Although light is the source of energy for photosynthetic organisms, it causes oxidative stress when in excess. Plants and algae prevent reactive oxygen species (ROS) formation by activation of nonphotochemical quenching (NPQ), which dissipates excess excitation energy as heat. Although NPQ is found in both algae and plants, these organisms rely on two different proteins for its activation, Light harvesting complex stress-related (LHCSR) and Photosystem II subunit S (PSBS). In the moss Physcomitrella patens, both proteins are present and active.
Several P. patens lines depleted in or over-expressing PSBS and/or LHCSR at various levels were generated by exploiting the ability of Physcomitrella to undergo homologous recombination.
The analysis of the transgenic lines showed that either protein is sufficient, alone, for NPQ activation independently of the other, supporting the idea that they rely on different activation mechanisms. Modulation of PSBS and/or LHCSR contents was found to be correlated with NPQ amplitude, indicating that plants and algae can directly modulate their ability to dissipate energy simply by altering the accumulation level of one or both of these proteins.
The availability of a large range of P. patens genotypes differing in PSBS and LHCSR content allowed comparison of their activation mechanisms and discussion of implications for the evolution of photoprotection during land colonization.
Photosynthetic organisms convert sunlight into chemical energy, generating molecular oxygen as a secondary product. When the first plants started colonizing land, c. 450 million yr ago, they evolved new morphophysiological traits to adapt to the conditions of the terrestrial environment, which included flooding/desiccation cycles, low and high temperatures, increased exposure to UV radiation and excess visible light, compared with water ecosystems (Waters, 2003; Becker & Marin, 2009). Oxygen reduces photosynthetic activity as a consequence of the oxygenase activity of ribulose-1,5-bisphosphate carboxylase/oxygenase (Maurino & Peterhansel, 2010) and it also leads to the generation of reactive species (Niyogi, 1999). It therefore acts as an inhibitor of photosynthesis, and it is more abundant and diffuses faster in the atmosphere than in water (Scott & Glasspool, 2006). In contrast, the availability of carbon dioxide, the final acceptor of electrons extracted from water by photosystems, is lower for land plants (Waters, 2003). These conditions increase the probability of harvesting light in excess with respect to the maximal rate of photochemical reactions, requiring optimization of photoprotection. Nonphotochemical quenching (NPQ) consists of the thermal dissipation of excited chlorophyll singlet states (1Chl*) and it is the fastest mechanism of protection, promptly activated upon exposure to excess light (Genty et al., 1989; Niyogi, 2000; Szabo et al., 2005). Two major components are commonly identified in NPQ, energy quenching (qE) and inhibitory quenching (qI). The former depends on lumen acidification and develops within seconds upon an increase in light intensity. qE also relaxes on a 1–2-min time-scale in the dark (Demmig-Adams et al., 1996; Horton et al., 1996). The slower component qI is dependent on more than one mechanism. It is in part due to photosynstem II (PSII) photoinhibition but also the accumulation of zeaxanthin, which requires c. 1h to be completely reconverted into violaxanthin, contributes to the slower NPQ components (Dall'Osto et al., 2005; Reinhold et al., 2008). The contribution attributed to zeaxanthin has recently been termed qZ, or NPQ AZ, to distinguish it from the photoinhibitory component, although it should be mentioned that the presence of zeaxanthin also enhances the qE component (Horton et al., 1996; Nilkens et al., 2010; Forster et al., 2011; Jahns & Holzwarth, 2012; Nichol et al., 2012).
In vascular plants, qE induction relies on the presence of PSBS (Photosystem II subunit S), a light harvesting complex (lhc)-like protein, whose activation is triggered by lumen acidification through the protonation of two glutamate residues (Li et al., 2000, 2004). Once PSBS has been activated, a decrease in the lifetime of Chl excited states is induced in the pigment-binding subunits of the antenna system, after segregation of PSII subunits to form distinct domains in the grana partitions of thylakoid membranes (Betterle et al., 2009; Johnson et al., 2011). The rapidly reversible quenching which occurs during this process is particularly important in highly variable environmental conditions, where plants lacking PSBS are quickly counter-selected (Kulheim et al., 2002).
Although genes encoding PSBS have been identified in the genomes of some green algae, the corresponding polypeptides have never been detected (Koziol et al., 2007; Bonente et al., 2008; Engelken et al., 2010), suggesting that algae rely on different proteins for NPQ activation. In the model Chlorophyta Chlamydomonas reinhardtii, NPQ activation has consistently been shown to depend on a distinct protein called LHCSR (Light harvesting complex stress-related, or Li818) (Peers et al., 2009), which is also an Lhc-like protein. As discussed in more detail in the last section, LHCSR isoforms have been found in many taxa such as brown and green algae, in which they are also involved in NPQ activation (Koziol et al., 2007; Bailleul et al., 2010; Engelken et al., 2010; Zhu & Green, 2010), but not in vascular plants. Although algae constitute a heterogeneous taxon and caution must be employed in making generalizations, all data currently available point to important differences in the mechanism for activation of feedback energy dissipation between algae and plants. In this context, the moss Physcomitrella patens, which occupies an evolutionarily intermediate position between plants and algae, is of particular interest, as both PSBS and LHCSR are present and catalyze the strong NPQ activity reported in this organism (Alboresi et al., 2010). In this species, the NPQ response is crucial in avoiding damage to the photosynthetic apparatus during excessive exposure to light (Gerotto et al., 2011).
The present work reports on the NPQ behavior of a large set of P. patens plants, either depleted in or over-accumulating LHCSR and/or PSBS to different levels. Their characterization showed that LHCSR and PSBS are active simultaneously and independently in P. patens and that they do not interact significantly. Results also indicated that the accumulation of PSBS and LHCSR controls the amplitude of NPQ, thus providing a powerful way of modulating its intensity to suit various environmental conditions.
Materials and Methods
Protonemal tissue of Physcomitrella patens Gransden wild-type (WT) strain (Alboresi et al., 2008), psbs and/or lhcsr knock-out (KO) lines (Alboresi et al., 2010) and PSBS or LHCSR over-expresser (OE) lines was grown in minimum PpNO3 medium (Asthon et al., 1979) solidified with 0.8% Plant Agar (Duchefa Biochemie, Noord-Holland, Netherlands). Plants were propagated in sterile conditions on 9-cm Petri dishes: protonema from 6- to 7-d-old plates was collected and tissue blended in water by homogenizing moss material with Polytron (IKA T25 Digital Ultra Turrax, IKA-Werke, Staufen, Germany). The suspension was then spread in new agar plates with medium overlaid with cellophane disks (A. A. Packaging Limited, Preston, UK) (Alboresi et al., 2008). Plates were placed in a growth chamber in controlled conditions: 24°C, with a 16-h light : 8-h dark photoperiod and a light intensity of 40 μmol m−2 s−1.
Protoplast transformation and KO and OE generation
Plasmids for targeted KO generation for PSBS or LHCSR1/2 polypeptides were available (Alboresi et al., 2010). In the present work, the same constructs and single KO lines previously generated were used to obtain double KO mutants. In particular, the lhcsr1 KO line was used as a background to obtain lhcsr1 lhcsr2 KO, while lhcsr2 KO was employed to generate psbs lhcsr2 double KO. For generation of OE mosses, cDNA obtained from P. patens protonema grown in control conditions was used as starting material. Coding sequences for PSBS (XM_001778511.1) and LHCSR1 (XM_001776900.1) were amplified by PCR (Supporting Information Table S1) and then cloned into pMAK1 vectors (kindly provided by Prof. Mitsuyasu Hasebe, National Institute for Basic Biology, Okazaki, Japan) with XhoI/HpaI and XhoI/ApaI restriction enzymes, respectively. As LHCSR1 and LHCSR2 isoforms have 91% sequence identity, we chose to over-express LHCSR1, as this isoform is strongly accumulated in Physcomitrella, according to our previous results (Alboresi et al., 2010). To obtain all the mutant lines presented in this work, P. patens transformation was performed as in Schaefer & Zryd (1997) with minor modifications. Five- to 6-d-old protonemal tissue was collected for protoplast generation and PEG-mediated transformation. Resistant colony selection was started 6 d after transformation by transferring regenerated plants to culture media supplemented with the antibiotic G418 (50 μg ml−1; Sigma-Aldrich) or zeocin (50 μg ml−1; Duchefa Biochemie). After 7–10 d of growth in selective media, resistant colonies were transferred to nonselective plates, on which they were maintained for a further 10 d. Stable mutant lines were then obtained with a further round of selection in culture media supplemented with the appropriate antibiotic. Genomic DNA from controls and mutant lines was obtained with EuroGOLD Plant DNA mini kits (EuroClone Pero (Mi), Italy) according to the manufacturer's instructions and used as templates to confirm DNA insertion by PCR (Table S1).
NPQ measurements and zeaxanthin production analysis
In vivo chlorophyll fluorescence of P. patens WT and mutant lines grown for 10 d in minimum medium was measured at room temperature on a Dual PAM-100 fluorometer (Walz, Effeltrich, Germany), with saturating light at 6000 μmol m−2 s−1 and actinic light at 830 μmol m−2 s−1. Before measurements, plates were dark-adapted for 40 min at room temperature. Parameters Fv/Fm and NPQ were calculated as (Fm – Fo)/Fm and (Fm – Fm′)/Fm′ (Demmig-Adams et al., 1996). Data are presented as means ± SD of at least three independent experiments. For zeaxanthin monitoring during NPQ, a small fragment of moss tissue was exposed to PAM actinic light for 8 min, as in NPQ induction curves. Immediately after the actinic light was switched off the pigments were extracted with 90% acetone and analyzed by HPLC (Gilmore & Yamamoto, 1991). The same analysis was also performed on samples dark-adapted for 40 min or illuminated in the presence of 5 mM dithiotreitol (DTT), a known inhibitor of violaxanthin de-epoxidase (VDE; Yamamoto & Kamite, 1972).
High light treatment
Ten-day-old plates of WT and mutant genotypes grown in control conditions (40 μmol m−2 s−1) were transferred to 800 μmol m−2 s−1 for 24 h. Photoinhibition was monitored by sampling a fragment of moss tissue from the plates and measuring the Fv/Fm parameter after 20 min of dark adaptation.
Thylakoid extraction, pigment content, SDS-PAGE and western blotting analyses
Thylakoids from protonemal tissue (10-d-old plants grown in minimum PpNO3) were prepared by means of an Arabidopsis protocol with minor modifications (Alboresi et al., 2008). Ten-day-old protonemal tissue was collected and homogenized with a potter homogenizer in a few milliliters of buffer 1 (0.4 M sorbitol, 0.1 M tricine, pH 7.8 (KOH), 10 mM NaCl, 5 mM MgCl2, 0.5% w/v milk powder, 0.2 mM phenylmethylsulphonyl fluoride, 0.2 mM benzamidine and 1 mM ε-aminocaproic acid). The suspension was then filtered through a cloth with a pore diameter of 20 μm to remove cell debris. The filtrate was centrifuged (1500 g for 10 min at 4°C) and the pellet containing the chloroplasts was re-suspended in buffer 2 (50 mM sorbitol, 5 mM tricine, pH 7.8 (KOH), and 10 mM EDTA, pH 8.0). The next centrifugation (10 000 g for 10 min at 4°C) allowed the thylakoids to be pelleted. The supernatant was withdrawn and the thylakoids were again washed in buffer 2, centrifuged at 10 000 g for 10 min at 4°C, and finally re-suspended in buffer 3 (50 mM HEPES KOH, pH 7.5, 50% glycerol, and 5 mM MgCl2), frozen in liquid nitrogen and stored at −80°C until used. Chl a : b and Chl : Carotenoids (Car) ratios were obtained by fitting the spectrum of 80% acetone pigment extracts with spectra of the individual purified pigments, as in Croce et al. (2002). For immunoblotting analysis, following SDS-PAGE, proteins were transferred to nitrocellulose membranes (Pall Corporation, Port Washington, NY, USA) and detected with specific home-made polyclonal antibodies.
The immunodetected bands for PSBS or LHCSR were then quantified by densitometric analysis with GelPro Analyzer software (Media Cybernetics, Rockville, MD, USA) software and normalized for protein content.
Modulation of PSBS and LHCSR contents in Physcomitrella patens by generation of KO and OE mosses
Reverse genetic studies of PSBS, LHCSR1 and LHCSR2 in Physcomitrella patens had previously shown that these gene products are all active in NPQ (Alboresi et al., 2010). Here, different KO mutations were combined to generate plants retaining only one of these subunits. As the triple psbs lhcsr1 lhcsr2 KO mutant lacks NPQ, the double KO mutants psbs lhcsr1 KO, psbs lhcsr2 KO and lhcsr1 lhcsr2 KO rely for their NPQ exclusively on LHCSR2, LHCSR1 and PSBS, respectively. These double mutants were obtained from single KO, as reported in Alboresi et al. (2010), exploiting the capacity of P. patens for homologous recombination. Transformations led to the isolation of several independent stably resistant colonies which, upon PCR analysis, were confirmed to carry an insertion at the expected target site. Individuals with single copy insertions were retained as described in Alboresi et al. (2010) and in the Supporting Information (Fig. S1). All PCR fragments were sequenced as further controls and at least two independent single insertion lines for each transformation were analyzed in detail for their phenotype. However, all lines showed indistinguishable phenotypes, so only results for one line for each mutation are reported here. In the selected lines, the accumulation of PSBS/LHCSR proteins was first determined by western blotting (Fig. 1a). As expected, only PSBS was detectable in lhcsr1 lhcsr2 KO plants, and only LHCSR was present in psbs lhcsr2 KO. In the case of psbs lhcsr1 KO, LHCSR2 accumulated at low levels, matching previous results (Alboresi et al., 2010).
In order to assess the dependence of NPQ amplitude on the relative amounts of PSBS and LHCSR, these KO plants were used as starting material for the generation of over-expressers for PSBS or LHCSR1. PSBS was over-expressed in lhcsr1 lhcsr2 KO and psbs KO backgrounds, to obtain lines with different PSBS contents in the absence or presence of LHCSR. LHCSR1 was also over-expressed in psbs KO to produce mosses over-accumulating LHCSR. All over-expression lines were generated by transformation with a construct driving the insertion of DNA in a region of the P. patens genome (called BS213) not encoding proteins, thus avoiding phenotypes caused by inactivation of endogenous genes (Schaefer & Zryd, 1997). In the plasmid employed (pMAK1), heterologous gene expression is under the control of a constitutive promoter derived from 35S (called 7113; Mitsuhara et al., 1996). At least five independent stable resistant lines were isolated for each transformation and retained for detailed analysis. The insertion of exogenous DNA was first verified by PCR (Fig. S2) and, in all lines showing positive PCR results, PSBS and LHCSR accumulation was evaluated with specific antibodies (Fig. 1b–d). Fig. 1(b) shows data from two independent lines over-expressing PSBS in the lhcsr1 lhcsr2 KO background (lhcsr1 lhcsr2 KO-PSBS OE), demonstrating that both these clones had increased PSBS accumulation with respect to WT and to the lhcsr1 lhcsr2 KO background which, instead, had a similar PSBS level. Over-accumulation is particularly strong in the case of line 2, also as a result of the lower amount of sample loaded on the gel. As transformation occurs by homologous recombination (Fig. S2), this strong over-expression is attributable to multiple insertions of the transgene during recombination, which are known to occur often in Physcomitrella patens (Kamisugi et al., 2006).
In order to assess possible differences in PSBS over-expression in the presence of LHCSR, psbs KO plants were also transformed with the same construct (psbs KO-PSBS OE). Fig. 1(c) shows two lines with different levels of PSBS over-expression. Despite their large differences in PSBS accumulation, the WT, KO and PSBS over-expressing lines showed no significant alterations in LHCSR accumulation, indicating that the latter is not affected by modulation of PSBS expression.
LHCSR1 was also over-expressed in a psbs KO line (psbs KO-LHCSR1 OE), generating lines where NPQ was totally dependent on LHCSR. Analysis of protein accumulation by western blotting showed increased LHCSR content with respect to the psbs KO background (Fig. 1d). In this case, however, no line with strong protein over-expression was identified, despite the analysis of a similar number of resistant lines. In P. patens, multiple insertions during homologous recombination are estimated to occur in c. 50% of cases (Kamisugi et al., 2005, 2006) and this number is consistent with our observations of PSBS over-expressing lines. It is thus likely that, out of the six independent lines over-expressing LHCSR, at least one carries a multiple insertion, as also indicated by PCR analysis (Fig. S2). Thus, failure to find lines with strong LHCSR accumulation is probably the result of an unidentified mechanism of post-translational regulation active for this protein.
Nonphotochemical quenching in mutants with altered PSBS/LHCSR contents
No mosses having variable accumulations of PSBS and/or LHCSR generated as described above, when grown in control conditions, showed a visible phenotype. Photochemical efficiency, estimated from Fv/Fm values, was indistinguishable from that of WT, as also observed previously for the triple mutant psbs lhcsr1 lhcsr2 KO (Table 1; Alboresi et al., 2010). No major differences were detectable in Chl a : b and Chl : Car ratios either (Table 1), which suggests that no substantial alterations in thylakoid composition or, at least, in the antenna/reaction center ratio took place.
Table 1. Photosystem II (PSII) quantum yield and pigment content of Physcomitrella patens wild type (WT), double knock-out (KO) and over-expresser (OE) lines grown in control conditions
Chl a : b
Chl : Car
PSII quantum yield (Fv/Fm) was evaluated in different genotypes grown in control light conditions (40 μmol m−2 s−1). Chl a : b and Chl : Car ratios were also evaluated in the same plants. Values reported are means of three to five independent experiments, each including multiple individual plants. SD was below 0.03 in Fv/Fm, below 0.1 for Chl a : b and 0.3 for Chl : Car.
psbs lhcsr1 KO
psbs lhcsr2 KO
lhcsr1 lhcsr2 KO
lhcsr1 lhcsr2 KO-PSBS OE line 1
lhcsr1 lhcsr2 KO-PSBS OE line 2
psbs KO-LHCSR1 OE line 1
psbs KO-LHCSR1 OE line 2
psbs KO-PSBS OE line 1
psbs KO-PSBS OE line 2
Instead, major alterations were observed in NPQ kinetics. All double KO mutants, which express only PSBS, LHCSR1 or LHCSR2, showed reduced NPQ with respect to WT and to the single KO mutant used as background, psbs lhcsr2 KO showing the most intense residual NPQ (Figs 2, S3). In all cases, residual NPQ rapidly relaxed in the dark and was thus identifiable as the qE component, indicating that PSBS and both LHCSR isoforms were all active, although to different extents, independently of the presence of the other proteins (Fig. 2).
It was observed that lhcsr1 lhcsr2 KO showed very low NPQ activity with respect to the other lines, raising the question as to whether P. patens PSBS protein is less active in triggering NPQ than vascular plant isoforms and/or whether the photosynthetic apparatus of P. patens is not able to fully undergo the structural rearrangements required for PSBS-dependent NPQ activation (Betterle et al., 2009; Johnson et al., 2011). Analysis of the lines over-expressing PSBS in the lhcsr1 lhcsr2 KO background (lhcsr1 lhcsr2 KO-PSBS OE) clarified this point. In both lhcsr1 lhcsr2 KO-PSBS OE lines a strong increase in NPQ capacity was observed with respect to the KO background of the transformation (Fig. 3a). Interestingly, NPQ amplitude depended on the level of PSBS accumulation and was stronger in over-expressing line 2, where PSBS was more abundant. In all over-expressing lines, most NPQ was rapidly relaxed in the dark and may thus qualify as the fast qE component, similar to the case of WT plants. These results clearly suggest that the low PSBS-dependent NPQ activity of lhcsr1 lhcsr2 KO is attributable to the low PSBS accumulation in Physcomitrella patens, while the protein is fully capable of activating strong NPQ even in the absence of LHCSR. It is also worth noting that this PSBS-only-dependent NPQ has a time-course very similar to that of Arabidopsis thaliana plants (Fig. S4), confirming that, once enough PSBS is present, quenching in mosses occurs as well as in vascular plants.
PSBS over-expression in psbs KO plants, which retain WT levels of LHCSR, yielded similar results. psbs KO-PSBS OE line 1, with a lower PSBS accumulation, also showed the smallest increase in NPQ induction (Fig. 3b). Instead, line 2, in which western blotting showed higher PSBS accumulation, was able to induce an NPQ response stronger than that of WT (Fig. 3b). As LHCSR protein content showed no significant differences in any of the lines tested (Fig. 1c), these results – the increased NPQ capacity of these clones – are consistent with the activity being attributable only to the larger accumulation of PSBS.
Both LHCSR isoforms alone were fully capable of inducing NPQ, as shown by the KO mutants' phenotype. The over-expression of LHCSR1 in psbs KO, where NPQ is only LHCSR-dependent, led to lines with increased protein accumulation which also showed increased NPQ capacity with respect to the background (Fig. 3c). The NPQ induced in the LHCSR over-expression lines could be identified as qE, owing to its fast relaxation in the dark. As observed previously, NPQ values were higher in genotypes with the highest LHCSR accumulation, indicating that the level of LHCSR influences NPQ in P. patens, as observed above in PSBS over-expression. In this case, however, the increase in protein accumulation observed with western blotting was lower, and also the increase in NPQ was not as strong as in mosses over-expressing PSBS.
Zeaxanthin accumulation is well known to enhance NPQ intensity (Niyogi et al., 1997, 1998). In order to verify if the differences observed in mutants above were at least partially attributable to alterations in zeaxanthin synthesis, carotenoid composition was monitored in mutant and OE lines during NPQ activation. As shown in Table 2 and Fig. S5, all lines, independently of their PSBS/LHCSR accumulation levels, showed a similar accumulation of zeaxanthin under illumination, showing that differences in NPQ levels are not attributable to alteration in xanthophyll cycle activation.
Table 2. Light-induced zeaxanthin formation in Physcomitrella patens
Light-treated + DTT
Zeaxanthin formation during nonphotochemical quenching (NPQ) determinations was monitored in wild type (WT) and selected mutant/over-expresser (OE) lines, chosen to represent the largest interval of NPQ values. De-epoxidation state was calculated as (Zea + 0.5 Ant)/(Zea + Ant + Vio) (Ruban et al., 1999), in dark-adapted and light-treated samples (with 800 μmol m−2 s−1) in the presence/absence of DTT.
0.15 ± 0.04
0.41 ± 0.05
0.13 ± 0.02
psbs KO-PSBS OE line2
0.11 ± 0.02
0.52 ± 0.06
0.16 ± 0.05
lhcsr1 lhcsr2 KO-PSBS OE line2
0.11 ± 0.02
0.47 ± 0.09
0.12 ± 0.03
psbs KO-LHCSR1 OE line1
0.13 ± 0.04
0.48 ± 0.15
0.11 ± 0.02
psbs lhcsr1 lhcsr2 KO
0.14 ± 0.02
0.49 ± 0.07
0.19 ± 0.05
In order to verify if this increased NPQ was effective in protecting P. patens from excess illumination, KO and OE lines were exposed to strong illumination, and PSII radiation damage was monitored using the Fv/Fm parameter, reductions in which are indicative of light-induced damage. psbs KO mutants showed no significant differences from the WT (Fig. 4a), suggesting that residual NPQ was sufficient to ensure resistance in the experimental conditions tested, as also observed in Alboresi et al. (2010). In contrast, lhcsr1 lhcsr2 KO, which showed a stronger reduction in NPQ, also showed increased light sensitivity (Fig. 4a). When PSBS was over-expressed in these lines (lhcsr1 lhcsr2 KO-PSBS OE), not only was NPQ intensity recovered but their capacity to withstand strong illumination treatments was also restored (Fig. 4b). Instead, psbs KO-PSBS OE and psbs KO-LHCSR1 OE showed a response to the light treatment similar to that of WT mosses and their backgrounds, as expected (Fig. S6).
Correlation of LHCSR and PSBS accumulation with NPQ values
Combining the various KO and over-expressing lines presented here and in Alboresi et al. (2010), a set of 20 moss lines with varying PSBS and/or LHCSR accumulation levels and altered NPQ values is now available. For a more general picture of the dependence of NPQ on accumulation of these proteins, the latter was more precisely quantified, by using specific antibodies (Fig. 5a). The same procedure was also performed for mutants acclimated to varying light conditions, in which PSBS and LHCSR were differentially accumulated (Gerotto et al., 2011). Only low-light-acclimated samples (grown at 10 μmol m−2 s−1) were considered in order to avoid the NPQ-enhancing effect of zeaxanthin, which is also accumulated in high light conditions (Niyogi et al., 1998).
Altogether, a total of 27 different samples with variable PSBS/LHCSR contents were analyzed, allowing us to explore the quantitative relationship between the accumulation of PSBS and LHCSR vs NPQ intensity. All data are reported in Fig. S7, whereas Fig. 5(b,c) shows details of the dependence of NPQ values on PSBS and LHCSR contents, respectively. In the case of PSBS (Fig. 5b), many samples with large differences in PSBS expression levels were available, ranging from 0 to c. 27-fold the WT level. The data distribution identifies three groups, differing in LHCSR accumulation: one group includes all lines without LHCSR, another those with WT LHCSR levels (1 ± 0.1), and a third, smaller group includes lhcsr1 KO lines, which have only residual LHCSR2 expression (0.08 ± 0.05 of WT level).
In all cases, the NPQ values showed a positive correlation with the amount of PSBS. In the range from 0 up to five times WT content, data could be fitted linearly, and the slope of the curve was clearly higher when LHCSR was present (Fig. 5b). In cases with the strongest PSBS accumulation, NPQ deviated from linearity, reaching saturation.
Similar results were obtained in the case of NPQ dependence on LHCSR content (Fig. 5c): two distinct groups of samples, depleted in or having WT levels of PSBS, were analyzed. As we found above for PSBS, LHCSR accumulation level also positively correlated with the NPQ capacity of the samples, both in genotypes where PSBS is absent and in those exhibiting a WT level of PSBS accumulation. As in the previous case, the slope of the linear fit was higher in samples accumulating both PSBS and LHCSR.
In view of the complex nature of NPQ, some caution should be employed to avoid over-interpreting this correlation between PSBS/LHCSR and NPQ. The data presented here, however, clearly indicate that PSBS and LHCSR accumulation levels are limiting for NPQ, also suggesting that a similar increase in PSBS/LHCSR is more effective when both subunits are present.
PSBS and LHCSR are active simultaneously and independently in Physcomitrella patens
Analysis of various KO and OE lines clearly demonstrates that PSBS and both LHCSR isoforms are fully able to induce NPQ in P. patens, independently of the presence of the other. Although LHCSR1 had a larger effect on NPQ in control growth conditions than the other two proteins, this could be attributed to a higher accumulation level, rather than to a higher specific activity. Consistent with this idea, several lines over-expressing PSBS showed NPQ as strong as that of WT plants, even in the complete absence of LHCSR.
Although either LHCSR or PSBS can activate NPQ independently of the other subunit, it is interesting to assess whether they influence each other. The first observation is that alteration of PSBS/LHCSR expression, ranging from 0 to 27 and 2.5 times the WT levels, respectively, did not significantly affect the content of the other protein. This suggests that there is no physical interaction between the two proteins or competition for a membrane compartment or binding site, as in this case some reciprocal influence of protein levels would be expected. This was particularly evident in PSBS-complemented mosses (psbs KO-PSBS OE), in which the strong increase in PSBS accumulation did not influence LHCSR content (Fig. 1c).
Another way of testing whether LHCSR and PSBS activities show some kind of interaction is to determine whether PSBS- and LHCSR-dependent NPQ can alone account for the total NPQ found in WT. NPQ values quantify a difference between dark and light states, and thus values from different plants (which may have different dark-adapted states) cannot in principle be summed. However, as Fig. 5(b,c) shows, there is at least one region where NPQ values correlate linearly with protein accumulation and the sum may be an acceptable approximation within this interval. After 1 min of illumination, WT NPQ was 3.2 ± 0.5, whereas the sum of lhcsr1 lhcsr2 KO and psbs KO was 3.1 ± 0.4 (data from Fig. 3). The difference increased with illumination time, and after 8 min the WT value was 4.2 ± 0.7 and the sum of lhcsr1 lhcsr2 KO and psbs KO 2.9 ± 0.4. Although this adding procedure has limitations, it nevertheless shows that there were no large differences between WT NPQ and the individual contributions of PSBS- and LHCSR-dependent components. In the case of major interactions between them, larger distortions would have been expected. These data thus support the hypothesis that the two proteins act independently in the activation of NPQ.
This hypothesis is consistent with the available data indicating a different mechanism for PSBS and LHCSR activity. In plants, PSBS triggers quenching by inducing reorganization of protein domains within the thylakoid membrane, triggering the formation of quenching sites located in the antenna complexes (Betterle et al., 2009; Johnson et al., 2011). In the case of algae, instead, LHCSR itself is a pigment-binding protein with an intrinsic capacity for heat dissipation, which is conserved in the isolated complex (Bonente et al., 2011). The present data thus suggest that LHCSR by itself acts as a light-harvesting and/or heat dissipation complex without the need to induce any reorganization of other antenna complexes.
Although the two types of protein independently act in NPQ induction, they are probably not energetically isolated from each other and are thus active in dissipating the same ‘pool’ of excitation energy. This hypothesis is supported by the observation of different slopes in the correlation between PSBS/LHCSR and NPQ, depending on the content of the other protein, as shown in Fig. 5(b,c). This means that, if LHCSR is present, the same increase in PSBS amount is more efficient in inducing NPQ. Instead, if LHCSR is absent, a much higher content of PSBS is required to reach NPQ saturation. The opposite case can be made for LHCSR.
Modulation of PSBS and LHCSR accumulation as a tool to adjust NPQ amplitude
The large set of different P. patens genotypes described here showed the existence of a correlation between NPQ and PSBS and/or LHCSR accumulation. This strongly suggests that NPQ amplitude is determined by the accumulation of these polypeptides in P. patens and confirms previous results on acclimated plants (Gerotto et al., 2011). Regulation of LHCSR and PSBS accumulation is thus the way in which P. patens plants modulate NPQ activity according to environmental conditions. NPQ is normally limited by the amount of PSBS or LHCSR but, if required by environmental conditions, its intensity can be increased by simply modulating the accumulation of the two proteins.
This is quite important for overall photosynthetic efficiency, as NPQ is essential for protecting photosynthetic organisms from light excess (Fig. 4; Kulheim et al., 2002; Li et al., 2002b), and yet its activation may also lead to unnecessary energy dissipation when light is limiting (Dall'Osto et al., 2005). Thus, finding the best compromise between light harvesting and energy dissipation is important for optimizing light-harvesting efficiency in varying conditions (Horton & Ruban, 2005). The data presented here show that P. patens plants can find such a ‘sweet spot’ by modulating the level of one NPQ ‘facilitator’, PSBS or LHCSR.
This possibility of modulating NPQ levels is not unique to P. patens. Arabidopsis thaliana plants with increased levels of PSBS also show stronger NPQ (Li et al., 2002a; Ballottari et al., 2007; Zia et al., 2011) and, when treated with strong light, they over-express the protein, with a corresponding increase in NPQ (Ballottari et al., 2007; Zia et al., 2011). In Chlamydomonas reinhardtii, where NPQ depends on LHCSR, the effect of acclimation is particularly evident. In dim light, LHCSR content and NPQ are low, but exposure to strong illumination induces an increase in protein expression and corresponding NPQ activity (Peers et al., 2009). In other algae, LHCSR-like proteins are constitutively present (Bailleul et al., 2010) but, upon sustained high light exposure, their increased content is accompanied by enhanced NPQ (Zhu & Green, 2010). Although data at the molecular level are still limited to a few species, all available information points to the hypothesis that plants and algae, independently of the activation mechanism in place, can modulate their NPQ ability through the levels of accumulation of these two proteins.
Evolution of PSBS- and LHCSR-dependent NPQ mechanisms in photosynthetic organisms
LHCSR and PSBS are differently distributed in algae and plants (Koziol et al., 2007; Alboresi et al., 2008; Bonente et al., 2008; Peers et al., 2009; Engelken et al., 2010), as summarized in Fig. S8. In vascular plants, NPQ depends only on PSBS (Li et al., 2000). The Chlorophyte Chlamydomonas reinhardtii, although carrying PSBS genes, has never been shown to accumulate the corresponding protein, while NPQ only depends on the level of LHCSR (Bonente et al., 2008; Peers et al., 2009). This is also the case of diatoms, which do not carry PSBS in their genome and rely only on LHCSR for NPQ induction (Bailleul et al., 2010; Engelken et al., 2010; Zhu & Green, 2010).
All data available in the literature thus suggest a picture of the evolution of the NPQ mechanism in photosynthetic organisms, proceeding from an LHCSR- to a PSBS-dependent mechanism along the phylogenetic tree of the Viridiplantae. Our results also suggest that, during the transition between alternative NPQ mechanisms, there was an overlapping phase in which the two mechanisms worked both together and independently. Although multiple protection mechanisms are present in algae as well as in plants (Forster et al., 2001; Li et al., 2009), the maintenance of an efficient NPQ was necessary for survival, as demonstrated by the photosensitivity of many mutants depleted in their NPQ capacity, including algae (Peers et al., 2009), mosses (Alboresi et al., 2010; Gerotto et al., 2011) and Arabidopsis plants (Kulheim et al., 2002). Such a strong NPQ capacity is very probably still important for a group of organisms such as mosses, which are normally exposed to low intensities of irradiance but which may be exposed to sudden increases in light intensity (Gerotto et al., 2011).
Although this picture is consistent with existing data, there is no clear explanation of why vascular plants lost LHCSR and retained only the PSBS-dependent mechanism. The answer probably does not lie in the amplitude of the NPQ response, which appears to be similar irrespective of the mechanism involved. Also, available data on comparison of various algal/plant species show no evolutionary trend in NPQ amplitude (Bonente et al., 2008).
Another explanation may be proposed based on the different activation mechanisms of PSBS- and LHCSR-dependent NPQ suggested by some recent works, as related to the different features of terrestrial and water environments. As mentioned above, in strong light irradiances PSBS triggers quenching sites within the antenna complexes, which instead are efficient light harvesters in low light conditions (Betterle et al., 2009; Johnson et al., 2011). This is substantially different from the case of LHCSR, a pigment-binding protein with intrinsic heat dissipation ability (Bonente et al., 2011), which probably thus dissipates some energy also in low light conditions. Therefore, the PSBS-dependent mechanism seems to be more efficient in modulating the dissipation of excitation energy, its activation being limited to occasions when it is really necessary. This ability to finely regulate energy dissipation of PSBS may be fundamental in the land environment, where fast, strong fluctuations in light irradiances occur (Kulheim et al., 2002). Also to be considered is that plants acquired a sessile form of life which prevented escape from unfavorable conditions by swimming away, as motile algae can do, and thus needed more efficient regulatory mechanisms. A future, more detailed evaluation of the productivity and light sensitivity of different genotypes presented here under different light conditions may help to test this hypothesis. It is worth mentioning, however, that at least some observations are consistent with this idea: first of all, the observation that psbs KO Arabidopsis plants were particularly sensitive to fluctuating light conditions (Kulheim et al., 2002). Also, it has been shown that Chlamydomonas reinhardtii in low light conditions does not accumulate LHCSR, and acclimative accumulation of the protein requires a significant amount of time, during which the cell is prone to photoinhibition (Peers et al., 2009), suggesting the presence of some pressure against the constitutive accumulation of a protein.
The mosses produced in this study with variable amounts of LHCSR and PSBS can be used to further explore these questions. For example, psbs KO and lhcsr1 lhcsr2 KO-PSBS OE line 1, remarkably, have similar NPQ levels, which in the two lines depend only on LHCSR and only on PSBS, respectively (Fig. 6a), and thus represent a good experimental system for comparing the two mechanisms. As shown in Fig. 6(b), some differences in their kinetics of NPQ generation are clear from the untreated fluorescence traces: PSBS protein (in lhcsr1 lhcsr2 KO-PSBS OE line 1) appears to induce strong quenching more rapidly than LHCSR (in psbs KO) during the first few seconds of illumination. PSBS may thus respond more rapidly to the sudden changes in illumination typical of the aerial environment. We intend to perform growth experiments in different light regimes in order to determine if a differential capacity to sustain growth is provided by PSBS- and LHCSR-catalyzed excess energy dissipation.
The authors are grateful to Fabien Nogué (Institut Jean-Pierre Bourgin (IJPB), Institut National de la Recherche Agronomique (INRA), Versailles, France) and Mitsuyasu Hasebe (National Institute for Basic Biology (NIBB), Okazaki, Japan) for kindly providing plasmids for moss transformation. The PhD grant received by C.G. was awarded by the ‘Cassa di Risparmio di Padova e Rovigo’ (CaRiPaRo) Foundation. T.M. acknowledges financial support from Università di Padova (grants CPDA089403 and CPDR104834). R.B. acknowledges support from EEC FP7 grant ‘Harvest’ and the ‘Sunbiopath’ EEC project for providing a postdoctoral salary to A.A. Anna Segalla, Stefania Basso (Università di Padova) and Dr Chiara Govoni (Università di Verona) are also thanked for their help.