Chlorophyll a phytylation is required for the stability of photosystems I and II in the cyanobacterium Synechocystis sp. PCC 6803

Authors

  • Alexey V. Shpilyov,

    1. Biology Division, Genetics Department, Lomonosov Moscow State University, Moscow, Russia
    2. Institut für Biologie/Pflanzenphysiologie, Humboldt-Universität zu Berlin, Berlin, Germany
    Search for more papers by this author
  • Vladislav V. Zinchenko,

    1. Biology Division, Genetics Department, Lomonosov Moscow State University, Moscow, Russia
    Search for more papers by this author
  • Bernhard Grimm,

    1. Institut für Biologie/Pflanzenphysiologie, Humboldt-Universität zu Berlin, Berlin, Germany
    Search for more papers by this author
  • Heiko Lokstein

    Corresponding author
    1. Institut für Biologie/Pflanzenphysiologie, Humboldt-Universität zu Berlin, Berlin, Germany
    2. Institut für Biologie III, Albert-Ludwigs-Universität Freiburg, Freiburg, Germany
    Current affiliation:
    1. Glasgow Biomedical Research Centre, Institute of Molecular, Cell and Systems Biology, University of Glasgow, Glasgow, UK
    • Biology Division, Genetics Department, Lomonosov Moscow State University, Moscow, Russia
    Search for more papers by this author

For correspondence (e-mail lokstein@gmx.de).

Summary

In oxygenic phototrophic organisms, the phytyl ‘tail’ of chlorophyll a is formed from a geranylgeranyl residue by the enzyme geranylgeranyl reductase. Additionally, in oxygenic phototrophs, phytyl residues are the tail moieties of tocopherols and phylloquinone. A mutant of the cyanobacterium Synechocystis sp. PCC 6803 lacking geranylgeranyl reductase, ΔchlP, was compared to strains with specific deficiencies in either tocopherols or phylloquinone to assess the role of chlorophyll a phytylatation (versus geranylgeranylation). The tocopherol-less Δhpt strain grows indistinguishably from the wild-type under ‘standard’ light photoautotrophic conditions, and exhibited only a slightly enhanced rate of photosystem I degradation under strong irradiation. The phylloquinone-less ΔmenA mutant also grows photoautotrophically, albeit rather slowly and only at low light intensities. Under strong irradiation, ΔmenA retained its chlorophyll content, indicative of stable photosystems. ΔchlP may only be cultured photomixotrophically (due to the instability of both photosystems I and II). The increased accumulation of myxoxanthophyll in ΔchlP cells indicates photo-oxidative stress even under moderate illumination. Under high-light conditions, ΔchlP exhibited rapid degradation of photosystems I and II. In conclusion, the results demonstrate that chlorophyll a phytylation is important for the (photo)stability of photosystems I and II, which, in turn, is necessary for photoautotrophic growth and tolerance of high light in an oxygenic environment.

Introduction

Chlorophyll (Chl) a is the key pigment involved in the primary reactions of oxygenic photosynthesis – the global biological process that provides primary biomass and energy for almost all living beings, and, additionally, supplies oxygen for respiration.

In Chl biosynthesis, the enzyme geranylgeranyl reductase (GGR, also designated ChlP) reduces either geranylgeranyl diphosphate to phytyl diphosphate or a side chain of geranylgeranylated Chl a (Chl aGG) to yield (phytylated) Chl a. In parallel, Chl synthase (ChlG) esterifies chlorophyllide a using either geranylgeranyl diphosphate or phytyl diphosphate, producing Chl aGG or Chl a, respectively (Figure S1) (Soll and Schultz, 1981; Keller et al., 1998; Shpilyov et al., 2005; Rüdiger, 2006). Additionally, Chls with di- and tetrahydrogeranylgeranyl ‘tails’ (Chl aDHGG and Chl aTHGG, respectively) may be formed due to incomplete reduction of geranylgeranyl residues by GGR (Table S1) (Maloney et al., 1989; Domanskii et al., 2003).

In oxygenic phototrophs, GGR is also involved in the synthesis of tocopherols (Figure S1) (Keller et al., 1998; Tanaka et al., 1999; Shibata et al., 2004a). Together with tocotrienols – analogs that have an unsaturated isoprenoid side chain – these compounds comprise a group of lipid-soluble antioxidants collectively referred to as tocochromanols (or vitamin E), with α–tocopherol being the predominant natural form (Table S1) (reviewed by Dörmann, 2007; Falk and Munné-Bosch, 2010; DellaPenna and Mène-Saffrané, 2011). The major role assumed for α–tocopherol is prevention of oxidation of membrane lipids triggered by reactive oxygen species. Additionally, α–tocopherol has been suggested to protect photosystem (PS) II from photoinhibition (Trebst, 2003; Krieger-Liszkay and Trebst, 2006; Inoue et al., 2011). Moreover, α–tocopherol is believed to be involved in the regulation of intracellular signaling, macronutrient homeostasis, osmotolerance, seed longevity, seedling and root development, growth rate, etc. (reviewed by Dörmann, 2007; Falk and Munné-Bosch, 2010; DellaPenna and Mène-Saffrané, 2011).

In addition to Chl and tocopherol synthesis, GGR participates in formation of phylloquinone (vitamin K1) (Figure S1 and Table S1) in oxygenic phototrophs. Phylloquinone functions as secondary electron acceptor at the A1 site of PSI (Keller et al., 1998; Johnson and Golbeck, 2004; Shibata et al., 2004b; Srinivasan and Golbeck, 2009; Ohashi et al., 2010).

In GRR-deficient plant and cyanobacterial mutants, Chl aGG accumulates (in some cases together with Chl aDHGG and Chl aTHGG) instead of phytylated Chl a. Chl aGG may be incorporated into photosynthetic pigment–protein complexes and even mediate light-induced electron transport in the mutants (Tanaka et al., 1999; Shibata et al., 2004a,b; Shpilyov et al., 2005). However, Chl aGG, Chl aDHGG and Chl aTHGG do not naturally occur in mature chloroplasts or cyanobacterial cells (Tamiaki et al., 2007). Moreover, full replacement of Chls with counterparts that have unsaturated tails abolishes photoautotrophic growth in plants (Shibata et al., 2004a), green algae (Henry et al., 1986) and cyanobacteria (Shpilyov et al., 2005). Altogether, these facts appear to indicate that Chl a species with only partially saturated tails may not be able to fulfill all the role(s) of phytylated Chl a, and may even have deleterious effects in oxygenic photosynthetic organisms.

However, more information is required for better understanding of the importance of the phytyl tail of Chl a and thus the significance of the GGR enzyme for oxygenic photosynthesis – also taking into account that the GGR-catalyzed reaction requires energy and redox equivalents (Schoch and Schäfer, 1978). The cyanobacterium Synechocystis sp. PCC 6803 (hereafter Synechocystis) is a suitable organism to study mutations that impair photosynthesis, as glucose-tolerant strain(s) are able to grow heterotrophically using glucose as external energy/carbon source (Ikeuchi and Tabata, 2001). We have previously described a GGR-deficient mutant (ΔchlP) of Synechocystis (Shpilyov et al., 2005). The goal of the present work was to distinguish the effects of replacement of (phytylated) Chl a with Chl aGG from possible effects of α–tocopherol and phylloquinone deficiency in the mutant. ΔchlP was compared to the Synechocystis Δhpt and ΔmenA mutants with inactivated homogentisate phytyl transferase (HPT) and 1,4–dihydroxy-2–naphthoate phytyl transferase (DHNA phytyl transferase, MenA), respectively, which specifically control tocopherol and phylloquinone formation (Figures S1 and S2) (Johnson et al., 2000; Collakova and DellaPenna, 2001).

Together with published data, the results demonstrate that, in the ΔchlP mutant, deficiency of neither tocopherol nor phylloquinone but instead accumulation of Chl aGG leads to instability of both PSI and PSII. Hence, Chl a phytylation appears to be crucial for photoautotrophic growth and prevention of photo-oxidative stress by ensuring the (photo)stability of photosynthetic pigment–protein complexes.

Results

Comparison of ΔchlP with the tocopherol-deficient Δhpt mutant

In Synechocystis, α–tocopherol accumulates as the only vitamin E species (Collakova and DellaPenna, 2001; Savidge et al., 2002; Shpilyov et al., 2005) (see also Figure S3a). As reported previously, inactivation of GGR in the ΔchlP mutant leads to accumulation of α–tocotrienol instead of α–tocopherol (Table S1) (Shpilyov et al., 2005). This provides evidence of the ability of the HPT enzyme to utilize geranylgeranyl diphosphate as a substrate for condensation with homogentisate in vivo in Synechocystis (Figure S1), corroborating results obtained in vitro (Collakova and DellaPenna, 2001). In accordance with the established pathway of vitamin E formation, no tocochromanols were detected in cells of the HPT-deficient Δhpt strain (Figures S1–S3b).

The ΔchlP mutant only grows photomixotrophically (Shpilyov et al., 2005). In contrast, Δhpt possesses a phenotype similar to the wild-type (WT) strain when grown photoautotrophically at a light intensity of 40 μmol photons m−2 sec−1 (‘standard’ light). This has also been documented for other Synechocystis mutants that are impaired in synthesis of vitamin E (Collakova and DellaPenna, 2001; Dähnhardt et al., 2002; Savidge et al., 2002; Sattler et al., 2003; Sakuragi et al., 2006). However, photomixotrophic cultivation at the same light intensity was found to have some adverse effects on the Δhpt strain. The mutant has a tendency to grow slightly more slowly than the WT, although the difference between the doubling times is not dramatic (Table 1). Both Δhpt and ΔchlP exhibit reduced Chl a and total carotenoid contents. However, the reduction of these pigments in ΔchlP is more pronounced than in Δhpt (Figure 1 and Table 1). Furthermore, phycobilisome (PBS) content is reduced in Δhpt but increased in ΔchlP (Figure 1).

Figure 1.

Room-temperature absorption spectra of WT (solid line), Δhpt (dotted line) and ΔchlP (dashed line) cells grown photomixotrophically at 40 μmol photons m−2 sec−1. Spectra were normalized on a per cell basis at 750 nm.

Table 1. Growth rates and pigment contents in the WT, Δhpt, ΔmenA and ΔchlP strainsa
 WTΔhptΔmenAΔchlP
  1. NG, no growth.

  2. a

    During growth, aeration was provided with ambient air by stirring. Data are mean values derived from three to five measurements.

  3. b

    Pigments were extracted using 90% v/v methanol/water from cells grown exponentially in the presence of 10 mm glucose.

Photoautotrophic growth, doubling time (h)
At 40 μmol photons m−2 sec−111.4 ± 1.011.6 ± 1.1NGNG
At 20 μmol photons m−2 sec−118.2 ± 3.1 42.28 ± 5.3NG
Photomixotrophic growth, doubling time (h)
At 40 μmol photons m−2 sec−19.3 ± 0.59.8 ± 0.7NG9.2 ± 0.4
At 20 μmol photons m−2 sec−112.1 ± 0.8 32.45 ± 4.1 
Chlorophyll contentb (μg ml−1 OD750−1)
At 40 μmol photons m−2 sec−12.59 ± 0.252.31 ± 0.23NG1.98 ± 0.20
At 20 μmol photons m−2 sec−12.56 ± 0.19 2.03 ± 0.12 
Total carotenoid contentb (μg ml−1 OD750−1)
At 40 μmol photons m−2 sec−10.65 ± 0.080.60 ± 0.06NG0.47 ± 0.04
At 20 μmol photons m−2 sec−10.63 ± 0.07 0.62 ± 0.06 

The 77 K fluorescence emission spectrum of Δhpt cells recorded upon Chl a excitation (at 435 nm) is essentially identical to that of the WT (Figure 2). The only difference found is a slight relative decrease in PSI as indicated by the decreased 725 nm peak (PSI fluorescence) and an unaffected 695 nm maximum (PSII). Thus, the reduction of the cellular Chl a level in Δhpt is due to a decreased PSI content. In contrast to Δhpt, the PSI emission peak in the fluorescence spectrum of ΔchlP is considerably decreased. Moreover, the PSI emission maximum is blue-shifted by 2 nm. A fluorescence band peaking at 684 nm is increased in the mutant (Figure 2). These spectral properties suggest that the structures of both PSI and PSII may be somewhat perturbed in the ΔchlP mutant. This was not observed for Δhpt.

Figure 2.

77 K fluorescence emission spectra of WT (solid line), Δhpt (dotted line) and ΔchlP (dashed line) cells grown photomixotrophically at 40 μmol photons m−2 sec−1. Excitation was at 435 nm.

Although it grows well at 40 μmol photons m−2 sec−1, ΔchlP cannot grow under light of approximately 100 μmol photons m−2 sec−1 and higher, intensities that are permissive for the WT (Shpilyov et al., 2005). To clarify whether α–tocopherol deficiency is responsible for light sensitivity in ΔchlP, we compared this mutant to Δhpt under a high light intensity of approximately 500 μmol photons m−2 sec−1 together with external glucose supply (see 'Experimental procedures'). Following transfer to increased illumination, growth ceased immediately in all strains (Figure 3a). Synechocystis sp. PCC 6803 was previously reported to be able to grow photoautotrophically at such a light intensity (Steiger et al., 1999; He et al., 2001). Hence, the observed growth arrest – particularly in the WT – was probably a result of glucose supply. For cyanobacteria, glucose is known to donate additional electrons to the photosynthetic electron transport chain (ETC) by reduction of plastoquinone–9 via respiratory dehydrogenases (Vermaas, 2001; Wang et al., 2002; Egorova et al., 2006). Under excess light conditions, glucose probably causes over-reduction of the plastoquinone pool and thereby the whole ETC, thus provoking strong photo-oxidative stress and abolishing cell division (Krieger-Liszkay, 2005; Ledford and Niyogi, 2005; Telfer, 2005). Note that glucose improves growth in glucose-tolerant Synechocystis strains at light intensities up to approximately 150–200 μmol photons m−2 sec−1 (Ikeuchi and Tabata, 2001; Johnson et al., 2001; Wang et al., 2002; Sakuragi, 2004; Shpilyov et al., 2005) (see also Table 1), but has an inhibitory effect under stronger irradiation (this study).

Figure 3.

Cell densities (a), Chl a (b) and total carotenoid (c) contents in WT (circles), Δhpt (triangles) and ΔchlP cultures (squares) during incubation at 500 μmol photons m−2 sec−1 in the presence of glucose. The 0 h time point corresponds to cells grown photomixotrophically at 40 μmol photons m−2 sec−1.

When subjected to high light intensity, all strains showed pigment degradation. Both WT and Δhpt exhibited similarly slow Chl degradation kinetics, as assessed by spectrophotometry and HPLC (Figures 3b and 4, respectively) and whole-cell absorption spectra (Figure 6a,b). After 10 h of strong light, WT and Δhpt still retained relatively high Chl levels, i.e. approximately 76 and 74% of the initial values, respectively (Figure 3b). Additionally, WT and Δhpt showed a transient increase in carotenoid contents during the first 4–7 h. Thereafter, total carotenoids returned to close to the initial levels, approximately 104 and 88% for WT and Δhpt, respectively (Figure 3c). The transient accumulation of carotenoids in the WT and Δhpt strains indicates that they were both able to combat photo-oxidative stress (at least for some time) by up-regulation of carotenoid synthesis, in accordance with published data (Steiger et al., 1999; Maeda et al., 2005). However, after 4–7 h, all pigments started to degrade. Degradation proceeded slightly faster in Δhpt than in the WT (Figures 3b,c and 4, Figure 6). The latter observation implies a somewhat increased light sensitivity for the Δhpt strain that lacks vitamin E.

Figure 4.

HPLC chromatograms of pigment extracts from WT, Δhpt and ΔchlP cells at various time points of incubation at 500 μmol photons m−2 sec−1 in the presence of glucose. Strains are indicated at the top, and time points (hours) of incubation are indicated on the right. The 0 h time point corresponds to cells grown photomixotrophically at 40 μmol photons m−2 sec−1. Peaks: 1, myxoxanthophyll; 2, zeaxanthin; 3a, (phytylated) Chl a; 3b, Chl aGG; 4, echinenone; 5, β–carotene.

However, the ΔchlP mutant displayed a rather different pigment profile (beyond replacement of Chl a with Chl aGG as reported previously; Shpilyov et al., 2005) as well as different degradation kinetics. ΔchlP eventually lost nearly 100% of its Chl over the course of the experiment (Figures 3b and 4, Figure 6d). Additionally, in contrast to Δhpt and WT, ΔchlP exhibited no increase in total carotenoid content, and the carotenoid level significantly decreased (to approximately 28%) in ΔchlP cells during high-light treatment (Figure 3c). A distinctive exception appears to be myxoxanthophyll. Increased levels of this carotenoid (exceeding those in WT and Δhpt) were found in ΔchlP cells even under standard light conditions, and remained relatively high until 4 h of light stress, when significant decay of all other pigments had already occurred (Figure 4). Myxoxanthophyll (myxol 2′–dimethyl-fucoside in Synechocystis sp. PCC 6803) is assumed to be a photo-protective carotenoid that is specific to cyanobacteria (Takaichi et al., 2001). An elevated level of myxoxanthophyll is considered to be symptomatic of photo-oxidative stress because it is commonly observed under photo-inhibitory conditions, e.g. high light, UV irradiation and low temperature (Ehling-Schulz et al., 1997; Steiger et al., 1999; Miśkiewicz et al., 2000; Takaichi et al., 2001; Maeda et al., 2005; Schäfer et al., 2005) (see also Figure 4 for WT and Δhpt). Thus, the increased amount of myxoxanthophyll in ΔchlP – especially under standard light conditions – indicates that the mutant is already stressed by moderate light.

The impact of excess light on both photosystems was also assessed using a series of 77 K fluorescence emission spectra recorded upon Chl a excitation (Figure 5). WT and Δhpt display essentially identical spectral patterns, revealing a slight decrease in both PSI and PSII contents during incubation under high light intensity. Again, Δhpt appears to be somewhat more affected by high light than the WT, at least with respect to the PSI content (Figure 5a,b). However, the decrease in PSI fluorescence in the ΔchlP spectra is much more dramatic (Figure 5c). Moreover, inversion of the 667/684 nm peak ratio with a sharp decrease in emission at 684 nm is also observed. These data indicate that both PSI and PSII underwent rapid degradation in ΔchlP. Interestingly, the decrease in emission at 684 nm, indicative of PSII decay, occurred after 4 h, which correlates with the kinetics of myxoxanthophyll decrease in the mutant during high-light exposure (Figures 4 and 5). This may indicate that, under photo-inhibitory conditions, myxoxanthophyll may specifically contribute to protection of PSII in cyanobacteria.

Figure 5.

77 K fluorescence emission spectra of WT (a), Δhpt (b) and ΔchlP (c) cells at various time points of incubation at 500 μmol photons m−2 sec−1 in the presence of glucose. Excitation was at 435 nm. The corresponding time points (hours) of incubation are given on the right. The 0 h time point corresponds to cells grown photomixotrophically at 40 μmol photons m−2 sec−1.

Comparison of ΔchlP with the phylloquinone-deficient ΔmenA mutant

In oxygenic phototrophs, GGR is additionally involved in formation of phylloquinone (Figure S1 and Table S1). Thus, ΔchlP was also compared to a Synechocystis mutant with specifically interrupted phylloquinone synthesis. The latter, designated ΔmenA, was constructed by genetic knockout of 1,4–dihydroxy-2–naphthoate (DHNA) phytyl transferase (MenA), an enzyme that is specific to the phylloquinone biosynthetic pathway (Figures S1 and S2).

Synechocystis mutants disrupted in MenA activity (menA) (Figure S1) and earlier committed steps of phylloquinone biosynthesis (menB, D and E mutants) have been previously described in detail, and appear to be rather similar to each other (reviewed by Johnson and Golbeck, 2004; Srinivasan and Golbeck, 2009). The phenotype of the ΔmenA strain is identical to the aforementioned mutants, particularly the analogous mutant menA, previously described by Johnson et al. (2000). ΔmenA shows a pale olive green coloration due to decreased Chl and PBS contents (Table 1 and Figure S4). Additionally, the mutant cannot grow under illumination exceeding approximately 30–35 μmol photons m−2 sec−1 (Table 1). Thus, our ΔmenA mutant shares an obvious light sensitivity consistent with the reported menA, B, D and E mutants. Also consistent with previous reports, ΔmenA exhibits slow but steady growth at reduced light intensity (e.g. 20 μmol photons m−2 sec−1), particularly under photoautotrophic conditions (Table 1). As established previously, photoautotrophy is retained due to recruitment of plastoquinone–9 (Table S1) into the A1 site of PSI instead of phylloquinone (Johnson et al., 2000; Zybailov et al., 2000). Upon incorporation of plastoquinone into PSI, the whole-chain electron transport rate is reduced by approximately 40% (Johnson et al., 2000; own observations). However, the ability of the men mutants to grow photoautotrophically suggests two remarkable conclusions for Synechocystis. First, phylloquinone is not strictly required to sustain electron transport through PSI and may be substituted for by other quinone(s). Second, even when binding a rather different quinone (e.g. plastoquinone–9; Table S1), PSI remains stable enough to retain photoautotrophy in cells. These conclusions are complemented and extended by numerous in vitro and in vivo studies (see 'Discussion'). In contrast, both PSI and PSII are unstable in ΔchlP, abolishing the capability for photoautotrophic growth in the mutant (Shpilyov et al., 2005).

To assess the resistance of the photosystems towards irradiation in ΔmenA, the mutant was also compared to the other strains in high-light experiments as described above. Interestingly, despite the light sensitivity of ΔmenA mentioned above, the mutant retained stable levels of pigments – particularly Chl – under strong illumination as shown in a series of absorption spectra (Figure 6c) and photometric measurements (Figure S5). These data clearly indicate the stability of both PSI and PSII in the ΔmenA mutant under high-light conditions, in contrast to ΔchlP.

Figure 6.

Room-temperature absorption spectra of WT (a), Δhpt (b), ΔmenA (c) and ΔchlP (d) cells at various time points of incubation at 500 μmol photons m−2 sec−1 in the presence of glucose. The corresponding time points (hours) of incubation are given on the right. The 0 h time point corresponds to cells grown photomixotrophically at 40 μmol photons m−2 sec−1, except ΔmenA, which was grown at 20 μmol photons m−2 sec−1 before incubation at 500 μmol photons m−2 sec−1.

Discussion

Effects of tocopherol deficiency

In Synechocystis, α–tocopherol has been proposed to have specific (probably regulatory) homeostatic roles beyond its antioxidant activity. A Synechocystis HPT-deficient mutant (Figure S1) was reported to be extremely glucose-sensitive (Sakuragi et al., 2006). However, our Δhpt strain is only slightly compromised under photomixotrophic conditions. The discrepancy between the data is resolved if the pH of the growth medium is considered. The previously described mutant (slr1736  ) was inhibited by added glucose at pH approximately 7.2 and lower, whereas at higher pH (e.g. 7.6 and 8.0), the mutant grew similarly to the WT (Sakuragi et al., 2006). In our experiments, the cultures were started at an initial pH approximately 8.0, not buffered and aerated with ambient air. The pH rose gradually to approximately 9.0 during growth until the stationary phase was reached. Good photomixotrophic growth was also observed by other researchers for an analogous mutant under a similar aeration regime with no buffer added (Sakuragi, 2004). The initial pH of their medium was also 8.0; hence, the growth conditions were similar to ours (Sakuragi, 2004). However, if the medium was bubbled with 3% CO2-enriched air and not buffered, the pH eventually decreased to approximately 7.0. Consequently, growth was inhibited, with concomitant strong bleaching of the culture (Sakuragi, 2004). However, if under the same conditions, the pH was kept at the initial value (8.0) using HEPES buffer, the mutant displayed persistent growth and phenotype (Sakuragi, 2004; Sakuragi et al., 2006). Thus, the cultivation conditions used in our study were permissive for the Δhpt strain. Under photoautotrophic conditions, pH does not influence the Synechocystis strains that lack α–tocopherol, e.g., slr1736, slr0089 and slr0090 (Sakuragi, 2004; Sakuragi et al., 2006).

Nevertheless, adverse effects of α–tocopherol deficiency, such as reduced PBS and PSI contents, were observed in the Δhpt strain grown photomixotrophically, even at high pH (Figures 1 and 2, and Table 1). These observations are consistent with results obtained by other researchers: in the presence of glucose at pH 8.0, the HPT-deficient mutant used by Sakuragi (2004) showed a pale green–yellow coloration, as did our Δhpt strain, due to slight de-pigmentation, particularly decreased PBS content (Figure 1). Thus, the present study corroborates previous data indicating that α–tocopherol may play specific homeostatic roles in Synechocystis (Sakuragi, 2004; Sakuragi et al., 2006). Consequently, although ΔchlP accumulates α–tocotrienol (Shpilyov et al., 2005), which also has a high antioxidant potency (Yoshida et al., 2003), loss of α–tocopherol was hypothesized to be critical in the mutant. Thus, one may expect the phenotypes of ΔchlP and Δhpt to be very similar, given that both strains lack α–tocopherol.

However, the comparative analyses revealed specific features for the ΔchlP mutant. The most striking trait is loss of photoautotrophic growth. Glucose appears to provide an additional energy/carbon source to maintain steady-state levels (through high rates of re-synthesis) of PSI and PSII in ΔchlP (Shpilyov et al., 2005). In contrast, the stability of PSI/II in Δhpt is comparable to that of the WT, as deduced from its robust photoautotrophic growth. Additionally, the 77 K fluorescence emission spectra imply that both PSI and PSII in ΔchlP are structurally perturbed, which is not observed in Δhpt (Figure 2).

The slightly increased light sensitivity of Δhpt (Figures 3-6) cannot be interpreted unambiguously. A possible reason is the absence of α–tocopherol as an antioxidant. Alternatively, this and other specific traits of the strain, e.g. somewhat different PBS, PSI and carotenoid contents (Figures 1-4 and Table 1), may be a side effect of glucose in a tocopherol-less background. Clarification of this issue requires further investigations beyond the scope of the present study.

The present study revealed extremely fast degradation of PSI and PSII induced by strong light in ΔchlP (Figures 3, 5 and 6). Evidently, α–tocopherol deficiency is not the determining factor in this regard, as, under the same conditions, Δhpt displayed only a slightly increased PSI instability in comparison to the WT (Figure 5). Furthermore, the abundant myxoxanthophyll in ΔchlP cells grown under standard light (Figure 4) indicates that the mutant experiences photo-oxidative stress, even under moderate illumination. This observation is corroborated by the more rapid degradation of PSI/II in the mutant under light of 40 μmol photons m−2 sec−1 than in darkness (Shpilyov et al., 2005). Under stronger irradiation (e.g. 500 μmol photons m−2 sec−1), photo-oxidative stress is further aggravated in the mutant, leading to very fast destruction of PSI and PSII. This type of enhanced light sensitivity was not observed in Δhpt and the other Synechocystis strains that lack vitamin E (Dähnhardt et al., 2002; Savidge et al., 2002; Sattler et al., 2003; Sakuragi, 2004; Sakuragi et al., 2006).

Thus, the present data indicate that both PSI and PSII are structurally altered in ΔchlP, unstable and very vulnerable to photodegradation, and these traits are not due to α-tocopherol deficiency in the mutant.

Effects of phylloquinone deficiency

The function of phylloquinone (vitamin K1) (Table S1) in the A1 site of PSI has been extensively explored for more than two decades. Approaches such as reconstitution of the A1 site with foreign quinones in vitro as well as in vivo were very informative. These studies revealed that diverse benzo-, naphtho- and anthraquinones – and even so-called ‘quinonoids’ – bind to the A1 site and function as efficient redox co-factors (Table S1) (reviewed by Ikegami et al., 2000; Itoh et al., 2001; Johnson and Golbeck, 2004; Srinivasan and Golbeck, 2009). Even the tail-less phylloquinone analog menadione (vitamin K3) (Table S1) may occupy the A1 site in the correct orientation and sustain A0→FX electron transfer in vitro (Iwaki and Itoh, 1989, 1994; Kumazaki et al., 1994; reviewed by Ikegami et al., 2000; Itoh et al., 2001). Thus, the ‘head’ group of a quinone (phylloquinone in most oxygenic phototrophs; see below also) is the main factor determining its binding, coordination and redox activity in PSI. However, the ‘tail’ is assumed to improve the quinone binding affinity through hydrophobic interaction with proteins (Iwaki and Itoh, 1989, 1994; Kumazaki et al., 1994; reviewed by Ikegami et al., 2000; Itoh et al., 2001), but the chemical nature of the tail appears not to be critical, as different long-chain isoprenoids, or even non-branched alkyl substituents, may also bind efficiently to the A1 site (Biggins, 1990; Srinivasan and Golbeck, 2009). For the phylloquinone-deficient Synechocystis mutants (menA, B, D and E), it has been shown in vivo that plastoquinone–9 (Table S1) functions at the A1 site but with diminished efficiency due to the more oxidizing redox potential (Johnson et al., 2000; Semenov et al., 2000; Zybailov et al., 2000; reviewed by Johnson and Golbeck, 2004; Srinivasan and Golbeck, 2009). Additionally, incorporation of plastoquinone into PSI leads to a decrease in PSI content (Johnson et al., 2000). Together, both effects result in a reduction of the whole ETC capacity by approximately 40% (Johnson et al., 2000; own observations). However, the capability for photoautotrophic growth is not abolished in these mutants (Johnson and Golbeck, 2004; Srinivasan and Golbeck, 2009; own observations). Altogether, the available data demonstrate that PSI possesses a considerable capacity to accommodate quinones of various structure and size at the A1 site (Table S1).

The incorporation of phylloquinone into PSI requires neither enzymatic activity nor de novo protein synthesis, and is not accompanied by disassembly/re-assembly of PSI in vitro (reviewed by Ikegami et al., 2000; Itoh et al., 2001) and in vivo (Johnson et al., 2001). Moreover, PSI containing a variant quinone, which may be considerably different in size and/or binding affinity, or even with an empty quinone-binding pocket, remains relatively stable in vitro (reviewed by Ikegami et al., 2000; Itoh et al., 2001) as well as in vivo (Johnson et al., 2000, 2001). The available data indicate that – in contrast to the role of Chl a in biogenesis of PSI (and PSII) (Eichacker et al., 1992, 1996; Adamska et al., 2001) – phylloquinone does not determine assembly and stabilization of PSI, at least not in cyanobacteria such as Synechocystis (Johnson and Golbeck, 2004; Srinivasan and Golbeck, 2009).

So far, the nature of the quinone in PSI of the ΔchlP mutant remains to be established. However, several lines of evidence suggest that this may be the geranylgeranylated phylloquinone analog menaquinone–4 (vitamin K2(20)) (Figure S1 and Table S1). First, there is no GGR activity in the mutant (Shpilyov et al., 2005). Hence, formation of (phytylated) phylloquinone is impossible (Figure S1). Second, if synthesis of the tailed vitamin K is interrupted at the MenA stage (Figure S1) or earlier, only plastoquinone–9 occupies the A1 site in the respective Synechocystis mutants (Johnson and Golbeck, 2004; Srinivasan and Golbeck, 2009). Third, the ΔchlP mutant possesses a fully functional ETC (Shpilyov et al., 2005) indicative of a highly active PSI, which appears to exclude the presence of plastoquinone–9 at the A1 site. Therefore, PSI in ΔchlP appears to contain a quinone with a redox capacity comparable to that of phylloquinone, but not phylloquinone itself.

According to the phylloquinone biosynthetic pathway, menaquinone–4 is indeed such a candidate (Figure S1). It possesses the same head group as phylloquinone and differs from phylloquinone only by three additional double bonds in the tail moiety, which appear not to be critical (see above). Indeed, in vitro reconstitution studies established menaquinone–4 to be a fully functional analog of phylloquinone, i.e. displaying comparable binding and electron-transfer properties at the A1 site (Iwaki and Itoh, 1989; Biggins, 1990; Iwaki and Itoh, 1991, 1994; Kumazaki et al., 1994; reviewed by Ikegami et al., 2000; Itoh et al., 2001). Furthermore, menaquinone–4 synthesis and efficient operation in PSI have been demonstrated using isolated spinach chloroplasts (Kaiping et al., 1984) and GGR-deficient rice mutants (Shibata et al., 2004b). Finally, menaquinone–4 was found to function at the A1 site in several oxygenic phototrophs, which, however, use phytylated Chls (Table S1).

Irrespective of the quinone species in PSI in the ΔchlP mutant, this cannot be the reason for PSI instability, as the quinone is not a factor influencing PSI assembly and stabilization, as mentioned above. 77 K fluorescence emission spectra of an analogous menA mutant indicated only depletion in PSI content, with no sign of structural alterations (Johnson et al., 2000). Moreover, the instability of PSII in ΔchlP is certainly not related to the quinone present in PSI, as the biogenesis and function of PSII are independent of those of PSI in cyanobacteria, as deduced from data obtained using phylloquinone-deficient (Johnson et al., 2000) and PSI-lacking Synechocystis mutants (Shen et al., 1993; Wu and Vermaas, 1995).

Hence, as in the case of tocopherol, impaired phylloquinone synthesis also cannot be the origin of the instability and high vulnerability of PSI and especially PSII to photodegradation in the ΔchlP mutant.

Requirement for chlorophyll phytylation

Active ETCs in GGR-deficient mutants of oxygenic phototrophs (Shibata et al., 2004a,b; Shpilyov et al., 2005) indicate that geranylgeranylated electron transfer co-factors, i.e. Chl aGG, pheophytin aGG and menaquinone–4 (see above), are functional. However, when assembled with Chl aGG, PSI and PSII complexes become unstable and tend to degrade spontaneously (i.e. in complete darkness; Shpilyov et al., 2005). The geranylgeranyl residue is more rigid than a phytyl residue due to three additional double bonds. This increased rigidity probably perturbs the association of ChlsGG with apoproteins, possibly also disturbing the interaction of protein subunits with each other. Reaction centre preparations from a GGR-deficient mutant of the purple bacterium Rhodobacter capsulatus synthesizing bacteriochlorophyll (BChl) aGG were also found to be much less stable than similar preparations from WT cells (Bollivar et al., 1994). Thus, phytylation of (B)Chls appears to be generally important for the stability of pigment–protein complexes among chlorophototrophs, with the exception of one purple bacterium, Rhodospirillum rubrum (see below), and heliobacteria. The latter accumulate BChl g with a fully unsaturated C15 isoprenoid (farnesyl) tail (Madigan, 2006). It should be noted that heliobacteria only grow photoheterotrophically or even heterotrophically (Madigan, 2006).

From an evolutionary perspective, (B)Chl tail saturation has apparently become more critical for oxygenic photosynthesis. For example, GGR-deficient mutants of anoxygenic photosynthetic bacteria are still capable of photoautotrophic growth, although less efficiently (Bollivar et al., 1994; Addlesee and Hunter, 1999; Harada et al., 2008a). Moreover, depending on growth stage and conditions, these organisms naturally accumulate various amounts (6–30%) of BChls with unsaturated tail moieties (Bollivar et al., 1994; Addlesee and Hunter, 1999; Mizoguchi et al., 2006; Harada et al., 2008ab). The purple bacterium Rhodospirillum rubrum even accumulates only BChl aGG – although bacteriopheophytin a in its reaction centre is phytylated due to the activity of a special bacteriopheophytin aGG reductase (Addlesee and Hunter, 2002). In striking contrast, Chls with only partially saturated tails never occur in the mature photosynthetic apparatus of oxygenic phototrophs (Tamiaki et al., 2007), and GGR deficiency in these organisms is lethal (Henry et al., 1986; Shibata et al., 2004a; Shpilyov et al., 2005).

Remarkably, similar light sensitivity as in the cyanobacterial ΔchlP mutant was observed in GGR-deficient plants, e.g. tobacco (Tanaka et al., 1999) and rice (Shibata et al., 2004a,b). The likely explanation is that perturbed binding and spatial orientation of Chl aGG in the pigment–protein complexes impedes interactions and thus efficient excitation energy transfer among Chls and possibly also to carotenoids. In turn, this may lead to the build-up of triplet excited states of Chls (3Chl*), leading to the generation of destructive reactive oxygen species (particularly 1O2) upon illumination. This is apparently the reason for aggravated photo-oxidative stress and the high vulnerability of pigment–protein complexes to light-induced degradation in GGR-deficient mutants of oxygenic phototrophic organisms. Thus, beyond stabilizing photosynthetic pigment–protein complexes by Chl phytylation, the second important role of the GGR enzyme in oxygenic phototrophs appears to be to ensure the optimal coordination of Chl molecules in these complexes that is necessary for efficient utilization of light energy and thus protection from photo-oxidative stress in an oxygenic environment.

Experimental procedures

Strains and growth conditions

Mutants were derived from the same glucose-tolerant non-motile Synechocystis sp. PCC 6803 strain (collection of the Genetics Department, Lomonosov Moscow State University, Russia) used as the wild-type (WT). The ΔchlP mutant was generated by disruption of the ggr (chlP) gene encoding the GGR (ChlP) enzyme using a kanamycin resistance (Kmr) cassette as described previously (Shpilyov et al., 2005). For construction of the Δhpt and ΔmenA mutants, the Kmr cassette was inserted inside the hpt and menA genes encoding the HPT and MenA enzymes, respectively. The biosynthetic pathways, Δhpt and ΔmenA genetic maps and mutant construction protocols are shown in Figures S1 and S2.

Cyanobacteria were cultivated in liquid BG–11 medium (Rippka et al., 1979) or solidified BG-11 containing 1% agar (Difco, www.bd.com) at 34°C under continuous illumination provided with white fluorescent lamps and aeration with ambient air by magnetic stirring (for liquid cultures). The WT and Δhpt strains were propagated under a light intensity of 40 μmol photons m−2 sec−1 (‘standard’ light) with no glucose supply (photoautotrophic conditions). Dim light of approximately 2–4 μmol photons m−2 sec−1 was used together with 10 mm glucose (photomixotrophic conditions) for propagation of the ΔchlP and ΔmenA mutants. Cyanobacteria were maintained on agar plates at room temperature in dim light in the presence of glucose in the case of ΔchlP and ΔmenA, and without glucose in case of WT and Δhpt. Kanamycin (40 μg ml−1) was supplied to propagate and maintain the mutants.

In the comparative studies, the strains were preliminary adapted to photomixotrophic conditions (unless indicated otherwise) through two re-inoculations. Cells were taken from plates used to maintain the strains and grown in liquid medium to mid-log phase in the presence of 10 mm glucose under aeration with ambient air and illumination of 40 μmol photons m−2 sec−1 for WT, Δhpt and ΔchlP and 20 μmol photons m−2 sec−1 for ΔmenA (and WT for comparison with ΔmenA). Thereafter, cells were re-inoculated into fresh BG–11 medium under the same conditions, and, when they had reached mid-log phase, were subjected to the experiments/measurements. In the high-light experiments, cells were diluted with fresh 10 mm glucose-containing BG–11 medium to an absorbance at 750 nm (OD750) of approximately 0.25, and exposed to illumination of approximately 500 μmol photons m−2 sec−1 under the same aeration and temperature conditions. Cell growth was monitored by measurements of OD750.

Biochemical and biophysical analyses

Contents of Chls, total carotenoids and PBS were estimated from the whole-cell absorption spectra recorded at room temperature. Furthermore, Chls and total carotenoids were quantified spectrophotometrically in cell extracts obtained using 90% v/v methanol/water. Individual pigments and tocochromanols were assayed by HPLC as described previously (Shpilyov et al., 2005). PSI/II levels were assessed by 77 K fluorescence emission spectroscopy. The spectra were recorded with 435 nm excitation for cell samples of OD750 = 1, i.e. Chl concentrations of 2.59, 1.98 and 2.31 μg ml−1, respectively, in the WT, ΔchlP and Δhpt liquid cultures growing exponentially under photomixotrophic standard light conditions. Cell samples of equal volumes and densities were analyzed during the high-light experiment. The measurement/estimation procedures and equipment used in the present work are the same as described previously (Shpilyov et al., 2005).

Acknowledgements

The authors thank Barbara Hickel for assistance with the HPLC measurements, and Annegret Wilde and Vadim Glazer for help with cultivation of the cyanobacteria. The work was supported by the Deutsche Forschungsgemeinschaft (SFB 429, TPs A2 and B9), a grant from the Russian Foundation of Basic Research (N 10-04-00840-a) and a grant from the European Science Foundation EuroCore (BB/J00823011).

Ancillary