The solar action spectrum of photosystem II damage

The production of oxygen and the supply of energy for life on earth rely on the process of photosynthesis using sunlight. Paradoxically, sunlight damages the photosynthetic machinery, primarily photosystem II (PSII), leading to photoinhibition and loss of plant performance. However, there is uncertainty about which wavelengths are most damaging to PSII under sunlight. In this work we examined this in a simple experiment where Arabidopsis leaves were exposed to different wavelengths of sunlight by dispersing the solar radiation across the surface of the leaf via a prism. In order to isolate only the process of photodamage, the repair of photodamaged PSII was inhibited by infiltration of chloramphenicol into the exposed leaves. The extent of photodamage was then measured as the decrease in the maximum quantum yield of PSII ( F v / F m ) using an imaging pulse amplitude modulation fluorometer. Under the experimental light conditions, photodamage to PSII occurred most strongly in regions exposed to UV or yellow light. The extent of UV photodamage under incident sunlight would be greater than we observed when one corrects for the optical efficiency of our system. Our results suggest that photodamage to PSII under sunlight is primarily associated with UV rather than photosynthetically active light wavelengths. PSII is caused by light absorbed by photosynthetic pigments and Cser, 2009). To identify which wavelengths of sunlight are most damaging to PSII, sunlight was spectrally dispersed via a prism onto an Arabidopsis leaf infiltrated with chloramphenicol and decrease in the maximum quantum yield of PSII ( F v / F m ) was measured using an imaging pulse amplitude modulation (PAM) fluorometer. This simple but powerful approach revealed the in vivo spectral dependence of photodamage which had two peaks at UV and yellow wavelengths. Since the spectral efficiency of our optical system decreased below 400 nm, we calculated photodamage to PSII under incident sunlight. Our results show that photodamage to PSII was primarily associated with UV wavelengths and secondarily with yellow light wavelengths. This finding indicates that photodamage to PSII is less associated with light absorbed by photosynthetic pigments under sunlight and suggest that most of photodamage to PSII is potentially avoidable during photosynthesis.


Introduction
Plants absorb sunlight to power the productive photochemical reactions of photosynthesis. Absorption of sunlight may also lead to deleterious photochemistry that damages the photosynthetic machinery. The photosystem II (PSII) protein complex is important in this regard as it seems to be most susceptible to photodamage that results in photoinhibition and ultimately suppresses photosynthetic CO 2 assimilation, growth and productivity (Long et al., 1994;Takahashi and Murata, 2008). Although plants have photoprotection mechanisms (Niyogi, 1999) and can effectively repair photodamaged PSII through the PSII repair cycle (Aro et al., 1993), photoinhibition still occurs under stressful environmental conditions (Murata et al., 2007;Nishiyama et al., 2006;Takahashi and Murata, 2008).
The onset of photoinhibition is strongly correlated with the absorption of excessive excitation energy for photosynthesis. Therefore, photodamage to PSII was most readily assumed to be attributed to the excess light absorbed by photosynthetic pigments (Melis, 1999). However, the extent of photodamage that is measured under conditions where the repair of photodamaged PSII is prevented by inhibiting chloroplast protein synthesis (i.e., lincomycin or chloramphenicol) is directly proportional to the intensity of light (Allakhverdiev and Murata, 2004;Chow et al., 2005;Mattoo et al., 1984;Nishiyama et al., 2004;Nishiyama et al., 2001;Tyystjärvi and Aro, 1996).
Furthermore, recent studies have demonstrated that interruption of the Calvin cycle (Hakala et al., 2005;Takahashi et al., 2007;Takahashi and Murata, 2005) and inhibition of electron transfer between Q A and Q B (Allakhverdiev et al., 2005;Jegerschöld et al., 1990;Kirilovsky et al., 1994) have no effect on the rate of photodamage to PSII, but in fact cause inhibition of the repair of photodamaged PSII due to suppression of the de www.plantphysiol.org on August 28, 2017 -Published by Downloaded from Copyright © 2010 American Society of Plant Biologists. All rights reserved. novo synthesis of PSII proteins (Allakhverdiev et al., 2005;Takahashi andMurata, 2005, 2006). Thus, photodamage to PSII is paradoxically not associated with the excess light absorbed by photosynthetic pigments (Murata et al., 2007;Nishiyama et al., 2006;Takahashi and Murata, 2008).
Studies of the effect of monochromatic light on the photodamage process have suggested that photodamage to PSII primarily occurs at the manganese cluster of the oxygen-evolving complex (OEC) through a direct photoexcitation of manganese (Hakala et al., 2005;Ohnishi et al., 2005). Release of manganese ions (Mn 2+ ) from thylakoid membranes is accompanied by photodamage to PSII (Hakala et al., 2005;Zsiros et al., 2006), suggesting that disruption of the manganese cluster upon absorption of light might be a primary event in photodamage. It is likely that the reaction center of PSII is secondarily damaged by light absorbed by photosynthetic pigments after inactivation of the oxygen-evolving complex (Hakala et al., 2005;Ohnishi et al., 2005), if an alternative electron transfer donor from lumenal ascorbate is not available (Mano et al., 2004;Tóth et al., 2009). These findings have lead to a recent photodamage model called the manganese (or two-step (Ohnishi et al., 2005)) mechanism of photoinhibition (Tyystjärvi, 2008).
Studies of the action spectrum of photodamage to PSII have shown that UV damages PSII more effectively than visible light (Hakala et al., 2005;Jones and Kok, 1966;Jung and Kim, 1990;Ohnishi et al., 2005). Thus, under identical light intensity, UV is the most damaging wavelength to PSII. However, inferring damage under natural sunlight is not straight forward as there is a need to account for the spectral distribution and intensity of sunlight. It is unclear which wavelengths of sunlight are most damaging to PSII and we cannot discount the premise that significant primary photodamage to www.plantphysiol.org on August 28, 2017 -Published by Downloaded from Copyright © 2010 American Society of Plant Biologists. All rights reserved. 5 PSII is caused by light absorbed by photosynthetic pigments (Vass and Cser, 2009). To identify which wavelengths of sunlight are most damaging to PSII, sunlight was spectrally dispersed via a prism onto an Arabidopsis leaf infiltrated with chloramphenicol and decrease in the maximum quantum yield of PSII (F v /F m ) was measured using an imaging pulse amplitude modulation (PAM) fluorometer. This simple but powerful approach revealed the in vivo spectral dependence of photodamage which had two peaks at UV and yellow wavelengths. Since the spectral efficiency of our

Results
We set out to examine the photodamage in Arabidopsis leaves caused by exposure to different wavelengths of sunlight. In order to investigate only the process of photodamage, the repair of PSII was inhibited by infiltration of chloramphenicol via the leaf petiole to block the synthesis of PSII proteins, primarily the D1 protein. To examine the relationship between photodamage to PSII and light absorbed by the leaf, the absorptance spectrum of Arabidopsis leaves (from 350 nm to 700 nm) was measured with a spectroradiometer interfaced with an integrating sphere attachment (Supplemental Fig. 2A). The absorptance spectrum of the Arabidopsis leaf showed maximum absorptance in the UV-blue spectral region (350-500 nm) and around the red spectral region (660-680 nm) (Supplemental Fig. 2A). Almost 90% of incident irradiance was absorbed in these two spectral regions. The region of green-yellow www.plantphysiol.org on August 28, 2017 -Published by Downloaded from Copyright © 2010 American Society of Plant Biologists. All rights reserved.
wavelengths (500-600 nm) showed the lowest absorptance (60% of incident irradiance was absorbed at 550 nm). The spectrum of light absorbed by the leaf during the photodamage experiment to the leaf in Figure 2 showed peaks in the blue and red wavebands (Supplemental Fig. 2B). Importantly, peak absorptance wavebands did not coincide with the peaks of photodamage to PSII, observed in UV (330 nm) and yellow (600 nm) wavebands (Fig. 3). These results indicate that photodamage to PSII is not associated with light absorbed by photosynthetic pigments.
Since the extent of photodamage is directly proportional to the intensity of light (Sarvikas et al., 2006;Tyystjärvi and Aro, 1996), the photodamage efficiency at different wavelengths can be calculated by dividing the extent of photodamage to PSII (the extent of decreased F v /F m ) by intensity of incident light at each wavelength (Fig. 4).
Since we do not know how much light is absorbed by the target of photodamage, we calculated photodamage efficiency using incident light intensity. The absorptance of the leaf (see Supplemental Fig. 2) is mainly due to absorption by photosynthetic pigments and may not represent the sites of photodamage. Photodamage efficiency was highest in UV and nearly constant through the visible waveband, consistent with data observed in Arabidopsis intact leaves (Sarvikas et al., 2006). Interestingly, our result showed that the quantum efficiency of photodamage has a small but apparent peak at yellow (600 nm) wavelength in the visible light region (Fig. 4).
Due to the optics of our experimental setup, the proportion of UV to visible light was lower in the spectrum of light reaching the leaf than that of incident sunlight (Supplemental Fig. 3). Consequently, the extent of photodamage by UV shown in Fig. 3 is an underestimate. To correct for this, we multiplied the quantum efficiency of photodamage by the intensity of incident sunlight at each wavelength (Fig. 5)

Discussion
The sunlight illumination results indicate that for an intact Arabidopsis leaf the PSII photodamage is primarily associated with UV wavelengths of sunlight (Fig. 5).
The spectrum of photodamage to PSII differed from that of light absorption by the leaf (Supplemental Fig. 2B). Therefore, the major photodamage to PSII under sunlight appears not to be associated with light absorbed by photosynthetic pigments (either via acceptor-or donor side photoinhibition mechanisms). Recent studies have hypothesized that initial photodamage to PSII occurs at the manganese cluster of the oxygen-evolving complex presumably via direct excitation of manganese (manganese mechanism of photoinhibition) (Tyystjärvi, 2008), with UV being more effective at inducing photodamage compared to visible light (Hakala et al., 2005;Ohnishi et al., 2005) (Fig.   4). The peak of photodamage in the UV region on the leaf (Fig. 5) is therefore consistent with this manganese-based mechanism. On the other hand, it is unclear why yellow light causes photodamage (Fig. 5).
In considering the adverse effects of the yellow wavelengths of sunlight, for a number of reasons we suggest this might also be attributed to the manganese mechanism. Firstly, although visible light excites manganese much less effectively than UV and blue wavelengths (Hakala et al., 2005;Sarvikas et al., 2006), yellow light is much more abundant in the solar spectrum than UV and blue wavelengths (Fig. 5).
Secondly, there is less absorption of yellow light by chlorophylls (primary absorption blue and red) and carotenoids (primary absorption blue and green) associated with the photosystems (Supplemental Fig. 2A). Thirdly, the non-absorbed light reflects and scatters within leaves, resulting in an increase in light intensity near the leaf surface and at depth (Vogelmann et al., 1996). Thus, the weakly absorbed yellow light may be more likely to excite manganese than other wavelengths of visible light, resulting in higher efficiency (Fig. 4) and extent of photodamage to PSII (Fig. 5). Finally, also consistent with this hypothesis, a decrease in leaf chlorophyll content has been demonstrated to enhance the sensitivity of PSII to photodamage in the presence of lincomycin that suppresses the repair of photodamaged PSII (Pätsikkä et al., 2002). The spectrum of photodamage efficiency shown in Arabidopsis intact leaves had no visible light peak (Sarvikas et al., 2006). The difference with our results may be due to a different content of leaf pigments (the leaf used in our study might have higher amount of photosynthetic pigments or lower amounts of green-yellow absorbing pigments such as carotenoids and phenolic compounds). Further studies are necessary to verify this hypothesis.
To prevent photoinhibition, photoprotective mechanisms are used by the plant to both suppress the photodamage to PSII and to minimize inhibition of the repair of photodamaged PSII. The reduced extent of photodamage at the light absorption peaks of photosynthetic pigments (Fig. 5) could therefore be associated with activation of photoprotective mechanisms. In the previous photodamage models, photodamage to PSII was proposed to be attributed to excess light absorbed by photosynthetic pigments (Melis, 1999). Therefore, utilization and dissipation of absorbed light energy through the photosynthetic carbon fixation, thermal energy dissipation, water-water cycle (hydrogen peroxide scavenging) and the photorespiratory pathway were assumed to suppress photodamage to PSII (Melis, 1999 Nishiyama et al., 2001;Takahashi et al., 2007;Takahashi et al., 2009). Thus, such photoprotection mechanisms have no influence on the extent of photodamage at any wavelengths of sunlight. However, chloroplast movement to avoid light might be partially associated with suppressing photodamage to PSII by blue light as chloroplast movement responds to strong blue light and prevents photodamage to PSII (Kasahara et al., 2002). Furthermore, proton gradients across the thylakoid membrane from linear and cyclic electron flows might be also associated with suppressing photodamage to PSII from light absorbed by photosynthetic pigments, primarily blue and red wavelengths (Fig. 5), as it prevents photodamage to PSII (Takahashi et al., 2009). We need to note that the prevention of photodamage to PSII by generation of proton gradients across thylakoid membranes is not associated with thermal energy dissipation and its mechanism has not yet been clarified (Takahashi et al., 2009).
Our results show that photodamage to PSII by sunlight is primarily associated with UV and yellow wavelengths (Fig. 5), suggesting that photodamage to PSII should be suppressed by mechanisms that attenuate these wavelengths, i.e., leaf (Satter and Galston, 1981) and chloroplast (Kasahara et al., 2002) movements and accumulation of compounds that absorb UV and/or yellow wavelengths (i.e., phenolic compounds in the epidermal cells (Booij-James et al., 2000;Landry et al., 1995;Li et al., 1993;Winkel-Shirley, 2001, 2002). The spectrum of sunlight photodamage might vary among plant species and growth conditions depending on the nature and the amount of light absorbing compounds. Indeed, plants grown under sunlight induce protective screening from UV wavelengths compared to those grown under artificial light lacking in UV, which lowers their photosynthetic quantum efficiency at and below 400nm (McCree, 1972). Furthermore, photodamage might also vary through the day and with www.plantphysiol.org on August 28, 2017 -Published by Downloaded from Copyright © 2010 American Society of Plant Biologists. All rights reserved.
1 2 changing weather and season through their influence on the spectrum of sunlight reaching leaves. Given that UV effectively damages PSII (Fig. 4), small increases in UV radiation through thinning of the stratospheric ozone layer caused by an artificial release of chlorofluorocarbons and other ozone antagonists might strongly enhance the extent of photodamage.
Photodamage to PSII has long been believed to be directly attributable to light absorbed by photosynthetic pigments. Therefore, photodamage to PSII was long assumed to be an unavoidable consequence for photosynthetic organisms. However, the present results suggest that artificial or in situ filtering of UV wavelengths could help in reducing photodamage to PSII that causes photoinhibition, with little detriment to photosynthetic CO 2 fixation. Consequently, the growth and productivity of plants under sunlight may be increased using these strategies. The emergent light entered into a quartz equilateral dispersing prism (DPSQ-30-10H, Sigma Koki, Japan). The leaf was placed 240 mm away from the prism and exposed perpendicularly to the dispersed light in the dark box (Fig. 1A). The temperature of the dark box was between 25-27 ºC. To define approximate wavelengths of dispersed light incident on different treated regions of the leaf, pairs of LS (short pass) and LL (long pass) cutoff filters (Corion, MA, USA) were used. The center of each band was defined as 424 nm, 525 nm and 625 nm with LL-400 / LS-450, LL-500 / LS-550, and LL-600 / LS-650 filter pairs respectively (Fig. 1B).

Measurement of photoinhibition.
The maximum quantum yield of PSII (F v /F m ) was measured after incubation in darkness for 15 min with an imaging pulse amplitude modulation fluorometer (Imaging-PAM; Walz, Germany). The F v /F m value shown in Fig. 2  Measurement of leaf absorptance. Leaf absorptance for Arabidopsis was calculated by measuring leaf reflectance and transmittance spectra with an LI-1800 spectroradiometer and the 1800-12S integrating sphere attachment (Li-Cor Inc.). The sample scan was divided by its corresponding reference scan from 350 nm to 800 nm.

Measurement of spectra of incident sunlight and light reaching the leaf.
The optical properties of the experimental setup were measured with a spectroradiometer (LI-1800, Li-Cor Inc.) using the small cosine corrected head. A 0.33 mm slit formed by two razor blades was positioned 10 mm above the centre of the head to enable the spectrum to be spatially resolved. A correction factor for the slit was made by multiplying the irradiance at each wavelength by the ratio of sunlight measured with and without the slit. Spectra were measured at one mm intervals along the position of the leaf in the light box. The frequency of the peak irradiance and half bandwidth frequency were linearly related to distance (half bandwidths increased from 26 nm at 359 nm, to 48 nm at 477 nm and 144 nm at 693 nm). The efficiency of the optical system was calculated at each position as the ratio of the irradiance at the peak wavelength to that of incident sunlight (Supplemental Fig. 3B). The efficiency, which increased curvilinearly from 0.24 at 361 nm to reach a maximum of 0.38 at 477 nm, was used to calculate the actual quantum dose at each position along the leaf.
www.plantphysiol.org on August 28, 2017 -Published by Downloaded from Copyright © 2010 American Society of Plant Biologists. All rights reserved.    Both spectra were normalized to a value of one at 400 nm. An Arabidopsis leaf was pre-incubated with 1mM chloramphenicol at 20 µmol photons m -2 s -1 for 4 h. The leaf was exposed to spectrally dispersed sunlight for 15 min. The maximum quantum yield of PSII (F v /F m ) was measured with an imaging-PAM before and after the light exposure. A, Photograph of Arabidopsis leaf exposed to sunlight dispersed using a prism. B, Image of the F v /F m value before exposure to dispersed sunlight. C, Image of the F v /F m value after exposure to dispersed sunlight for 15 min.
www.plantphysiol.org on August 28, 2017 -Published by Downloaded from Copyright © 2010 American Society of Plant Biologists. All rights reserved.