides an apoprotein peak, a second peak is visible that elutes later than the tetramer peak of the reconstitution with Chl a. Again, the apoprotein peak and the apoprotein band in the native PAGE reveal a lower reconstitution yield of LvWSCP with Mg-Proto IX than with Chl a. No experimental data support the binding of Proto IX to LvWSCP. In addition to SEC and native PAGE, CD and absorption spectroscopy (Fig. 9) also support binding of PChlide and MgProto IX to LvWSCP. Binding of the pigments to LvWSCP changes the absorption properties of the pigments notably compared to free pigments (Fig. S5C & D†). Purified LvWSCP PChlide has a Qy absorption maximum at 635 nm and a Soret absorption maximum at 440 nm (Fig. 9A). Purified LvWSCP Mg-Proto IX shows only minor light absorption in the Qy and Qx regions, but strong absorption in the Soret region with a maximum at 420 nm (Fig. 9A). Both purified complexes show strong CD signals (Fig. 9B), which supports the idea that both PChlide and Mg-Proto IX are arranged as excitonically coupled dimers when bound to LvWSCP and thus are similarly bound to LvWSCP as Chl. Purified LvWSCP PChlide shows a CD signal in both Qy and Soret region. However, the spectrum is dominated by the Soret minimum at 445 nm. The CD spectrum of purified LvWSCP Mg-Proto IX only has a band in the Soret region with a maximum at 423 nm and a minimum at 411 nm. WSCP binds heme Proto IX is the common precursor of both Chl and heme synthesis. Insertion of Mg2+ is the first step in the Chl branch and gives rise to Mg-Proto IX, whereas insertion of Fe2+ initiates heme synthesis and gives rise to heme b.3 Since WSCP promiscuously binds different tetrapyrroles including Mg-Proto IX, heme b (hereafter only heme) binding to WSCP was tested. First, binding of heme was investigated with CD and absorption spectroscopy (Fig. 10). The addition of LvWSCP to heme gives rise to a strong CD signal with a maximum at 430 nm and a minimum at 407 nm (Fig. 10A), which indicates excitonic coupling of heme molecules. Thus, the heme molecules are most likely bound to WSCP as dimers, similar to Chl. Furthermore, the addition of LvWSCP to heme strongly changes the absorption properties of heme (Fig. 10B). Whereas heme in buffer has an absorption maximum at 390 nm with a shoulder at 360 nm, the maximum is shifted to 410 nm by the addition of LvWSCP. For further confirmation of heme binding to WSCP, the reconstitution mixture was analyzed with SEC and native PAGE (Fig. 10C & E). In the SEC elution diagram (Fig. 10C) of the reconstitution of LvWSCP with heme two peaks are present. The first peak at 12 ml is pigmented (see Fig. S6A†) and elutes earlier than LvWSCP Chl a tetramers (at 13 ml). The second peak is unpigmented and thus consists of apoprotein. The native PAGE (Fig. 10E) confirms the lower yield of the reconstitution of LvWSCP with heme compared to Chl a. In the coomassie stained PAGE, a significant amount of apoprotein is left when LvWSCP is reconstituted with heme. However, a red band in the heme reconstitution is visible in the unstained gel that runs at the same height as the green LvWSCP Chl a tetramer band, which suggests tetramer formation with heme. In contrast to the LvWSCP Chl a tetramer band the heme tetramer band does not fluoresce because heme does not show any auto-fluorescence. Tetramer formation with heme is additionally supported by the coomassie staining of the gel. LvWSCP apoprotein already tends to form unpigmented tetramers, but tetramers are enriched by the addition of heme, which can be seen by comparing the tetramer bands of LvWSCP apoprotein and LvWSCP heme. Furthermore, heme-binding was investigated by fluorescence spectroscopy (Fig. 10D). Heme is a strong fluorescence quencher. Thus, the intrinsic Trp fluorescence of SEC purified LvWSCP heme was compared to that of LvWSCP apoprotein. In LvWSCP heme the Trp fluorescence is strongly quenched, only 24% of the fluorescence of LvWSCP apoprotein is left. Overall, the spectroscopic data, the SEC and native PAGE point to binding of heme to LvWSCP. Furthermore, LvWSCP forms tetramers with heme and the heme molecules are most likely organized as dimers. Yet, the reconstitution yield with heme is low. Discussion Chl a/b affinities of WSCPs Already in the 1970s, WSCPs isolated from plants of different Brassicaceae species were discovered to have different Chl a/b ratios.43,44 BobWSCP (Chl a/b >6)43 has a higher Chl a content than LvWSCP (Chl a/b <3.5).44 Over 40 years later, it was shown that BobWSCP has no clear preference for Chl a or b,19 whereas LvWSCP prefers Chl b binding over Chl a binding.20 However, it was not known how Chl preference in WSCP is established until recently we found a loop in close contact to the C7 side chain of the chlorin macrocycle, where Chl a differs structurally from Chl b.12,36 For different WSCP versions relative Chl a/b binding affinities were obtained, which unfortunately does not allow to conclude, whether Chl preferences are controlled thermodynamically by KD values or kinetically by different reaction rates. Even though some experimental data like exchange of WSCP-bound Chl a with Chl b suggest that Chl binding to WSCP is a true equilibrium reaction, this is not completely clear since Chl dissociation from the WSCP Chl complex at low concentrations is hardly detectable. If Chlbinding to WSCP is irreversible or has an irreversible component, then the apparent relative affinities of Chl a and Chl b competing for Chl-binding sites in WSCP may not be caused by the according binding constants of Chl a and Chl b but rather by different binding kinetics. To resolve this issue, in this study KD values as well as time constants for the pigment binding to WSCP were determined. Differences in the binding kinetics between Chl/Chlide a and Chl/Chlide b are not sufficient to explain the previously determined relative Chl a/b affinities in LvWSCP, BobWSCP, and LvWSCP PCPS. For instance, LvWSCP and LvWSCP PCPS bind Chl a and b with about the same time constants, which certainly cannot explain why LvWSCP PCPS preferentially binds Chl a. Thus, the reaction kinetics are not primarily responsible for different Chl a/b selectivity of WSCPs. Unfortunately, accurate KD values for the Chl binding to the WSCPs were not determined since the KD values are <500 nM, which is below the detection limit of the CD titration. However, KD values with Chlide were determined. The determined Chlide a/b affinities for LvWSCP PCPS are in good agreement with the relative Chl a/b binding affinity, confirming that LvWSCP PCPS preferentially binds chlorins without a C7 formyl group in its macrocycle. This is further supported by the fact that LvWSCP PCPS does not bind Pheo b but can bind Pheo a. BobWSCP binds Chlide a/b with about the same affinity, supporting that BobWSCP has no clear preference for Chl a/b. However, it is noteworthy that the big error margins to not allow to conclude this with certainty. LvWSCP shows a slightly higher affinity for Chlide a than for Chlide b, which contrasts with the known relative Chl a/b affinities. Yet, a higher Chlide b affinity over Chlide a is in the error range of both KD values. A higher Chl/Chlide b affinity over Chl/Chlide a for LvWSCP is supported by the observation that LvWSCP is still able to bind Pheoide b but not Pheoide a. Thus, the affinities are the main selectivity factor in Chl a vs. Chl b binding in WSCP. Influence of tetrapyrrole structural features on binding to WSCP WSCPs are known to bind a variety of different tetrapyrroles including chlorins, bacteriochlorins as well as porphyrins.14 However, it is unknown how some structural features of tetrapyrroles like different side chains at the macrocycle, the phytol chain, the size of the π electron system, and the central ion influence the binding to WSCP. So far, only the C7 side chain of the chlorin macrocycle was found to affect the interaction between tetrapyrrole and WSCP (see above). KD values and time constants for Chlide and Pheo binding to WSCP were determined, which allowed direct comparison of the affinities and binding kinetics between Chl, Chlide, and Pheo. In WSCP, Chl is bound in a cavity in the center of the protein.29,30 The four Chl molecules in a WSCP tetramer are arranged as two dimers, each in a so-called open-sandwich conformation, where the chlorin macrocycles of the Chl molecules within a dimer face each other. The phytol chains of all four Chl molecules interact with each other via hydrophobic interactions in the center of the protein.29,30 Thus, WSCP Chlide tetramers show a reduced stability, which suggests lower affinities for Chlide binding than Chl binding.14,35 Here, we show that indeed the missing hydrophobic interactions between the phytol chains of bound Chl reduce the affinity of WSCP for Chlide at least by a factor of 10. Furthermore, the binding kinetics are altered between Chlide and Chl binding. Chlide – with the exception of Chlide b to LvWSCP binding – in general is bound faster than Chl, most likely because it takes more time to arrange the phytols properly in the protein environment since the phytol chains of a Chl dimer are intertwined.29,30 Another major interaction site between WSCP and Chl is the central Mg2+ ion of Chl, which is ligated by the backbone carbonyl of a conserved proline residue in WSCP.29,30 The central Mg2+ ion is ligated from the energetically unfavorable β-side.29 Pheo a, which lacks the central Mg2+ ion, is bound to WSCP with a lower affinity than Chl a. BobWSCP has about the same affinity for Pheo a and Chlide a, which implies that both phytol chain and central Mg2+ ion have the same importance for the pigment binding in BobWSCP. LvWSCP binds Pheo a with a 30-fold lower affinity than Chlide a. Thus, the central Mg2+ ion is more important for the pigment binding in LvWSCP than the phytol chain. Also, the kinetics of the Pheo binding to LvWSCP and BobWSCP are altered compared to Chl binding. Pheo binding reactions (except Pheo b to LvWSCP) are notably slower than binding of the corresponding Chl to WSCP. This indicates that the ligation of the central Mg2+ ion by the protein is a fast process, which might suggest that the ligation of the central Mg2+ ion is the initial step in the Chl binding. Both phytol and central Mg2+ ion are missing in Pheoide. Only LvWSCP was found to be able to bind Pheoide b, where the protein can form a hydrogen bond with the formyl group at the C7 atom of the macrocycle. This interaction seems sufficient to compensate for the missing hydrophobic interactions of the phytols and the missing ligation of the central Mg2+ ion. Thus, for BobWSCP and LvWSCP PVCNEL both the phytol chain and the central Mg2+ ion are important interaction sites of pigments bound. For LvWSCP, also the hydrophobic interactions of the phytols and ligation of the central Mg2+ ion are important, but a third contact, the hydrogen bonding of the C7 side group can compensate for both other interactions. For LvWSCP, other structural features of the tetrapyrroles such as the size of the π electron system and the fifth ring that influence the pigment binding to LvWSCP were additionally investigated. However, KD values for those pigments were not obtained, and the pigments often have several alterations compared to Chl, which makes it difficult to correlate individual features with differences in pigment binding. The size of π electron systems seems to affect pigment binding only marginally. LvWSCP is able to bind porphyrins, chlorins, and bacteriochlorins ranging from 22 to 18 π electrons.14 Bacteriochlorophyll a, which has a different side chain at the C3 atom of the macrocycle in addition to the reduced π electron system, is bound with a somewhat lower yield to LvWSCP than Chl a (ref. 14 and unpublished data). Furthermore, LvWSCP binds PChlide, Mg-Proto IX and heme, which are all porphyrins. In addition, all lack the phytol chain and have at least one modification at the macrocycle, which may explain why the reconstitution yield with those are lower than with Chl. However, the only tested porphyrin that did not bind to LvWSCP is Proto IX, which has no phytol chain, no central Mg2+ ion and no formyl group at the C7 position of the macrocycle, again underlining the importance of these three interaction sites for the binding of a pigment to LvWSCP. Since LvWSCP is able to bind heme with a central Fe2+ ion and also Zn-Chl with a central Zn2+ ion,14 the kind of central metal ion does not seem to matter. However, the kind of coordination of the central ion does seem to be important. Because WSCP is not able to bind chlorins with a central Cu2+ ion.14 Cu2+ prefers a tetracoordinated ligation in tetrapyrrole complexes, whereas Mg2+ and Zn2+ form preferentially pentacoordinated tetrapyrrole complexes and Fe2+ is able to form both pentaand hexacoordinated tetrapyrrole complexes.45,46 Also the fifth ring of the macrocycle, which is present in Chl, Chlide, Pheo, Pheoide, PChlide and BChl, does not affect the binding of a pigment to WSCP severely, since heme and Mg-Proto IX lack it and can still be bound to WSCP. However, the fifth ring does seem to influence the oligomerization of WSCP because with Mg-Proto IX only pigmented dimers were observed in a native PAGE. In a previous study,14 WSCP did not show tetramer bands with Chlide on a native PAGE. Recently, it has been established that WSCP does form tetramers with Chlide27,35 albeit at lower stability and the occurrence of WSCP Chlide dimer bands in a native PAGE has been linked to mildly destabilizing conditions in native PAGE.40 Besides the main tetramer band, a weaker band of pigmented dimers was also observed with Chlide and PChlide in the native PAGE, but with Mg-Proto IX no tetramer band was detectable. The lack of WSCP Mg-Proto IX tetramers is supported by SEC, where no tetramer peak occurs in the elution. Thus, the fifth ring as the phytol chain might be important for the pigment dimer–dimer interaction in WSCP. Yet, the structure of the WSCP Chl complex does not support this idea. The fifth ring is not located at the Chl dimer–dimer interface, but rather at the Chl monomer–monomer interface. Furthermore, for heme we see tetramer but no dimer formation, even though heme lacks both phytol chain and the fifth ring, as does Mg-Proto IX. To sum up, in this study interactions sites were identified that significantly influence the binding of a tetrapyrrole to WSCP. A central ion like Mg2+, Zn2+ or Fe2+, which all allow ligation by WSCP, and the phytol chain are major interaction sites between pigment and protein. However, if one of these is missing, WSCP is still able to bind the corresponding pigment with a reduced affinity. When both central ion and phytol are missing, only LvWSCP is able to compensate the missing interaction, when it is possible to form a hydrogen bond with a formyl residue at the C7 position of the macrocycle. How does WSCP bind heme? Even though Chl and heme are both tetrapyrroles that share the same structural backbone, to our knowledge no plant protein has been identified so far that can bind both Chl and heme. Here, we report that WSCP is also able to bind heme. WSCP forms tetrameric complexes with heme as known for the Chl complex, but the tetramer yield with heme is much lower than with Chl, which indicates a lower affinity for heme than for Chl. In contrast to Chl, heme has no phytol chain, which is one important interaction site between Chl and WSCP that influences the affinity for a pigment (see above). Thus, this at least partially explains the differences in the tetramer yield. Other differences between heme and Chl like the different sizes of the π electron system and the missing fifth ring in heme are most likely neglectable factors to explain the different tetramer yields. The influence of the Fe2+ ligation on the binding affinity cannot be evaluated on basis of the current data (see above). However, it should be noted that the used TX-based reconstitution for Chl is an optimized and wellestablished method, which leads to a tetramer yield of nearly 100%,12 whereas the methanol-based reconstitution used for heme is a first approach and thus might not be the best-suited method. The induction of a CD, the shift of the heme absorption and the quenching of the Trp fluorescence of LvWSCP are strong evidence for heme binding to LvWSCP. Upon binding of heme to LvWSCP the absorption properties of heme change due to the changed surrounding of the heme molecules. The protein binding-induced CD of heme can in principal result from two phenomena: either from dissymmetrical distortion of the bound heme molecule or from excitonic coupling of the heme molecules either with other heme molecules or with aromatic amino acid side chains of the protein. It has been consensus since studies by Hsu and Woody in the 1970s47,48 that excitonic coupling of heme with aromatic amino acids is largely responsible for the rise of the CD of protein-bound hemes. However, recent studies on heme binding to globins point out that torsions of the vinyl side chains of the macrocycle are the major determinants for the heme CD.49–51 Without a high-resolution structure of the WSCP-heme complex, it is almost impossible to predict what gives rise to the CD when heme is bound to WSCP. However, if we assume a similar binding mode for heme as for Chl, then heme is bound as a dimer and the heme molecules of the dimers would be excitonically coupled, which is then responsible for the CD of heme. The arrangement of protein-bound heme as a dimer is unusual but has been observed before.52–54 In the heme transport protein ChaN from Campylobacter jejuni heme is bound in a homodimeric complex, where each subunit binds one heme molecule.53 The two heme molecules face each other and form a dimer in the protein center, which is reminiscent of the Chl arrangement in a WSCP dimer. In contrast to the Chl dimer in WSCP, in which the Chl molecules are arranged head-to-head, in the ChaN heme dimer the heme molecules are arranged head-to-tail. However, some cytochrome c proteins have heme bound as dimers in head-tohead position, but in these cases the heme dimer is bound in a binding pocket of a single protein chain.52 Unfortunately, the CD spectrum of the LvWSCP heme complex does not allow us to decide whether or not heme is bound as an excitonically coupled dimer. Even though the CD spectrum of the heme dimer bound to ChaN (maximum at 428 nm and minimum at 387 nm)53 reminds of the CD spectrum of the LvWSCP heme complex, other proteins such as hemoplexin55 or the β and γ subunits of human hemoglobin,56 which only have one heme molecule bound, show similar CD spectra. Yet, the occurrence of heme containing LvWSCP tetramers strongly suggests that heme is bound in the Chl binding pocket similarly to Chl as excitonically coupled dimers. The ligation of the central Fe2+ ion of heme by proteins is well investigated. The central Fe2+ ion can either be penta- or hexacoordinated by axial ligands provided from the protein.46 According to the Heme Protein Database,46 the most frequent ligands of the protein for central Fe2+ ions are histidine, cysteine, tyrosine and methionine side chains. In the structure of the WSCP-Chl complex, none of these potential ligands is in close proximity to the central Mg2+ ion of Chl, which would allow ligation. However, all groups that contain either oxygen, nitrogen or sulfur atoms can potentially act as ligands.57 Thus, ligation of the central Fe2+ ion by a backbone carbonyl as it is observed for the ligation of Chl in WSCP is in principle possible, even though no protein has been described so far that ligates heme with a backbone carbonyl. As discussed before, for binding of a tetrapyrrole to LvWSCP, one of the following interactions is necessary: hydrogen bond formation with the residue of the C7 side chain, hydrophobic interactions of the phytol chains and/or ligation of the central ion. Since heme has neither a phytol chain nor a group at the C7 position of its macrocycle that would allow hydrogen bonding, the central Fe2+ ion, according to these criteria, must be ligated to form a stable LvWSCP heme complex. To sum up, heme binds to WSCP. Especially the induction of WSCP tetramerization by heme and the induction of a heme CD signal with LvWSCP support the idea that heme is bound in the Chl binding site in a similar arrangement as Chl molecules. However, this needs further evaluation, ultimately by obtaining structural information of the WSCP heme complex. Implications for the biological function A half-century after the discovery of WSCPs, their biological function remains unclear. When Satoh et al.18 noticed for the first time that WSCP is able to extract Chl from the thylakoid membrane they speculated that WSCP might act as a carrier for Chl or its derivatives in reorganization processes of the photosynthetic apparatus under stress, or in Chl metabolism. Yet, it is still unknown how WSCP – if at all – enters the plastid, since all known WSCPs have a signal peptide that targets them to the endoplasmic reticulum (ER).15 Moreover, no WSCP has been found in plastids so far but rather in a special ER-derived compartment19,20,58 that is only found in a few plant families belonging to the Brassicales order: the ER bodies.59 Here, we present data that WSCP also binds heme. In plants, heme is solely synthesized in plastids, but is an essential cofactor for many enzymes and thus needed in all cellular compartments.2 Therefore, heme is exported from the plastids,60,61 yet the mechanism and transporters remain unknown.4 Since WSCP binds heme, which under most physiological conditions is poorly soluble, WSCP can in principle be envisioned as a water-soluble heme carrier that picks up heme at the plastid envelope membrane and transports it to its destination point. However, such a role as heme carrier is unlikely because WSCP is only found in Brassicaceae and it is expected that such a pivotal process like heme transport would be conserved throughout higher plants. Furthermore, ROS production upon illumination by heme should be suppressed in the carrier-heme complex,4 to protect the bound heme as well as other cellular components from photooxidation. Yet, the WSCP-Chl complex is an efficient producer of ROS, which is even highly photostable due to a special conformation of the phytol chains restricting access of ROS to the atoms at the chlorin macrocycle that are most vulnerable to oxidation.27,35 Consistently, Chl derivatives without phytol are not photostable anymore.27 Thus, WSCP-bound heme is expected to produce ROS upon illumination. Furthermore, other watersoluble heme-binding proteins have been identified that would be better suited for a function as heme carriers like cytosolic heme-binding proteins (cHBPs). cHBPs of the p22HBP/SOUL family are water-soluble proteins conserved in green algae and land plants62 that are able to bind heme and transfer it to apo-enzymes that need heme as cofactor.63 Thus, it was proposed that cHBPs might act as heme carriers that transport heme from the plastids to other cellular compartments,2,63 but this still awaits experimental proof. The efficient production of ROS by the WSCP Chl complex provoked speculations that WSCP might act as a ROS source for ROS signaling,27 which is a key regulator of responses to biotic and abiotic stresses as well as of programmed cell death.64,65 The hypothesis that WSCP is involved in stress response is further strengthened by its stress-induced expression. Different biotic and abiotic stress conditions like herbivore attacks,24,66 jasmonic acid treatment,25 drought,67,68 salt stress,69 leaf abscission70 and heat71 induce WSCP expression. However, it is still unclear how WSCP, which is present in the ER bodies, is exposed to Chl, which resides in the thylakoid membrane. For a function as ROS source, Chl binding would be essential. Takahashi et al.19,20 hypothesized that WSCP binds Chl after severe damage to the cell when the organelles are dismantled. However, under the described circumstances when the thylakoid membrane is severely damaged the Chl molecules of the photosynthetic apparatus are already uncoupled and produce ROS upon illumination. Thus, the benefit of Chl-binding to WSCP remains unclear at this point. In line with a potential involvement of WSCP in signaling, heme, Mg-Proto IX as well as other Chl precursors are involved in retrograde plastid-to-nucleus signaling to regulate the tetrapyrrole metabolism and to induce stress response mechanisms,4,72–74 but the mechanisms of tetrapyrrole signaling are poorly understood. Even though it has been proposed in the past that also Chl precursors are transported as signals to the nucleus,4 the current model by Terry and Bampton75 proposes two tetrapyrrole linked signals. Based on multiple experimental observations (for reviews see ref. 73 and 75) heme was proposed as a positive signal, which promotes the expression of photosynthesis-linked genes in the nucleus. However, it is unclear whether heme itself is transported as signal to the nucleus (and if so, how this might work), or whether a secondary messenger conveys the heme signal from the plastid to the nucleus. The other proposed tetrapyrrole retrograde signal in this model is the tetrapyrrole-produced ROS singlet oxygen (1O2), which is thought to be an inhibitor of the expression of photosynthesis-linked genes and an activator of stress- and cell death-related genes.75 Again, it is unknown how this signal is transferred from the plastid to the nucleus. Since 1O2 has a short lifetime, it is impossible that it diffuses from the plastid to the nucleus. Therefore, the plastid proteins Executer1 and 2, which mediate the induction of cell death when tetrapyrroles are overproduced in plastids,76 and carotenoid breakdown products, which have been suspected to signal O2 production by the photosynthetic apparatus under stress to the nucleus,77,78 have been suggested as 1O2 signal mediators. However, the exact tetrapyrrole signaling pathways for retrograde plastid-to-nucleus signaling and even the actual signal still remain unclear.75 Since WSCP binds tetrapyrroles and produces 1O2 upon illumination,27 it is possible that WSCP is involved in plastid-to-nucleus signaling. In addition, the low photostability of phytol-less tetrapyrroles27,35 appears to be beneficial for signaling, because it only allows a limited production of 1O2 before the tetrapyrrole itself is bleached and thus cannot act as photosensitizer anymore. Unfortunately, we do not know when tetrapyrrole and WSCP meet each other. WSCP is located in the ER bodies, whereas the tetrapyrroles, which can be synthesized in the thylakoid membrane, the stroma and the inner envelope membrane,3 are found in the plastids. In contrast to other tetrapyrroles, it has been shown that heme is exported from the plastid,60,61 but even then WSCP is still separated from heme by the ER membrane. Although glycosylated plastid proteins are targeted to the ER and then enter the plastid via the Golgi apparatus,79,80 this is unlikely for WSCP because of the lack of glycosylation sites. On the other hand, plants have permanent plastid-ER contact sites, where the plastid and ER membranes are presumably hemi-fused.81 Even though no protein transport mechanism is known, hydrophobic molecules are constantly transported at those contact sites from the ER to the plastid (phospholipids) and vice versa (amino acids, tocopherol, carotenoids, and fatty acids).82 Mehrshahi et al.81 even proposed a metabolic continuum between ER and plastid, because they could restore the activity of plastid-located enzymes involved in tocopherol and carotenoid metabolism in vivo when they retargeted the enzymes to the ER. Thus, also tetrapyrroles could be transported through the plastid-ER contact sites into the ER, where they then would bind to WSCP. However, it is unlikely that tetrapyrroles are transported as free molecules, as this might lead to uncontrolled production of ROS. Another previously discussed biological function of WSCP is that of a Chl-dependent protease inhibitor because all known WSCPs belong to the Künitz soybean trypsin inhibitor (STI) family.15 However, it remains inconclusive whether or not WSCP acts as a protease inhibitor. For the two most-studied WSCPs, LvWSCP and BobWSCP, a protease inhibitor activity was never observed, whereas the WSCP from Arabidopsis thaliana (AtWSCP) inhibited the papain-like cysteine proteases papain, proaleurain maturation protease from cauliflower and RD21A from Arabidopsis thaliana.21,23 Furthermore, Reinbothe and colleagues observed a complex between AtWSCP and the protease RD21 in etiolated Arabidopsis thaliana seedlings, which dissociates upon illumination and subsequent greening of the seedling probably due to Chl or PChlide binding to AtWSCP.23,26 Thus, they proposed that AtWSCP controls the activity of the protease RD21 in a Chl-dependent manner. First, RD21 and AtWSCP form a complex, in which RD21 is inhibited, then when the complex is exposed to Chl (or a precursor), AtWSCP binds Chl. Upon Chl-binding, AtWSCP detaches from RD21, which consequently becomes active. Protease inhibitors of the STI family and thus WSCPs probably as well inhibit their target proteases via the canonical standard mechanism, also known as Laskowski mechanism83 by reversibly binding to the active site of proteases in a substrate-like manner. The inhibitor resembles the substrate and binds to the protease in competition to the substrate. Thus, the inhibitor relies on a high affinity for the protease, and the proteaseinhibitor-complex is formed with KD values between 10−7 and 10−13 M.83 In good agreement, AtWSCP binds to RD21 with an apparent KD of 10−8 M.23 As a proposed regulator of the AtWSCP-RD21 interaction, Chl can either bind to AtWSCP at the interaction site between AtWSCP and the protease to block the interaction or somewhere else on the protein, which presumably induces a conformation change in AtWSCP and consequently reduces the affinity between AtWSCP and RD21. So far, all models for the interaction support that the Chl binding pocket in AtWSCP contains the interaction site between AtWSCP and RD21.22,26 Consequently, AtWSCP cannot bind Chl and RD21 at the same time. Thus, Chl replaces RD21 in the RD21-AtWSCP complex by either being present at a higher concentration or by binding to AtWSCP with a higher affinity than RD21. Here, we estimate that the KD values for the binding of Chl to WSCP are below 10−5 M. Thus, the affinity of AtWSCP for RD21 is probably higher than for Chl, which means that the Chl concentration has to be higher than the RD21 concentration to bind to AtWSCP. Since other tetrapyrroles are much lower concentrated in plant cells and tetrapyrroles lacking their phytol chain and central ion have even lower affinities for WSCP than Chl, they can be excluded as regulators for the AtWSCP-RD21 interaction. Yet again, it remains unknown when and where WSCP might be exposed to Chl. The hypothesis that a high Chl concentration is required for the activation of RD21, leaves only two options: either the AtWSCP-RD21 complex enters the chloroplast or the complex is exposed to Chl when the organelles are dismantled as it was proposed by Takahashi et al.19,20 The latter possibility would raise the question again why it is advantageous to activate a protease, when the cell is already bound to die after severe cell damage. Conclusions In this study, the binding of different tetrapyrroles to different versions of WSCPs was characterized. The determined affinities of Chl a/b binding for those WSCP versions are in good agreement with previously determined relative Chl a/b binding affinities, demonstrating that the Chl a/b selectivity is thermodynamically controlled. An investigation of the reaction kinetics revealed that all investigated WSCPs bind Chl a faster than Chl b. Moreover, affinities and time constants for Chlide and Pheo binding to WSCP were determined. WSCP binds both with reduced affinities compared to Chl, which suggests both the phytol chain and central Mg2+ ion to be important for the interaction between pigment and WSCP. Chlide is faster bound to WSCP than Chl, and Pheo is bound about as fast as Chl. WSCP is able to bind other tetrapyrroles like PChlide, MgProto IX, and heme, which implies that the size of the π electron system of the macrocycle, the fifth ring of the macrocycle and the type of central ion have little impact for the binding to WSCP. References 1 A. R. Battersby, Tetrapyrroles: the pigments of life, Nat. Prod. Rep., 2000, 17(6), 507–526. 2 N. Mochizuki, R. Tanaka, B. Grimm, T. Masuda, M. Moulin, A. G. Smith, A. Tanaka and M. J. Terry, The cell biology of tetrapyrroles: a life and death struggle, Trends Plant Sci., 2010, 15(9), 488–498. 3 R. Tanaka, K. Kobayashi and T. Masuda, Tetrapyrrole Metabolism in Arabidopsis thaliana, The arabidopsis book, 2011, vol. 9, p. e0145. 4 P. Brzezowski, A. S. Richter and B. Grimm, Regulation and function of tetrapyrrole biosynthesis in plants and algae, Biochim. Biophys. Acta, Bioenerg., 2015, 1847(9), 968–985. 5 N. Frankenberg-Dinkel and M. J. Terry, Synthesis and Role of Bilins in Photosynthetic Organisms, in Tetrapyrroles. Birth, Life and Death, ed. M. J. Warren and A. G. Smith, Springer-Verlag New York, New York, NY, 2009, pp. 208–220. 6 M. O. Senge, S. A. MacGowan and J. M. O’Brien, Conformational Protoporphyrin IX control of cofactors in nature – the influence of protein-induced macrocycle distortion on the biological function of tetrapyrroles, Chem. Commun., 2015, 51(96), 17031–17063.
7 R. G. L. op den Camp, D. Przybyla, C. Ochsenbein, C. Laloi, C. Kim, A. Danon, D. Wagner, E. Hideg, C. Göbel, I. Feussner, M. Nater and K. Apel, Rapid induction of distinct stress responses after the release of singlet oxygen in Arabidopsis, Plant Cell, 2003, 15(10), 2320–2332.
8 L. Fiedor, M. Zbyradowski and M. Pilch, Tetrapyrrole pigments of photosynthetic antennae and reaction centers of higher plants: Structures, biophysics, functions, biochemistry, mechanisms of regulation, applications, in Metabolism, structure and function of plant tetrapyrroles. Introduction, microbial and eukaryotic chlorophyll synthesis and catabolism, ed. B. Grimm, Elsevier, 2019, vol. 90, pp. 1–33.
9 H. Scheer, An Overview of Chlorophylls and Bacteriochlorophylls: Biochemistry, Biophysics, Functions and Applications, in Chlorophylls and Bacteriochlorophylls. Biochemistry, Biophysics, Functions and Applications, ed. B. Grimm, R. J. Porra, W. Rüdiger and H. Scheer, Springer, Dordrecht, 2006, pp. 1–26.
10 T. Mirkovic, E. E. Ostroumov, J. M. Anna, R. van Grondelle, Govindjee and G. D. Scholes, Light Absorption and Energy Transfer in the Antenna Complexes of Photosynthetic Organisms, Chem. Rev., 2017, 117(2), 249–293.
11 A. Krieger-Liszkay, C. Fufezan and A. Trebst, Singlet oxygen production in photosystem II and related protection mechanism, Photosynth. Res., 2008, 98(1), 551–564.
12 D. M. Palm, A. Agostini, V. Averesch, P. Girr, M. Werwie, S. Takahashi, H. Satoh, E. Jaenicke and H. Paulsen,
Chlorophyll a/b binding-specificity in water-soluble chlorophyll protein, Nat. Plants, 2018, 4(11), 920.
13 D. R. Ort, S. S. Merchant, J. Alric, A. Barkan, R. E. Blankenship, R. Bock, R. Croce, M. R. Hanson, J. M. Hibberd, S. P. Long, T. A. Moore, J. Moroney, K. K. Niyogi, M. A. J. Parry, P. P. Peralta-Yahya, R. C. Prince, K. E. Redding, M. H. Spalding, K. J. van Wijk, W. F. J. Vermaas, S. v. Caemmerer, A. P. M. Weber, T. O. Yeates, J. S. Yuan and X. G. Zhu, Redesigning photosynthesis to sustainably meet global food and bioenergy demand, Proc. Natl. Acad. Sci. U. S. A., 2015, 112(28), 8529– 8536.
14 K. Schmidt, C. Fufezan, A. Krieger-Liszkay, H. Satoh and H. Paulsen, Recombinant water-soluble chlorophyll protein from Brassica oleracea var. Botrys binds various chlorophyll derivatives, Biochemistry, 2003, 42(24), 7427–7433.
15 H. Satoh, A. Uchida, K. Nakayama and M. Okada, Watersoluble chlorophyll protein in Brassicaceae plants is a stress-induced chlorophyll-binding protein, Plant Cell Physiol., 2001, 42(9), 906–911.
16 Y. Kamimura, T. Mori, T. Yamasaki and S. Katoh, Isolation, properties and a possible function of a water-soluble chlorophyll a/b-protein from brussels sprouts, Plant Cell Physiol., 1997, 38(2), 133–138.
17 I. Bektas, C. Fellenberg and H. Paulsen, Water-soluble chlorophyll protein (WSCP) of Arabidopsis is expressed in the gynoecium and developing silique, Planta, 2012, 236(1), 251–259.
18 H. Satoh, K. Nakayama and M. Okada, Molecular cloning and functional expression of a water-soluble chlorophyll protein, a putative carrier of chlorophyll molecules in cauliflower, J. Biol. Chem., 1998, 273(46), 30568–30575.
19 S. Takahashi, H. Yanai, Y. Nakamaru, A. Uchida, K. Nakayama and H. Satoh, Molecular cloning, characterization and analysis of the intracellular localization of a water-soluble Chl-binding protein from Brussels sprouts (Brassica oleracea var. gemmifera), Plant Cell Physiol., 2012, 53(5), 879–891.
20 S. Takahashi, H. Yanai, Y. Oka-Takayama, A. Zanma- Sohtome, K. Fujiyama, A. Uchida, K. Nakayama and H. Satoh, Molecular cloning, characterization and analysis of the intracellular localization of a water-soluble chlorophyll-binding protein (WSCP) from Virginia pepperweed (Lepidium virginicum), a unique WSCP that preferentially binds chlorophyll b in vitro, Planta, 2013, 238(6), 1065– 1080.
21 C. E. Halls, S. W. Rogers, M. Oufattole, O. Østergard, B. Svensson and J. C. Rogers, A Kunitz-type cysteine protease inhibitor from cauliflower and Arabidopsis, Plant Sci., 2006, 170(6), 1102–1110.
22 E. Boex-Fontvieille, S. Rustgi, D. von Wettstein, S. Reinbothe and C. Reinbothe, Water-soluble chlorophyll protein is involved in herbivore resistance activation during greening of Arabidopsis thaliana, Proc. Natl. Acad. Sci. U. S. A., 2015, 112(23), 7303–7308.
23 E. Boex-Fontvieille, S. Rustgi, S. Reinbothe and C. Reinbothe, A Kunitz-type protease inhibitor regulates programmed cell death during flower development in Arabidopsis thaliana, J. Exp. Bot., 2015, 66(20), 6119–6135.
24 E. Boex-Fontvieille, S. Rustgi, D. von Wettstein, S. Pollmann, S. Reinbothe and C. Reinbothe, An EthyleneProtected Achilles’ Heel of Etiolated Seedlings for Arthropod Deterrence, Front. Plant Sci., 2016, 7, 1246.
25 E. Boex-Fontvieille, S. Rustgi, D. von Wettstein, S. Pollmann, S. Reinbothe and C. Reinbothe, Jasmonic acid protects etiolated seedlings of Arabidopsis thaliana against herbivorous arthropods, Plant Signaling Behav., 2016, 11(8), e1214349.
26 S. Rustgi, E. Boex-Fontvieille, C. Reinbothe, D. von Wettstein and S. Reinbothe, Serpin1 and WSCP differentially regulate the activity of the cysteine protease RD21 during plant development in Arabidopsis thaliana, Proc. Natl. Acad. Sci. U. S. A., 2017, 114(9), 2212–2217.
27 A. Agostini, D. M. Palm, F.-J. Schmitt, M. Albertini, M. Di Valentin, H. Paulsen and D. Carbonera, An unusual role for the phytyl chains in the photoprotection of the chlorophylls bound to Water-Soluble Chlorophyll-binding Proteins, Sci. Rep., 2017, 7(1), 7504.
28 S. Takahashi, M. Ono, A. Uchida, K. Nakayama and H. Satoh, Molecular cloning and functional expression of a water-soluble chlorophyll-binding protein from Japanese wild radish, J. Plant Physiol., 2013, 170(4), 406–412.
29 D. Horigome, H. Satoh, N. Itoh, K. Mitsunaga, I. Oonishi, A. Nakagawa and A. Uchida, Structural mechanism and photoprotective function of water-soluble chlorophyllbinding protein, J. Biol. Chem., 2007, 282(9), 6525–6531.
30 D. Bednarczyk, O. Dym, V. Prabahar, Y. Peleg, D. H. Pike and D. Noy, Fine Tuning of Chlorophyll Spectra by ProteinInduced Ring Deformation, Angew. Chem., Int. Ed., 2016, 55(24), 6901–6905.
31 V. Prabahar, L. Afriat-Jurnou, I. Paluy, Y. Peleg and D. Noy, New homologues of Brassicaceae water-soluble chlorophyll proteins shed light on chlorophyll binding, spectral tuning and molecular evolution, FEBS J., 2020, 287(5), 991–1004.
32 G. N. Ramachandran, A. V. Lakshminarayanan, R. Balasubramanian and G. Tegoni, Studies on the conformationof amino acids XII. Energy calculations on prolyl residue, Biochim. Biophys. Acta, Protein Struct., 1970, 221(2), 165–181.
33 J. L. Hughes, R. Razeghifard, M. Logue, A. Oakley, T. Wydrzynski and E. Krausz, Magneto-optic spectroscopy of a protein tetramer binding two exciton-coupled chlorophylls, J. Am. Chem. Soc., 2006, 128(11), 3649–3658.
34 A. Agostini, D. M. Palm, H. Paulsen and D. Carbonera, Optically Detected Magnetic Resonance of Chlorophyll Triplet States in Water-Soluble Chlorophyll Proteins from Lepidium virginicum: Evidence for Excitonic Interaction among the Four Pigments, J. Phys. Chem. B, 2018, 122(23), 6156–6163.
35 D. M. Palm, A. Agostini, A.-C. Pohland, M. Werwie, E. Jaenicke and H. Paulsen, Stability of Water-Soluble Chlorophyll Protein (WSCP) Depends on Phytyl Conformation, ACS Omega, 2019, 4(5), 7971–7979.
36 A. Agostini, E. Meneghin, L. Gewehr, D. Pedron, D. M. Palm, D. Carbonera, H. Paulsen, E. Jaenicke and E. Collini, How water-mediated hydrogen bonds affect chlorophyll a/b selectivity in Water-Soluble Chlorophyll Protein, Sci. Rep., 2019, 9(1), 18255.
37 P. J. Booth and H. Paulsen, Assembly of light-harvesting chlorophyll a/b complex in vitro. Time-resolved fluorescence measurements, Biochemistry, 1996, 35(16), 5103– 5108.
38 P. H. Hyninnen, Chemistry of chlorophylls: Modifications, in Chlorophylls, ed. H. Scheer, CRC Press, Boca Raton, Fla., 1991, pp. 145–209.
39 J. Kruk and B. Myśliwa-Kurdziel, Separation of Monovinyl and Divinyl Protochlorophyllides Using C30 Reverse Phase High Performance Liquid Chromatography Column: Analytical and Preparative Applications, Chromatographia, 2004, 60(1), 117–123.
40 D. M. Palm, A. Agostini, S. Tenzer, B. M. Gloeckle, M. Werwie, D. Carbonera and H. Paulsen, Water-Soluble Chlorophyll Protein (WSCP) Stably Binds Two or Four Chlorophylls, Biochemistry, 2017, 56(12), 1726–1736.
41 G. S. Collier, J. M. Pratt, C. R. de Wet and C. F. Tshabalala, Studies on haemin in dimethyl sulphoxide/water mixtures, Biochem. J., 1979, 179(2), 281–289.
42 F. Zsila, Z. Bikádi and M. Simonyi, Probing the binding of the flavonoid, quercetin to human serum albumin by circular dichroism, electronic absorption spectroscopy and molecular modelling methods, Biochem. Pharmacol., 2003, 65(3), 447–456.
43 T. Murata, F. Toda, K. Uchino and E. Yakushiji, Watersoluble chlorophyll protein of Brassica oleracea var. Botrys (cauliflower), Biochim. Biophys. Acta, Bioenerg., 1971, 245(1), 208–215.
44 T. Murata and N. Murata, Water-soluble chlorophyll-proteins from Brassica nigra and Lepidium virginicum L, Carnegie Institution Year Book, 1971, 70, 504–507.
45 M. Helfrich and W. Rüdiger, Various Metallopheophorbides as Substrates for Chlorophyll Synthetase, Z. Naturforsch., C: J. Biosci., 1992, 47(3–4), 231– 238.
46 C. J. Reedy, M. M. Elvekrog and B. R. Gibney, Development of a heme protein structure-electrochemical function database, Nucleic Acids Res., 2008, 36(Database issue), D307– D313.
47 M. C. Hsu and R. W. Woody, The origin of the heme Cotton effects in myoglobin and hemoglobin, J. Am. Chem. Soc., 1971, 93(14), 3515–3525.
48 M.-C. Hsu and R. W. Woody, Origin of the rotational strength of heme transitions of myoglobin, J. Am. Chem. Soc., 1969, 91(13), 3679–3681.
49 M. Nagai, Y. Nagai, Y. Aki, K. Imai, Y. Wada, S. Nagatomo and Y. Yamamoto, Effect of reversed heme orientation on circular dichroism and cooperative oxygen binding of human adult hemoglobin, Biochemistry, 2008, 47(2), 517– 525.
50 R. W. Woody and G. Pescitelli, The Role of Heme Chirality in the Circular Dichroism of Heme Proteins, Z. Naturforsch., A: Phys. Sci., 2014, 69(7), 313–325.
51 M. Nagai, N. Mizusawa, T. Kitagawa and S. Nagatomo, A role of heme side-chains of human hemoglobin in its function revealed by circular dichroism and resonance Raman spectroscopy, Biophys. Rev., 2018, 10(2), 271–284.
52 A. Brigé, D. Leys, T. E. Meyer, M. A. Cusanovich and J. J. van Beeumen, The 1.25 A resolution structure of the diheme NapB subunit of soluble nitrate reductase reveals a novel cytochrome c fold with a stacked heme arrangement, Biochemistry, 2002, 41(15), 4827–4836.
53 A. C. K. Chan, B. Lelj-Garolla, F. I. Rosell, K. A. Pedersen, A. G. Mauk and M. E. P. Murphy, Cofacial Heme Binding is Linked to Dimerization by a Bacterial Heme Transport Protein, J. Mol. Biol., 2006, 362(5), 1108–1119.
54 N. Chim, A. Iniguez, T. Q. Nguyen and C. W. Goulding, Unusual Diheme Conformation of the Heme-Degrading Protein from Mycobacterium tuberculosis, J. Mol. Biol., 2010, 395(3), 595–608.
55 N. Shipulina, A. Smith and W. T. Morgan, Heme Binding by Hemopexin: Evidence for Multiple Modes of Binding and Functional Implications, J. Protein Chem., 2000, 19(3), 239–248.
56 Y. Sugita, M. Nagai and Y. Yoneyama, Circular Dichroism of Hemoglobin in Relation to the Structure Surrounding the Heme, J. Biol. Chem., 1971, 246(2), 383–388.
57 C. Fufezan, J. Zhang and M. R. Gunner, Ligand preference and orientation in b- and c-type heme-binding proteins, Proteins, 2008, 73(3), 690–704.
58 S. Takahashi, K. Aizawa, K. Nakayama and H. Satoh, Watersoluble chlorophyll-binding proteins from Arabidopsis thaliana and Raphanus sativus target the endoplasmic reticulum body, BMC Res. Notes, 2015, 8, 365.
59 R. T. Nakano, K. Yamada, P. Bednarek, M. Nishimura and I. Hara-Nishimura, ER bodies in plants of the Brassicales order: biogenesis and association with innate immunity, Front. Plant Sci., 2014, 5, 73.
60 J. Thomas and J. D. Weinstein, Measurement of heme efflux and heme content in isolated developing chloroplasts, Plant Physiol., 1990, 94(3), 1414–1423.
61 R. van Lis, A. Atteia, L. A. Nogaj and S. I. Beale, Subcellular localization and light-regulated expression of protoporphyrinogen IX oxidase and ferrochelatase in Chlamydomonas reinhardtii, Plant Physiol., 2005, 139(4), 1946–1958.
62 A. Grossman, E. Sanz-Luque, H. Yi and W. Yang, Building the GreenCut2 suite of proteins to unmask photosynthetic function and regulation, Microbiology, 2019, 165(7), 697– 718.
63 S. Takahashi, T. Ogawa, K. Inoue and T. Masuda, Characterization of cytosolic tetrapyrrole -binding proteins in Arabidopsis thaliana, Photochem. Photobiol., 2008, 7(10), 1216–1224.
64 R. Mittler, S. Vanderauwera, N. Suzuki, G. Miller, V. B. Tognetti, K. Vandepoele, M. Gollery, V. Shulaev and F. van Breusegem, ROS signaling: the new wave?, Trends Plant Sci., 2011, 16(6), 300–309.
65 A. Baxter, R. Mittler and N. Suzuki, ROS as key players in plant stress signalling, J. Exp. Bot., 2014, 65(5), 1229–1240.
66 H. Vogel, J. Kroymann and T. Mitchell-Olds, Different transcript patterns in response to specialist and generalist herbivores in the wild Arabidopsis relative Boechera divaricarpa, PLoS One, 2007, 2(10), e1081.
67 W. L. Downing, F. Mauxion, M. O. Fauvarque, M. P. Reviron, D. de Vienne, N. Vartanian and J. Giraudat, A Brassica napus transcript encoding a protein related to the Künitz protease inhibitor family accumulates upon water stress in leaves, not in seeds, Plant J., 1992, 2(5), 685– 693.
68 M. P. Reviron, N. Vartanian, M. Sallantin, J. C. Huet, J. C. Pernollet and D. de Vienne, Characterization of a Novel Protein Induced by Progressive or Rapid Drought and Salinity in Brassica napus Leaves, Plant Physiol., 1992, 100(3), 1486–1493.
69 F. Lopez, G. Vansuyt, P. Fourcroy and F. Casse-Delbart, Accumulation of a 22 kDa protein and its mRNA in the leaves of Raphanus sativus in response to salt stress or water deficit, Physiol. Plant., 1994, 91(4), 605–614.
70 N. Nishio and H. Satoh, A water-soluble chlorophyll protein in cauliflower may be identical to BnD22, a droughtinduced, 22-kilodalton protein in rapeseed, Plant Physiol., 1997, 115(2), 841–846.
71 G. Ilami, C. Nespoulous, J.-C. Huet, N. Vartanian and J.-C. Pernollet, Characterization of BnD22, a droughtinduced protein expressed in Brassica napus leaves, Phytochemistry, 1997, 45(1), 1–8.
72 D. S. K. Nagahatenna, P. Langridge and R. Whitford, Tetrapyrrole-based drought stress signalling, Plant Biotechnol. J., 2015, 13(4), 447–459.
73 R. M. Larkin, Tetrapyrrole Signaling in Plants, Front. Plant Sci., 2016, 7, 1586.
74 J. E. Cornah, M. J. Terry and A. G. Smith, Green or red: what stops the traffic in the tetrapyrrole pathway?, Trends Plant Sci., 2003, 8(5), 224–230.
75 M. J. Terry and J. Bampton, The role of tetrapyrroles in chloroplast-to-nucleus retrograde signaling, in Metabolism, Structure and Function of Plant Tetrapyrroles: Control Mechanisms of Chlorophyll Biosynthesis and Analysis of Chlorophyll-Binding Proteins, Elsevier, 2019, vol. 91, pp. 225–246.
76 K. P. Lee, C. Kim, F. Landgraf and K. Apel, EXECUTER1and EXECUTER2-dependent transfer of stress-related signals from the plastid to the nucleus of Arabidopsis thaliana, Proc. Natl. Acad. Sci. U. S. A., 2007, 104(24), 10270– 10275.
77 F. Ramel, S. Birtic, C. Ginies, L. Soubigou-Taconnat, C. Triantaphylides and M. Havaux, Carotenoid oxidation products are stress signals that mediate gene responses to singlet oxygen in plants, Proc. Natl. Acad. Sci. U. S. A., 2012, 109(14), 5535–5540.
78 L. Shumbe, R. Bott and M. Havaux, Dihydroactinidiolide, a High Light-Induced β-Carotene Derivative that Can Regulate Gene Expression and Photoacclimation in Arabidopsis, Mol. Plant, 2014, 7(7), 1248–1251.
79 A. Villarejo, S. Buren, S. Larsson, A. Dejardin, M. Monne, C. Rudhe, J. Karlsson, S. Jansson, P. Lerouge, N. Rolland, G. von Heijne, M. Grebe, L. Bako and G. Samuelsson, Evidence for a protein transported through the secretory pathway en route to the higher plant chloroplast, Nat. Cell Biol., 2005, 7(12), 1224–1231.
80 R. N. Radhamony and S. M. Theg, Evidence for an ER to Golgi to chloroplast protein transport pathway, Trends Cell Biol., 2006, 16(8), 385–387.
81 P. Mehrshahi, G. Stefano, J. M. Andaloro, F. Brandizzi, J. E. Froehlich and D. DellaPenna, Transorganellar complementation redefines the biochemical continuity of endoplasmic reticulum and chloroplasts, Proc. Natl. Acad. Sci. U. S. A., 2013, 110(29), 12126–12131.
82 J. Pérez-Sancho, J. Tilsner, A. L. Samuels, M. A. Botella, E. M. Bayer and A. Rosado, Stitching Organelles: Organization and Function of Specialized Membrane Contact Sites in Plants, Trends Cell Biol., 2016, 26(9), 705–717.
83 M. Laskowski and I. Kato, Protein inhibitors of proteinases, Annu. Rev. Biochem., 1980, 49, 593–626.