# Source paper archive - Source title: PSI photoinhibition and protection at low temperatures and in fluctuating light: An 30 ecophysiological vista - Author(s): TAKEUCHI, K., OGUCHI, R., SUGIMOTO, K., TERASHIMA, I., KONO, M. - Source landing URL: https://www.biorxiv.org/content/10.64898/2026.11.02.763831v1 - PDF URL: https://www.biorxiv.org/content/10.64898/2026.11.02.763831v1.full.pdf - PDF SHA-256: 70a4b8f2758608f38e547dfdc7b9f5a9c9474013d181dfc4694f2a1780986d9a - Converter: pymupdf4llm 0.0.27 - Limits: Automated PDF text extraction; reading order, equations, tables and formatting may be incomplete or incorrect. Images are omitted, and no OCR was performed. Markup and links are neutralized for safe downloading. Consult the original PDF for the authoritative version. --- [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.02.763831; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.02.763831) preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 1 **Corresponding authors** 2 3 **Ichiro TERASHIMA** 4 Institute of Molecular Biology, National Chung Hsing University 5 145 Xingda Rd., South Dist., Taichung 40227, Taiwan 6 TEL: +886-958-127-704 7 terashimaichiro@nchu.edu.tw, terashimaichiro6@gmail.com 8 9 **Masaru KONO** 10 Exo-Life Search Project Office, Astrobiology Center, National Institutes of Natural Sciences (NINS) 11 2-21-1 Osawa, Mitaka, Tokyo 181-8588, Japan 12 TEL: +81-422-34-4072 13 Laboratory for Chemistry and Life Science, Institute of Integrated Research, Institute of Science Tokyo 14 4259 Nagatsuta-cho, Midori-ku, Yokohama, Kanagawa 226-8501, Japan 15 TEL: +81-45-924-5246 16 [konom07@bs.s.u-tokyo.ac.jp](mailto:konom07@bs.s.u-tokyo.ac.jp) 17 18 19 Subject Area 20 (2) Ecophysiology / (5) Physiology/Biochemistry 21 22 Number of black and white figures, color figures and table: 23 Black and white figure: 1 24 Color figure: 4 25 Table: 2 26 Supplementary figure: 0 27 Supplementary table: 3 28 Supplementary text: 3 29 1 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.02.763831; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.02.763831) preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 30 **PSI photoinhibition and protection at low temperatures and in fluctuating light: An** 31 **ecophysiological vista** 32 33 **Ko TAKEUCHI** **[1]** **, Riichi OGUCHI** **[2]** **, Kazuhiko** **SUGIMOTO** **[3]** **, Ichiro TERASHIMA** **[3]** ***, Masaru** **KONO** **[4, 5]** ***** 34 35 1 Graduate School for Agricultural Science, Kobe University, 1-1 Rokkodai, Nada, Kobe, 657-8501, Japan 36 2 Graduate School of Science, Osaka Metropolitan University, 2000, Kisaichi, Katano-shi, Osaka, 576-0004, Japan 37 3 Institute of Molecular Biology, National Chung Hsing University, 145 Xingda Rd., South Dist., Taichung 40227, 38 Taiwan 39 4 Exo-Life Search Project Office, Astrobiology Center, National Institutes of Natural Sciences (NINS), 2-21-1 Osawa, 40 Mitaka, Tokyo 181-8588, Japan 41 5 Laboratory for Chemistry and Life Science, Institute of Integrated Research, Institute of Science Tokyo, 4259 42 Nagatsuta-cho, Midori-ku, Yokohama, Kanagawa 226-8501, Japan 43 *Corresponding authors: terashimaichiro@nchu.edu.tw 44 [konom07@bs.s.u-tokyo.ac.jp](mailto:konom07@bs.s.u-tokyo.ac.jp) 45 46 47 48 ORCID 49 Ko Takeuchi https://orcid.org/0009-0008-6810-5331 50 Riichi Oguchi https://orcid.org/0000-0003-1845-1558 51 Kazuhiko Sugimoto 52 Ichiro Terashima https://orcid.org/0000-0001-7680-9867 53 Masaru Kono https://orcid.org/0009-0000-6927-8930 54 55 56 2 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.02.763831; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.02.763831) preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 57 **PSI photoinhibition and protection at low temperatures and in fluctuating light: An** 58 **ecophysiological vista** 59 60 **Abstract** 61 In green leaves _in vivo_, photosystem I (PSI) was thought to be resistant to photoinhibition. In 1994, it was found that, 62 in chilling-sensitive plants such as cucumber ( _Cucumis sativus_ L.), exposing leaves to moderate constant light at 63 temperatures below 10 °C induces PSI photoinhibition with little effect on PSII. Since the beginning of this century, 64 PSI photoinhibition induced by fluctuating light (alternating high and low light, FL) or repetitive saturating pulses 65 (rSPs) has been reported in many species. In this review, we compare the causes and consequences of PSI 66 photoinhibition in cucumber at chilling temperatures with those of PSI photoinhibition induced by FL or rSPs. 67 Changes in chloroplasts during these photoinhibitory treatments are followed, focusing on where and how reactive 68 oxygen species are produced and which parts of PSI are damaged. We also discuss the roles of two pathways of cyclic 69 electron flow around PSI, the NDH and PGR5/PGRL1 pathways, particularly in suppression of PSI photoinhibition. 70 Lastly, we discuss problems with methods for measuring PSI parameters and future research directions from an 71 ecophysiological viewpoint. 72 73 **Keywords** Cyclic electron flow · NDH pathway · PGR5/PGRL1 pathway · Proton motive force (pmf) 74 75 **Abbreviations** 76 A 0 primary electron acceptor of PSI (Chl _a_ ) 77 A 1 secondary electron acceptor of PSI (phylloquinone) 78 APX ascorbate peroxidase 79 CBB cycle Calvin–Benson–Bassham cycle 80 CEF-PSI cyclic electron flow around PSI 81 CF 1 catalytic portion (coupling factor 1) of the chloroplast ATP synthase 82 Chl chlorophyll 83 Cyt _b_ 6 _f_ cytochrome _b_ 6 _f_ complex 84 DAD 3,6-diaminodurene 85 DCMU 3-(3,4-dichlorophenyl)-1,1-dimethylurea 86 ECS electrochromic shift 87 ETR(I) electron transport rate through PSI 88 F A, F B terminal iron–sulfur clusters of PSI, bound to PsaC 89 Fd ferredoxin 90 Fe–S iron–sulfur 91 FL fluctuating light 92 FLV flavodiiron protein 93 FNR ferredoxin–NADP [+] reductase 94 FR far-red light 95 FTR ferredoxin–thioredoxin reductase 96 F v /F m maximum quantum yield of PSII photochemistry 97 F X iron–sulfur cluster of PSI, bound to PsaA and PsaB 98 _g_ H [+] proton conductivity of the thylakoid ATP synthase 99 HL high light 100 _K_ pi rate constant of PSI photoinhibition 101 _K_ pi / _I_ rate constant of PSI photoinhibition divided by irradiance 102 LED light-emitting diode 103 LHCI light-harvesting complex I 104 LL low light 3 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.02.763831; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.02.763831) preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 105 MV methyl viologen 106 NDH chloroplast NADH dehydrogenase-like complex 107 NPQ non-photochemical quenching of chlorophyll fluorescence 108 NTRC NADPH-dependent thioredoxin reductase C 109 1 O 2 singlet oxygen 110 O 2 [•] ⁻ superoxide radical anion 111 P700 reaction center chlorophyll of PSI 112 PAR photosynthetically active radiation 113 PC plastocyanin 114 PGR5 PROTON GRADIENT REGULATION 5 115 PGRL1 PGR5-LIKE PHOTOSYNTHETIC PHENOTYPE 1 116 PIFR post-illumination fluorescence rise 117 P m maximum amount of photooxidizable P700 118 pmf proton motive force 119 PQ plastoquinone 120 PQH 2 plastoquinol 121 PSI photosystem I 122 PSII photosystem II 123 PTOX plastid terminal oxidase 124 RISE reduction-induced suppression of electron flow 125 ROS reactive oxygen species 126 rSPs repetitive saturating pulses 127 SOD superoxide dismutase 128 SP saturating pulse 129 Trx thioredoxin 130 WT wild type 131 Y(I) quantum yield of PSI photochemistry 132 Y(II) effective quantum yield of PSII photochemistry 133 Y(NA) quantum yield of non-photochemical energy dissipation due to acceptor-side limitation of PSI 134 Y(ND) quantum yield of non-photochemical energy dissipation due to donor-side limitation of PSI 135 Y(NO) quantum yield of non-regulated energy dissipation in PSII 136 Y(NPQ) quantum yield of regulated non-photochemical energy dissipation in PSII 137 Δ _E_ h difference in redox potential between the electron acceptor (PQ) and donor (Fd) of NDH 138 Δ _p_ proton motive force expressed in mV 139 ΔpH pH difference across the thylakoid membrane 140 ΔΨ membrane potential across the thylakoid membrane 141 142 4 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.02.763831; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.02.763831) preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 143 **Introduction** 144 Photosystem I (PSI) is a plastocyanin/ferredoxin photooxidoreductase. The core complex is formed as a p seudo145 homodimer (PsaA and PsaB) having a reaction center chlorophyll dimer (Chl _a_ and Chl _a_ ’). Each of PsaB and PsaA 146 has cofactors, namely Trp, three Chl _a_ molecules including (A 0 ), and phylloquinone (A 1 ). Unlike photosystem II 147 (PSII), charge separation occurs on both sides of the pseudo-homodimer. The two routes merge at the iron sulfur 148 center F X sandwiched between the PsaA and PsaB proteins, while F A and F B are cofactors in the subunit PsaC protein 149 (Rutherford et al. 2012). PSI reduces ferredoxin (Fd) and thereby provides reducing power for carbon assimilation 150 and other metabolic processes. 151 Although there were pioneering studies showing photoinhibition of PSI _in vitro_ (Inoue et al. 1986, 1989; Satoh 152 1970a, b, c), there were few studies showing PSI photoinhibition _in vivo_ (Öquist et al. 1987; Powles 1984) until 1994. 153 After the discovery that PSI in a chilling-sensitive plant, _Cucumis sativus_ L. (cucumber), is readily damaged by 154 moderate constant light when the leaf temperature is kept below 10 °C (Terashima et al. 1994), chilling-induced PSI 155 photoinhibition was reported in a wide range of species originating from warm areas. 156 PSI photoinhibition can also be induced by fluctuating light (FL, alternating high light (HL) and low light (LL)) 157 (for reviews, see Kono and Terashima 2014; Yamori 2016). Tikkanen et al. (2010) grew several mutant lines of 158 _Arabidopsis thaliana_ (L.) Heynh. in the FL alternating 500/50 µmol photons m [–2] s [–1] for 1 min/5 min for 12 h daytime 159 (500/50 | 1 min/5 min, 12-h photoperiod) and found that the _pgr5_ mutant (the mutant of _PROTON GRADIENT_ 160 _REGULATION 5_, a component of the PGR5 cyclic electron flow pathway around PSI, CEF-PSI) died. The lethal 161 phenotype was due to PSI photoinhibition (Suorsa et al. 2012). PSI photoinhibition induced by FL occurs within 162 much shorter periods, not only in the _pgr5_ mutant but also in the wild type. Kono et al. (2014) treated leaves of _A._ 163 _thaliana_ with the FL from red LEDs (240/30 | 2 min/2 min | 42 min). It has been also known that treatment of the 164 leaves with repetitive saturating pulses (rSPs) induces remarkable PSI photoinhibition in _A. thaliana_ (Tsuyama and 165 Kobayashi 2009), _Helianthus annuus_ L. (sunflower, Sejima et al. 2014), and _Triticum aestivum_ L. (wheat, Takagi et 166 al. 2019; Zivcak et al. 2015). For example, Sejima et al. (2014) used the regime of (20,000/0 | 300 ms/9.7 s | 240 167 min). Distinction between the FL and rSPs is not clear. In this article, the FL using very strong short pulses (less than 168 1 s duration) are called rSPs according to Sonoike (2025). Because it matters whether the LL phase of FL or rSPs 169 consists of darkness or nonzero low irradiance, we treat these as distinct lighting regimes in § 3. 170 Photodamage to PSI imposes high cost on plants because repair of photodamaged PSI or PSI _de novo_ synthesis 171 is slow and thereby PSI photoinhibition adversely affects plant growth (Amin et al. 2024; Kudoh and Sonoike 2002). 172 To avoid photoinhibition, plants must tightly control the redox state of PSI by balancing electron flows into and out 173 of PSI, and by dissipating excitation energy through safe charge recombination. This protection depends on several 174 coordinated processes, including photosynthetic control and non-photochemical quenching (NPQ), both induced by 175 acidification of the thylakoid lumen, regulation of photosystem II (PSII) activity, alternative electron flow pathways 176 including the Asada (water to water) pathway and two CEF-PSI pathways, the state transitions, spillover of 177 excitations from PSII to PSI, etc. 178 We should also pay attention to light spectra used in the photoinhibition studies. If the FL in natural 179 environments exerted the same effect as the FL from red LEDs, many plants in the field would suffer from 180 photoinhibition. However, in natural environments, it does not appear that many plants suffer from the PSI 181 photoinhibition. In the presence of sufficient far-red light, PSI photoinhibition by the red FL is largely suppressed 182 (Kono et al. 2017, 2020). From an ecophysiological viewpoint as well, the role of far-red light in PSI photoinhibition 183 should be properly recognized. 184 In this article, we review recent progress in the PSI photoinhibition studies in the model chilling-sensitive plant, 185 cucumber, and those using the FL or rSPs. Then, the mechanisms underlying PSI photoinhibition and its 186 photoprotection are argued. In particular, the roles of two CEF-PSI pathways are discussed in detail. Angiosperms 187 have two pathways of cyclic electron flow around PSI that transfer electrons from Fd to the PQ pool: the 188 PGR5/PGRL1-dependent pathway (hereafter PGR5 pathway), which is sensitive to antimycin A, and the NDH189 dependent pathway (hereafter NDH pathway), which is catalyzed by the chloroplast NADH dehydrogenase-like 190 complex (Munekage et al. 2002, 2004; Shikanai 2007; Yamori and Shikanai 2016). 191 We also consider technical problems in measurements of PSI parameters. In stress physiology studies, 5 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.02.763831; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.02.763831) preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 192 researchers often expose plants to drastic and unrealistic stress conditions. Although plants respond to such treatments 193 and we are fully aware that such studies often deepen our knowledge of plant biological machinery, the plant 194 responses in such studies should be interpreted with caution. We also have to seek ecological realism, because plant 195 biological machinery has been mostly shaped by natural selection occurring in natural environments. The contents 196 of this review are as follows: 197 198 1. PSI photoinhibition in cucumber, the chilling-sensitive model plant. 199 2. PSI photoinhibition in the fluctuating light and repetitive saturating pulses. 200 3. Meta-analyses of PSI photoinhibition studies 201 4. How do plants protect PSI? 202 5. Roles of NDH and PGR5/PGRL1 in protection from PSI photoinhibition 203 6. ROS: their generation and scavenging 204 7. Roles of far-red light and allocation of excitation between PSI and PSII 205 8. Technical problems in quantification of PSI parameters. 206 9. Future directions in PSI photoinhibition research 207 208 **1. PSI photoinhibition in cucumber, the chilling-sensitive model plant.** 209 **1.1 Chilling stress inhibits photosystem I.** 210 In 1994, it was found that illuminating cucumber leaves with weak light at temperatures below 10 °C decreased the 211 quantum yield of PSI with little effect on PSII (Sonoike and Terashima 1994; Terashima et al. 1994). This 212 phenomenon was not observed under anaerobic conditions or in darkness, indicating that it is oxidative damage 213 requiring both light and oxygen (Bukhov et al. 2004; Sonoike 1996b; Terashima et al. 1998). By 2002, studies using 214 cucumber as a model had shown that chilling weakens association of CF 1 of the thylakoid ATP synthase with the 215 thylakoid membrane, thereby impairing ΔpH formation and photosynthetic control (Peeler and Naylor 1988; 216 Terashima et al. 1989, 1991a, b); degrades the PSI core protein PsaB (Kudoh and Sonoike 2002; Sonoike 1996a; 217 Sonoike and Terashima 1994; Sonoike et al. 1997); and damages the F A /F B and F X Fe–S clusters (Sonoike et al. 1995). 218 More recently, Takeuchi, Ifuku, and their colleagues demonstrated that destabilization of the PSI–NDH supercomplex 219 and PSI over-reduction occur at an early stage of light treatment at chilling temperatures in cucumber (Takeuchi et 220 al. 2022, 2025a, 2026). 221 After its discovery in cucumber, chilling-induced PSI photoinhibition was reported in a wide range of species, 222 including _Solanum lycopersicum_ L. (tomato), _Oryza sativa_ L. (rice), _Zea mays_ L. (maize), _Gossypium hirsutum_ L. 223 (cotton), _Saccharum officinarum_ L. (sugarcane), _Secale cereale_ L. (rye) etc. (Chen et al. 2024; Govindachary et al. 224 2004; Ivanov et al. 1998; Kornyeyev et al. 2003; Takeuchi et al. 2026; Tjus et al. 1999; Zhang Y et al. 2025). Except 225 for rye, these plants all originate from warm-climate regions. 226 227 **1.2 How is PSI photoinhibited?** 228 PSI photoinhibition is not an immediate consequence of exposing leaves to low temperatures. Instead, it develops 229 gradually as chilling weakens the array of mechanisms that protect PSI. The key to understanding this process is to 230 determine the sequence of events leading to PSI photoinhibition. The time course outlined below is based primarily 231 on studies of cucumber leaves chilled at 4 °C under irradiances of 100–250 µmol photons m⁻ [2] s⁻ [1 ] (400-700 nm). The 232 timing of individual events varies with plant species or varieties, leaf age, temperature, and the irradiance level. 233 234 **Phase 1 (<30 min).** Immediately upon exposure to chilling treatment, the Calvin-Benson-Bassham (CBB) cycle, 235 photorespiration, and other metabolic processes that consume reducing power downstream of PSI slow down (Zhou 236 et al. 2007). As NADP⁺ regeneration slows, oxidation of Fd by ferredoxin–NADP⁺ reductase (FNR) becomes 6 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.02.763831; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.02.763831) preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 237 restricted. Irreversible PSI photoinhibition is not yet detectable, but transient over-reduction of the electron transfer 238 chain can promote reactive oxygen species (ROS) formation. ROS generated within PSI are normally scavenged by 239 the Asada cycle (water-water cycle) (Asada 1999; Miyake 2010; Miyake and Asada 1992; Miyake et al. 1998). At 240 low temperatures, however, the activity of thylakoid-bound ascorbate peroxidase (APX) decreases markedly 241 (Terashima et al. 1998), limiting the capacity for ROS scavenging. Insufficient scavenging during this early phase – 242 may contribute to the destabilization of the chloroplast H⁺-ATP synthase and PSI NDH supercomplex observed in 243 Phase 2. 244 In parallel with these early metabolic and redox responses, the physical state of the thylakoid membrane may 245 also change. The phase transition temperature of polar lipids extracted from cucumber thylakoids is approximately 246 10 °C (Hodgson et al. 1987; Orr and Raison 1987, 1990). Notably, this transition temperature is strikingly close to 247 the approximately 10 °C threshold below which photoinhibition occurs in intact cucumber leaves (Hodgson et al. 248 1987; Sonoike 2011; Terashima et al. 1994). Cucumber thylakoids also show lower lipid fluidity and a larger fraction 249 of ‘motion-restricted lipids associated with proteins’ than thylakoids from _Pisum sativum_ L., a chilling tolerant 250 species (Li et al. 1990). Exposure to temperatures below the transition may therefore rapidly reduce local membrane 251 fluidity and affect lipid–protein interactions involved in the regulation of photosynthetic electron transport and P700 252 oxidation. However, no study has directly examined both the membrane phase transition and PSI photoinhibition 253 under the same experimental conditions. Thus, it remains unclear whether thylakoid lipid properties contribute to PSI 254 photoinhibition in cucumber. Despite this uncertainty, changes in thylakoid membrane fluidity would contribute, at – 255 least in part, to the destabilization of the thylakoid ATP synthase and PSI NDH supercomplex observed in Phase 2. 256 257 **Phase 2 (30 min to 2 h).** Between 30 min and 2 h of chilling in the light, the CF 1 sector of the thylakoid ATP synthase 258 becomes more loosely associated with the thylakoid membrane, potentially weakening the coupling between proton 259 translocation and ATP synthesis (see § 1.3). This impairs the formation and maintenance of ΔpH across the thylakoid 260 membrane, thereby weakening both NPQ and photosynthetic control. Electron flow from PSII to PSI consequently 261 continues even though the CBB cycle is suppressed. During the same period, NDH activity also declines (see § 1.3). 262 By accepting electrons from reduced Fd and returning them to the PQ pool, NDH provides a route for Fd oxidation 263 and can relieve acceptor-side over-reduction when photosynthetic control is weakened. Destabilization of the PSI– 264 NDH supercomplex within the first hour of chilling, however, abolishes NDH activity and eliminates this route. 265 Taken together, loose association of thylakoid ATP synthase and NDH inactivation allow the transient PSI over266 reduction observed in Phase 1 to become enhanced. 267 268 **Phase 3 (2–5 h).** At this stage, the P700 oxidation capacity substantially decreases, as indicated by decreased Y(ND), 269 and reduced Fd accumulates. This persistent PSI acceptor side over-reduction accelerates ROS production within PSI 270 and promotes irreversible PSI photoinhibition. The maximum amount of photooxidizable P700 (P m ) decreases 271 markedly while the maximum quantum yield of PSII photochemistry (F v /F m ) is relatively well maintained, indicating 272 preferential photoinhibition of PSI. Pronounced damage to the Fe–S clusters and degradation of core proteins such 273 as PsaB are also observed. Once this stage is reached, subsequent plant growth is severely inhibited. With further 274 chilling, the loss of PSI activity restricts electron transport downstream of PSII, eventually causing F v /F m to decline 275 markedly. 276 Taken together, chilling-induced PSI photoinhibition proceeds in the following sequence (Fig. 1): (1) reduced 277 electron consumption by enzymatic reactions and a decrease in thylakoid membrane fluidity; (2) transient over278 reduction of the PSI acceptor side and ROS production; (3) weakened association of CF 1 of the ATP synthase with 279 the thylakoid membrane and destabilization of the PSI–NDH supercomplex; (4) impairment of photosynthetic control 280 and Fd oxidation capacity; and (5) persistent PSI over-reduction and damage to the Fe–S clusters and PSI reaction 281 center. 282 283 **1.3 Dysfunction of protection mechanisms against chilling-induced PSI photoinhibition.** 7 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.02.763831; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.02.763831) preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 284 **Donor-side.** Damage to the thylakoid ATP synthase is a key event in the early progression of chilling-induced PSI 285 photoinhibition. When chilling-sensitive cucumber leaves were illuminated at low temperatures, the association of 286 the catalytic portion (CF 1 ) of ATP synthase with the membrane loosened, resulting in uncoupling of ATP synthesis 287 (Terashima et al. 1989, 1991a, b, 1998). Accordingly, when thylakoids were isolated immediately after the treatment 288 of the leaves at chilling temperatures in the light, substantial amounts of CF 1 polypeptides were recovered in the 289 supernatant. Such loose association would leave the CF 0 proton channel open to the stroma, allowing H⁺ to leak from 290 the lumen. This proton leakage impairs formation of the trans-thylakoid pH gradient (ΔpH), thereby weakening ΔpH291 dependent NPQ and photosynthetic control at Cyt _b_ 6 _f_ complex (Kono et al. 2022; Takagi et al. 2017a). Indeed, when 292 cucumber leaves were exposed to 4 °C and 100 µmol photons m⁻² s⁻¹, NPQ formation was markedly suppressed 293 within 60 min after the onset of treatment (Fig. 2). Using the electrochromic shift (ECS) at 515 nm, Kono et al. (2022) 294 showed that proton leakage associated with CF 1 destabilization or dissociation is slower than protonophore-mediated 295 proton transport but sufficiently rapid to impede ΔpH formation. Proton leakage was detectable after only 30 min of 296 chilling and can therefore precede PSI photoinhibition. 297 **Acceptor-side.** A comparative study among cucumber cultivars revealed substantial variation in susceptibility to 298 chilling-induced PSI photoinhibition (Takeuchi et al. 2022). Notably, a chilling-tolerant cucumber cultivar was 299 relatively resistant to PSI photoinhibition despite a decrease in ΔpH (Takeuchi et al. 2025a). This observation points 300 to an important role for acceptor-side pathways in protecting PSI when ΔpH-dependent control is weakened. Takeuchi 301 et al. (2025a) identified NDH activity as the major difference on the PSI acceptor side. NDH is a large thylakoid 302 membrane complex that accepts electrons from reduced Fd and reduces PQ (Ifuku et al. 2011; Shikanai 2016; 303 Shikanai et al. 2025). In land plants, NDH associates with two PSI units through Lhca5 and Lhca6 to form the PSI– 304 NDH supercomplex (Kato et al. 2018b, 2021; Otani et al. 2017, 2018). Fd binds to the subcomplex E (SubE) of the 305 PSI–NDH supercomplex, and electrons pass through three internal Fe–S clusters before reaching PQ (Introini et al. 306 2025; Schuller et al. 2019; Shen et al. 2022; Su et al. 2022). The close association between PSI and NDH may 307 facilitate electron transfer from Fd to NDH. Despite this structural arrangement, the physiological role of NDH in C 3 308 plants remained unclear for many years because NDH-deficient mutants generally show no obvious phenotype under 309 stable growth conditions. We discuss the functions of NDH in C 3 plants in more detail in §§ 2 and 5. 310 In a chilling-sensitive cucumber cultivar, NDH activity was lost within 1 h of chilling treatment (Takeuchi et al. 311 2025a, 2026). This loss was not observed in leaves chilled in the dark, consistent with photooxidative inactivation of 312 PSI–NDH supercomplex (Takeuchi et al. 2026). During chilling in the light, the PSI–NDH supercomplex was 313 destabilized and NDH became monomeric. Because monomeric NDH is inactive (Kato et al. 2018a; Otani et al. 314 2018), maintenance of the PSI–NDH supercomplex during chilling is important for limiting PSI over-reduction and 315 subsequent photoinhibition. It has been further shown that an NDH-deficient rice mutant was more susceptible to 316 chilling-induced PSI photoinhibition than the wild type (Takeuchi et al. 2026). Taken together, these results show 317 that NDH protects PSI in C 3 plants at low temperatures. It is noteworthy that Terashima, Sonoike, and colleagues 318 luckily chose a very chilling-sensitive cultivar ‘Nanshin’ for their studies started in 1990s. 319 How does NDH prevent over-reduction of the PSI acceptor side? NDH accepts and transfers electrons to the PQ 320 pool. At the same time, NDH pumps protons into the lumen (Burnap 2026; Miller et al. 2021; Richardson et al. 2021; 321 Strand et al. 2017). ΔpH thus formed would strengthen photosynthetic control. A highly reduced PQ pool would 322 restrict electron transfer by the reduction-induced suppression of electron flow (RISE) (Shaku et al. 2016; Shimakawa 323 et al. 2018; Tikhonov 2024). NDH contains three internal Fe–S clusters that may act as transient electron sinks and 324 thereby shorten the lifetime of reduced states on the PSI acceptor side (Richardson et al. 2021). The physiological 325 importance of NDH during chilling is supported by the reports that NDH abundance increases under chilling in the 326 light (Teicher et al. 2000) and that reduced NDH activity in maize is associated with lower chilling tolerance 327 (Urzinger _et al._ 2025). However, large ΔpH and highly reduced PQ pool could suppress NDH activity (Strand et al. 328 2017). We will discuss regulation of NDH activity in §§ 5.1 and 5.3. 329 The importance of relieving acceptor-side reduction is further illustrated by flavodiiron proteins (FLVs). FLV 330 provides a terminal electron sink by transferring electrons to O 2 and reducing it to water. Rice plants expressing FLV 331 are highly resistant to chilling-induced PSI photoinhibition (Takeuchi et al. 2026). Conifers such as _Pinus sylvestris_ 332 L. (Scots pine) and _Picea abies_ (L.) Karst. (Norway spruce) have lost chloroplast NDH during evolution and instead 8 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.02.763831; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.02.763831) preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 333 rely on FLV to mitigate early-spring chilling damage (Bag et al. 2023; Wakasugi et al. 1994). These observations 334 suggest that NDH and FLV provide biochemically distinct but functionally overlapping routes for protecting PSI 335 during chilling stress. Concerning another CEF-PSI pathway involving PGR5/PGRL1, which is widely regarded as 336 more active than NDH in C₃ plants (Yamori and Shikanai 2016), information on its role in PSI photoinhibition at low 337 temperatures is scarce. 338 339 **2. PSI photoinhibition in the fluctuating light and repetitive saturating pulses.** 340 Fluctuating light (FL) and repetitive saturating pulses (rSPs) have been used to clarify the mechanisms underlying 341 PSI photoinhibition and the mechanisms that protect PSI against photoinhibition. PGR5- and NDH-pathways both 342 generate proton motive force (pmf) without net production of NADPH, and both have been linked to the protection 343 of PSI from photoinhibition. However, there are contrasting views on which pathway is more important. 344 In the paper that first reported the _A. thaliana pgr5_ mutant, the plants grown at 50 µmol photons m⁻² s⁻¹ were 345 used. When the plants were exposed to a constant HL of 1,500 µmol photons m⁻² s⁻¹ for 10 min, up to 60% of PSI 346 was damaged in _pgr5_, whereas less than 10% was damaged in WT (Munekage et al. 2002). This pioneering study 347 showed that PSI in the _pgr5_ mutant is susceptible to photoinhibition, although the _pgr5_ mutant does not show a 348 marked phenotype when grown in constant light. 349 When _A. thaliana pgr5_ seedlings were grown in white fluctuating light (FL) alternating between 500 µmol 350 photons m⁻² s⁻¹ for 1 min and 50 µmol photons m⁻² s⁻¹ for 5 min (500/50 | 1 min/5 min), they died due to PSI 351 photoinhibition, whereas the WT and _ndho_, an NDH mutant, showed no visible phenotype (Suorsa et al. 2012). PSI 352 photoinhibition occurs in milder FL treatments for much shorter periods. In the red FL for only 42 min (240/30 | 2 353 min/2 min | 42 min), P m decreased to 89% in the wild type (WT) and 62% in _pgr5_, whereas in _crr2-2_, an NDH mutant, 354 the decrease in P m was comparable to that in WT (Kono et al. 2014). In the harsher red FL (1,200/30 | 2 min/2 min | 355 162 min), P m decreased to 70% in WT, 20% in _pgr5_ and 60% in _crr2-2_ (Kono and Terashima 2016). These results 356 clearly indicate that PSI is readily photoinhibited by FL and that PGR5 is the main actor in suppression of PSI 357 photoinhibition. When _Oryza sativa_ L. WT and mutants grown at 28 °C were treated with the FL (1,500/200 | 10 358 min/10 min | 5 h), P m decreased to 81%, 61% and 33% of the initial levels in WT, _crr6_ (an NDH mutant) and PGR5359 knockdown lines, respectively (Yamori et al. 2016), indicating that PGR5 contributes to protection much more than 360 NDH. 361 When far-red light (FR, peaking at 740 nm, 25.6 W m⁻²) was added to both HL and LL phases of the red FL 362 regime (1,200/30 | 800 ms/10 s | 120 min), the conclusion reversed. Without FR, P m declined to 70% in all genotypes, 363 whereas with FR in both phases, P m in WT and _pgr5_ remained at 97% and 92% of the initial level, and P m in _crr2-2_ 364 and _crr4-3_, NDH mutants, declined to 80% (Kono et al. 2017). With FR, therefore, PSI protection depended on the 365 NDH pathway rather than on the PGR5 pathway. Similar results were obtained in the tobacco ( _Nicotiana tabacum_ 366 L.) _ndhB_ -disrupted line and in the rice _crr6_ line in the FL with a HL phase of 2,000 µmol photons m⁻² s⁻¹: FR protected 367 PSI in WT, but the protection was weaker in the NDH-deficient lines (M. Kono, unpublished data). 368 In rice lines grown at a low temperature of 20 °C, the NDH-deficient mutant _crr6_ showed lower ETR(I), CO₂ 369 assimilation rate and relative growth rate than WT. When grown at 28 or 35 °C, the growth of _crr6_ was similar to 370 that of WT, but ETR(I) and CO₂ assimilation rate were still lower below the growth irradiance, 800 µmol photons 371 m⁻² s⁻¹ (Yamori et al. 2011). A recent comprehensive study by Kodama and Yamori (2026) using rice WT and _crr6_, 372 grown in the field, showed that NDH deficiency markedly decreases ETR(I) and CO₂ assimilation rate, at low 373 temperatures and/or in low irradiances. In the liverwort _Marchantia polymorpha_ L., disruption of _ndhB_ increased the 374 reduction state of the PQ pool at low irradiances (Ueda et al. 2012), probably due to PSI acceptor-side limitation. 375 As already discussed, NDH plays a crucial role in the protection of PSI at low temperatures (§ 1). Taken together, 376 NDH is important at low temperatures, at low irradiances, and in FL in the presence of FR, whereas PGR5 functions 377 in ordinary HL and in FL of high irradiances with no FR supplementary light. 378 9 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.02.763831; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.02.763831) preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 379 **3. Meta-analyses of PSI photoinhibition studies.** 380 The number of studies on PSI photoinhibition has been increasing since the 1990s. Recently, studies examining the 381 effects of fluctuating-light or intermittent strong light, and those using the mutants related to CEF-PSI have been 382 increasing. These studies have used various plant species. To conduct meta-analysis and discuss the effects of 383 photoinhibitory conditions on the rate of PSI photoinhibition, we collected data from 107 experiments from 32 papers 384 (Table S1). In photoinhibition studies, the extent of photoinhibition is related both to the photon-exposure (fluence) 385 and to the irradiance level (fluence rate) (Kornyeyev et al. 2003; Oguchi et al. 2021), although this approach may not 386 be simply applicable to the FL or rSPs regimes. Since the repair process of PSI is slow, we may neglect the repair 387 process. Using the equation proposed by Kok (1956), the rate of PSI photoinhibition is described by a simple 388 differential equation: 389 390 d _a_ /d _t_ = - _K_ pi - _a_ ( _t_ ) 391 392 where _a_ ( _t_ ) is the active fraction of PSI reaction centers and _K_ pi is the rate constant for PSI photoinhibition. This is 393 readily solved: 394 395 _a_ ( _t_ ) = _a_ (0)·exp (- _K_ pi - _t_ ) 396 397 With _I_ and X, denoting irradiance (fluence rate) in mol m [-2] s [-1] and photon-exposure (fluence) in mol m [-2], respectively, 398 the solution can be expressed as: 399 400 Y = exp[-( _K_ pi / _I_ ) ·X] 401 402 where Y is the ratio of PSI activity at the end of photoinhibitory treatment to that before the treatment. The ratio of 403 P m after the photoinhibition treatment to that before the treatment can also be used as Y. _K_ pi / _I_ in mol [-1] m [2] is the 404 coefficient of PSI photoinhibition, which is a useful parameter to consider the effect of irradiance and fluence (Table 405 S1; Oguchi et al. 2009; Tyystjärvi and Aro 1996). 406 Light at very high irradiances would give damage to PSI in a manner different from light at moderate irradiances 407 (see § 6; Shimakawa et al. 2024; Sonoike 2025), thus we divided the categories of the FL and intermittent pulses 408 based on whether the irradiance was below or above 10,000 µmol photons m [-2] s [-1] . It has also been suggested that the 409 LL phase of the FL regime affects the extent of PSI photoinhibition, particularly depending on whether this phase is 410 darkness or low light (Kono et al. 2017; Sejima et al. 2014). Therefore, we also divided the category of FL and flashes 411 depending on whether the LL phase was actual LL or darkness. Intermittent HL/pulses denote that the HLs or 412 saturating pulses were given with no background light. We group the experimental conditions into 8 categories. 413 414 1. isolated thylakoids, constant light 415 2. leaves, at chilling temperatures, constant light 416 3. leaves, at moderate temperatures, constant light 417 4. leaves of _A. thaliana pgr5_, at moderate temperatures, constant light 418 5. leaves, FL (< 10,000 µmol photons m [-2] s [-1] ) 419 6. leaves, intermittent HL/pulses (< 10,000 µmol photons m [-2] s [-1] ) 420 7. leaves, FL (≥ 10,000 µmol photons m [-2] s [-1] ) 421 8. leaves, intermittent HL/pulses (≥ 10,000 µmol photons m [-2] s [-1] ) 422 423 In early studies of PSI photoinhibition, isolated thylakoids were usually treated with constant light. Although _K_ pi / _I_ 424 varied widely among these studies, its mean value was higher than that obtained for leaves exposed to constant light 425 at chilling temperatures (1 _vs_ . 2 in Fig. 3a). The _K_ pi / _I_ of the leaves photoinhibited at room temperature in constant 426 light was very low, although not zero, and was significantly lower than that of leaves photoinhibited at chilling 427 temperatures (3 _vs_ . 2). Even in the constant light, _K_ pi / _I_ in the leaves of _pgr5_ (the mutant of PGR5/PGRL1 CEF-PSI) 10 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.02.763831; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.02.763831) preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 428 were higher than WT plants (3 _vs_ . 4, because the number of reports with _pgr5_ mutant is small, the difference is not 429 statistically significant). In these experiments, _pgr5-1_ was used. It should be noted that _pgr5-1_ also carries another 430 mutation, which enhances the phenotype (Wada et al. 2021). 431 When the irradiance in the HL was lower than 10,000 µmol photons m [-2] s [-1], the _K_ pi / _I_ of leaf samples was 432 significantly higher than that of leaf samples photoinhibited in constant light, both in the FL and intermittent HL/pulse 433 regimes (3 vs. 5, and 3 vs. 6). The difference between the FL and intermittent HL/pulse regimes was not clear (5 _vs_ . 434 6). When the irradiance of HL/pulse was 10,000 µmol photons m [-2] s [-1] or higher, the _K_ pi / _I_ of leaves in the FL regime 435 with actual LL was not higher than that of leaves in constant light, whereas leaves in the intermittent HL/pulse showed 436 higher but highly variable _K_ pi / _I_ values (category 3 _vs_ . 7, and category 3 _vs_ . 8). The large variation among the reported 437 values and the limited number of data points in very HL/pulse may have prevented the detection of a statistically 438 significant difference. To see the effect of very high HL/pulse on _K_ pi / _I_ of PSI, increment of the number of future 439 studies under the same conditions would be necessary. We should note that the _K_ pi / _I_ and the extent of PSI 440 photoinhibition can be also affected by differences in leaf traits, including leaf thickness and chlorophyll content per 441 leaf area, among material species and growth environments, as the extent of PSII photoinhibition shows a gradient 442 in a leaf and is influenced by leaf thickness and Chl content per leaf area (Oguchi et al. 2011; Pätsikkä et al. 2002). 443 As shown in the boxplot, _K_ pi / _I_ of PSI varies considerably among studies conducted in the FL or intermittent 444 HL/pulse regime, which may be attributable to differences in the redox state around PSI during the exposure. The 445 variation of redox state around PSI can be caused by different conditions among the photoinhibition experiments and 446 by different materials. As already mentioned, suppression of the electron flow from PSII to PSI and the increment of 447 the proportion of oxidized P700 (P700 [+] ) are important to reduce the extent of PSI photoinhibition (Shimakawa and 448 Miyake 2018; Sonoike 2011; Sonoike and Terashima 1994; Terashima et al. 1994; Trissl 1997), therefore we plotted 449 the relationship between the proportion of P700 [+] to total P700 (Y(ND)) and the _K_ pi / _I_ of PSI, using data from the 450 meta-analysis (Fig. 3b). We only used the data from reports, which measured P700 [+] during the HL phase of 451 fluctuating or intermittent-light conditions. The negative correlation between Y(ND) and the _K_ pi / _I_ of PSI would 452 support the suggestion that keeping higher Y(ND) in fluctuating-light conditions is important to reduce the extent of 453 PSI photoinhibition. 454 In his recent review, Sonoike (2025) claims that, although an increase in [P700 [+] ] or Y(ND) is a necessary 455 condition for protection of PSI, oxidation of the acceptor side, or the decrease in Y(NA), is essential for PSI protection 456 (see §§ 4 and 5). Although the studies in which Y(NA) was measured during the HL phase are few, we plotted such 457 data (Fig. 3c). Although the data points are not many, we may see the trend indicating that _K_ pi / _I_ is inversely related 458 to Y(NA). 459 A recent paper by Tanigawa et al. (2026) is worth mentioning here. They used _A. thaliana_ overexpression lines 460 of PGR5. When they grew these lines in a moderate FL regime, growth of the 2X over-expressor outcompeted WT 461 and the over-expressors with more PGR5, indicating that maintaining an appropriate level of PGR5 is important for 462 efficient photosynthetic regulation in white FL. 463 464 **4. How do plants protect PSI?** 465 Studies in cucumber have revealed how chilling-induced PSI photoinhibition develops. Chilling-tolerant plants, by 466 contrast, retain mechanisms that limit electron flow into PSI and relieve over-reduction of its acceptor side (Kono et 467 al. 2022; Miyake 2020). How plants protect PSI from photoinhibition induced by chilling temperatures, FL and rSPs 468 is discussed below. 469 470 **4.1 Photosynthetic control.** 471 The decreased capacity to regenerate NADP⁺ and the decline in ATP consumption under chilling stress lead to 472 acidification of the thylakoid lumen. This restricts PQH 2 oxidation at the Cyt _b_ 6 _f_ complex and thereby suppresses 473 electron flow into PSI. This ΔpH-dependent restriction of electron flow at the Cyt _b_ 6 _f_ complex is known as the 474 photosynthetic control (see Degen and Johnson 2024 for a recent review). Lumen acidification also promotes NPQ, 11 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.02.763831; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.02.763831) preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 475 further decreasing electron flow into PSI. In cucumber, however, uncoupling of the thylakoid ATP synthase dissipates 476 ΔpH and weakens both photosynthetic control and NPQ. 477 478 **4.2 Suppression of the electron flow to PSI.** 479 PSII has a greater repair capacity than PSI. Under various environmental stresses, PSII photoinhibition occurs first, 480 limiting electron flow from PSII to PSI and protecting PSI (Ivanov et al. 1998; Kim et al. 2005; Lima‐Melo et al. 481 2019; Messant et al. 2024; Napaumpaiporn et al. 2024; Obara et al. 2023; Ozawa et al. 2024; Sonoike 1996b; 482 Takeuchi et al. 2025b; Tikkanen et al. 2014; Zhang Z et al. 2016). As has been discussed in § 1.2, this is not the case 483 in cucumber. PSII photoinhibition appears to be a sacrificial defense, although it is difficult to prove. In the chilling484 tolerant plant _A. thaliana,_ exposure to chilling in the light induced PSII photoinhibition, as indicated by a decrease 485 in F v /F m, whereas P m was maintained. By varying the duration of chilling treatment, Takeuchi et al. (2025b) obtained 486 _Arabidopsis_ plants with F v /F m values ranging from approximately 0.2 to 0.8 and subsequently exposed them to rSPs 487 that induce PSI photoinhibition. Plants with lower initial F v /F m maintained P700 in a more oxidized state during the 488 SPs and showed a slower decline in PSI activity. F v /F m before the pulse treatment was strongly correlated with 489 susceptibility to PSI photoinhibition ( _r_ = 0.97) (Takeuchi et al. 2025b). Thus, PSII photoinhibition protects PSI. Even 490 in the plants in which PSII photoinhibition maintained P700 in an oxidized state, however, Fd became reduced during 491 the SPs. Steady-state oxidation of P700 alone therefore cannot completely prevent acceptor-side reduction during 492 intense excitation. Nevertheless, reoxidation of Fd after each short pulse was faster in plants that had undergone PSII 493 photoinhibition. Faster Fd reoxidation indicates that electrons were removed more efficiently from the PSI acceptor 494 side in plants in which PSII was photoinhibited. By keeping the PQ pool relatively oxidized, PSII photoinhibition 495 may also favor CEF-PSI and thereby contribute to rapid Fd reoxidation (Allen 2003; Satoh et al. 2025). 496 We have dealt with several examples in § 2. However, is susceptibility of PSI to red FL a general phenomenon? 497 PSI photoinhibition was indeed readily induced by FL from red LEDs in most of the plants we tested (Kono et al. 498 2017), whereas _Alocasia odora_ (Lodd.) Spach, a shade-tolerant plant, showed considerable resistance. Among the _A._ 499 _odora_ plants grown at several different irradiances, the plants grown at the lowest irradiance, 10 µmol photons m⁻² 500 s⁻¹, were most resistant to fluctuating-light-induced PSI photoinhibition. Their leaves showed the highest Chl content 501 per leaf area (Chl area ) and P700 area and the lowest Cyt _f_ /Chl and Cyt _f_ /P700. The electron flow rate per P700 was 502 lowest in such leaves (Terashima et al. 2021). On the other hand, concerning PSII photoinhibition per se, it is 503 established that high light-grown leaves are more resistant (Anderson and Aro 1993; Miyata et al. 2012, 2015; Öquist 504 et al. 1987), although there are some exceptions (Öquist et al. 1992). 505 Kono et al. (2025) investigated PSI tolerance in the leaves from transgenic tobacco ( _Nicotiana tabacum_ L _._ ) lines 506 with varying Cyt _b_ 6 _f_ levels and found that a marked reduction in Cyt _b_ 6 _f_ levels conferred substantial PSI 507 photoprotection against the FL‐induced PSI damage, even under extremely high light pulses. In such leaves of low 508 levels of Cyt _b_ 6 _f_, however, the PSII quantum yield, Y(II), was lowest, indicating a trade-off between PSI 509 photoprotection and overall photosynthetic productivity. 510 Taken together, suppression of the electron transport flow to PSI is effective in alleviating PSI photoinhibition. 511 The decrease in the content of Cyt _b_ 6 _f_ per Chl with the decrease in the growth irradiance has been long known 512 (Anderson 1986; Boardman 1977; Hikosaka and Terashima 1995; Schöttler and Tóth 2014). P700/Chl in these reports 513 were rather stable in most of the experiments. In natural environment, the decrease in irradiance in most cases occurs 514 due to vegetation cover. In such cases, the fraction of far-red light increases with decreasing irradiance. The effects 515 of an increased fraction of far-red light on thylakoid composition have been examined in a few studies (Chow et al. 516 1990; Ibrahim et al. 2022; Tullberg et al. 2000). In these studies, however, P700/Chl decreased with far-red light 517 enrichment. Since Cyt _f_ /P700 increased in such plants, the electron flow to PSI would not be suppressed sufficiently. 518 Alternatively, abundant far-red light in natural environment could still protect PSI. More ecophysiology-oriented 519 studies are needed in this research field. 520 521 **4.3 Flavodiiron proteins (FLVs).** 522 Using electrons from Fd, FLV reduces O 2 to water. Conifers use FLV to prevent photoinhibition at low temperatures 12 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.02.763831; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.02.763831) preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 523 in early-spring (Bag et al. 2023), whereas angiosperms have lost FLVs. For FLV-expressing rice plants at low 524 temperature, see § 1.3. Yamamoto et al. (2016) expressed FLV in _A. thaliana_ WT and _pgr5_ mutant. FLV expression 525 significantly increased rates of pseudo-CEF in the mutant plants, particularly at high light intensities, and partially 526 restored the photosynthetic phenotype. In WT plants, FLV did not compete with PGR5-dependent CEF. Wada et al. 527 (2018) expressed FLV in rice single mutant of PGR5, that of NDH, and a double mutant, and exposed these lines to 528 the FL and found that PSI photoinhibition was not detected in any of these lines. 529 530 **5. Roles of NDH and PGR5/PGRL1 in protection from PSI photoinhibition.** 531 As already detailed in § 1, NDH transfers electrons from reduced Fd to the PQ pool rather than passing them to O 2 532 as FLVs do (Richardson et al. 2021; Takeuchi et al. 2025a, 2026). NDH also limits acceptor-side over-reduction. 533 Although PGR5 contributes to proton motive force formation and photosynthetic control, direct evidence that PGR5 534 or PGRL1 prevents chilling-induced PSI photoinhibition is lacking. On the other hand, in the PSI photoinhibition 535 studies using FL and rSPs, the roles of these two pathways have been compared. 536 537 **5.1 Differences between the two CEF-PSI pathways.** 538 Although both NDH and PGR5 transfer electrons from Fd to the PQ pool, they play different roles. Table 1 539 summarizes their properties. 540 **Abundance and velocity.** In _A. thaliana_, the NDH complex is present at 0.02 copies per PSI, compared with 0.07 541 for PGR5 (Ermakova et al. 2024a). Assuming a turnover similar to that of respiratory complex I (approximately 200 542 electrons s⁻¹; Strand et al. 2017), an NDH content of 0.05 per PSI, as reported for tobacco (Strand et al. 2017), would 543 support a cyclic flux of approximately 10 electrons s⁻¹ per PSI, and the lower content of 0.02 per PSI in _A. thaliana_ 544 approximately 4 electrons s⁻¹ per PSI. The direct Fd–Cyt _b_ ₆ _f_ route, reported for _Chlamydomonas reinhardtii_ 545 P.A.Dang., is the fastest way to recycle electrons, and the NDH-dependent pathway is slower by one to two orders of 546 magnitude (Ermakova et al. 2024a; Nawrocki et al. 2019). In _A. thaliana_ in constant light, the mutants with reduced 547 NDH activity showed only an 8–14% reduction in the light-induced pmf, whereas _pgr5_ showed a 30–47% reduction 548 at irradiances ≥ 138 µmol photons m⁻² s⁻¹ (Wang et al. 2015). 549 **Proton stoichiometry.** The chloroplast NDH complex is a homologue of respiratory complex I. Strand et al. (2017) 550 provided evidence for its proton pumping: coupled thylakoids from _Spinacia oleracea_ L. (spinach) synthesized ATP 551 in the dark when Fd, NADPH and decyl-plastoquinone were supplied, and the ATP synthesis was abolished by 552 uncouplers but insensitive to antimycin A. They estimated a stoichiometry of 2 H⁺ per electron transferred from Fd 553 to PQ (2 H⁺/e⁻). When the Q-cycle at the Cyt _b_ 6 _f_ is included, each electron passing through the NDH pathway deposits 554 4 H⁺ in the lumen (4 H⁺/e⁻), whereas each electron passing through the PGR5 pathway deposits 2 H⁺ (2 H⁺/e⁻). In 555 terms of ATP production, the NDH pathway would therefore produce twice as much ATP per electron (Ermakova et 556 al. 2024a; Strand et al. 2017). PGR5/PGRL1 has no known proton-pumping module, and whether these proteins act 557 as electron carriers is still debated (Ermakova et al. 2024a; see Table S2, row 8). 558 **Regulation by ATP.** Both pathways are inhibited by stromal ATP in a rapid and reversible manner (Fisher et al. 559 2019). The inhibition constants differ: 670 µM for NDH and 240 µM for the antimycin A-sensitive PGR5 pathway, 560 both within the physiological range. Because the inhibition of NDH is only partially competitive, NDH retains some 561 activity even when ATP is high, whereas the PGR5 pathway is switched off more completely. These measurements 562 were made in uncoupled thylakoids from spinach, _Amaranthus hybridus_ L. and _A. thaliana_ . The inhibition by ATP is 563 not, however, settled: in an earlier assay of the antimycin A-sensitive Fd–PQ reductase in washed pea thylakoids, 564 ATP and ADP had no effect up to 1 mM, whereas NADPH was inhibitory (Cleland and Bendall 1992). The reason 565 for this difference is not known. This weakens the ATP feedback hypothesis as the explanation for division of their 13 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.02.763831; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.02.763831) preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 566 roles (see Table S2, row 3). 567 **Sensitivity to the proton motive force.** Because NDH pumps protons, it must work against the proton gradient 568 across the thylakoid membrane. The proton motive force, pmf or Δp (in mV), has an electrical component, the 569 membrane potential ΔΨ, and a chemical component, the ΔpH across the thylakoid membrane: **𝑟** **𝑚** - [𝑅] ∆𝑝𝑝= ΔΨ + 2.3 × 10 [3] **𝑟** **𝑚** 570 - [𝑅] ∆𝑝𝑝= ΔΨ + 2.3 × 10 [3] 𝐹𝐹 [ · ΔpH] **𝑟** **𝑚** 571 **𝑟** **𝑚** 572 where _R_, _F_ and _T_ are the gas constant, the Faraday constant and the absolute temperature. At 25 °C (298 K), 2.3 × 573 10³ · ( _RT_ / _F_ ) equals 59 mV. We should also consider Δ _E_ h, the difference in redox potential between the electron 574 acceptor and the electron donor of NDH, that is, between PQ/PQH 2 and Fd ox /Fd red (Δ _E_ h = _E_ h (PQ) − _E_ h (Fd)). The _E_ h 575 of component A is expressed as **𝑟** **𝑚** 𝑟 [𝑜] **𝑟** 𝑟𝑟 **𝑚** 𝐴𝐴 576 h m (A) + 2.3 × 10 [3] - ~~)~~ **𝑟** **𝑚** [ · log (] 577 **𝑟** **𝑚** h m (A) + 2.3 × 10 [3] - [𝑅] 𝑚 𝐸𝐸 (A) = 𝐸𝐸 **𝑟** **𝑚** [𝑅] [𝐴𝐴] [𝑜] 𝑚 𝑟 **𝑟** **𝑚** [ · log (] 𝐴𝐴 **𝑟** **𝑚** 578 where _E_ m (A) is the midpoint potential of A, at which [A ox ] = [A red ], and _m_ is the number of electrons involved in the 579 redox reaction of A. Δ _E_ h is, therefore, large when Fd is highly reduced and/or the PQ pool is oxidized, and small 580 when Fd is oxidized and/or the PQ pool is reduced. From Δ _E_ h in mV, the energy available per mol of electrons is 10⁻³ 581 - Δ _E_ h - _F_ . A pump that moves _n_ protons per electron must spend energy corresponding to _n_ - Δp against the gradient. 582 We therefore call Δ _E_ h − _n_ - Δp the net driving force of NDH. The forward reaction (Fd → PQ) can occur only when 583 the net driving force is positive; when it is zero, the net rate is zero, and when it is negative, the reverse reaction is 584 favored. Strand et al. (2017) used midpoint potentials of −430 mV for Fd and +80 mV for PQ at pH 7.5. With these 585 values and treating PQ as a two-electron couple ( _m_ = 2), Δ _E_ h for 90% reduced Fd and PQ pools is ~ 540 mV (520 586 mV in Strand et al. 2017), so NDH can operate as a 2 H⁺/e⁻ pump against a Δp of 180 mV. However, if Fd equilibrates 587 with NADPH (midpoint −335 mV), or if the PQ pool becomes highly reduced while Δp stays high, the net driving 588 force shrinks and the reaction approaches a standstill (Strand et al. 2017). In thylakoids from spinach and _A. hybridus_, 589 collapsing the pmf with uncouplers doubled both the amplitude and the rate of the NDH-dependent post-illumination 590 fluorescence rise, which is the expected behavior of a pump (Strand et al. 2017). The PGR5 pathway, which pumps 591 no protons of its own, should not be limited in this way. Whether the back pressure of the pmf actually limits NDH 592 _in vivo_, however, is still elusive (see § 9.1 and Table S2, row 1). 593 **Thioredoxins (Trxs).** Trxs receive electrons from reduced Fd through Fd–Trx reductase (FTR) and regulate many 594 chloroplast proteins (for a review, see Yoshida and Hisabori 2023). The m-type Trxs appear to restrain both CEF-PSI 595 pathways. In _A. thaliana_, loss of Trx m4 increased NDH activity, overexpression of Trx m in tobacco suppressed it, 596 and Trx m4 partially inhibited the PGR5-dependent pathway in vitro (Courteille et al. 2013). Okegawa and Motohashi 597 (2020) further showed that Trx m4 forms a disulfide-linked complex with PGRL1 and suppresses the PGR5598 dependent pathway; during photosynthetic induction, the complex temporarily dissociates, and CEF-PSI is activated. 599 By contrast, the x- and y-type Trxs work together with CEF-PSI. In the _trx x_ and _trx x trx y1 trx y2_ mutants of _A._ 600 _thaliana_, the PSI acceptor side was more reduced than in WT, particularly under FL, and PSI photoinhibition was 601 more severe, although NDH activity was unchanged (Okegawa et al. 2023). In the _pgr5 trx x trx y1 trx y2_ quadruple 602 mutant, the acceptor side was even more reduced, and the leaves bleached and died under FL (Okegawa and Sakamoto 603 2026). Other Trxs also affect photosynthesis under FL: Trx m1 and m2 are needed for full photosynthetic activity in 604 the HL phases, and NADPH-dependent Trx reductase C (NTRC) for the dynamic responses of NPQ and of the PQ 605 redox state (Thormählen et al. 2017); the combined loss of Trx f1 and NTRC also makes PSI core proteins more 606 vulnerable to light (Thormählen et al. 2015). Thus, Trxs can both restrain and support CEF-PSI, and how they regulate 607 the two pathways under red and FR light deserves closer examination. 14 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.02.763831; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.02.763831) preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 608 **Interaction with Fd.** Both pathways accept electrons from reduced Fd, so the way each binds Fd could, in principle, 609 decide which pathway is used. For NDH, the binding site is known. The subunit NdhS has an SH3-like fold, and its 610 conserved Arg-193 is required for the high-affinity, electrostatic binding of Fd: in ruptured _A. thaliana_ chloroplasts, 611 NDH-dependent PQ reduction was almost saturated at 3 µM Fd in WT, whereas in NdhS-deficient mutants it 612 remained low even at 20 µM Fd (Yamamoto and Shikanai 2013; Yamamoto et al. 2011). For the PGR5-dependent 613 pathway, the site is less clear. PGRL1 interacts with Fd, PGR5, PSI, the Cyt _b_ 6 _f_ complex and FNR (DalCorso et al. 614 2008), and recombinant PGRL1 reduces quinones with electrons from Fd in an antimycin A-sensitive manner (Hertle 615 et al. 2013). Whether PGRL1 is the Fd–PQ reductase itself is still debated (Maekawa et al. 2024; see Table S2, row 616 8). Measured in the same system, the two pathways differ little in their affinity for Fd: the apparent _K_ m was 617 approximately 2 µM for NDH, measured in _A. hybridus_ with antimycin A, and 1 µM in spinach, in which most of 618 the activity is antimycin A-sensitive (Fisher et al. 2019); the value for NDH is less certain, because the rate was not 619 saturated at the highest Fd concentration tested (approximately 5.5 µM). Older assays of the antimycin A-sensitive 620 pathway required more Fd, with an optimum of about 10 µM in air (Arnon and Chain 1975) and saturation at about 621 75 µM reduced Fd under anaerobic conditions (Cleland and Bendall 1992). These assays did not measure NDH and 622 used other species and conditions, so their values cannot be compared directly with those of Fisher et al. (2019). All 623 these values are similar to those of the other Fd-dependent enzymes of the stroma: the dissociation constant for Fd is 624 about 1–10 µM for FNR at physiological ionic strength, about 1 µM for nitrite reductase and about 10 µM for Fd625 dependent glutamate synthase (Knaff and Hirasawa 1991; see Text S1). The two pathways therefore compete with 626 these enzymes for reduced Fd on roughly equal terms, and FNR, which is present at about half the amount of Fd 627 (Böhme 1978; Terashima and Inoue 1985), is the largest competitor. Total Fd is unlikely to limit CEF-PSI, because 628 chloroplasts contain roughly one to six Fd per PSI, or 3–6 mmol Fd per mol Chl (Antal et al. 2013; Böhme 1978; 629 Terashima and Inoue 1985), which, with a chlorophyll concentration of 30–60 mM in the chloroplast (Antal et al. 630 2013), corresponds to about 0.1–0.4 mM (see Text S1), more than ten times these _K_ m values. What can limit CEF631 PSI is the fraction of Fd that is reduced, and this is set by the balance between the supply of electrons from PSI and 632 their use by FNR and the other enzymes. Tagawa et al. (1963) already observed in isolated chloroplasts that cyclic 633 photophosphorylation began when NADP⁺ ran out, and in leaves, Fd-dependent CEF-PSI increases with the reduction 634 level of Fd and appears when linear electron flow is saturated (Ohnishi et al. 2023; Satoh et al. 2025). Here, kinetics 635 and thermodynamics should be kept apart: the rate of each pathway depends on the concentration of reduced Fd 636 (Cleland and Bendall 1992), whereas Δ _E_ h in the condition Δ _E_ h - _n_ - Δp depends on the ratio of reduced to oxidized 637 Fd. Under pure FR, PSI receives few electrons from PSII: the oxidized PQ pool favors the forward reaction of NDH, 638 but little Fd is reduced, which is why CEF-PSI under FR needs a small supply of electrons from PSII (Arnon and 639 Chain 1975; Joët et al. 2002; see Table S2, row 2). Because both pathways draw on the same pool of reduced Fd with 640 similar affinity, affinity alone cannot explain why one pathway is favored under one light quality and the other under 641 another; the explanation must be sought in the pmf, ATP and the redox state of the chain (see §§ 5.2–5.4). 642 **Physical organization.** NDH forms a supercomplex with PSI through the minor antenna proteins Lhca5 and Lhca6; 643 the stromal loop of Lhca6 is required for binding to NDH (Otani et al. 2017). The supercomplex stabilizes NDH. In 644 chilling-sensitive cucumber, the PSI–NDH supercomplex is partially disassembled and NDH activity is lost within 1 645 h of chilling in the light, and this loss accompanies PSI photoinhibition (Takeuchi et al. 2026; see § 1.3). However, 646 in the liverwort _M. polymorpha_, the supercomplex is not required for NDH activity: NDH functions without Lhca5 647 and Lhca6 (Ueda et al. 2012). Cyanobacterial NDH-1 has two Fd-binding site candidates (Laughlin et al. 2019), and 648 the supercomplex may place NDH close to the Fd reduced by PSI; however, as noted above, both pathways draw on 649 the same pool of reduced Fd, so physical proximity alone cannot explain the division of roles between light qualities. 650 **Proton conductivity of the ATP synthase.** It has been argued that PGR5 may lower the proton conductivity of the 651 ATP synthase, _g_ H⁺ (Rantala et al. 2020; Suorsa et al. 2012; Wang et al. 2015), whereas _g_ H⁺ is unaffected in NDH 652 mutants (Wang et al. 2015; see Table 1). 653 654 **5.2 Red FL: How does the PGR5 pathway protect PSI?** 15 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.02.763831; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.02.763831) preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 655 In red FL, PSI is protected by a sequence of events (Kono and Terashima 2016). When the irradiance is increased 656 from 30 to 1,200 µmol photons m⁻² s⁻¹, electrons gush into PSI at a rate faster than the rate of electron consumption 657 by CBB cycle, photorespiration and other sinks. Thus, the acceptor side of PSI becomes reduced (high Y(NA)). CEF658 PSI removes a part of the excess electrons and transfers them to the PQ pool. The protons deposited in the lumen by 659 CEF-PSI build up ΔpH, which suppresses the oxidation of PQH 2 at the Cyt _b_ 6 _f_ complex (photosynthetic control), and 660 the flow of electrons into PSI is suppressed. As a result, the acceptor side is relieved and P700 is oxidized: Y(NA) 661 decreases and Y(ND) increases. Kono and Terashima (2016) argued that the relief of the acceptor side must come 662 first: because Y(I) + Y(ND) + Y(NA) = 1, Y(ND) can become large only when Y(NA) has become small. This 663 depends on PGR5. When the PGR5 pathway was blocked, either genetically in _pgr5_ or pharmacologically with 664 antimycin A, Y(NA) rose step by step during the FL treatment, and the large Y(ND) of the HL phase never developed. 665 P m of _pgr5_ and of antimycin A-treated WT leaves fell to 20% and 40% of the initial value, respectively, compared 666 with 70% in untreated WT (Kono and Terashima 2016). The large Y(ND) in the HL phase was observed only when 667 the PGR5 pathway was functional, either in WT or in _crr2-2_ . 668 Two terms describe the restriction of electron flow at the Cyt _b_ 6 _f_ complex. "Photosynthetic control" is an 669 old concept (Rumberg and Siggel 1969; Tikhonov et al. 1981; West and Wiskich 1968): when the lumen becomes 670 acidic, protonation of the histidine ligand of the Rieske Fe–S protein, whose p _K_ a lies between about 6 and 6.5 in the 671 oxidized state, slows the oxidation of PQH 2 at the Q o site (Tikhonov 2024; Vilyanen and Kozuleva 2025). "Reduction672 induced suppression of electron flow" (RISE) describes the similar bottleneck from the PQ side: a highly reduced PQ 673 pool limits the turnover of the Q-cycle even without a change in lumen pH (Miyake 2020; Shimakawa and Miyake 674 2018). Degen and Johnson (2024) treat the two as a pH input and a redox input on the same step. In leaves, the two 675 inputs usually occur together, because the conditions that acidify the lumen also reduce the PQ pool. They differ, 676 however, in what they require: the pH input requires a ΔpH, whereas the redox input requires a reduced PQ pool, a 677 difference that becomes important under FR (see § 5.3). An increase in Y(ND) alone does not show that electron flow 678 is restricted at the Cyt _b_ 6 _f_ complex by either input; the rate constant of P700⁺ re-reduction is a more direct indicator 679 (Degen and Johnson 2024). 680 The _pgr5_ phenotype also shows that the mutant fails to build and hold the pmf. In white FL, _pgr5_ leaves generate 681 a normal pmf during LL, but the pmf collapses almost completely at the transition to HL, whereas in WT, it increases 682 sharply (Suorsa et al. 2012). The consequence is that P700 stays reduced throughout the HL phase in _pgr5_ (Suorsa et 683 al. 2012), and the acceptor side is also reduced. The fact that the damage originates on the acceptor side is shown by 684 two observations: methyl viologen, an artificial electron acceptor, restores P700 oxidation in _pgr5_ in high light 685 (Munekage et al. 2002), and lowering the O₂ concentration to 2.7% or 0% during the red FL enhanced CEF-PSI in 686 _pgr5_, released the acceptor-side limitation, and abolished PSI photoinhibition (Kono et al. 2014). In red FL, the 687 protective effect of CEF-PSI comes from relieving the acceptor side and building the pmf, and the pathway that can 688 do this fast enough is the PGR5 pathway. 689 Why is this role assigned to PGR5 instead of NDH? In the HL phase, a large pmf must be built quickly from a 690 large electron flux. This exactly fits the characteristics of the PGR5 pathway (Table 1). It can also be understood why 691 NDH is largely dispensable in red FL, at least in _A. thaliana_ . NDH can recycle only a few percent of the maximum 692 linear flux (Table 1), which is negligible in the HL phase. When the linear electron flow is fast, the PQ pool becomes 693 strongly reduced and the pmf is high. Both changes diminish the net driving force of NDH (Δ _E_ h − _n_ - Δp; Table 1) and 694 push NDH toward a standstill. 695 However, this conclusion should not be generalized. Photosynthetic responses of the rice NDH mutant _crr6_, 696 grown at 28 °C, to HL in FL (1,500/200 | 10 min/10 min) were almost identical to those of WT, whereas the mutant 697 phenotype appeared in the LL phase, where NPQ was lower and the PQ pool was more reduced than in WT (Yamori 698 et al. 2016). NDH plays a larger physiological role in rice (Yamori et al. 2011, 2016) than in _A. thaliana_ or tobacco, 699 whose NDH mutants grow normally and show little change in NPQ or electron transport, even at low irradiances 700 (Munekage et al. 2004; Shikanai et al. 1998). 701 702 **5.3 Far-red light: Why does the NDH pathway take over?** 16 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.02.763831; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.02.763831) preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 703 **What does FR change?** Adding far-red light (FR) to the FL changed the state of PSI in both phases (Kono et al. 704 2017). In the LL phase, 10 s after the 100th HL pulse, Y(ND) rose from 0.006 without FR to 0.389 with FR, and 705 Y(NA) fell from 0.432 to 0.224. In the HL phase, Y(ND) rose from 0.253 to 0.528 and Y(NA) fell from 0.428 to 706 0.232. FR thus both increased P700 oxidation and decreased acceptor-side limitation. The high Y(ND) of the LL 707 phase cannot be explained by ΔpH-dependent photosynthetic control, because at the irradiance of 30 µmol photons 708 m⁻² s⁻¹ the pmf is small, and in an FL regime with much longer phases the total pmf measured by the electrochromic 709 shift was not clearly changed by FR (Kono et al. 2020). Instead, FR excites PSI selectively, so that P700 is oxidized 710 faster than PSII can supply electrons to re-reduce it. This corresponds to the "linear mode" of Joliot and Johnson 711 (2011): under FR, the electrons released by P700 are passed on to acceptors and few return, so that P700 is oxidized 712 rapidly and the PQ pool stays largely oxidized. It follows that the redox input of photosynthetic control (see § 5.2) is 713 not engaged in FR; only the pH input can act, but this is also weak in the LL phase. Consistent with this, when FR is 714 given alone after darkness, the PGR5-dependent control of P700 remains in its inactive dark state (Rantala et al. 715 2020). In short, FR changes the main cause of P700 oxidation. In red FL, P700 is oxidized mainly by the pmf built 716 by CEF-PSI; with FR, it is oxidized by the balance of excitation between the two photosystems. This is the basis of 717 what may be called the masking hypothesis. In the sequence mentioned in § 5.2, the task of the PGR5 pathway was 718 to relieve the acceptor side and build the pmf that oxidizes P700. Addition of FR reaches the same two endpoints, 719 low Y(NA) and high Y(ND), without going through CEF-PSI. The specific contribution of PGR5, a fast, pmf-driven 720 oxidation of P700, is therefore no longer required, and _pgr5_ recovers to 92% of P m when FR is present (Kono et al. 721 2017). This does not mean that FR acts "before" or "instead of" CEF-PSI. Every oxidation of P700 releases one 722 electron to Fd, and these electrons must be recycled; P700 oxidation and CEF-PSI run at the same time and support 723 each other. What FR removes is the need for the fast and large pmf-generating step. Whether P700⁺ is protective itself 724 has been questioned (Rantala et al. 2020; Sonoike 2025); we regard high Y(ND) as protective only when it is 725 accompanied by an oxidized acceptor side, a condition that FR meets (see Table S2, row 5). 726 **Why NDH? A working hypothesis.** If the pmf-generating step no longer must be fast, why does the protection by 727 FR still require a CEF-PSI pathway, and why NDH in particular? Our working hypothesis, which may be called the 728 redox–Δp gating, is that the redox state of the electron transport chain and the size of the pmf gate the two pathways 729 differently. A proton pump such as NDH can run forward only while the redox driving force from Fd to PQ exceeds 730 the back-pressure of the proton gradient, that is, while the net driving force (Δ _E_ h − _n_ - Δp) is positive (Strand et al. 731 2017; see § 5.1). In FR, the PQ pool is oxidized, so Δ _E_ h is large, and the pmf is moderate, so _n_ - Δp is small. Thus, the 732 forward reaction of NDH is thermodynamically favored. Defining this hypothesis by the net driving force, rather 733 than by Δp alone, makes it consistent with the observation that NDH-dependent CEF-PSI operates in the H 2 O 2 734 accumulating line GO5 despite the pmf about twice that of WT, because in GO5 the stroma is strongly reduced and 735 Δ _E_ h is correspondingly large (Strand et al. 2015). That NDH supplies electrons to the PQ pool under FR is indicated 736 by leaves of _A. thaliana_ : when actinic light was turned off with FR kept on, FR re-oxidized P700 more slowly in WT 737 than in plants with impaired NDH activity, as in _ndhB_ -disrupted tobacco (Shikanai et al. 1998, 2025). In tobacco 738 leaves under FR, cyclic PSI activity is weak in air, because the intersystem chain is too oxidized; it becomes fast 739 when the PQ and NADP pools are partly reduced under anaerobiosis, and the fast cycling requires NDH (Joët et al. 740 2002). This observation indicates two important points. First, in FR, NDH is the pathway that carries out electron 741 cycling. Second, the rate of cycling depends on an adequate supply of reduced stromal donors, and not on the light. 742 The redox–Δp gating determines whether the forward reaction of NDH is possible, but how fast it runs also depends 743 on how much reduced Fd and NADPH is available. In our FR-supplemented FL (or rSPs), the red light of the LL 744 phase continues to drive PSII (Kono et al. 2017). In an FL regime with much longer phases (800/30 | 10 min/15 min; 745 Kono et al. 2020), adding FR to low red light oxidized the PQ pool while increasing the PSII yield. FR thus keeps 746 the chain on the oxidized side without cutting off the supply of electrons from PSII, the condition NDH needs. 747 **High proton efficiency at low flux.** In the presence of FR, the role of CEF-PSI is not to drive a large electron flux 748 but to provide a safe return route for the electrons released by P700 oxidation, together with just enough pmf to keep 749 the Cyt _b_ ₆ _f_ complex under minimal photosynthetic control. NDH is suited to this role because its H⁺/e⁻ ratio, including 750 the Q-cycle, is 4, whereas that of the PGR5 pathway is 2 (Table 1). Therefore, a small flux through NDH has a 751 disproportionate effect on the pmf. A natural objection is that proton pumping activity should show up in NPQ. 17 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.02.763831; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.02.763831) preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 752 However, this is not observed in _A. thaliana_ under steady red light: FR changed the PSI parameters below 500 µmol 753 photons m⁻² s⁻¹, but left Y(II), Y(NO) and Y(NPQ) unchanged (Kono et al. 2017). The total proton input through 754 NDH is small, which would be enough to support the small photosynthetic control needed for P700 oxidation but not 755 enough to raise NPQ. In addition, FR increases the proton conductivity and proton flux through the ATP synthase 756 (Kono et al. 2020), so that additional proton input can be used for ATP synthesis rather than stored as pmf. This 757 applies to _A. thaliana_ in steady light and to tobacco, in which the disruption of _ndhB_ did not change NPQ at 20–2,000 758 µmol photons m⁻² s⁻¹ (Shikanai et al. 1998); in rice, whose NDH content is higher, _crr6_ shows a clearly lower NPQ 759 in low light (Yamori et al. 2011, 2016). The same reasoning explains the asymmetry seen with FR: _pgr5_ still has the 760 pathway that the FR regime needs and recovers to the WT level, whereas P m in _crr2-2_ and _crr4-3_, which lack it, 761 remained at 80% of the initial level even with FR (Kono et al. 2017). 762 **The operating window of NDH.** The FL (or rSPs) regime of Kono et al. (2017) spends 92.6% of its time in the LL 763 phase (1,200/30 | 800 ms/10 s). Low, steady electron cycling is the operating domain of NDH; Joliot and Johnson 764 (2011) noted that, although the NDH content is far too low to sustain the cyclic rates required under strong light, 765 NDH can carry a significant fraction of CEF-PSI under low light. Because NDH pumps protons, its own activity 766 raises the pmf, until _n_ - Δp approaches Δ _E_ h and the forward reaction stops (Table 1). The upper limit of NDH operation 767 is therefore set by the pmf. In the LL phase of FR-supplemented FL, the pmf is moderate and stays below this limit, 768 whereas in the HL phase of red FL it is high enough to stop NDH. The lower limit is set not by the pmf but by the 769 supply of electrons: the PQ pool must be oxidized enough to keep Δ _E_ h large, but enough reduced Fd and NADPH 770 must remain to supply the reaction, because cycling almost stops when the electron transport chain is too oxidized 771 (Joët et al. 2002). Similarly, in isolated chloroplasts, cyclic photophosphorylation under 715-nm light required a small 772 supply of electrons from PSII and stopped when PSII was blocked by DCMU (Arnon and Chain 1975). In addition, 773 in the LL phase the stromal ATP level falls while the demand for ATP persists, and a moderate fall in ATP activates 774 NDH first (Fisher et al. 2019); ATP feedback would thus favor NDH under the same conditions, although the ATP 775 dynamics under FR have not been measured. 776 **Maintaining redox poise across both phases.** FR added to the FL protected PSI fully only when it was present in 777 both phases of the FL; FR in the HL phase alone or in the LL phase alone gave only partial protection (Kono et al. 778 2017). This can be explained as follows. In the LL phase, FR keeps P700 and the donor-side carriers in an oxidized 779 state. These carriers (plastocyanin, the Cyt _b_ 6 _f_ complex and the PQ pool) together hold a few tens of electron 780 equivalents per P700. This oxidized pool is small, but it prevents P700 from becoming fully reduced at the onset of 781 the next HL pulse. During the HL pulse, FR keeps P700 oxidized while CEF-PSI returns the released electrons. 782 Therefore, the protective state is not a stock of P700⁺ built in advance, but a continuously maintained redox poise of 783 the whole chain. If ΔpH, photosynthetic control and P700 oxidation are relaxed in the LL phase, the chain must 784 rebuild them in the next HL pulse, and the acceptor side is transiently over-reduced during this "cold start". FR 785 prevents this cold start. This also resolves an apparent paradox in the NDH literature: the loss of NDH makes the PQ 786 pool more reduced under low light (Ueda et al. 2012; Yamori et al. 2011, 2016), although NDH reduces PQ. If the 787 pmf and ATP generated by NDH sustain the CBB cycle and minimal photosynthetic control when the light is low, 788 then NDH acts as an enzyme that keeps the intersystem carriers poised, not merely as a PQ reductase (Yamori et al. 789 2011). Under FR, this poising function is what the plant needs, and the NDH pathway provides it. 790 791 **5.4 Summary and outlook: light quality sets the redox poise and the pmf, and these gate the two pathways.** 792 **Two regimes, one rule.** Table 2 places the two regimes side by side. In the high-light phase of red FL, a reduced PQ 793 pool and a high pmf drive the net driving force of NDH (Δ _E_ h − _n_ - Δp) close to zero, and P700 is oxidized by the pmf 794 that the PGR5-dependent pathway builds within seconds (Fig. 4a, c). Under FR-supplemented FL, an oxidized PQ 795 pool and a moderate pmf make the net driving force large, and P700 is oxidized by the excitation balance, so that 796 NDH can run forward with high proton efficiency (Fig. 4b, c). One rule covers both regimes: light quality sets the 797 redox state of the PQ pool and the pmf, and these two variables decide which pathway can run. It explains the result 798 of Kono and Terashima (2016), its reversal in Kono et al. (2017), and why the reversal stayed hidden until FR was 18 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.02.763831; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.02.763831) preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 799 added: red LED light produces the one regime in which NDH has the least to contribute. 800 **Two separate questions.** Two questions have often been conflated in the literature (Fig. 4d). The first is by what 801 means P700 is kept oxidized, and three means are now documented. One is the pmf-driven restriction at the Cyt _b_ 6 _f_ 802 complex, built mainly by the PGR5-dependent pathway under red high light (see § 5.2). Another is a restriction of 803 the electron flux into PSI from upstream, either by a low content of the Cyt _b_ 6 _f_ complex (Kono et al. 2025) or by PSII 804 photoinhibition, which can itself protect PSI (Sonoike 2025): in _A. thaliana_ leaves whose F v /F m had been lowered by 805 prior chilling, the residual PSI after rSPs increased as F v /F m decreased, independently of ΔpH (Takeuchi et al. 2025b). 806 Both protect PSI at the cost of photosynthetic capacity. The third is the excitation balance imposed by FR, which 807 oxidizes P700 without a high pmf and, in the low-light phase of FL, raises, instead of lowers, CO 2 assimilation (Kono 808 et al. 2020). The second question is how the pmf and ATP are supplied when linear flow is low. Here, the two pathways 809 share the task according to their capacity and efficiency: a large flux at two protons per electron through the PGR5810 dependent pathway, or a small flux at four through NDH (Table 1). Seen this way, FR and a low Cyt _b_ 6 _f_ content act 811 at the same point: both remove the need for a fast, pmf-driven oxidation of P700 and leave the plant with the second 812 question, which the NDH pathway answers efficiently at low flux. 813 **A general rule for low light.** If the redox–Δp gating is correct, NDH should matter whenever the light is low enough 814 to keep the PQ pool oxidized and the pmf moderate, whether or not FR is present. This is what the NDH mutants 815 show: the Burren _ndhG_ variant of _Arabidopsis_ affects PSII efficiency only below approximately 350 µmol photons 816 m⁻² s⁻¹ (Theeuwen et al. 2022, preprint), the loss of NDH in _M. polymorpha_ increases PQ reduction only < 100 µmol 817 photons m⁻² s⁻¹ (Ueda et al. 2012), and the rice _crr6_ phenotype appears below the growth irradiance and in the low818 light phase of FL (Yamori et al. 2011, 2016). In each case—paradoxically, since NDH reduces PQ—the intersystem 819 chain is more reduced without NDH when the light is low, a paradox that the poising function of NDH resolves (§ 820 5.3). FR is the extreme case of this condition, a light that excites PSI while driving little electron flow from PSII. 821 Seen in this way, the "low light versus high light" rule in § 2 and the "red versus far-red" reversal in § 5.3 are two 822 expressions of the same gating. The rule is therefore about redox poise and pmf, not about irradiance as such, and C₄ 823 plants illustrate the difference. In the bundle-sheath cells of NADP-malic enzyme species, which operate in full 824 sunlight, the chloroplasts contain little PSII, the stroma receives NADPH from malate decarboxylation, and the CBB 825 cycle needs ATP that linear flow cannot supply; it is in these cells that NDH accumulates to its highest levels 826 (Takabayashi et al. 2005). What they share with an FR-illuminated leaf is not low light but a chain fed by little PSII 827 and a reduced stroma, which should keep Δ _E_ h large enough for NDH to run even against a substantial pmf, as in GO5 828 (see Table S2, row 1), together with a demand for ATP that only a proton-efficient cycle can meet. The high NDH 829 content serves the CO₂-concentrating mechanism rather than PSI protection (Takabayashi et al. 2005), but the 830 condition that selects NDH is the same as under FR. 831 832 **6. ROS: their generation and scavenging.** 833 **6.1 Routes of ROS generation.** 834 Electrons are transferred within PSI from the P700 to Fd via A 0, phylloquinone A 1, and the Fe–S clusters F X, F A, and 835 F B . Because these electron carriers have highly negative midpoint potentials (A 0A /A 0B, approximately −1200 mV; 836 PhQ A 1A /A 1B, −800 to −600 mV; F X /F A /F B, −600 to −500 mV), their reduced forms can transfer electrons to O 2 to 837 generate the superoxide radical anion (O 2 [•] ⁻). For comparison, the reduction potential of the O 2 /O 2 [•] ⁻ couple is 838 approximately −160 mV in water, while −600 to −500 mV in N,N-dimethylformamide (DMF) (Cherepanov et al. 839 2017; Degen and Johnson 2024; Kozuleva and Ivanov 2010; Ptushenko et al. 2008; Wardman 1990). O 2 [•] ⁻ is converted 840 to H 2 O 2 through non-enzymatic dismutation or through dismutation catalyzed by chloroplast superoxide dismutase 841 (SOD) localized near PSI on the stromal surface of the thylakoid membrane (Asada 2006; Ogawa et al. 1995). H 2 O 2 842 can generate •OH through Fenton reactions involving the reduced PSI Fe–S clusters or free Fe [2+], further amplifying 843 PSI damage (Šnyrychová et al. 2006; Sonoike 2011). In addition, O 2 [•] ⁻ is known to react directly with solvent-exposed 19 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.02.763831; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.02.763831) preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 844 [4Fe–4S] clusters, inducing their breakdown through cluster oxidation and release of Fe (Bulteau et al. 2003; Flint et 845 al. 1993; Gardner 1997; Petronek et al. 2021). The Fe–S clusters of PSI may therefore be direct targets of O 2 [•] ⁻. 846 Severe over-reduction of the PSI acceptor side provides an additional pathway for ROS production. Charge 847 recombination of the P700⁺/A 0 ⁻ or P700⁺/A 1 ⁻ radical pair can form triplet P700* ( [3] P700*), from which energy transfer 848 to O 2 produces singlet oxygen ( [1] O 2 ) (Kim et al. 2001; Rajagopal et al. 2005; Rutherford et al. 2012; Shimakawa et 849 al. 2024; Shuvalov et al. 1986). Not all charge recombination leads to [3] P700* formation, however; an alternative 850 pathway returns the charge-separated state directly to the ground state without forming [3] P700* (see § 6.3). ROS can 851 therefore be generated in PSI by at least two pathways: O 2 [•] ⁻ generation through electron transfer to O 2 and [1] O 2 852 generation through charge recombination (Sonoike 2025). O 2 [•] ⁻ is less reactive than •OH and [1] O 2 and therefore does 853 not rapidly attack most proteins or lipids. By contrast, [1] O 2 and •OH are highly reactive but short-lived, so their effects 854 are largely confined to pigments and proteins close to their sites of formation. 855 856 **6.2 Sites of O** **2** **[•]** **⁻ generation within PSI.** 857 Which PSI electron carrier reduces O 2 during chilling-induced PSI photoinhibition has long been debated. Chilling858 induced PSI photoinhibition was barely detectable in darkness or under anoxic conditions, directly implying 859 involvement of ROS in the damage. The early loss of Fe–S cluster activity in cucumber chilled in the light initially 860 pointed to these clusters as the principal sites of both ROS production and damage (Shimakawa et al. 2024; Sonoike 861 et al. 1995). However, the site where damage is first detected is not necessarily the site where ROS are generated. 862 O 2 [•] ⁻ reacts weakly with most biomolecules but selectively attacks Fe–S clusters (Flint et al. 1993; Furutani et al. 863 2023). O 2 [•] ⁻ generated closer to the PSI core, for example at A 1, could therefore damage the Fe–S clusters even if the 864 clusters themselves are not its site of production. Because A 0 and A 1 have more negative redox potentials than F A /F B, 865 electron transfer from their reduced forms to O 2 is thermodynamically more favorable. Under normal electron 866 transport conditions, however, the reduced states of A 0 and A 1 are short-lived, making electron accumulation at these 867 cofactors unlikely. By contrast, the reduced states of F A /F B are considerably longer-lived. The site of O 2 reduction 868 within PSI is therefore determined not only by the redox potential of each carrier but also by the lifetime of its reduced 869 state and its accessibility to O 2 (Khorobrykh et al. 2020). 870 Although A 1 lies within the PSI complex, structural analyses have revealed internal protein cavities near A 1A 871 and A 1B that connect to the external aqueous phase, providing possible routes for O 2 to reach these sites (Cherepanov 872 et al. 2017). Access to A 1 alone, however, does not determine the dominant site of O 2 reduction, because the relative 873 contributions of the PSI cofactors vary with light intensity. Kozuleva and colleagues isolated PSI from 874 _Chlamydomonas_ grown under approximately 70 µmol photons m⁻ [2] s⁻ [1] and measured O 2 photoreduction under 50– 875 2,200 µmol photons m⁻ [2] s⁻ [1] . At irradiances of 300 µmol photons m⁻ [2] s⁻ [1] or higher, phylloquinone A 1 was the major 876 site of O 2 reduction, whereas F A /F B made a greater contribution at lower irradiances (Kozuleva 2022; Kozuleva and 877 Ivanov 2023; Kozuleva et al. 2021). Because the reduced states of F A /F B persist longer than those of A 1 and F X, 878 electron transfer from F A /F B to O 2 is favored at lower irradiances. As irradiance increases, repeated excitation supplies 879 electrons to the PSI acceptor chain faster than they can be removed from the terminal clusters, allowing reduced F X 880 and eventually reduced A 1 to accumulate. A 1 may therefore make a greater contribution to O 2 reduction at higher 881 irradiances. In intact leaves, however, photosynthetic control normally restricts electron flow into PSI, limiting the 882 sustained accumulation of reduced F X and A 1 . O 2 reduction at A 1 is therefore expected to become appreciable only 883 when photosynthetic control is impaired or overwhelmed and the PSI acceptor side becomes severely over-reduced 884 (Miyake 2020). 885 Fd is also the site of O 2 [•] ⁻ formation (Badger 1985). Although Fd could become a victim of O 2 [•] ⁻, O 2 [•] ⁻ formed 886 at Fd in the stroma would not attack PSI Fe–S centers. Similarly, addition of methyl viologen (MV), which is an 887 electron acceptor from PSI and produces O 2 [•] ⁻ through autoxidation upon reduction, suppressed PSI photoinhibition 888 induced by constant light in isolated thylakoid membranes (Sonoike 1996a) or by rSPs in isolated chloroplasts 889 (Takagi et al. 2016). 890 891 **6.3** **[1]** **O** **2** **generation through charge recombination.** 20 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.02.763831; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.02.763831) preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 892 Among the possible fates of electrons accumulated on the PSI acceptor side are electron transfer to O 2 and charge 893 recombination with P700⁺. The latter can return electrons safely to P700⁺, but some recombination pathways generate 894 3 P700* and thereby promote the formation of 1 O 2, potentially damaging PSI. Whether charge recombination protects 895 or damages PSI may depend on the pathway it follows. Formation of [3] P700* from P700⁺A 1 ⁻ is thought to involve re896 formation of the primary P700⁺A 0 ⁻ radical pair, followed by population of the triplet state of P700*. Alternatively, 897 the P700⁺A 1 ⁻ can return to the ground state through a recombination pathway without forming [3] P700*. Rutherford et 898 al. (2012) proposed that these contrasting outcomes reflect the different energetics of A 1A and A 1B . In their model, a 899 back-reaction from F X ⁻ or another downstream acceptor may avoid [3] P700* formation by proceeding through A 1A, 900 whereas recombination through the B-branch may be more likely to re-form the primary P700⁺A 0B ⁻ radical pair and 901 produce [3] P700* (Bindra et al. 2024; Cherepanov et al. 2017; Degen and Johnson 2024; Ishikita and Knapp 2003; 902 Rutherford et al. 2012). 903 Evidence for a protective role of charge recombination comes from high-light experiments with intact leaves. 904 Tiwari and colleagues showed that continuous high-light exposure caused 10–35% photodamage to F A /F B in wild905 type _A. thaliana_, while having little effect on F X function or P700 oxidation capacity (Tiwari et al. 2024). They 906 proposed that, following partial inactivation of F A /F B, electrons accumulating at F X can recombine safely with P700⁺ 907 if photosynthetic control maintains a sufficient population of P700⁺. In the _pgr5_ mutant, by contrast, impaired 908 photosynthetic control limited the accumulation of P700⁺. Reduction then spread to F X and was accompanied by 909 damage to this cluster, resulting in severe PSI damage (Degen et al. 2026; Tiwari et al. 2024). Thus, charge 910 recombination between F X ⁻ and P700⁺ may limit the spread of damage from F A /F B to F X and the PSI core. 911 A different pattern of photo-oxidative damage to PSI is observed in rSPs. Shimakawa et al. (2024) found that 912 such pulses damaged the PSI core near P700 more severely than the Fe–S clusters. Under these conditions, P700⁺ 913 and reduced A 0 /A 1 may coexist more frequently, increasing the likelihood of charge recombination that produces 914 3 P700* and, consequently, 1 O 2 . The short lifetime and diffusion distance of 1 O 2 could then account for the preferential 915 damage near the PSI reaction center rather than the Fe–S clusters. It remains unclear, however, why recombination 916 pathways that form [3] P700* would be favored over pathways that return the charge-separated state safely to the ground 917 state during rSPs, or to what extent the B-branch is directly involved in this process. 918 A major defense against triplet-mediated [1] O 2 formation in photosynthetic organisms is triplet–triplet energy 919 transfer from triplet chlorophyll to carotenoids, which quenches the chlorophyll triplet before it can sensitize O 2 920 (Cazzaniga et al. 2012; Qin et al. 2025; Shi et al. 2026). A comparable process has been demonstrated in the reaction921 center complex of green sulfur bacteria, where triplet energy transfer from core-antenna BChl _a_ to carotenoids 922 produces triplet carotenoids ( [3] Car) (Azai et al. 2020; Inagaki et al. 2026). Direct experimental evidence for such 923 quenching of [3] P700* by a carotenoid in land-plant PSI is, however, still lacking. 924 925 **6.4 Summary of ROS generation in PSI under different environmental conditions.** 926 The mechanisms of ROS generation in PSI are summarized in Fig. 5. Under continuous high light, photosynthetic 927 control can maintain P700⁺. Thus, even when a fraction of the F A /F B clusters is damaged, electrons accumulated at 928 F X may be safely removed by charge recombination with P700⁺ through the A-branch. Although this reaction does 929 not restore forward electron transfer to Fd, it may prevent damage from spreading to F X and the PSI core (Tiwari et 930 al. 2024). When photosynthetic control does not function adequately, as in _pgr5_, P700⁺ cannot be maintained, and 931 over-reduction and damage spread to F X (Fig. 5a), leading to severe PSI photoinhibition (Mazor et al. 2017; Tiwari 932 et al. 2016, 2024). 933 Under rSPs, and possibly, though to a much lesser extent, under FL, P700⁺ and reduced A 0 /A 1 coexist more 934 frequently. Consequently, charge recombination through the B-branch may more readily generate [3] P700* and thereby 935 increase [1] O 2 formation, potentially accounting for preferential damage to the region near the PSI reaction center. By 936 contrast, the reduced states of the Fe–S clusters are thought to be relatively short-lived under rSPs, which would limit 937 O 2 [•] ⁻ production from these cofactors (Fig. 5b). 938 Finally, chilling-induced PSI photoinhibition in cucumber progresses even under relatively weak light of 939 approximately 100–250 µmol photons m⁻² s⁻¹ (Fig. 5c). At low temperatures, NDH activity declines, while weakened 940 association of CF 1 with the thylakoid membrane impairs ΔpH formation, ATP synthesis, and photosynthetic control. 21 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.02.763831; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.02.763831) preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 941 Reduced ATP production, together with the temperature-dependent decreases in enzymatic reactions, strongly 942 suppresses the CBB cycle and photorespiration, further decreasing electron consumption downstream of PSI. 943 Consequently, even under weak light, electron flow into PSI may exceed the electron-consuming capacity of 944 downstream reactions, allowing the reduced state to extend beyond F A /F B to F X and A 1 . Impaired photosynthetic 945 control also suppresses the accumulation of P700⁺, making safe charge recombination from F X [–] to P700⁺ less likely. 946 Under chilling conditions, electrons may therefore leak to O 2 from multiple acceptors, including A 1 and F A /F B, 947 promoting O 2 [•] ⁻ production. The resulting O 2 [•] ⁻ can damage the Fe–S clusters. In addition, dismutation of O 2 [•] ⁻ produces 948 H 2 O 2, which can generate •OH through Fenton reactions involving the Fe–S clusters or free Fe [2+], causing further 949 oxidative damage (Hegedűs et al. 2008; Nakano et al. 2006; Šnyrychová et al. 2006). Consistent with a role of Fe in 950 light-induced oxidative damage, cucumber grown under Fe-rich conditions is more susceptible to oxidative injury 951 under light-chilling stress (Takeuchi et al. 2024). 952 Thus, the ROS species generated within PSI and the sites of production are likely to differ among stress 953 conditions, depending on which acceptors become reduced, how long they remain reduced, and whether P700⁺ 954 remains available for charge recombination. 955 Chilling injury caused by ROS generation involves the chain reaction of lipid peroxyl radicals that eventually 956 causes membrane disruption and cell death. In chilling-tolerant plants, levels of antioxidants including ascorbic acid 957 and tocopherols are high. In particular, α-tocopherol, a lipophilic antioxidant localized in the membrane, acts as 958 scavenger of singlet oxygen and lipid peroxyl radicals. For reviews, see Asada (2006) and Mesa and Munné-Bosch 959 (2023). 960 961 **7. Roles of far-red light and allocation of excitation between PSI and PSII.** 962 To keep P700 in its oxidized form, excitation of PSI is prerequisite. In natural light environments in the terrestrial 963 ecosystems, far-red light, which preferentially excites PSI, is abundant. When spillover of excitation from PSII to 964 PSI occurs, PSI is further excited. Under fluctuating-light conditions, PSI may be most vulnerable when a leaf is 965 suddenly exposed to high light after an extended period of low light. NPQ in such leaves is low and does not increase 966 spontaneously. In contrast, the spillover of the excitation from PSII to PSI occurs almost instantaneously. 967 Concerning spillover, especially in shade-type chloroplasts, the occurrence of spillover itself is often doubted, 968 because PSII is in the appressed membranes, whereas PSI is mostly in the non-appressed membranes (Andersson and 969 Anderson 1980), and thus PSII and PSI are too far apart. A study of _Pinus sylvestris_ L. (Scots pine) using time970 resolved fluorescence analysis showed that the spillover increased in early spring with the decrease in the number of 971 thylakoids per granum in the overwintered needles (Bag et al. 2020). These results support the notion that the spillover 972 from PSII to PSI would not occur in chloroplasts with highly stacked grana. 973 We recently examined the degree of spillover in the leaf discs at 77 K using various plants, grown at high 974 and low light. We measured the fluorescence induction at 760 nm following the model by Kitajima and Butler (1975). 975 They assumed that there is no induction in intrinsic PSI fluorescence and that the fluorescence at 760 nm contains no 976 PSII fluorescence. Then, the induction observed at 760 nm would be largely attributed to spillover from PSII. 977 Although we confirmed that the intrinsic PSI fluorescence excited at 700 nm did not show any induction, the second 978 assumption is not valid. Even at 760 nm, there was some vibronic PSII fluorescence. We estimated the leftover PSII 979 fluorescence at 760 nm to be around 15% of the maximum fluorescence, _F_ _m_, at 760 nm (Terashima et al. 2025). After 980 illumination with weak far-red light and subsequent dark treatment for less than 2 min, the leaf discs in state 1 were 981 frozen and their fluorescence induction was analyzed at 690 and 760 nm. Taking account of the PSII contamination, 982 spillover of excitation was calculated to range from 16 to 28% of the excitation initially allocated to PSII. These 983 values were not lower in low-light grown plants, indicating that spillover is rather substantial irrespective of the 984 species and growth light conditions. The discrepancy between the results of studies using time-resolved fluorescence 985 analysis and ours should be resolved. It is noteworthy that our electron microscopy analyses revealed that there is 986 enough space for the PSII-PSI megacomplexes (Kim et al. 2023; Yokono et al. 2024) in chloroplasts of _A. odora_ 987 irrespective of the growth environments, which have vast inter-grana thylakoids. The margins of appressed thylakoids 988 would be able to accommodate such megacomplexes (Terashima et al. 2025). 22 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.02.763831; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.02.763831) preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 989 When the spillover from PSII to PSI is active, PSII may be protected because Chl* near the PSII reaction center 990 decreases considerably. It is interesting to examine whether shade-tolerant plants grown in low light generally 991 construct thylakoid membranes so as to be resistant to FL-induced photoinhibition. Given that shade-tolerant plants 992 are frequently exposed to strong sunflecks in their natural habitats, such a thylakoid construction would be adaptive. 993 994 **8. Technical problems in quantification of PSI parameters.** 995 _**In vitro**_ **electron transport measurement with methyl viologen.** When electron transport activities of thylakoid 996 membranes are measured _in vitro_, methyl viologen (MV) is widely used as an electron acceptor from PSI (Izawa 997 1980). However, this reagent accepts electrons not only from F A /F B at the membrane/protein surface, but also from 998 F X in the PSI complex in the membrane because of its hydrophobic/lipophilic nature (Fujii et al. 1990; Sonoike and 999 Terashima 1994). In our earlier studies using cucumber, we isolated thylakoids after tr **e** ating the leaves with light at 1000 low temperatures and measured the PSI electron transport rate from ascorbate-reduced 3,6-diaminodurene (DAD) to 1001 MV at a concentration of 1 mM. Worse, only the electron transport rates at saturating light were measured (Terashima 1002 et al. 1989, 1991b). Thus, we failed to detect PSI photoinhibition at F A /F B . By contrast, in Terashima et al. (1994), 1003 we assessed the electron transport rate from ascorbate-reduced 3,6-diaminodurene to NADP [+] in the presence of 1004 ferredoxin by measuring the absorbance change at 340 nm. Moreover, in this study, the initial slopes obtained at low 1005 irradiances were compared. In this way, PSI photoinhibition was found. 1006 **Assessing the PSI redox state** _**in vivo**_ **.** Accurate assessment of the PSI redox state is essential for elucidating the 1007 mechanism of PSI photoinhibition. In conventional P700 analyses, Y(ND), Y(I), and Y(NA) are measured by 1008 monitoring concentration of P700 [+] with measuring pulses at two wavelengths, at 810 or 830 nm (sample) and 870 1009 nm (reference) with the PAM 101, and at 830 (sample) and 870 nm (reference) with the Dual-PAM (both from Walz, 1010 Germany). At first, the sample is kept in the dark to reduce P700, and P m level was measured with a rather intense 1011 far-red light and a short saturating flash (SP, for example for 30 ms). Next, actinic light is turned on, and, at the steady 1012 state, the P700 [+] level is measured (P). Then by adding the far-red light and a SP, P m ’ is monitored. Using these data, 1013 Y(ND), Y(I) and Y(NA) are obtained as P/P m, (P m ’ – P)/P m, and (P m – P m ’)/P m, respectively (Klughammer and 1014 Schreiber 1994). Two main concerns for this method have been raised (Fan et al. 2016; Furutani et al. 2022). 1015 First, the components near the P700, plastocyanin (PC) and ferredoxin, show absorbance changes in near infra1016 red waveband as well. The absorption coefficient of P700 [+] was reported to be 12 mM [-1] cm [-1] at 815 nm for spinach 1017 chloroplasts (Inoue et al. 1973). Oxidation minus reduced differential absorption coefficient of P700 [+] at (830 nm 1018 870 nm) calculated from Inoue et al. (1973) is 8.6 mM [-1] cm [-1], while those of spinach plastocyanin and ferredoxin 1019 would be 0.31 and 0.07 mM [-1] cm [-1 ] (calculated from Schreiber 2017). Although the coefficient of Fd is small, that of 1020 PC is considerable. Therefore, the redox levels of PC and Fd influence the apparent redox status of P700 [+] /P700 (see 1021 below). The Dual KLAS/NIR (Walz) deconvolutes redox levels of these three components, using four wavelength 1022 pairs of differential measurements (Schreiber 2017). Oja et al. (2003) measured the redox level of PC using “far-red 1023 light titration.” 1024 Second, the method to determine Y(I) itself has been challenged (Furutani et al. 2022). Manipulating the redox 1025 level of PC before the intense far-red light and a SP application, they showed that Y(I) values under the similar 1026 conditions varied to a considerable extent. During the SP, not only P700 photooxidation but also electron donation to 1027 PSI and electron transfer from its acceptor side proceed rapidly. Values calculated using the SP therefore do not 1028 necessarily reflect the PSI redox state even under the steady illumination conditions (Furutani et al. 2022). Thus, Y(I) 1029 should be regarded as an operational index that reflects the balance between electron donation to PSI and electron 1030 transfer from its acceptor side during the SP rather than direct measure of the steady-state PSI quantum yield. The 1031 estimation of Y(NA) therefore suffers from the same problem. 1032 Here, we raise another concern. It is now widely recognized that fluorescence signals from the leaf do not 1033 represent photosynthetic activity of the whole leaf, because only the fluorescence signal from the shallow part of the 1034 leaf is detected, especially when the measuring beam is blue or red (Oguchi et al. 2011; Terashima et al. 2009). 1035 Concerning the near infra-red light for PSI measurements, leaf absorptance in this waveband is only a few percent 23 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.02.763831; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.02.763831) preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 1036 (Gates 1980), because there are few pigments absorbing light in near infra-red. Due to its diffusive nature of 1037 mesophyll structure, the near infra-red light is well scattered in the leaf sample, and it is believed that absorbance 1038 change in P700 [+] is rather evenly detected throughout the leaf (Fan et al. 2016). 1039 Let us assume that P700 [+] is evenly detected. However, the actinic light for photosynthesis is absorbed unevenly. 1040 The chloroplasts near the illuminated side of the leaf absorb more light on a unit chlorophyll basis than those located 1041 deep in the leaf. Although the photosynthetic properties of chloroplasts change from sun-type to shade-type with the 1042 leaf depth, this gradient of sun-shade chloroplast properties is usually less steep than the light absorption gradient 1043 (Terashima et al. 2009). Then, we can expect that Y(ND) is greater and Y(I) is smaller in the chloroplasts near the 1044 illuminated surface than in those deep in the leaf. On the absorbed photon basis, therefore, Y(I) tends to be 1045 overestimated and Y(ND) tends to be underestimated. This would explain the tendency that the electron transport 1046 rate through PSI, ETR(I), is overestimated, as has been warned (Fan et al. 2016). 1047 When PSI photoinhibition proceeds progressively, Y(I), and thereby ETR(I), tend to be further overestimated. 1048 Thus, the changes in P m should be incorporated (Grebe et al. 2024; Kono et al. 2014; Takeuchi et al. 2025b). 1049 The redox state of Fd obtained by deconvolution of near infra-red measurement data with systems like the Dual 1050 KLAS/NIR (Walz) provides a comprehensive view of PSI acceptor-side reduction (Klughammer and Schreiber 2016; 1051 Schreiber 2017). Fd redox state can be followed continuously under steady illumination without applying saturating 1052 flashes. After the chilling treatment, Fd reoxidation became markedly slower in the chilling-sensitive cucumber 1053 cultivar than in the chilling-tolerant cultivar (Takeuchi et al. 2025a). Fd reoxidation kinetics provide an integrated 1054 measure of electron efflux from reduced Fd (Furutani et al. 2023; Kadota et al. 2019). Fd oxidation reflects the 1055 combined contributions of the CBB cycle, photorespiration, FLV, CEF-PSI, and other alternative electron transport 1056 pathways. The very small oxidized minus reduced absorption coefficient of Fd in near infra-red currently prevents 1057 detailed analysis of Fd use. However, in principle, the oxidation kinetics may be expressed by addition of several 1058 exponential decay components. 1059 The capacities for P700 and Fd oxidation are sensitive indicators of PSI redox regulation at low temperatures 1060 and can be used for early assessment of chilling sensitivity. Takeuchi et al. (2022) reported that the accumulation of 1061 P700⁺ after 3 h of chilling stress predicted subsequent chilling injury among cucumber cultivars. The predictive value 1062 of these measurements, however, depends on physiological context: in cucumber grown under different Fe supplies, 1063 oxidative injury was more pronounced under Fe-rich conditions despite comparable levels of P700 oxidation, 1064 consistent with amplification of ROS-mediated damage by Fe-dependent Fenton chemistry (Takeuchi et al. 2024). 1065 It should be noted again that calculation of the electron flux through Fd suffers from the above-mentioned 1066 problem. Even when the average oxidation rate throughout the leaf can be measured rather correctly upon cessation 1067 of the actinic light, it is expected that the Fd reduction level and oxidation rate are lower and slower in the chloroplasts 1068 far from the illuminated surface. However, the electron flux through Fd is greater in the chloroplasts located near the 1069 illuminated surface. Thus, it is problematical to multiply the average Fd reduction level and reoxidation rate obtained 1070 for Fd with the photon flux density allocated to PSI. 1071 1072 **9. Future directions in PSI photoinhibition research.** 1073 **9.1 Species differences and unresolved molecular mechanisms.** 1074 **PSI photoinhibition at low temperatures.** Research on chilling-induced PSI photoinhibition has advanced 1075 substantially using cucumber as a model. An important question is whether the sequence of events leading to chilling1076 induced PSI photoinhibition described here also occurs in plant species other than cucumber. As has been discussed, 1077 PSI protection at low temperature depends on different combinations of processes that restrict electron flow to PSI 1078 and relieve reduction on its acceptor side. These include PSII photoinhibition, photosynthetic control, CEF-PSI, FLV, 1079 etc. The relative importance of these processes is likely to vary among plant species and growth environments. In 1080 cucumber, the chilling-induced weak association of CF 1 with the thylakoid membrane uncouples ATP synthesis and 1081 impairs ΔpH formation and thereby photosynthetic control. The resulting loss of donor-side regulation makes PSI 1082 protection increasingly dependent on acceptor-side processes, particularly NDH (Takeuchi et al. 2025a, 2026). By 24 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.02.763831; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.02.763831) preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 1083 contrast, in plants that rapidly downregulate PSII activity, electron flow into PSI is suppressed, potentially reducing 1084 their dependence on acceptor-side processes such as NDH. Future comparative studies should therefore examine how 1085 different species combine donor-side regulation with acceptor-side electron dissipation, rather than evaluating each 1086 protective process in isolation. 1087 Further studies are needed to determine how changes in NDH activity at low temperature affect chloroplast 1088 function. NDH is embedded in a broader chloroplast redox network: its activity is modulated by several regulatory 1089 factors (Ermakova et al. 2024b; He et al. 2026; Kambakam et al. 2016; Kohzuma et al. 2026, preprint; Lascano et al. 1090 2003; Nikkanen et al. 2018, 2020; Wang et al. 2006; Zhang Q et al. 2024; Zheng et al. 2025), and changes in its 1091 activity can, in turn, affect chloroplast metabolism, including NADPH homeostasis (Nashilevitz et al. 2010; Zhang 1092 Q et al. 2024; Zhang Y et al. 2023). Understanding the consequences of altered NDH activity during chilling will 1093 therefore require examining both its regulation and its metabolic effects within this network. 1094 Studies of PSI photoinhibition under stresses other than chilling have identified diverse redox-regulatory and 1095 electron-consuming processes that contribute to PSI protection, some of which operate beyond the linear 1096 photosynthetic electron transport chain. These include FTR, nitrogen and sulfur assimilation, the malate valve, 1097 mitochondrial respiration, and plastid terminal oxidase (PTOX) (Backhausen et al. 1994; Hani et al. 2025; Hanke et 1098 al. 2008; Höhner et al. 2021; Konishi and Noguchi 2025; Krämer et al. 2024; Messant et al. 2024; Yamada et al. 1099 2024). It remains to be determined to what extent these pathways protect PSI at low temperature, whether they relieve 1100 acceptor-side reduction directly or indirectly, and how they interact with the protective mechanisms described above. 1101 Another important question is which component of PSI is damaged first during chilling-induced photoinhibition. 1102 Recent evidence indicates that light-harvesting complex I (LHCI) may be one of the early targets of PSI 1103 photoinhibition (Takagi et al. 2026). Oxidative damage to LHCI has been reported during chilling-induced PSI 1104 photoinhibition (Alboresi et al. 2009; Rajagopal et al. 2005). Oxidatively modified amino acid residues have also 1105 been identified in LHCI proteins, including residues near bound chlorophylls and on the protein surface (Kale et al. 1106 2020, 2022). Disruption of excitation-energy transfer from LHCI to the PSI core could prolong chlorophyll excitation 1107 within LHCI, increasing the probability of [3] Chl* formation and subsequent [1] O 2 generation (Carbonera et al. 2005; 1108 Kale et al. 2020; Rajagopal et al. 2005). Collectively, these observations suggest that LHCI damage may be an early 1109 event during chilling-induced PSI photoinhibition (Sonoike 1999; Takeuchi et al. 2026). This possibility may also 1110 help explain the early destabilization of the PSI–NDH supercomplex. Lhca5 and Lhca6 mediate the association 1111 between PSI and NDH, and damage to these subunits during chilling stress has been suggested to contribute to 1112 destabilization of the PSI–NDH supercomplex (Takeuchi et al. 2026). 1113 **FL and rSPs induced PSI photoinhibition: Points of tension concerning the division of labor between the two** 1114 **CEF-PSI pathways.** Several unresolved issues are summarized in Table S2 and discussed in Text S2. Three points 1115 deserve mentioning here. The first concerns gating itself (see § 5.3). Two observations weigh against the idea that the 1116 pmf holds NDH back in a leaf: a light-induced pmf did not inhibit NDH in the way that added ATP did (Fisher et al. 1117 2019), and NDH-dependent CEF-PSI operates in the H 2 O 2 -accumulating line GO5 despite a pmf about twice that of 1118 the wild type (Strand et al. 2015). By contrast, collapsing the pmf with uncouplers increased the NDH-dependent 1119 post-illumination fluorescence rise in thylakoids from spinach, _A. hybridus_ and _A. thaliana_ ; the increase was smaller 1120 in _A. thaliana_, but this reflects the larger share of the antimycin A-sensitive pathway in its fluorescence rise (Strand 1121 et al. 2017). Neither of the two observations is decisive: gating defined by the net driving force rather than by Δp 1122 alone accommodates GO5, and Fisher et al. (2019) themselves expect NDH to become limited at a higher pmf (Table 1123 S2, row 1). We therefore adopt gating as a working hypothesis. It adds to, rather than replaces, the established reasons 1124 why NDH contributes little under red high light (low abundance, slow turnover and competition with PSII for 1125 oxidized PQ; Ermakova et al. 2024a), and it explains why the balance should shift with light quality as well as with 1126 irradiance. Two mutant studies point the same way: in _hope2_, the NDH pathway contributed most under low light, 1127 where back-pressure from the pmf is small (Degen et al. 2023), and in _crr2-2_ under red FL, NDH relieved the acceptor 1128 side of PSI in the low-light phases but not in the high-light phases, possibly because a large ΔpH and a reduced PQ 1129 pool favor its reverse reaction (Zhou et al. 2023). In a recent review, Shikanai et al. (2025) also proposed that back1130 pressure from the pmf, possibly through a reverse reaction, limits NDH, so that NDH contributes to pmf formation 1131 mainly under low light. Two qualifications remain: an oxidized chain favors NDH only while enough reduced donors 25 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.02.763831; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.02.763831) preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 1132 are left (Joët et al. 2002; Table S2, row 2), and ATP feedback could produce a similar division of roles (Fisher et al. 1133 2019; Table S2, row 3). The sign of the net driving force has never been determined _in vivo_, and a direct test is 1134 overdue (Table S3, predictions 1–3). 1135 The second point concerns the pulse-type regime. Under repeated 1-s pulses of approximately 3,000 µmol 1136 photons m⁻² s⁻¹, _pgr5_ and _ndho_ lose P m to a similar extent as the wild type (Rantala et al. 2020), as in our FR-free 1137 controls, in which P m fell to approximately 70% in all genotypes (Kono et al. 2017): the loss of either pathway does 1138 not increase PSI photoinhibition under this kind of burst. In pulse regimes, genotype dependence appears only when 1139 FR is added, and it then runs through NDH, whereas in the 2-min regime it runs through PGR5. This is what both the 1140 masking hypothesis and the gating hypothesis predict, although the mechanism of photoinhibition may differ between 1141 the two types of regimes (Sonoike 2025; Table S2, row 6). The third point concerns the molecular role of PGR5, 1142 which remains uncertain (Table S2, row 8). A cycle at steady state returns to PSI exactly as many electrons as it 1143 removes and therefore cannot keep P700 oxidized by itself; its protective effect must come from the protons it moves, 1144 either as extra ATP that relieves the acceptor side or as lumen acidification that engages photosynthetic control 1145 (Vilyanen and Kozuleva 2025). Our account in § 5.2 uses both routes and requires only that the PGR5-dependent 1146 pathway builds the pmf quickly from a large electron flux without a proton pump of its own. If PGR5 proves to be a 1147 regulator rather than a carrier, the "capacity" entries of Table 1 describe the pathway rather than the protein, and the 1148 rescue of _pgr5_ by FR is read as FR removing the need for that regulation; the conclusion is unchanged. 1149 **Testable predictions.** The hypotheses make predictions that can be tested with existing methods (Table S3). The 1150 central one is that FR shifts the thermodynamic balance of NDH toward the forward reaction: the pmf from the 1151 electrochromic shift and the redox states of P700, plastocyanin and the PQ pool, measured together in leaves, should 1152 show that the net driving force is more positive under FR than in the high-light phase of red FL. The Δp dependence 1153 of NDH itself can be measured in coupled thylakoids under FR when the pmf is set without ATP, for example by an 1154 acid–base transition (Jagendorf and Uribe 1966); on the genetic side, FR should fail to rescue the _pgr5 crr_ double 1155 mutant. If predictions (1) to (5) fail to show a wavelength-dependent constraint on NDH, the gating hypothesis should 1156 be abandoned, and the division of roles should be explained by capacity, ATP feedback and activation alone. The 1157 masking hypothesis, however, would survive: FR oxidizes P700 by the excitation balance, this removes the specific 1158 need for PGR5, and NDH provides the low-flux recycling that remains necessary. 1159 1160 **9.2 Need of ecophysiological studies.** 1161 **Ecological implications.** Two lines of evidence bear on which pathway matters in nature. The first concerns light 1162 intensity. Under FL that reproduced the irradiance, but not the spectrum, measured inside a maize canopy at 100-ms 1163 resolution, the NDH-deficient mutants _ndho_ and _ndhm_ lost on average approximately 63% of their shoot dry weight 1164 and the Burren _ndhG_ variant approximately 10%, whereas neither was affected under constant light or slow square1165 wave fluctuations (Theeuwen et al. 2022, preprint). When the same lines were grown under a similar canopy-like 1166 regime, the NDH knockouts showed lower linear electron flow, pmf and biomass than under constant light, whereas 1167 Fv/Fm was unaffected (L.A. Vorbach, unpublished report). PSI photoinhibition was not measured in either study. The 1168 second concerns light quality. Canopy and understory light is enriched in FR, the red:far-red photon ratio falling from 1169 about 1.2 in daylight to about 0.1 under a dense canopy (Casal 2013; Franklin and Quail 2010; Smith 1982), and in 1170 field-grown _Erigeron annuus_ (L.) Pers. and _Commelina communis_ L., adding FR to red FL largely prevented PSI 1171 photoinhibition (Kono et al. 2017). The two lines support two different statements, that NDH matters for growth 1172 under canopy-like fluctuations and that NDH-dependent protection is engaged by FR; their combination is a 1173 hypothesis: in the FR-rich, rapidly fluctuating light of canopies, the NDH-dependent pathway may be the ecologically 1174 relevant route of PSI protection, and the red LED regimes of most laboratory studies may have overestimated the 1175 role of PGR5 and underestimated that of NDH; testing this requires PSI photoinhibition to be measured under natural 1176 spectra (Table S3, prediction 11). 1177 To understand PSI photoinhibition and its suppression mechanisms in extant plants, we should not forget that 1178 plants have evolved under natural environmental conditions. Studies of PSI photoinhibition under FL have become 26 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.02.763831; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.02.763831) preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 1179 increasingly active, and PSI is now considered to be vulnerable to FL. However, we have found that PSI 1180 photoinhibition is almost completely suppressed by far-red light at a level comparable to that present in natural 1181 sunlight. Despite this clear result, the extent of PSI photoinhibition under fluctuating natural light has scarcely been 1182 examined. Careful field studies are needed to demonstrate the suppressive effect of far-red light on PSI 1183 photoinhibition under natural conditions. Another important problem is that we do not know the detailed roles of far1184 red light in photosynthesis. We may call this situation “Far-red drop.” In addition to the problems discussed above, 1185 far-red light is not included in the actinic light of most conventional photosynthetic instruments. The latest LI-6800 1186 (LI-COR) allows far-red light to be added, but this is not the default setting. With such excitation light, insufficient 1187 excitation of PSI can cause over-reduction of the PQ pool, resulting in a decrease in Y(II) (Kono et al. 2022). 1188 Improvement of the spectral composition of excitation light is therefore essential. 1189 Based on the effects of far-red light supplementation, it has been proposed that photosynthetically active 1190 radiation (PAR), conventionally defined as 400–700 nm, should be extended to 750 nm (Zhen and Bugbee 2020). 1191 The conventional definition of PAR as 400–700 nm was partly based on the fact that green leaves strongly absorb 1192 light in this wavelength range. The absorptance of green leaves at 700 nm is approximately comparable to that at 550 1193 nm, and decreases sharply at longer wavelengths and becomes a few percent at 750 nm. Considering the first law of 1194 photochemistry, “only absorbed photons can do photochemical work,” the low absorptance of green leaves between 1195 700 and 750 nm suggests that a simple extension of PAR to 750 nm is not appropriate (Levine et al. 2026). 1196 In this review, we have examined the mechanisms of PSI photoinhibition in green leaves and the mechanisms 1197 that suppress it. In plant factories and other controlled environments, we can now freely use light sources with a wide 1198 variety of spectral compositions. In a “controlled” light environment, any wavelength composition can be selected 1199 as long as the light is absorbed by plants. However, whenever plant functions are considered, it is important to keep 1200 in mind that extant plants have evolved under natural environmental conditions. 1201 1202 1203 1204 1205 1206 27 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.02.763831; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.02.763831) preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 1207 **Supplementary Information** 1208 Table S1. Summary of PSI photoinhibition studies, with explanatory notes 1209 Table S2. Points of tension in the proposed division of roles between the CEF-PSI pathways 1210 Table S3. Testable predictions of the masking and redox–Δp gating hypotheses 1211 Text S1. Affinity of CEF-PSI pathways to Fd 1212 Text S2. Points of tension and open questions 1213 Text S3. Experiments to test the masking and redox–Δp gating hypotheses 1214 1215 1216 **Acknowledgement** 1217 We thank Professor Chikahiro MIYAKE for closely reading some parts of the text and providing useful suggestions. 1218 We dedicate this review paper to the late Professor Shizuo YOSHIDA (who passed away on January 9, 2024), a 1219 pioneer in plant membrane biology at low temperatures, particularly known for his studies on the role of plasma 1220 membrane H [+] -ATPase, and to the late Professor Sakae KATOH (who passed away on May 21, 2025), a pioneer in 1221 photosynthesis research, particularly known for purification of plastocyanin and determination of its role. One of the 1222 authors (I.T.) was enlightened by Prof. Yoshida’s ecophysiological approach. Prof. Katoh taught him how to approach 1223 thylakoid reactions, with incredible patience. 1224 1225 1226 **Declarations** 1227 **Author contributions** 1228 IT and MK conceived the review. RO performed the literature search and data analysis for the meta-analysis (§ 3, 1229 Table S1 and Fig. 3). KT, MK, IT and RO searched the literature for their sections and drafted the manuscript: KT 1230 drafted §§ 1, 4 and 6 and the paragraphs on chilling in § 9.1; MK drafted §§ 2 and 5, the rest of § 9.1, the paragraph 1231 “Ecological implications” in § 9.2, Supplementary Tables S2 and S3, and Supplementary Texts S1–S3; IT drafted the 1232 Introduction, §§ 7 and 8, and the rest of § 9.2. KS and IT compiled the drafts. All authors critically revised the 1233 manuscript and read and approved the final version. 1234 1235 1236 **Funding** 1237 This work was supported by JSPS KAKENHI Grant Numbers JP24K09493 to M.K., JP23KJ1357 to K.T., 1238 JP25K02047 to R.O., and JP17H05718, JP19H04718, JP22H02640 and JP23K23903 to I.T.; by the Yushan Fellow 1239 Program of the Ministry of Education (MOE), Taiwan (MOE-113-YSFAG-0006-001-P1) to I.T.; and by the National 1240 Science and Technology Council (NSTC), Taiwan (114-2311-B-005-003) to I.T. 1241 1242 1243 **Competing interests** 1244 The authors declare no competing interests. 1245 1246 1247 **Data availability** 1248 Data in this article are available from the corresponding authors upon request. 1249 1250 1251 28 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.02.763831; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.02.763831) preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 1252 1253 **References** 1254 1255 Alboresi A, Ballottari M, Hienerwadel R, Giacometti GM, Morosinotto T (2009) Antenna complexes protect 1256 photosystem I from photoinhibition. BMC Plant Biol 9:71. [https://doi.org/10.1186/1471-2229-9-71](https://doi.org/10.1186/1471-2229-9-71) 1257 Allen JF (2003) Cyclic, pseudocyclic and noncyclic photophosphorylation: new links in the chain. Trends Plant Sci 1258 8:15–19. [https://doi.org/10.1016/S1360-1385(02)00006-7](https://doi.org/10.1016/S1360-1385(02)00006-7) 1259 Amin B, Atif MJ, Kandegama WMW, Nasar J, Alam P, Fang Z, Cheng Z (2024) Low temperature and high humidity 1260 affect dynamics of chlorophyll biosynthesis and secondary metabolites in cucumber. BMC Plant Biol 24:903. 1261 [https://doi.org/10.1186/s12870-024-05615-2](https://doi.org/10.1186/s12870-024-05615-2) 1262 Anderson JM (1986) Photoregulation of the composition, function, and structure of thylakoid membranes. Annu Rev 1263 [Plant Physiol 37:93–136. https://doi.org/10.1146/annurev.pp.37.060186.000521](https://doi.org/10.1146/annurev.pp.37.060186.000521) 1264 Anderson JM, Aro EM (1993) Grana stacking and protection of photosystem II in thylakoid membranes of higher 1265 plant leaves under sustained high irradiance: an hypothesis. Photosynth Res 41:315–326. 1266 [https://doi.org/10.1007/BF00019409](https://doi.org/10.1007/BF00019409) 1267 Andersson B, Anderson JM (1980) Lateral heterogeneity in the distribution of chlorophyll-protein complexes of the 1268 thylakoid membranes of spinach chloroplasts. Biochim Biophys Acta 593:427–440. https://doi.org/10.1016/00051269 2728(80)90078-X 1270 Antal TK, Kovalenko IB, Rubin AB, Tyystjärvi E (2013) Photosynthesis-related quantities for education and 1271 modeling. Photosynth Res 117:1–30. [https://doi.org/10.1007/s11120-013-9945-8](https://doi.org/10.1007/s11120-013-9945-8) 1272 Arnon DI, Chain RK (1975) Regulation of ferredoxin-catalyzed photosynthetic phosphorylations. Proc Natl Acad 1273 Sci USA 72:4961–4965. [https://doi.org/10.1073/pnas.72.12.4961](https://doi.org/10.1073/pnas.72.12.4961) 1274 Asada K (1999) The water-water cycle in chloroplasts: scavenging of active oxygens and dissipation of excess 1275 [photons. Annu Rev Plant Physiol Plant Mol Biol 50:601–639. https://doi.org/10.1146/annurev.arplant.50.1.601](https://doi.org/10.1146/annurev.arplant.50.1.601) 1276 Asada K (2006) Production and scavenging of reactive oxygen species in chloroplasts and their functions. Plant 1277 Physiol 141:391–396. [https://doi.org/10.1104/pp.106.082040](https://doi.org/10.1104/pp.106.082040) 1278 Azai C, Harada J, Fujimoto S, Masuda S, Kosumi D (2020) Anaerobic energy dissipation by glycosylated carotenoids 1279 in the green sulfur bacterium _Chlorobaculum tepidum_ . J Photochem Photobiol A Chem 403:112828. 1280 [https://doi.org/10.1016/j.jphotochem.2020.112828](https://doi.org/10.1016/j.jphotochem.2020.112828) 1281 Backhausen JE, Kitzmann C, Scheibe R (1994) Competition between electron acceptors in photosynthesis: regulation 1282 of the malate valve during CO2 fixation and nitrite reduction. Photosynth Res 42:75–86. 1283 [https://doi.org/10.1007/BF00019060](https://doi.org/10.1007/BF00019060) 1284 Badger MR (1985) Photosynthetic oxygen exchange. Annu Rev Plant Physiol 36:27–53. 1285 https://doi.org/10.1146/annurev.pp.36.060185.000331 1286 Bag P, Chukhutsina V, Zhang Z, Paul S, Ivanov AG, Shutova T, Croce R, Holzwarth AR, Jansson S (2020) Direct 1287 energy transfer from photosystem II to photosystem I confers winter sustainability in Scots pine. Nat Commun 1288 11:6388. [https://doi.org/10.1038/s41467-020-20137-9](https://doi.org/10.1038/s41467-020-20137-9) 1289 Bag P, Shutova T, Shevela D, Lihavainen J, Nanda S, Ivanov AG, Messinger J, Jansson S (2023) Flavodiiron1290 mediated O 2 photoreduction at photosystem I acceptor-side provides photoprotection to conifer thylakoids in early 1291 [spring. Nat Commun 14:3210. https://doi.org/10.1038/s41467-023-38938-z](https://doi.org/10.1038/s41467-023-38938-z) 1292 Bindra JK, Malavath T, Teferi MY, Kretzschmar M, Kern J, Niklas J, Utschig LM, Poluektov OG (2024) Light1293 induced charge separation in photosystem I from different biological species characterized by multifrequency 1294 electron paramagnetic resonance spectroscopy. Int J Mol Sci 25:8188. [https://doi.org/10.3390/ijms25158188](https://doi.org/10.3390/ijms25158188) 1295 Boardman NK (1977) Comparative photosynthesis of sun and shade plants. Annu Rev Plant Physiol 28:355–377. 1296 https://doi.org/10.1146/annurev.pp.28.060177.002035 1297 Böhme H (1978) Quantitative determination of ferredoxin, ferredoxin-NADP [+] reductase and plastocyanin in spinach 1298 chloroplasts. Eur J Biochem 83:137–141. [https://doi.org/10.1111/j.1432-1033.1978.tb12077.x](https://doi.org/10.1111/j.1432-1033.1978.tb12077.x) 1299 Bukhov NG, Govindachary S, Rajagopal S, Joly D, Carpentier R (2004) Enhanced rates of P700+ dark-reduction in 1300 leaves of _Cucumis_ _sativus_ L. photoinhibited at chilling temperature. Planta 218:852–861. 1301 [https://doi.org/10.1007/s00425-003-1165-6](https://doi.org/10.1007/s00425-003-1165-6) 1302 Bulteau A-L, Ikeda-Saito M, Szweda LI (2003) Redox-dependent modulation of aconitase activity in intact 1303 mitochondria. Biochemistry 42:14846–14855. [https://doi.org/10.1021/bi0353979](https://doi.org/10.1021/bi0353979) 1304 Burnap RL (2026) Energy coupling in cyanobacterial NDH-1 complexes. Biochim Biophys Acta Bioenerg 1305 1868:149609. [https://doi.org/10.1016/j.bbabio.2026.149609](https://doi.org/10.1016/j.bbabio.2026.149609) 1306 Carbonera D, Agostini G, Morosinotto T, Bassi R (2005) Quenching of chlorophyll triplet states by carotenoids in 1307 reconstituted Lhca4 subunit of peripheral light-harvesting complex of photosystem I. Biochemistry 44:8337–8346. 1308 [https://doi.org/10.1021/bi050260z](https://doi.org/10.1021/bi050260z) 29 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.02.763831; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.02.763831) preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 1309 Casal JJ (2013) Photoreceptor signaling networks in plant responses to shade. Annu Rev Plant Biol 64:403–427. 1310 [https://doi.org/10.1146/annurev-arplant-050312-120221](https://doi.org/10.1146/annurev-arplant-050312-120221) 1311 Cazzaniga S, Li Z, Niyogi KK, Bassi R, Dall’Osto L (2012) The _Arabidopsis_ _szl1_ mutant reveals a critical role of β1312 [carotene in photosystem I photoprotection. Plant Physiol 159:1745–1758. https://doi.org/10.1104/pp.112.201137](https://doi.org/10.1104/pp.112.201137) 1313 Chen S, Zheng Q, Qi Z, Ding J, Song X, Xia X (2024) Stress-induced delay of the I-P rise of the fast chlorophyll _a_ 1314 fluorescence transient in tomato. Sci Hortic 326:112741. [https://doi.org/10.1016/j.scienta.2023.112741](https://doi.org/10.1016/j.scienta.2023.112741) 1315 Cherepanov DA, Milanovsky GE, Petrova AA, Tikhonov AN, Semenov AY (2017) Electron transfer through the 1316 acceptor side of photosystem I: interaction with exogenous acceptors and molecular oxygen. Biochemistry (Mosc) 1317 82:1249–1268. https://doi.org/10.1134/S0006297917110037 1318 Chow WS, Melis A, Anderson JM (1990) Adjustments of photosystem stoichiometry in chloroplasts improve the 1319 quantum efficiency of photosynthesis. Proc Natl Acad Sci USA 87:7502–7506. 1320 Cleland RE, Bendall DS (1992) Photosystem I cyclic electron transport: measurement of ferredoxin-plastoquinone 1321 reductase activity. Photosynth Res 34:409–418. [https://doi.org/10.1007/BF00029815](https://doi.org/10.1007/BF00029815) 1322 Courteille A, Vesa S, Sanz-Barrio R, Cazalé A-C, Becuwe-Linka N, Farran I, Havaux M, Rey P, Rumeau D (2013) 1323 Thioredoxin m4 controls photosynthetic alternative electron pathways in _Arabidopsis_ . Plant Physiol 161:508–520. 1324 https://doi.org/10.1104/pp.112.207019 1325 DalCorso G, Pesaresi P, Masiero S, Aseeva E, Schünemann D, Finazzi G, Joliot P, Barbato R, Leister D (2008) A 1326 complex containing PGRL1 and PGR5 is involved in the switch between linear and cyclic electron flow in 1327 _Arabidopsis_ . Cell 132:273–285. [https://doi.org/10.1016/j.cell.2007.12.028](https://doi.org/10.1016/j.cell.2007.12.028) 1328 Degen GE, Johnson MP (2024) Photosynthetic control at the cytochrome _b_ 6 _f_ complex. Plant Cell 36:4065–4079. 1329 [https://doi.org/10.1093/plcell/koae133](https://doi.org/10.1093/plcell/koae133) 1330 Degen GE, Jackson PJ, Proctor MS, Zoulias N, Casson SA, Johnson MP (2023) High cyclic electron transfer via the 1331 PGR5 pathway in the absence of photosynthetic control. Plant Physiol 192:370–386. 1332 https://doi.org/10.1093/plphys/kiad084 1333 Degen GE, Schwarzländer M, Johnson MP (2026) Disequilibrium between chloroplast proton motive force and ATP 1334 levels in _Arabidopsis_ . Nat Plants 12:1679–1687. [https://doi.org/10.1038/s41477-026-02374-w](https://doi.org/10.1038/s41477-026-02374-w) 1335 Ermakova M, Fitzpatrick D, Larkum AWD (2024a) Cyclic electron flow and photosystem II-less photosynthesis. 1336 Funct Plant Biol 51:FP24185. [https://doi.org/10.1071/FP24185](https://doi.org/10.1071/FP24185) 1337 Ermakova M, Woodford R, Fitzpatrick D, Nix SJ, Zwahlen SM, Farquhar GD, von Caemmerer S, Furbank RT 1338 (2024b) Chloroplast NADH dehydrogenase-like complex-mediated cyclic electron flow is the main electron 1339 transport route in C4 bundle sheath cells. New Phytol 243:2187–2200. [https://doi.org/10.1111/nph.19982](https://doi.org/10.1111/nph.19982) 1340 Fan D-Y, Fitzpatrick D, Oguchi R, Ma W, Kou J, Chow WS (2016) Obstacles in the quantification of the cyclic 1341 electron flux around photosystem I in leaves of C 3 plants. Photosynth Res 129:239–251. 1342 https://doi.org/10.1007/s11120-016-0223-4 1343 Fisher N, Bricker TM, Kramer DM (2019) Regulation of photosynthetic cyclic electron flow pathways by adenylate 1344 status in higher plant chloroplasts. Biochim Biophys Acta Bioenerg 1860:148081. 1345 [https://doi.org/10.1016/j.bbabio.2019.148081](https://doi.org/10.1016/j.bbabio.2019.148081) 1346 Flint DH, Tuminello JF, Emptage MH (1993) The inactivation of Fe-S cluster-containing hydro-lyases by superoxide. 1347 J Biol Chem 268:22369–22376. 1348 Franklin KA, Quail PH (2010) Phytochrome functions in _Arabidopsis_ development. J Exp Bot 61:11–24. 1349 [https://doi.org/10.1093/jxb/erp304](https://doi.org/10.1093/jxb/erp304) 1350 Fujii T, Yokoyama E-I, Inoue K, Sakurai H (1990) The sites of electron donation of photosystem I to methyl viologen. 1351 Biochim Biophys Acta 1015:41–48. https://doi.org/10.1016/0005-2728(90)90213-N 1352 Furutani R, Ohnishi M, Mori Y, Wada S, Miyake C (2022) The difficulty of estimating the electron transport rate at 1353 photosystem I. J Plant Res 135:565–577. [https://doi.org/10.1007/s10265-021-01357-6](https://doi.org/10.1007/s10265-021-01357-6) 1354 Furutani R, Wada S, Ifuku K, Maekawa S, Miyake C (2023) Higher reduced state of Fe/S-signals, with the suppressed 1355 oxidation of P700, causes PSI inactivation in _Arabidopsis_ _thaliana_ . Antioxidants 12:21. 1356 [https://doi.org/10.3390/antiox12010021](https://doi.org/10.3390/antiox12010021) 1357 Gardner PR (1997) Superoxide-driven aconitase Fe-S center cycling. Biosci Rep 17:33–42. 1358 [https://doi.org/10.1023/A:1027383100936](https://doi.org/10.1023/A:1027383100936) 1359 Gates DM (1980) Biophysical ecology. Springer, New York. 1360 Govindachary S, Bukhov NG, Joly D, Carpentier R (2004) Photosystem II inhibition by moderate light under low 1361 temperature in intact leaves of chilling-sensitive and chilling-tolerant plants. Physiol Plant 121:322–333. 1362 Grebe S, Porcar-Castell A, Riikonen A, Paakkarinen V, Aro E-M (2024) Accounting for photosystem I 1363 photoinhibition sheds new light on seasonal acclimation strategies of boreal conifers. J Exp Bot 75:3973–3992. 1364 [https://doi.org/10.1093/jxb/erae145](https://doi.org/10.1093/jxb/erae145) 1365 Hani U, Naranjo B, Shimakawa G, Espinasse C, Vanacker H, Sétif P, Rintamäki E, Issakidis-Bourguet E, Krieger 30 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.02.763831; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.02.763831) preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 1366 Liszkay A (2025) A complex and dynamic redox network regulates oxygen reduction at photosystem I in 1367 _Arabidopsis_ . Plant Physiol 197:kiae501. https://doi.org/10.1093/plphys/kiae501 1368 Hanke GT, Endo T, Satoh F, Hase T (2008) Altered photosynthetic electron channelling into cyclic electron flow and 1369 nitrite assimilation in a mutant of ferredoxin(H) reductase. Plant Cell Environ 31:1017–1028. 1370 He Y, Li R, Li J, Zheng S, Jiang D, Li X, Liu Q, Zhou X, Lu J, Chen J (2026) OsRNase ZS2 degrades the transcripts 1371 of NdhD and NdhF subunits of the NDH complex to modulate the balance between cyclic and linear 1372 photosynthetic electron transport. Plant Commun 7:101791. [https://doi.org/10.1016/j.xplc.2026.101791](https://doi.org/10.1016/j.xplc.2026.101791) 1373 Hegedűs A, Janda T, Horváth GV, Dudits D (2008) Accumulation of overproduced ferritin in the chloroplast provides 1374 protection against photoinhibition induced by low temperature in tobacco plants. J Plant Physiol 165:1647–1651. 1375 Hertle AP, Blunder T, Wunder T, Pesaresi P, Pribil M, Armbruster U, Leister D (2013) PGRL1 is the elusive 1376 ferredoxin-plastoquinone reductase in photosynthetic cyclic electron flow. Mol Cell 49:511–523. 1377 [https://doi.org/10.1016/j.molcel.2012.11.030](https://doi.org/10.1016/j.molcel.2012.11.030) 1378 Hikosaka K, Terashima I (1995) A model of the acclimation of photosynthesis in the leaves of C3 plants to sun and 1379 shade with respect to nitrogen use. Plant Cell Environ 18:605–618. [https://doi.org/10.1111/j.1365-](https://doi.org/10.1111/j.1365-3040.1995.tb00562.x) 1380 [3040.1995.tb00562.x](https://doi.org/10.1111/j.1365-3040.1995.tb00562.x) 1381 Hodgson RAJ, Orr GR, Raison JK (1987) Inhibition of photosynthesis by chilling in the light. Plant Sci 49:75–79. 1382 Höhner R, Day PM, Zimmermann SE, Lopez LS, Krämer M, Giavalisco P, Correa Galvis V, Armbruster U, Schöttler 1383 MA, Jahns P, Krueger S, Kunz H-H (2021) Stromal NADH supplied by PHOSPHOGLYCERATE 1384 DEHYDROGENASE3 is crucial for photosynthetic performance. Plant Physiol 186:142–167. 1385 [https://doi.org/10.1093/plphys/kiaa117](https://doi.org/10.1093/plphys/kiaa117) 1386 Ibrahim IM, McKenzie SD, Chung J, Aryal UK, Leon-Salas WD, Puthiyaveetil S (2022) Photosystem stoichiometry 1387 adjustment is a photoreceptor-mediated process in _Arabidopsis_ . Sci Rep 12:10982. 1388 [https://doi.org/10.1038/s41598-022-14967-4](https://doi.org/10.1038/s41598-022-14967-4) 1389 Ifuku K, Endo T, Shikanai T, Aro E-M (2011) Structure of the chloroplast NADH dehydrogenase-like complex: 1390 nomenclature for nuclear-encoded subunits. Plant Cell Physiol 52:1560–1568. https://doi.org/10.1093/pcp/pcr098 1391 Inagaki T, Kida M, Kosumi D, Azai C (2026) Energy quenching via triplet-excited state formation of glycosylated 1392 carotenoids in the photosynthetic reaction center complex of the green sulfur bacterium _Chlorobaculum tepidum_ . 1393 Photosynth Res 164:38. [https://doi.org/10.1007/s11120-026-01226-0](https://doi.org/10.1007/s11120-026-01226-0) 1394 Inoue K, Sakurai H, Hiyama T (1986) Photoinactivation sites of photosystem I in isolated chloroplasts. Plant Cell 1395 Physiol 27:962–968. [https://doi.org/10.1093/oxfordjournals.pcp.a077213](https://doi.org/10.1093/oxfordjournals.pcp.a077213) 1396 Inoue K, Fujii T, Yokoyama E, Matsuura K, Hiyama T, Sakurai H (1989) The photoinhibition site of photosystem I 1397 in isolated chloroplasts under extremely reducing conditions. Plant Cell Physiol 30:65–71. 1398 Inoue Y, Ogawa T, Shibata K (1973) Light-induced spectral changes of P700 in the 800 nm region in _Anacystis_ and 1399 spinach lamellae. Biochim Biophys Acta 305:483–487. https://doi.org/10.1016/0005-2728(73)90194-1 1400 Introini B, Hahn A, Kühlbrandt W (2025) Cryo-EM structure of the NDH–PSI–LHCI supercomplex from _Spinacia_ 1401 _oleracea_ . Nat Struct Mol Biol 32:968–978. [https://doi.org/10.1038/s41594-024-01478-1](https://doi.org/10.1038/s41594-024-01478-1) 1402 Ishikita H, Knapp E-W (2003) Redox potential of quinones in both electron transfer branches of photosystem I. J 1403 Biol Chem 278:52002–52011. 1404 Ivanov AG, Morgan RM, Gray GR, Velitchkova MY, Huner NPA (1998) Temperature/light-dependent development 1405 of selective resistance to photoinhibition of photosystem I. FEBS Lett 430:288–292. 1406 https://doi.org/10.1016/S0014-5793(98)00681-4 1407 Izawa S (1980) Acceptors and donors and chloroplast electron transport. Methods Enzymol 69:413–434. 1408 [https://doi.org/10.1016/S0076-6879(80)69041-7](https://doi.org/10.1016/S0076-6879(80)69041-7) 1409 Jagendorf AT, Uribe E (1966) ATP formation caused by acid–base transition of spinach chloroplasts. Proc Natl Acad 1410 Sci USA 55:170–177. [https://doi.org/10.1073/pnas.55.1.170](https://doi.org/10.1073/pnas.55.1.170) 1411 Joët T, Cournac L, Peltier G, Havaux M (2002) Cyclic electron flow around photosystem I in C3 plants: in vivo 1412 control by the redox state of chloroplasts and involvement of the NADH-dehydrogenase complex. Plant Physiol 1413 128:760–769. [https://doi.org/10.1104/pp.010775](https://doi.org/10.1104/pp.010775) 1414 Joliot P, Johnson GN (2011) Regulation of cyclic and linear electron flow in higher plants. Proc Natl Acad Sci USA 1415 108:13317–13322. [https://doi.org/10.1073/pnas.1110189108](https://doi.org/10.1073/pnas.1110189108) 1416 Kadota K, Furutani R, Makino A, Suzuki Y, Wada S, Miyake C (2019) Oxidation of P700 induces alternative electron 1417 flow in photosystem I in wheat leaves. Plants 8:152. [https://doi.org/10.3390/plants8060152](https://doi.org/10.3390/plants8060152) 1418 Kale R, Sallans L, Frankel LK, Bricker TM (2020) Natively oxidized amino acid residues in the spinach PS I-LHC I 1419 supercomplex. Photosynth Res 143:263–273. [https://doi.org/10.1007/s11120-019-00698-7](https://doi.org/10.1007/s11120-019-00698-7) 1420 Kale RS, Seep JL, Sallans L, Frankel LK, Bricker TM (2022) Oxidative modification of LHC II associated with 1421 photosystem II and PS I-LHC I-LHC II membranes. Photosynth Res 152:261–274. 1422 [https://doi.org/10.1007/s11120-022-00902-1](https://doi.org/10.1007/s11120-022-00902-1) 31 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.02.763831; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.02.763831) preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 1423 Kambakam S, Bhattacharjee U, Petrich J, Rodermel S (2016) PTOX mediates novel pathways of electron transport 1424 in etioplasts of _Arabidopsis_ . Mol Plant 9:1240–1259. 1425 Kato Y, Odahara M, Fukao Y, Shikanai T (2018a) Stepwise evolution of supercomplex formation with photosystem 1426 I is required for stabilization of chloroplast NADH dehydrogenase-like complex: Lhca5-dependent supercomplex 1427 formation in _Physcomitrella patens_ . Plant J 96:937–948. [https://doi.org/10.1111/tpj.14080](https://doi.org/10.1111/tpj.14080) 1428 Kato Y, Sugimoto K, Shikanai T (2018b) NDH-PSI supercomplex assembly precedes full assembly of the NDH 1429 complex in chloroplast. Plant Physiol 176:1728–1738. 1430 Kato Y, Odahara M, Shikanai T (2021) Evolution of an assembly factor-based subunit contributed to a novel NDH1431 PSI supercomplex formation in chloroplasts. Nat Commun 12:3685. https://doi.org/10.1038/s41467-021-240651432 0 1433 Khorobrykh S, Havurinne V, Mattila H, Tyystjärvi E (2020) Oxygen and ROS in photosynthesis. Plants 9:91. 1434 https://doi.org/10.3390/plants9010091 1435 Kim E, Yokono M, Tsugane T, Ishii A, Noda C, Minagawa J (2023) Formation of a stable PSI–PSII megacomplex in 1436 rice that conducts energy spillover. Plant Cell Physiol 64:858–865. https://doi.org/10.1093/pcp/pcad037 1437 Kim JH, Kim SJ, Sung HC, Chow WS, Lee CH (2005) Photosystem I acceptor side limitation is a prerequisite for 1438 the reversible decrease in the maximum extent of P700 oxidation after short-term chilling in the light in four plant 1439 species with different chilling sensitivities. Physiol Plant 123:100–107. [https://doi.org/10.1111/j.1399-](https://doi.org/10.1111/j.1399-3054.2005.00443.x) 1440 [3054.2005.00443.x](https://doi.org/10.1111/j.1399-3054.2005.00443.x) 1441 Kim S-J, Lee C-H, Hope A, Chow WS (2001) Inhibition of photosystems I and II and enhanced back flow of 1442 photosystem I electrons in cucumber leaf discs chilled in the light. Plant Cell Physiol 42:842–848. 1443 Kitajima M, Butler WL (1975) Excitation spectra for photosystem I and photosystem II in chloroplasts and the 1444 spectral characteristics of the distribution of quanta between the two photosystems. Biochim Biophys Acta 1445 408:297–305. 1446 Klughammer C, Schreiber U (1994) An improved method, using saturating light pulses, for the determination of 1447 photosystem I quantum yield via P700+-absorbance changes at 830 nm. Planta 192:261–268. 1448 https://doi.org/10.1007/BF01089043 1449 Klughammer C, Schreiber U (2016) Deconvolution of ferredoxin, plastocyanin, and P700 transmittance changes in 1450 intact leaves with a new type of kinetic LED array spectrophotometer. Photosynth Res 128:195–214. 1451 Knaff DB, Hirasawa M (1991) Ferredoxin-dependent chloroplast enzymes. Biochim Biophys Acta 1056:93–125. 1452 Kodama H, Yamori W (2026) Cyclic electron transport via the NDH complex sustains photosynthesis and 1453 productivity under fluctuating and sub-optimal environments. Physiol Plant 178:e71004. 1454 [https://doi.org/10.1111/ppl.71004](https://doi.org/10.1111/ppl.71004) 1455 Kohzuma K, Murai M, Imaizumi K, Miura K, Kimura A, Yoshida K, Che Y, Ishikawa N, Hisabori T, Ifuku K (2026) 1456 PIFI stabilizes chloroplast NDH-PSI supercomplex to maintain plastoquinone redox balance and PSII efficiency. 1457 bioRxiv. [https://doi.org/10.64898/2026.03.22.713156](https://doi.org/10.64898/2026.03.22.713156) – 1458 Kok B (1956) On the inhibition of photosynthesis by intense light. Biochim Biophys Acta 21:234 258. 1459 Konishi T, Noguchi K (2025) The mitochondrial respiratory chain is important as an electron sink to avoid over1460 reduction of chloroplasts in _Arabidopsis thaliana_ leaves under high-light stress. Plant Cell Physiol 66:1141–1154. 1461 Kono M, Terashima I (2014) Long-term and short-term responses of the photosynthetic electron transport to 1462 fluctuating light. J Photochem Photobiol B 137:89–99. https://doi.org/10.1016/j.jphotobiol.2014.02.016 1463 Kono M, Terashima I (2016) Elucidation of photoprotective mechanisms of PSI against fluctuating light 1464 photoinhibition. Plant Cell Physiol 57:1405–1414. [https://doi.org/10.1093/pcp/pcw103](https://doi.org/10.1093/pcp/pcw103) 1465 Kono M, Noguchi K, Terashima I (2014) Roles of the cyclic electron flow around PSI (CEF-PSI) and O 2 -dependent 1466 alternative pathways in regulation of the photosynthetic electron flow in short-term fluctuating light in 1467 _Arabidopsis thaliana_ . Plant Cell Physiol 55:990–1004. https://doi.org/10.1093/pcp/pcu033 1468 Kono M, Yamori W, Suzuki Y, Terashima I (2017) Photoprotection of PSI by far-red light against the fluctuating 1469 light-induced photoinhibition in _Arabidopsis thaliana_ and field-grown plants. Plant Cell Physiol 58:35–45. 1470 https://doi.org/10.1093/pcp/pcw215 1471 Kono M, Kawaguchi H, Mizusawa N, Yamori W, Suzuki Y, Terashima I (2020) Far-red light accelerates 1472 photosynthesis in the low-light phases of fluctuating light. Plant Cell Physiol 61:192–202. 1473 https://doi.org/10.1093/pcp/pcz191 1474 Kono M, Oguchi R, Terashima I (2022) Photoinhibition of PSI and PSII in nature and in the laboratory: ecological 1475 approaches. Prog Bot 84:241–292. https://doi.org/10.1007/124_2022_67 1476 Kono M, Kodama H, Tanigawa K, Terashima I, Yamori W (2025) Drastic reduction in cytochrome _b_ 6 / _f_ complex 1477 confers robust PSI photoprotection under fluctuating light at the expense of photosynthetic capacity. Physiol Plant 1478 177:e70483. https://doi.org/10.1111/ppl.70483 1479 Kornyeyev D, Logan BA, Allen RD, Holaday AS (2003) Effect of chloroplastic overproduction of ascorbate 32 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.02.763831; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.02.763831) preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 1480 peroxidase on photosynthesis and photoprotection in cotton leaves subjected to low temperature photoinhibition. 1481 Plant Sci 165:1033–1041. https://doi.org/10.1016/S0168-9452(03)00294-2 1482 Kozuleva M (2022) Recent advances in the understanding of superoxide anion radical formation in the photosynthetic 1483 electron transport chain. Acta Physiol Plant 44:92. 1484 Kozuleva M, Petrova A, Milrad Y, Semenov A, Ivanov B, Redding KE, Yacoby I (2021) Phylloquinone is the 1485 principal Mehler reaction site within photosystem I in high light. Plant Physiol 186:1848–1858. 1486 Kozuleva MA, Ivanov BN (2010) Evaluation of the participation of ferredoxin in oxygen reduction in the 1487 photosynthetic electron transport chain of isolated pea thylakoids. Photosynth Res 105:51–61. 1488 Kozuleva MA, Ivanov BN (2023) Superoxide anion radical generation in photosynthetic electron transport chain. 1489 Biochemistry (Mosc) 88:1045–1060. 1490 Krämer M, Blanco NE, Penzler J-F, Davis GA, Brandt B, Leister D, Kunz H-H (2024) Cyclic electron flow 1491 compensates loss of PGDH3 and concomitant stromal NADH reduction. Sci Rep 14:29274. 1492 [https://doi.org/10.1038/s41598-024-80836-x](https://doi.org/10.1038/s41598-024-80836-x) 1493 Kudoh H, Sonoike K (2002) Irreversible damage to photosystem I by chilling in the light: cause of the degradation 1494 of chlorophyll after returning to normal growth temperature. Planta 215:541–548. [https://doi.org/10.1007/s00425-](https://doi.org/10.1007/s00425-002-0790-9) 1495 [002-0790-9](https://doi.org/10.1007/s00425-002-0790-9) 1496 Lascano HR, Casano LM, Martín M, Sabater B (2003) The activity of the chloroplastic Ndh complex is regulated by 1497 phosphorylation of the NDH-F subunit. Plant Physiol 132:256–262. [https://doi.org/10.1104/pp.103.020321](https://doi.org/10.1104/pp.103.020321) 1498 Laughlin TG, Bayne AN, Trempe JF, Savage DF, Davies KM (2019) Structure of the complex I-like molecule NDH 1499 of oxygenic photosynthesis. Nature 566:411–414. [https://doi.org/10.1038/s41586-019-0921-0](https://doi.org/10.1038/s41586-019-0921-0) 1500 Levine CP, Tanigawa K, Wakabayashi Y, Guo W, Qu Y, Terashima I, Yamori W (2026) Far-red light in early growth 1501 stages boosts lettuce biomass and preserves anthocyanins. Ann Bot 137:1215–1232. 1502 https://doi.org/10.1093/aob/mcag031 1503 Li G, Knowles PF, Murphy DJ, Marsh D (1990) Lipid-protein interactions in thylakoid membranes of chilling1504 resistant and -sensitive plants studied by spin label electron spin resonance spectroscopy. J Biol Chem 265:16867– 1505 16872. 1506 Lima-Melo Y, Gollan PJ, Tikkanen M, Silveira JAG, Aro E-M (2019) Consequences of photosystem-I damage and 1507 repair on photosynthesis and carbon use in _Arabidopsis_ _thaliana_ . Plant J 97:1061–1072. 1508 [https://doi.org/10.1111/tpj.14177](https://doi.org/10.1111/tpj.14177) 1509 Maekawa S, Ohnishi M, Wada S, Ifuku K, Miyake C (2024) Enhanced reduction of ferredoxin in PGR5-deficient 1510 mutant of _Arabidopsis thaliana_ stimulated ferredoxin-dependent cyclic electron flow around photosystem I. Int J 1511 Mol Sci 25:2677. [https://doi.org/10.3390/ijms25052677](https://doi.org/10.3390/ijms25052677) 1512 Mazor Y, Borovikova A, Caspy I, Nelson N (2017) Structure of the plant photosystem I supercomplex at 2.6 Å 1513 resolution. Nat Plants 3:17014. [https://doi.org/10.1038/nplants.2017.14](https://doi.org/10.1038/nplants.2017.14) 1514 Mesa T, Munné-Bosch S (2023) α-Tocopherol in chloroplasts: nothing more than an antioxidant? Curr Opin Plant 1515 Biol 74:102400. https://doi.org/10.1016/j.pbi.2023.102400 1516 Messant M, Hani U, Lai T-L, Wilson A, Shimakawa G, Krieger-Liszkay A (2024) Plastid terminal oxidase (PTOX) 1517 protects photosystem I and not photosystem II against photoinhibition in _Arabidopsis thaliana_ and _Marchantia_ 1518 _polymorpha_ . Plant J 117:669–678. [https://doi.org/10.1111/tpj.16520](https://doi.org/10.1111/tpj.16520) 1519 Miller NT, Vaughn MD, Burnap RL (2021) Electron flow through NDH-1 complexes is the major driver of cyclic 1520 electron flow-dependent proton pumping in cyanobacteria. Biochim Biophys Acta Bioenerg 1862:148354. 1521 [https://doi.org/10.1016/j.bbabio.2021.148354](https://doi.org/10.1016/j.bbabio.2021.148354) 1522 Miyake C (2010) Alternative electron flows (water–water cycle and cyclic electron flow around PSI) in 1523 photosynthesis: molecular mechanisms and physiological functions. Plant Cell Physiol 51:1951–1963. 1524 Miyake C (2020) Molecular mechanism of oxidation of P700 and suppression of ROS production in photosystem I 1525 in response to electron-sink limitations in C3 plants. Antioxidants 9:230. [https://doi.org/10.3390/antiox9030230](https://doi.org/10.3390/antiox9030230) 1526 Miyake C, Asada K (1992) Thylakoid-bound ascorbate peroxidase in spinach chloroplasts and photoreduction of its 1527 primary oxidation product monodehydroascorbate radicals in thylakoids. Plant Cell Physiol 33:541–553. 1528 Miyake C, Schreiber U, Hormann H, Sano S, Asada K (1998) The FAD-enzyme monodehydroascorbate radical 1529 reductase mediates photoproduction of superoxide radicals in spinach thylakoid membranes. Plant Cell Physiol 1530 39:821–829. 1531 Miyata K, Noguchi K, Terashima I (2012) Cost and benefit of the repair of photodamaged photosystem II in spinach 1532 [leaves: roles of acclimation to growth light. Photosynth Res 113:165–180. https://doi.org/10.1007/s11120-012-](https://doi.org/10.1007/s11120-012-9767-0) 1533 [9767-0](https://doi.org/10.1007/s11120-012-9767-0) 1534 Miyata K, Ikeda H, Nakaji M, Kanel DR, Terashima I (2015) Rate constants of PSII photoinhibition and its repair, 1535 and PSII fluorescence parameters in field plants in relation to their growth light environments. Plant Cell Physiol 1536 56:1841–1854. https://doi.org/10.1093/pcp/pcv107 33 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.02.763831; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.02.763831) preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 1537 Munekage Y, Hojo M, Meurer J, Endo T, Tasaka M, Shikanai T (2002) PGR5 is involved in cyclic electron flow 1538 around photosystem I and is essential for photoprotection in _Arabidopsis_ . Cell 110:361–371. 1539 [https://doi.org/10.1016/S0092-8674(02)00867-X](https://doi.org/10.1016/S0092-8674(02)00867-X) 1540 Munekage Y, Hashimoto M, Miyake C, Tomizawa K, Endo T, Tasaka M, Shikanai T (2004) Cyclic electron flow 1541 around photosystem I is essential for photosynthesis. Nature 429:579–582. [https://doi.org/10.1038/nature02598](https://doi.org/10.1038/nature02598) 1542 Nakano R, Ishida H, Makino A, Mae T (2006) In vivo fragmentation of the large subunit of ribulose-1,5-bisphosphate 1543 carboxylase by reactive oxygen species in an intact leaf of cucumber under chilling-light conditions. Plant Cell 1544 Physiol 47:270–276. 1545 Napaumpaiporn P, Ogawa T, Sonoike K, Nishiyama Y (2024) Improved capacity for the repair of photosystem II via 1546 reinforcement of the translational and antioxidation systems in _Synechocystis_ sp. PCC 6803. Plant J 117:1165– 1547 1178. 1548 Nashilevitz S, Melamed-Bessudo C, Izkovich Y, Rogachev I, Osorio S, Itkin M, Pankratov I, Shiboleth YM, Cohen 1549 S, Bino RJ, Adato A, Aharoni A (2010) An orange ripening mutant links plastid NAD(P)H dehydrogenase 1550 complex activity to central and specialized metabolism during tomato fruit maturation. Plant Cell 22:1977–1997. 1551 [https://doi.org/10.1105/tpc.110.074716](https://doi.org/10.1105/tpc.110.074716) 1552 Nawrocki WJ, Bailleul B, Picot D, Cardol P, Rappaport F, Wollman FA, Joliot P (2019) The mechanism of cyclic 1553 electron flow. Biochim Biophys Acta Bioenerg 1860:433–438. https://doi.org/10.1016/j.bbabio.2018.12.005 1554 Nikkanen L, Toivola J, Trotta A, Guinea Diaz M, Tikkanen M, Aro E-M, Rintamäki E (2018) Regulation of cyclic 1555 electron flow by chloroplast NADPH-dependent thioredoxin system. Plant Direct 2:e00093. 1556 [https://doi.org/10.1002/pld3.93](https://doi.org/10.1002/pld3.93) 1557 Nikkanen L, Santana Sánchez A, Ermakova M, Rögner M, Cournac L, Eisenhut M, Zhang P, Aro E-M (2020) 1558 Functional redundancy between flavodiiron proteins and NDH-1 in _Synechocystis_ sp. PCC 6803. Plant J 1559 103:1460–1476. https://doi.org/10.1111/tpj.14812 1560 Obara A, Ogawa M, Oyama Y, Suzuki Y, Kono M (2023) Effects of high irradiance and low water temperature on 1561 photoinhibition and repair of photosystems in marimo ( _Aegagropila linnaei_ ) in Lake Akan, Japan. Int J Mol Sci 1562 24:60. 1563 Ogawa K, Kanematsu S, Takabe K, Asada K (1995) Attachment of CuZn-superoxide dismutase to thylakoid 1564 membranes at the site of superoxide generation (PSI) in spinach chloroplasts: detection by immuno-gold labeling 1565 after rapid freezing and substitution method. Plant Cell Physiol 36:565–573. 1566 Oguchi R, Terashima I, Chow WS (2009) The involvement of dual mechanisms of photoinactivation of photosystem 1567 II in _Capsicum annuum_ L. plants. Plant Cell Physiol 50:1815–1825. 1568 Oguchi R, Douwstra P, Fujita T, Chow WS, Terashima I (2011) Intra-leaf gradients of photoinhibition induced by 1569 different color lights: implications for the dual mechanisms of photoinhibition and for the application of 1570 conventional chlorophyll fluorometers. New Phytol 191:146–159. https://doi.org/10.1111/j.14691571 8137.2011.03669.x 1572 Oguchi R, Terashima I, Chow WS (2021) The effect of different spectral light quality on the photoinhibition of 1573 photosystem I in intact leaves. Photosynth Res 149:83–92. https://doi.org/10.1007/s11120-020-00805-z 1574 Ohnishi M, Maekawa S, Wada S, Ifuku K, Miyake C (2023) Evaluating the oxidation rate of reduced ferredoxin in 1575 _Arabidopsis thaliana_ independent of photosynthetic linear electron flow: plausible activity of ferredoxin1576 dependent cyclic electron flow around photosystem I. Int J Mol Sci 24:12145. 1577 [https://doi.org/10.3390/ijms241512145](https://doi.org/10.3390/ijms241512145) 1578 Oja V, Eichelmann H, Peterson RB, Rasulov B, Laisk A (2003) Deciphering the 820 nm signal: redox state of donor 1579 side and quantum yield of photosystem I in leaves. Photosynth Res 78:1–15. 1580 Okegawa Y, Motohashi K (2020) M-type thioredoxins regulate the PGR5/PGRL1-dependent pathway by forming a 1581 disulfide-linked complex with PGRL1. Plant Cell 32:3866–3883. https://doi.org/10.1105/tpc.20.00304 1582 Okegawa Y, Sakamoto W (2026) The thioredoxin system protects PSI from photoinhibition in coordination with PSI 1583 cyclic electron transport under fluctuating light conditions. Plant Cell Physiol 67:872–883. 1584 https://doi.org/10.1093/pcp/pcaf172 1585 Okegawa Y, Sato N, Nakakura R, Murai R, Sakamoto W, Motohashi K (2023) x- and y-type thioredoxins maintain 1586 redox homeostasis on photosystem I acceptor side under fluctuating light. Plant Physiol 193:2498–2512. 1587 https://doi.org/10.1093/plphys/kiad466 1588 Öquist G, Greer DH, Ögren E (1987) Light stress at low temperature. In: Kyle DJ, Osmond CB, Arntzen CJ (eds) 1589 _Photoinhibition_ . Elsevier, Amsterdam, pp 67–87. 1590 Öquist G, Anderson JM, McCaffery S, Chow WS (1992) Mechanistic differences in photoinhibition of sun and shade 1591 plants. Planta 188:422–431. 1592 Orr GR, Raison JK (1987) Compositional and thermal properties of thylakoid polar lipids of _Nerium oleander_ L. in 1593 relation to chilling sensitivity. Plant Physiol 84:88–92. [https://doi.org/10.1104/pp.84.1.88](https://doi.org/10.1104/pp.84.1.88) 34 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.02.763831; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.02.763831) preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 1594 Orr GR, Raison JK (1990) The effect of changing the composition of phosphatidylglycerol from thylakoid polar 1595 lipids of oleander and cucumber on the temperature of the transition related to chilling injury. Planta 181:137– 1596 143. 1597 Otani T, Yamamoto H, Shikanai T (2017) Stromal loop of Lhca6 is responsible for the linker function required for 1598 the NDH–PSI supercomplex formation. Plant Cell Physiol 58:851–861. https://doi.org/10.1093/pcp/pcx009 1599 Otani T, Kato Y, Shikanai T (2018) Specific substitutions of light-harvesting complex I proteins associated with 1600 photosystem I are required for supercomplex formation with chloroplast NADH dehydrogenase-like complex. 1601 Plant J 94:122–130. https://doi.org/10.1111/tpj.13846 1602 Ozawa S-I, Zhang G, Sakamoto W (2024) Dysfunction of chloroplast protease activity mitigates _pgr5_ phenotype in 1603 the green algae _Chlamydomonas reinhardtii_ . Plants 13:606. [https://doi.org/10.3390/plants13050606](https://doi.org/10.3390/plants13050606) 1604 Pätsikkä E, Kairavuo M, Sersen F, Aro E-M, Tyystjärvi E (2002) Excess copper predisposes photosystem II to 1605 photoinhibition in vivo by outcompeting iron and causing decrease in leaf chlorophyll. Plant Physiol 129:1359– 1606 1367. [https://doi.org/10.1104/pp.004788](https://doi.org/10.1104/pp.004788) 1607 Peeler TC, Naylor AW (1988) A comparison of the effects of chilling on thylakoid electron transfer in pea ( _Pisum_ 1608 _sativum_ L.) and cucumber ( _Cucumis sativus_ L.). Plant Physiol 86:147–151. 1609 Petronek MS, Spitz DR, Allen BG (2021) Iron–sulfur cluster biogenesis as a critical target in cancer. Antioxidants 1610 10:1458. [https://doi.org/10.3390/antiox10091458](https://doi.org/10.3390/antiox10091458) 1611 Powles SB (1984) Photoinhibition of photosynthesis induced by visible light. Annu Rev Plant Physiol 35:15–44. 1612 [https://doi.org/10.1146/annurev.pp.35.060184.000311](https://doi.org/10.1146/annurev.pp.35.060184.000311) 1613 Ptushenko VV, Cherepanov DA, Krishtalik LI, Semenov AY (2008) Semi-continuum electrostatic calculations of 1614 redox potentials in photosystem I. Photosynth Res 97:55–74. 1615 Qin Q, Gao R-Y, Li Y-Q, Zhang R, Lu Y, Yamano N, Qin X-C, Li D-H, Zhang J-P (2025) Effect of oxygen exposure 1616 on the triplet excitation dynamics of the monomeric LHCII complex from spinach. J Phys Chem B 129:9312– 1617 9321. 1618 Rajagopal S, Joly D, Gauthier A, Beauregard M, Carpentier R (2005) Protective effect of active oxygen scavengers 1619 on protein degradation and photochemical function in photosystem I submembrane fractions during light stress. 1620 FEBS J 272:892–902. 1621 Rantala S, Lempiäinen T, Gerotto C, Tiwari A, Aro E-M, Tikkanen M (2020) PGR5 and NDH-1 systems do not 1622 function as protective electron acceptors but mitigate the consequences of PSI inhibition. Biochim Biophys Acta 1623 Bioenerg 1861:148154. [https://doi.org/10.1016/j.bbabio.2020.148154](https://doi.org/10.1016/j.bbabio.2020.148154) 1624 Richardson KH, Wright JJ, Šimėnas M, Thiemann J, Esteves AM, McGuire G, Myers WK, Morton JJL, Hippler M, 1625 Nowaczyk MM, Hanke GT, Roessler MM (2021) Functional basis of electron transport within photosynthetic 1626 [complex I. Nat Commun 12:5387. https://doi.org/10.1038/s41467-021-25527-1](https://doi.org/10.1038/s41467-021-25527-1) 1627 Rumberg B, Siggel U (1969) pH changes in the inner phase of the thylakoids during photosynthesis. 1628 [Naturwissenschaften 56:130–132. https://doi.org/10.1007/BF00601025](https://doi.org/10.1007/BF00601025) 1629 Rutherford AW, Osyczka A, Rappaport F (2012) Back-reactions, short-circuits, leaks and other energy wasteful 1630 reactions in biological electron transfer: redox tuning to survive life in O 2 . FEBS Lett 586:603–616. 1631 Satoh H, Ohara Y, Hanke G, Ifuku K, Shimakawa G, Suzuki Y, Makino A, Morigaki K, Miyake C (2025) The 1632 regulation of PSI cyclic electron transport by both plastoquinone and ferredoxin redox states: correlation with the 1633 rate of proton motive force utilization. Front Plant Sci 16:1626163. https://doi.org/10.3389/fpls.2025.1626163 1634 Satoh K (1970a) Mechanism of photoinactivation in photosynthetic systems. I. The dark reaction in photoinactivation. 1635 Plant Cell Physiol 11:15–27. [https://doi.org/10.1093/oxfordjournals.pcp.a074487](https://doi.org/10.1093/oxfordjournals.pcp.a074487) 1636 Satoh K (1970b) Mechanism of photoinactivation in photosynthetic systems. II. The occurrence and properties of 1637 two different types of photoinactivation. Plant Cell Physiol 11:29–38. 1638 [https://doi.org/10.1093/oxfordjournals.pcp.a074493](https://doi.org/10.1093/oxfordjournals.pcp.a074493) 1639 Satoh K (1970c) Mechanism of photoinactivation in photosynthetic systems. III. Site and mode of photoinactivation 1640 in photosystem I. Plant Cell Physiol 11:187–197. [https://doi.org/10.1093/oxfordjournals.pcp.a074500](https://doi.org/10.1093/oxfordjournals.pcp.a074500) 1641 Schöttler MA, Tóth SZ (2014) Photosynthetic complex stoichiometry dynamics in higher plants: environmental 1642 acclimation and photosynthetic flux control. Front Plant Sci 5:188. [https://doi.org/10.3389/fpls.2014.00188](https://doi.org/10.3389/fpls.2014.00188) 1643 Schreiber U (2017) Redox changes of ferredoxin, P700, and plastocyanin measured simultaneously in intact leaves. 1644 Photosynth Res 134:343–360. 1645 Schuller JM, Birrell JA, Tanaka H, Konuma T, Wulfhorst H, Cox N, Schuller SK, Thiemann J, Lubitz W, Sétif P, 1646 Ikegami T, Engel BD, Kurisu G, Nowaczyk MM (2019) Structural adaptations of photosynthetic complex I enable 1647 ferredoxin-dependent electron transfer. Science 363:257–260. [https://doi.org/10.1126/science.aau3613](https://doi.org/10.1126/science.aau3613) 1648 Sejima T, Takagi D, Fukayama H, Makino A, Miyake C (2014) Repetitive short-pulse light mainly inactivates – 1649 photosystems I in sunflower leaves. Plant Cell Physiol 55:1184 1193. https://doi.org/10.1093/pcp/pcu061 1650 Shaku K, Shimakawa G, Hashiguchi M, Miyake C (2016) Reduction-induced suppression of electron flow (RISE) 35 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.02.763831; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.02.763831) preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 1651 in the photosynthetic electron transport system of _Synechococcus elongatus_ PCC 7942. Plant Cell Physiol 1652 57:1443–1453. https://doi.org/10.1093/pcp/pcv198 1653 Shen L, Tang K, Wang W, Wang C, Wu H, Mao Z, An S, Chang S, Kuang T, Shen J-R, Han G, Zhang X (2022) 1654 Architecture of the chloroplast PSI–NDH supercomplex in _Hordeum vulgare_ . Nature 601:649–654. 1655 [https://doi.org/10.1038/s41586-021-04277-6](https://doi.org/10.1038/s41586-021-04277-6) 1656 Shi Y-P, Gao R-Y, Wang H-Y, Zou J-W, Yamano N, Dong F-Q, Zheng J, Zhang J-P (2026) Lipid membrane regulation 1657 of chlorophyll triplet excitation quenching in photosynthetic light-harvesting complex LHCII. J Phys Chem B 1658 130:690–701. [https://doi.org/10.1021/acs.jpcb.5c06547](https://doi.org/10.1021/acs.jpcb.5c06547) 1659 Shikanai T (2007) Cyclic electron transport around photosystem I: genetic approaches. Annu Rev Plant Biol 58:199– 1660 217. [https://doi.org/10.1146/annurev.arplant.58.091406.110525](https://doi.org/10.1146/annurev.arplant.58.091406.110525) 1661 Shikanai T (2016) Chloroplast NDH: a different enzyme with a structure similar to that of respiratory NADH 1662 dehydrogenase. Biochim Biophys Acta Bioenerg 1857:1015–1022. 1663 Shikanai T, Endo T, Hashimoto T, Yamada Y, Asada K, Yokota A (1998) Directed disruption of the tobacco _ndhB_ 1664 gene impairs cyclic electron flow around photosystem I. Proc Natl Acad Sci USA 95:9705–9709. 1665 [https://doi.org/10.1073/pnas.95.16.9705](https://doi.org/10.1073/pnas.95.16.9705) 1666 Shikanai T, Ieda H, Kobayashi Y, Tamura MN (2025) The chloroplast NADH dehydrogenase-like complex: 1667 evolutionary considerations. Plant Cell Physiol 66:1525–1535. 1668 Shimakawa G, Miyake C (2018) Oxidation of P700 ensures robust photosynthesis. Front Plant Sci 9:1617. 1669 [https://doi.org/10.3389/fpls.2018.01617](https://doi.org/10.3389/fpls.2018.01617) 1670 Shimakawa G, Shaku K, Miyake C (2018) Reduction-induced suppression of electron flow (RISE) is relieved by 1671 non-ATP-consuming electron flow in _Synechococcus elongatus_ PCC 7942. Front Microbiol 9:886. 1672 Shimakawa G, Müller P, Miyake C, Krieger-Liszkay A (2024) Photo-oxidative damage of photosystem I by repetitive 1673 flashes and chilling stress in cucumber leaves. Biochim Biophys Acta Bioenerg 1865:149490. 1674 Shuvalov V, Nuijs A, van Gorkom HJ, Smit H, Duysens LNM (1986) Picosecond absorbance changes upon selective 1675 excitation of the primary electron donor P-700 in photosystem I. Biochim Biophys Acta Bioenerg 850:319–323. 1676 [https://doi.org/10.1016/0005-2728(86)90187-8](https://doi.org/10.1016/0005-2728(86)90187-8) 1677 Smith H (1982) Light quality, photoperception, and plant strategy. Annu Rev Plant Physiol 33:481–518. 1678 [https://doi.org/10.1146/annurev.pp.33.060182.002405](https://doi.org/10.1146/annurev.pp.33.060182.002405) 1679 Šnyrychová I, Pospíšil P, Nauš J (2006) Reaction pathways involved in the production of hydroxyl radicals in 1680 thylakoid membrane: EPR spin-trapping study. Photochem Photobiol Sci 5:472–476. 1681 Sonoike K (1996a) Degradation of _psaB_ gene product, the reaction center subunit of photosystem I, is caused during 1682 photoinhibition of photosystem I: possible involvement of active oxygen species. Plant Sci 115:157–164. 1683 Sonoike K (1996b) Photoinhibition of photosystem I: its physiological significance in the chilling sensitivity of plants. 1684 Plant Cell Physiol 37:239–247. https://doi.org/10.1093/oxfordjournals.pcp.a028938 1685 Sonoike K (1999) The different roles of chilling temperatures in the photoinhibition of photosystem I and 1686 photosystem II. J Photochem Photobiol B Biol 48:136–141. https://doi.org/10.1016/S1011-1344(99)00030-5 1687 Sonoike K (2011) Photoinhibition of photosystem I. Physiol Plant 142:56–64. 1688 Sonoike K (2025) Photoinhibition and protection of photosystem I. Plant Cell Physiol 66:1562–1574. 1689 https://doi.org/10.1093/pcp/pcaf079 1690 Sonoike K, Terashima I (1994) Mechanism of photosystem I photoinhibition in leaves of _Cucumis sativus_ L. Planta 1691 194:287–293. 1692 Sonoike K, Terashima I, Iwaki M, Itoh S (1995) Destruction of photosystem I iron–sulfur centers in leaves of _Cucumis_ 1693 _sativus_ L. by weak illumination at chilling temperatures. FEBS Lett 362:235–238. 1694 Sonoike K, Kamo M, Hihara Y, Hiyama T, Enami I (1997) The mechanism of the degradation of _psaB_ gene product, 1695 one of the photosynthetic reaction center subunits of photosystem I, upon photoinhibition. Photosynth Res 53:55– 1696 63. https://doi.org/10.1023/A:1005852330671 1697 Strand DD, Livingston AK, Satoh-Cruz M, Froehlich JE, Maurino VG, Kramer DM (2015) Activation of cyclic 1698 electron flow by hydrogen peroxide in vivo. Proc Natl Acad Sci USA 112:5539–5544. 1699 [https://doi.org/10.1073/pnas.1418223112](https://doi.org/10.1073/pnas.1418223112) 1700 Strand DD, Fisher N, Kramer DM (2017) The higher plant plastid NAD(P)H dehydrogenase-like complex (NDH) is 1701 a high efficiency proton pump that increases ATP production by cyclic electron flow. J Biol Chem 292:11850– 1702 11860. https://doi.org/10.1074/jbc.M116.770792 1703 Su X, Cao D, Pan X, Shi L, Liu Z, Dall’Osto L, Bassi R, Zhang X, Li M (2022) Supramolecular assembly of 1704 chloroplast NADH dehydrogenase-like complex with photosystem I from _Arabidopsis thaliana_ . Mol Plant 1705 15:454–467. [https://doi.org/10.1016/j.molp.2022.01.020](https://doi.org/10.1016/j.molp.2022.01.020) 1706 Suorsa M, Järvi S, Grieco M, Nurmi M, Pietrzykowska M, Rantala M, Kangasjärvi S, Paakkarinen V, Tikkanen M, 1707 Jansson S, Aro E-M (2012) PROTON GRADIENT REGULATION5 is essential for proper acclimation of 36 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.02.763831; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.02.763831) preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 1708 _Arabidopsis_ photosystem I to naturally and artificially fluctuating light conditions. Plant Cell 24:2934–2948. 1709 [https://doi.org/10.1105/tpc.112.097162](https://doi.org/10.1105/tpc.112.097162) 1710 Tagawa K, Tsujimoto HY, Arnon DI (1963) Role of chloroplast ferredoxin in the energy conversion process of 1711 photosynthesis. Proc Natl Acad Sci USA 49:567–572. [https://doi.org/10.1073/pnas.49.4.567](https://doi.org/10.1073/pnas.49.4.567) 1712 Takabayashi A, Kishine M, Asada K, Endo T, Sato F (2005) Differential use of two cyclic electron flows around 1713 photosystem I for driving CO₂-concentration mechanism in C₄ photosynthesis. Proc Natl Acad Sci USA 1714 102:16898–16903. [https://doi.org/10.1073/pnas.0507095102](https://doi.org/10.1073/pnas.0507095102) 1715 Takagi D, Takumi S, Hashiguchi M, Sejima T, Miyake C (2016) Superoxide and singlet oxygen produced within the 1716 thylakoid membranes both cause photosystem I photoinhibition. Plant Physiol 171:1626–1634. 1717 https://doi.org/10.1104/pp.16.00246 1718 Takagi D, Amako K, Hashiguchi M, Fukaki H, Ishizaki K, Goh T, Fukao Y, Sano R, Kurata T, Demura T, Sawa S, 1719 Miyake C (2017a) Chloroplastic ATP synthase builds up a proton motive force preventing production of reactive 1720 oxygen species in photosystem I. Plant J 91:306–324. [https://doi.org/10.1111/tpj.13566](https://doi.org/10.1111/tpj.13566) 1721 Takagi D, Ishizaki K, Hanawa H, Mabuchi T, Shimakawa G, Yamamoto H, Miyake C (2017b) Diversity of strategies 1722 for escaping reactive oxygen species production within photosystem I among land plants: P700 oxidation system 1723 is prerequisite for alleviating photoinhibition in photosystem I. Physiol Plant 161:56–74. 1724 https://doi.org/10.1111/ppl.12562 1725 Takagi D, Ihara H, Takumi S, Miyake C (2019) Growth light environment changes the sensitivity of photosystem I 1726 photoinhibition depending on common wheat cultivars. Front Plant Sci 10:686. 1727 [https://doi.org/10.3389/fpls.2019.00686](https://doi.org/10.3389/fpls.2019.00686) 1728 Takagi D, Kishie A, Ifuku K (2026) Photosystem I photoinhibition attenuates LHCI-dependent light-harvesting 1729 activity in spinach leaves. Plant Physiol 201:kiag251. [https://doi.org/10.1093/plphys/kiag251](https://doi.org/10.1093/plphys/kiag251) 1730 Takeuchi K, Che Y, Nakano T, Miyake C, Ifuku K (2022) The ability of P700 oxidation in photosystem I reflects 1731 [chilling stress tolerance in cucumber. J Plant Res 135:681–692. https://doi.org/10.1007/s10265-022-01404-w](https://doi.org/10.1007/s10265-022-01404-w) 1732 Takeuchi K, Ochiai K, Kobayashi M, Kuroda K, Ifuku K (2024) Light-chilling stress causes hyper-accumulation of 1733 iron in shoot, exacerbating leaf oxidative damage in cucumber. Plant Cell Physiol 65:1873–1887. 1734 https://doi.org/10.1093/pcp/pcae111 1735 Takeuchi K, Harimoto S, Che Y, Kumazawa M, Satoh H, Maekawa S, Miyake C, Ifuku K (2025a) The protective 1736 role of chloroplast NADH dehydrogenase-like complex (NDH) against PSI photoinhibition under chilling stress. 1737 New Phytol 248:2262–2279. [https://doi.org/10.1111/nph.70573](https://doi.org/10.1111/nph.70573) 1738 Takeuchi K, Harimoto S, Maekawa S, Miyake C, Ifuku K (2025b) PSII photoinhibition as a protective strategy: 1739 maintaining an oxidative state of PSI by suppressing PSII activity under environmental stress. Physiol Plant 1740 177:e70392. [https://doi.org/10.1111/ppl.70392](https://doi.org/10.1111/ppl.70392) 1741 Takeuchi K, Harimoto S, Ifuku K (2026) The PSI–NDH supercomplex prevents chilling-induced PSI photoinhibition. 1742 New Phytol. https://doi.org/10.1111/nph.71594 1743 Tanigawa K, Kodama H, Okegawa Y, Shikanai T, Yamori W (2026) Moderate overexpression of _PROTON_ 1744 _GRADIENT REGULATION 5_ improves photosynthetic performance and plant growth under fluctuating low light 1745 in _Arabidopsis thaliana_ . Plant J 127:e71120. https://doi.org/10.1111/tpj.71120 1746 Teicher HB, Møller BL, Scheller HV (2000) Photoinhibition of photosystem I in field-grown barley ( _Hordeum_ 1747 _vulgare_ L.): induction, recovery and acclimation. Photosynth Res 64:53–61. 1748 [https://doi.org/10.1023/A:1026524302191](https://doi.org/10.1023/A:1026524302191) 1749 Terashima I, Inoue Y (1985) Palisade tissue chloroplasts and spongy tissue chloroplasts in spinach: biochemical and 1750 ultrastructural differences. Plant Cell Physiol 26:63–75. 1751 Terashima I, Huang L-K, Osmond CB (1989) Effects of leaf chilling on thylakoid functions, measured at room 1752 temperature, in _Cucumis sativus_ L. and _Oryza sativa_ L. Plant Cell Physiol 30:841–850. 1753 Terashima I, Kashino Y, Katoh S (1991a) Exposure of leaves of _Cucumis sativus_ L. to low temperatures in the light 1754 causes uncoupling of thylakoids I. Studies with isolated thylakoids. Plant Cell Physiol 32:1267–1274. 1755 Terashima I, Sonoike K, Kawazu T, Katoh S (1991b) Exposure of leaves of _Cucumis sativus_ L. to low temperatures 1756 in the light causes uncoupling of thylakoids II. Non-destructive measurements with intact leaves. Plant Cell 1757 Physiol 32:1275–1283. 1758 Terashima I, Funayama S, Sonoike K (1994) The site of photoinhibition in leaves of _Cucumis sativus_ L. at low 1759 temperatures is photosystem I, not photosystem II. Planta 193:300–306. https://doi.org/10.1007/BF00192544 1760 Terashima I, Noguchi K, Itoh-Nemoto T, Park Y-M, Kubo A, Tanaka K (1998) The cause of PSI photoinhibition at 1761 low temperatures in leaves of _Cucumis sativus_, a chilling-sensitive plant. Physiol Plant 103:295–303. 1762 [https://doi.org/10.1034/j.1399-3054.1998.1030301.x](https://doi.org/10.1034/j.1399-3054.1998.1030301.x) 1763 Terashima I, Fujita T, Inoue T, Chow WS, Oguchi R (2009) Green light drives leaf photosynthesis more efficiently 1764 than red light in strong white light: revisiting the enigmatic question of why leaves are green. Plant Cell Physiol 37 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.02.763831; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.02.763831) preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 1765 50:684–697. https://doi.org/10.1093/pcp/pcp034 1766 Terashima I, Matsuo M, Suzuki Y, Yamori W, Kono M (2021) Photosystem I in low light-grown leaves of _Alocasia_ 1767 _odora_, a shade-tolerant plant, is resistant to fluctuating light-induced photoinhibition. Photosynth Res 149:69–82. 1768 https://doi.org/10.1007/s11120-021-00832-4 1769 Terashima I, Oguchi R, Atsuzawa K, Kaneko Y, Kono M (2025) Excitation spillover from PSII to PSI measured in 1770 leaves at 77 K. Plant Cell Physiol 66:358–373. https://doi.org/10.1093/pcp/pcaf002 1771 Theeuwen TPJM, Lawson AW, Tijink D, Fornaguera F, Becker FFM, Caracciolo L, Fisher N, Kramer DM, Wijnker 1772 E, Harbinson J, Aarts MGM (2022) The NDH complex reveals a trade-off that constrains maximising 1773 photosynthesis in _Arabidopsis thaliana_ . bioRxiv. [https://doi.org/10.1101/2022.11.13.516254](https://doi.org/10.1101/2022.11.13.516254) 1774 Thormählen I, Meitzel T, Groysman J, Öchsner AB, von Roepenack-Lahaye E, Naranjo B, Cejudo FJ, Geigenberger 1775 P (2015) Thioredoxin f1 and NADPH-dependent thioredoxin reductase C have overlapping functions in regulating 1776 photosynthetic metabolism and plant growth in response to varying light conditions. Plant Physiol 169:1766– 1777 1786. https://doi.org/10.1104/pp.15.01122 1778 Thormählen I, Zupok A, Rescher J, Leger J, Weissenberger S, Groysman J, Orwat A, Chatel-Innocenti G, Issakidis1779 Bourguet E, Armbruster U, Geigenberger P (2017) Thioredoxins play a crucial role in dynamic acclimation of 1780 photosynthesis in fluctuating light. Mol Plant 10:168–182. https://doi.org/10.1016/j.molp.2016.11.012 1781 Tikhonov AN (2024) The cytochrome _b_ ₆ _f_ complex: plastoquinol oxidation and regulation of electron transport in 1782 chloroplasts. Photosynth Res 159:203–227. https://doi.org/10.1007/s11120-023-01034-w 1783 Tikhonov AN, Khomutov GB, Ruuge EK, Blumenfeld LA (1981) Electron transport control in chloroplasts. Effects 1784 of photosynthetic control monitored by the intrathylakoid pH. Biochim Biophys Acta Bioenerg 637:321–333. 1785 [https://doi.org/10.1016/0005-2728(81)90171-7](https://doi.org/10.1016/0005-2728(81)90171-7) 1786 Tikkanen M, Grieco M, Kangasjärvi S, Aro E-M (2010) Thylakoid protein phosphorylation in higher plant 1787 chloroplasts optimizes electron transfer under fluctuating light. Plant Physiol 152:723–735. 1788 https://doi.org/10.1104/pp.109.150250 1789 Tikkanen M, Mekala NR, Aro E-M (2014) Photosystem II photoinhibition-repair cycle protects photosystem I from 1790 irreversible damage. Biochim Biophys Acta Bioenerg 1837:210–215. 1791 https://doi.org/10.1016/j.bbabio.2013.10.001 1792 Tiwari A, Mamedov F, Grieco M, Suorsa M, Jajoo A, Styring S, Tikkanen M, Aro E-M (2016) Photodamage of iron– 1793 sulphur clusters in photosystem I induces non-photochemical energy dissipation. Nat Plants 2:16035. 1794 [https://doi.org/10.1038/nplants.2016.35](https://doi.org/10.1038/nplants.2016.35) 1795 Tiwari A, Mamedov F, Fitzpatrick D, Gunell S, Tikkanen M, Aro E-M (2024) Differential FeS cluster photodamage 1796 plays a critical role in regulating excess electron flow through photosystem I. Nat Plants 10:1592–1603. 1797 https://doi.org/10.1038/s41477-024-01780-2 1798 Tjus SE, Møller BL, Scheller HV (1999) Photoinhibition of photosystem I damages both reaction centre proteins 1799 PSI-A and PSI-B and acceptor-side located small photosystem I polypeptides. Photosynth Res 60:75–86. 1800 [https://doi.org/10.1023/A:1006283618695](https://doi.org/10.1023/A:1006283618695) 1801 Trissl HW (1997) Determination of the quenching efficiency of the oxidized primary donor of photosystem I, P700⁺: 1802 implications for the trapping mechanisms. Photosynth Res 54:237–240. 1803 [https://doi.org/10.1023/A:1005981016835](https://doi.org/10.1023/A:1005981016835) 1804 Tsuyama M, Kobayashi Y (2009) Reduction of the primary donor P700 of photosystem I during steady-state 1805 photosynthesis under low light in _Arabidopsis_ . Photosynth Res 99:37–47. [https://doi.org/10.1007/s11120-008-](https://doi.org/10.1007/s11120-008-9379-x) 1806 [9379-x](https://doi.org/10.1007/s11120-008-9379-x) 1807 Tullberg A, Alexciev K, Pfannschmidt T, Allen JF (2000) Photosynthetic electron flow regulates transcription of the 1808 _psaB_ gene in pea ( _Pisum sativum_ L.) chloroplasts through the redox state of the plastoquinone pool. Plant Cell 1809 [Physiol 41:1045–1054. https://doi.org/10.1093/pcp/pcd031](https://doi.org/10.1093/pcp/pcd031) 1810 Tyystjärvi E, Aro E-M (1996) The rate constant of photoinhibition, measured in lincomycin-treated leaves, is directly 1811 proportional to light intensity. Proc Natl Acad Sci USA 93:2213–2218. [https://doi.org/10.1073/pnas.93.5.2213](https://doi.org/10.1073/pnas.93.5.2213) 1812 Ueda M, Kuniyoshi T, Yamamoto H, Sugimoto K, Ishizaki K, Kohchi T, Nishimura Y, Shikanai T (2012) Composition 1813 and physiological function of the chloroplast NADH dehydrogenase-like complex in _Marchantia polymorpha_ . 1814 Plant J 72:683–693. [https://doi.org/10.1111/j.1365-313X.2012.05115.x](https://doi.org/10.1111/j.1365-313X.2012.05115.x) 1815 Urzinger S, Avramova V, Frey M, Urbany C, Scheuermann D, Presterl T, Reuscher S, Ernst K, Mayer M, Marcon C, 1816 Hochholdinger F, Brajkovic S, Ordas B, Westhoff P, Ouzunova M, Schön CC (2025) Embracing native diversity 1817 to enhance the maximum quantum efficiency of photosystem II in maize. Plant Physiol 197:kiae670. 1818 [https://doi.org/10.1093/plphys/kiae670](https://doi.org/10.1093/plphys/kiae670) 1819 Vilyanen DV, Kozuleva MA (2025) Photosynthetic control and its role in protection of photosystem I against 1820 photoinhibition. Biochemistry (Mosc) 90:915–936. [https://doi.org/10.31857/S0320972525070034](https://doi.org/10.31857/S0320972525070034) 1821 Wada S, Yamamoto H, Suzuki Y, Yamori W, Shikanai T, Makino A (2018) Flavodiiron protein substitutes for cyclic 38 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.02.763831; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.02.763831) preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 1822 electron flow without competing CO2 assimilation in rice. Plant Physiol 176:1509–1518. 1823 https://doi.org/10.1104/pp.17.01335 1824 Wada S, Amako K, Miyake C (2021) Identification of a novel mutation exacerbated the PSI photoinhibition in 1825 _pgr5_ / _pgrl1_ mutants; caution for overestimation of the phenotypes in _Arabidopsis pgr5-1_ mutant. Cells 10:2884. 1826 https://doi.org/10.3390/cells10112884 1827 Wakasugi T, Tsudzuki J, Ito S, Nakashima K, Tsudzuki T, Sugiura M (1994) Loss of all _ndh_ genes as determined by 1828 sequencing the entire chloroplast genome of the black pine _Pinus thunbergii_ . Proc Natl Acad Sci USA 91:9794– 1829 9798. [https://doi.org/10.1073/pnas.91.21.9794](https://doi.org/10.1073/pnas.91.21.9794) 1830 Wang C, Yamamoto H, Shikanai T (2015) Role of cyclic electron transport around photosystem I in regulating proton 1831 motive force. Biochim Biophys Acta Bioenerg 1847:931–938. [https://doi.org/10.1016/j.bbabio.2014.11.013](https://doi.org/10.1016/j.bbabio.2014.11.013) 1832 Wang P, Duan W, Takabayashi A, Endo T, Shikanai T, Ye JY, Mi H (2006) Chloroplastic NAD(P)H dehydrogenase 1833 in tobacco leaves functions in alleviation of oxidative damage caused by temperature stress. Plant Physiol 1834 141:465–474. https://doi.org/10.1104/pp.105.070490 1835 Wardman P (1990) Bioreductive activation of quinones: redox properties and thiol reactivity. Free Radic Res 1836 Commun 8:219–229. 1837 West KR, Wiskich JT (1968) Photosynthetic control by isolated pea chloroplasts. Biochem J 109:527–532. 1838 [https://doi.org/10.1042/bj1090527](https://doi.org/10.1042/bj1090527) 1839 Yamada Y, Suzuki K, Yanagishita H, Noguchi K (2024) Roles of mitochondrial alternative oxidase in photosynthetic 1840 electron transport in illuminated leaves of _Arabidopsis thaliana_ at low temperature. J Biosci 49:61. 1841 Yamamoto H, Shikanai T (2013) In planta mutagenesis of Src homology 3 domain-like fold of NdhS, a ferredoxin1842 binding subunit of the chloroplast NADH dehydrogenase-like complex in _Arabidopsis_ : a conserved Arg-193 plays 1843 a critical role in ferredoxin binding. J Biol Chem 288:36328–36337. [https://doi.org/10.1074/jbc.M113.511584](https://doi.org/10.1074/jbc.M113.511584) 1844 Yamamoto H, Peng L, Fukao Y, Shikanai T (2011) An Src homology 3 domain-like fold protein forms a ferredoxin 1845 binding site for the chloroplast NADH dehydrogenase-like complex in _Arabidopsis_ . Plant Cell 23:1480–1493. 1846 https://doi.org/10.1105/tpc.110.080291 1847 Yamamoto H, Takahashi S, Badger MR, Shikanai T (2016) Artificial remodelling of alternative electron flow by 1848 flavodiiron proteins in _Arabidopsis_ . Nat Plants 2:16012. [https://doi.org/10.1038/nplants.2016.12](https://doi.org/10.1038/nplants.2016.12) 1849 Yamori W (2016) Photosynthetic response to fluctuating environments and photoprotective strategies under abiotic – 1850 stress. J Plant Res 129:379 395. https://doi.org/10.1007/s10265-016-0816-1 1851 Yamori W, Shikanai T (2016) Physiological functions of cyclic electron transport around photosystem I in sustaining – 1852 photosynthesis and plant growth. Annu Rev Plant Biol 67:81 106. https://doi.org/10.1146/annurev-arplant 1853 043015-112002 1854 Yamori W, Sakata N, Suzuki Y, Shikanai T, Makino A (2011) Cyclic electron flow around photosystem I via 1855 chloroplast NAD(P)H dehydrogenase (NDH) complex performs a significant physiological role during 1856 photosynthesis and plant growth at low temperature in rice. Plant J 68:966–976. https://doi.org/10.1111/j.13651857 313X.2011.04747.x 1858 Yamori W, Makino A, Shikanai T (2016) A physiological role of cyclic electron transport around photosystem I in 1859 sustaining photosynthesis under fluctuating light in rice. Sci Rep 6:20147. https://doi.org/10.1038/srep20147 1860 Yokono M, Noda C, Minagawa J (2024) Spillover in the direct-type PSI-PSII megacomplex isolated from 1861 _Arabidopsis_ _thaliana_ is regulated by pH. Biochim Biophys Acta Bioenerg 1865:149012. 1862 https://doi.org/10.1016/j.bbabio.2023.149012 1863 Yoshida K, Hisabori T (2023) Current insights into the redox regulation network in plant chloroplasts. Plant Cell – 1864 Physiol 64:704 715. https://doi.org/10.1093/pcp/pcad049 1865 Zhang Q, Tian S, Chen G, Tang Q, Zhang Y, Fleming AJ, Zhu X, Wang P (2024) Regulatory NADH dehydrogenase‐ 1866 like complex optimizes C 4 photosynthetic carbon flow and cellular redox in maize. New Phytol 241:82–101. 1867 Zhang Y, Fan Y, Lv X, Zeng X, Zhang Q, Wang P (2023) Deficiency in NDH-cyclic electron transport retards heat 1868 acclimation of photosynthesis in tobacco over day and night shift. Front Plant Sci 14:1267191. 1869 https://doi.org/10.3389/fpls.2023.1267191 1870 Zhang Y, He S, Chen G (2025) Metabolomics of chloroplasts combined with photosynthetic properties reveal low1871 temperature stress responses in sugarcane ( _Saccharum officinarum_ L.). J Plant Growth Regul 1:17. 1872 Zhang Z-S, Jin L-Q, Li Y-T, Tikkanen M, Li Q-M, Ai X-Z, Gao H-Y (2016) Ultraviolet-B radiation (UV-B) relieves 1873 chilling-light-induced PSI photoinhibition and accelerates the recovery of CO 2 assimilation in cucumber 1874 ( _Cucumis sativus_ L.) leaves. Sci Rep 6:34455. https://doi.org/10.1038/srep34455 1875 Zhen S, Bugbee B (2020) Far-red photons have equivalent efficiency to traditional photosynthetic photons: 1876 implications for redefining photosynthetically active radiation. Plant Cell Environ 43:1259–1272. 1877 https://doi.org/10.1111/pce.13730 39 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.02.763831; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.02.763831) preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 1878 Zheng M, Jiang Y, Ran Z, Liang S, Xiao T, Li X, Ma W (2025) A cyanobacteria-derived intermolecular salt bridge 1879 stabilizes photosynthetic NDH-1 and prevents oxidative stress. Commun Biol 8:172. 1880 https://doi.org/10.1038/s42003-025-07556-4 1881 Zhou Q, Yamamoto H, Shikanai T (2023) Distinct contribution of two cyclic electron transport pathways to P700 1882 oxidation. Plant Physiol 192:326–341. https://doi.org/10.1093/plphys/kiac557 1883 Zhou YH, Huang LF, Zhang Y, Shi K, Yu JQ, Nogués S (2007) Chill-induced decrease in capacity of RuBP 1884 carboxylation and associated H 2 O 2 accumulation in cucumber leaves are alleviated by grafting onto figleaf gourd. 1885 Ann Bot 100:839–848. 1886 Zivcak M, Brestic M, Kunderlikova K, Sytar O, Allakhverdiev SI (2015) Repetitive light pulse-induced 1887 photoinhibition of photosystem I severely affects CO2 assimilation and photoprotection in wheat leaves. 1888 Photosynth Res 126:449–463. https://doi.org/10.1007/s11120-015-0121-1 1889 1890 1891 1892 1893 1894 1895 1896 40 1897 **Table 1** Properties of the two CEF-PSI pathways relevant to their division of roles **Property** **PGR5/PGRL1-dependent pathway** **NDH-dependent pathway** **Key references** Inhibitor antimycin A-sensitive antimycin A-insensitive Munekage et al. 2002, 2004 Protons per electron, Fd → PQ none known ~2 H [+] /e [−] (proton pump) Laughlin et al. 2019; Strand et al. 2017 step Protons per electron, including the Ermakova et al. 2024a; Strand et al. 2 H [+] /e [−] ~4 H [+] /e [−] cytochrome _b_ 6 _f_ Q-cycle 2017 Abundance (copies per PSI, ~0.07 (PGR5) ~0.02 Ermakova et al. 2024a _Arabidopsis_ ) 1–2 orders of magnitude slower than the direct Fd– Relative speed [a] fast cytochrome _b_ 6 _f_ route; ~4–10 e [−] s [−1] PSI [−1] (estimate) Ermakova et al. 2024a; Nawrocki et al. 2019; Strand et al. 2017 Contribution to pmf under steady red light ( _Arabidopsis_ ) Proton conductivity of the ATP synthase ( _g_ H [+] ) large (30–47% lower pmf in _pgr5_ at ≥ 138 µmol small (8–14% lower pmf in NDH mutants) Wang et al. 2015 photons m [−2] s [−1] ) elevated _g_ H [+] in _pgr5_ under high light suggests that PGR5 lowers _g_ H [+] ; normalized by FLV in _pgr5_, so the effect may be indirect Degen et al. 2023; Rantala et al. 2020; unaffected in NDH mutants Suorsa et al. 2012; Wang et al. 2015 within seconds of high light; relaxes in the dark Activation with a half-time of ~20 s by H 2 O 2 over minutes (half-time ~20 min); protein content increases under prolonged H 2 O 2 stress and after PSI damage ~670 µM, partially competitive; activity retained at high ATP 41 Joliot and Johnson 2011; Rantala et al. 2020; Strand et al. 2015 Cleland and Bendall 1992; Fisher et al. 2019 Inhibition by ATP ( _K_ i ) [b] ~240 µM, competitive with Fd; no effect of ATP or ADP up to 1 mM in an earlier assay in pea thylakoids Apparent _K_ m for Fd [b] ~1 µM; higher in older assays (optimum ~10 µM in air; saturation at ~75 µM reduced Fd under anaerobic conditions) ~2 µM; high affinity requires NdhS (Arg-193) Arnon and Chain 1975; Cleland and Bendall 1992; Fisher et al. 2019; Yamamoto and Shikanai 2013; Yamamoto et al. 2011 forward reaction requires Δ _E_ h - _n_ - Δp; shown _in_ Thermodynamic limit by Δp none expected (no proton pump) Fisher et al. 2019; Strand et al. 2017 _vitro_, uncertain _in vivo_ Supercomplex with PSI — formed via Lhca5 and Lhca6; stabilizes NDH (partially disassembled in chilled cucumber) but is not required for its activity Otani et al. 2017; Shikanai et al. 2025; Takeuchi et al. 2026; Ueda et al. 2012 1898 a Calculated from the NDH content (approximately 0.02 per PSI in _Arabidopsis_ and 0.05 per PSI in tobacco; Ermakova et al. 2024a; Strand et al. 2017) and a maximal turnover similar to that of 1899 respiratory complex I (approximately 200 e [−] s [−1] ; Strand et al. 2017); the lower value applies to _Arabidopsis_ . The flux has not been measured directly. 1900 b Measured in uncoupled thylakoids, that is, in the absence of a pmf, from _A. hybridus_ with antimycin A (NDH) and from spinach, in which most of the activity is antimycin A-sensitive (Fisher 1901 et al. 2019). The _K_ m of NDH is less certain, because the rate was not saturated at the highest Fd concentration tested (~5.5 µM). The older assays measured only the antimycin A-sensitive pathway, 1902 under other conditions, and cannot be compared directly (Text S1). 1903 1904 1905 1906 1907 42 1908 **Table 2** Two light regimes and the resulting division of roles between the CEF-PSI pathways **Red FL (high-light phase unless stated)** **FR-supplemented FL (both phases)** Main cause of P700 oxidation pmf-driven restriction at cytochrome _b_ 6 _f_ (photosynthetic control) excitation balance (selective PSI excitation) PQ pool reduced oxidized Inputs of photosynthetic control pH input and redox input both engaged pH input only, weak in the low-light phase pmf high in the high-light phase; collapses in _pgr5_ moderate; total pmf not clearly changed by FR Δ _E_ h − _n_ - Δp for NDH [a] near zero; NDH held back large; NDH forward reaction favored PGR5/PGRL1 pathway rate-limiting: builds and holds the pmf not required for P700 oxidation NDH pathway minor in _Arabidopsis_ ; larger in rice main route for recycling electrons at low flux, with high proton efficiency low-light phase: 0.389 / 0.224 high-light phase: 0.528 / 0.232 Y(ND) / Y(NA), wild type low-light phase: 0.006 / 0.432 high-light phase: 0.253 / 0.428 Decisive observation _pgr5_ P m ~20%; _crr2-2_ ~10% worse than the wild type WT P m ~97%; _pgr5_ ~92%; _crr2-2_, _crr4-3_ ~80% Main data sources Kono and Terashima 2016; Kono et al. 2017 Kono et al. 2017, 2020 1909 Y(ND) and Y(NA) are values for wild-type _Arabidopsis_ 10 s after the 100th high-light pulse (low-light phase) and during the pulse (high-light phase) in the FL regime of Kono et al. (2017), 1910 without FR (left) and with FR (right). P m values are those after the FL treatment, expressed relative to the value before the treatment. Other entries summarize §§ 5.2 and 5.3. 1911 a Inferred from the redox–Δp gating hypothesis; not measured _in vivo_ (§ 5.4; Table S2, row 1). 1912 43 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.02.763831; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.02.763831) preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 1913 **Fig. 1** 1914 1915 1916 **Fig. 1 Temporal progression of chilling-induced PSI photoinhibition.** 1917 Within the first 0.5 h of the onset of chilling stress, suppression of the CBB cycle and possible decrease in thylakoid 1918 membrane fluidity cause transient over-reduction of the photosynthetic electron transport chain and ROS production 1919 (1, 2) (Phase 1). These early redox changes and changes in membrane physics would weaken the association of CF1 1920 with the thylakoid membrane, resulting in uncoupling of the chloroplast ATP synthase (3-1), as well as destabilization 1921 of the PSI–NDH supercomplex (3-2). Both changes occur within the first hour of chilling (upper panel). The 1922 weakened association of CF₁ allows H⁺ to leak from the lumen into the stroma, thereby dissipating ΔpH and 1923 weakening photosynthetic control at the Cyt b6fcomplex (4-1). At the same time, destabilization of the PSI–NDH 1924 supercomplex leaves NDH in an inactive monomeric form, reducing NDH-mediated Fd oxidation (4-2) (Phase 2). 1925 The combined impairment of photosynthetic control and Fd oxidation leads to a marked decline in P700 oxidation 1926 capacity between 2 and 5 h after the onset of chilling, accompanied by further ROS production in PSI and damage to 1927 the Fe–S clusters and PSI core proteins (5; lower panel) (Phase 3). 1928 44 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.02.763831; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.02.763831) preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 1929 **Fig. 2** 1930 1931 1932 **Fig. 2 Quenching analysis curves in the cucumber leaf segments after the light-chilling treatment.** 1933 Cucumber ‘Nanshin’ leaf segments were treated at 4 °C and an irradiance of 100 µmol photons m [-2] s [-1] . a, control leaf 1934 segment kept in the dark at 23 °C (RTD); b, leaf segment treated at 4 °C in the light for 30 min (LTL-30 min); c, 1935 LTL-60 min; and d, LTL-120 min. Fluorescence was measured with a Dual-PAM (Walz, Germany) at 23 °C. The 1936 irradiance level of the actinic light (red LEDs at 635 nm) was 100 µmol photons m [-2] s [-1] . As _F_ _m_ ’ levels were nearly 1937 constant in c and d, development of NPQ was suppressed by the chilling treatment in the light for 60 to 120 min. This 1938 may be explained by the loose association of CF 1 of ATP synthase with the thylakoid membrane. The quantum yield 1939 of PSII photochemistry, _Y(II)_ = ( _F_ m ’– _F_ s ’)/ _F_ m ’, measured at 23 °C decreased from 0.68 (control) to 0.1 at 2 h with 1940 the increasing damage to PSI. Reproduced from Kono et al. (2022) with permission from the publisher. 1941 45 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.02.763831; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.02.763831) preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 1942 1943 **Fig. 3** 1944 1945 1946 1947 **Fig. 3 Boxplot of** _**K**_ **pi** **/** _**I**_ **(a) and scatter plots between Y(ND) and** _**K**_ **pi** **/** _**I**_ **(b) and between Y(NA) and** _**K**_ **pi** **/** _**I**_ **(c)** 1948 **produced from meta-analysis of recent PSI photoinhibition studies.** Categories from 1 to 8 in panel (a) indicate 1949 different types of photoinhibition environments as follows: 1, 2, 3 and 4 indicate the data gained under steady-light 1950 conditions using isolated thylakoids, leaves in chilling condition, leaves in moderate temperature, and leaves of _A_ . 1951 _thaliana pgr5_ in moderate temperature, respectively. 5, 6, 7 and 8 indicate the data with leaves in fluctuating light 1952 condition, leaves in intermittent light condition, leaves in fluctuating light condition with very strong light and leaves 1953 in intermittent light condition with very strong light, respectively. We classified treatments with an irradiance of 1954 10,000 µmol photons m⁻² s⁻¹ or higher during the high-light phase as very-strong-light treatments. The data of _A_ . 1955 _thaliana pgr5_ obtained under FL or intermittent light are not included in the boxes of categories 5 and 8 but are shown 1956 as red points. Different letters indicate statistically significant differences among the categories according to the 1957 Games-Howell test. In panel (b), Y(ND) was calculated as proportion of P700 [+] at 1 s from the start of high-light 1958 phase in the fluctuating-light conditions to the maximum P700 [+] . In the case of Takagi et al. 2017b, the maximum 1959 P700 [+] was measured under pulse light (5,000 µmol photons m [-2] s [-1] ) condition without far-red light. In the case of 1960 Kono et al. 2017, Y(ND) was calculated with the data measured at 100th high-light phase, in which we considered 1961 the PSI photoinhibition was still small. In panel (c), Y(NA) was calculated as (P m – P m ’)/P m, where P m means total 1962 P700 measured under saturating pulse with far-red light and P m ’ means the P700 [+] measured under saturating pulse at 1963 a steady state of high-light phase in the fluctuating-light conditions. We should note that the timing of the 1964 measurement was different in panel (b): at 1 s and in panel (c): at a steady state. In panel (b) and (c), the data of 1965 references in categories from 1 to 4 and references in which Y(ND) or Y(NA) was measured at a light intensity 1966 different from that of the high-light phase were not included. 1967 1968 46 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.02.763831; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.02.763831) preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 1969 **Fig. 4** 1970 1971 1972 **Fig. 4 Light quality sets the redox poise and the pmf, and these gate the two CEF-PSI pathways.** ( **a**, **b** ) Schemes 1973 of the electron transport chain in the high-light phase of red FL ( **a** ) and under FR-supplemented FL ( **b** ); only the light 1974 quality differs between the two panels. Filled PQ, reduced pool; open PQ, oxidized pool; arrow thickness indicates 1975 electron flux; dashed arrows and outlines indicate a pathway that is held back or not required. The 2 H⁺/e⁻ shown at 1976 NDH is the stoichiometry of the NDH pump alone. ( **c** ) Position of the two regimes on the axis Δ _E_ ₕ − _n_ - Δp that 1977 governs the NDH forward reaction (redox–Δp gating hypothesis). The high-light phase of red FL is placed at the 1978 static head (Δ _E_ ₕ − _n_ - Δp ≈ 0) rather than far below it, because the loss of NDH still lowers the light-induced pmf by 1979 ~10% under steady red light (Wang et al. 2015). Two reference states are shown for comparison: leaves with elevated 1980 H₂O₂ (the GO5 line), in which NDH-dependent CEF-PSI operates despite a pmf ~2 times that of the wild type, 1981 because the stroma is strongly reduced (Strand et al. 2015); and the dark period after illumination, in which the post1982 illumination fluorescence rise (PIFR) is measured at Δp ≈ 0. The axis is schematic; positions indicate only the sign 1983 of the term and relative distances. ( **d** ) Two-axis summary: the means by which P700 is kept oxidized (rows) and the 1984 pathway that supplies the pmf and ATP when linear flow is low (columns); the H⁺/e⁻ values include the Q-cycle of 1985 the cytochrome _b_ ₆ _f_ complex. Data from Kodama and Yamori (2026), Kono and Terashima (2016), Kono et al. (2017, 1986 2025), Theeuwen et al. (2022), Wang et al. (2015) and Yamori et al. (2011, 2016). CEF-PSI, cyclic electron flow 1987 around PSI; Cyt, cytochrome; Fd, ferredoxin; FL, fluctuating light; FR, far-red light; NDH, NADH dehydrogenase1988 like complex; PC, plastocyanin; pmf, proton motive force; PQ, plastoquinone. 1989 47 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.02.763831; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.02.763831) preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 1990 **Fig. 5** 1991 1992 1993 **Fig. 5** **Proposed mechanisms of ROS production in PSI under different environmental conditions.** ( **a** ) Under 1994 continuous high light, partial damage to F A /F B may occur without leading to irreversible PSI core damage. 1995 Photosynthetic control maintains P700 in its oxidized state (P700⁺), allowing excess electrons to be safely removed 1996 through charge recombination from F X to P700⁺ via the A-branch, even when F A /F B is partially damaged. ( **b** ) Under 1997 repetitive saturating pulses, P700⁺ and reduced A 0 /A 1 are more likely to coexist. Consequently, charge recombination 1998 via the B-branch may generate [3] P700* more frequently and thereby increase [1] O 2 formation, potentially accounting 1999 for preferential damage near the PSI reaction center. ( **c** ) In cucumber, chilling-induced PSI photoinhibition progresses 2000 even under relatively weak light. At low temperature, the decline in NDH activity, together with uncoupling of the 2001 chloroplast H⁺-ATP synthase, impairs ΔpH formation and photosynthetic control. Consequently, over-reduction may 2002 extend beyond F A /F B to F X and A 1 . Because P700⁺ accumulation is also suppressed, safe removal of excess electrons 2003 through charge recombination from F X [–] to P700⁺ becomes less effective. Electrons may therefore leak to O 2 from 2004 multiple acceptors, including A 1 and F A /F B, promoting O 2 [•] ⁻ production. The resulting O 2 [•] ⁻ may directly damage the 2005 Fe–S clusters or undergo nonenzymatic or SOD-catalyzed dismutation to form H 2 O 2 . The H 2 O 2 thus produced can 2006 generate •OH through Fenton reactions involving the Fe–S clusters or free Fe [2+], leading to further oxidative damage. 48 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.02.763831; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.02.763831) preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 1 _Journal of Plant Research_ 2 **Supplementary Information** 3 **PSI photoinhibition and protection at low temperatures and in fluctuating light: An ecophysiological vista** 4 Ko Takeuchi [1], Riichi Oguchi [2], Kazuhiko Sugimoto [3], Ichiro Terashima [3] *, Masaru Kono [4, 5] 5 1 Graduate School for Agricultural Science, Kobe University, 1-1 Rokkodai, Nada, Kobe, 657-8501, Japan 6 2 Graduate School of Science, Osaka Metropolitan University, 2000, Kisaichi, Katano-shi, Osaka, 576-0004, Japan 7 3 Institute of Molecular Biology, National Chung Hsing University, 145 Xingda Rd., South Dist., Taichung 40227, 8 Taiwan 9 4 Exo-Life Search Project Office, Astrobiology Center, National Institutes of Natural Sciences (NINS), 2-21-1 Osawa, 10 Mitaka, Tokyo 181-8588, Japan 11 5 Laboratory for Chemistry and Life Science, Institute of Integrated Research, Institute of Science Tokyo, 4259 12 Nagatsuta-cho, Midori-ku, Yokohama, Kanagawa 226-8501, Japan 13 14 *Corresponding authors 15 Ichiro Terashima [3] : terashimaichiro@nchu.edu.tw 16 Masaru Kono [4, 5] : konom07@bs.s.u-tokyo.ac.jp 1 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.02.763831; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.02.763831) preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 17 **Supplementary Materials** 18 Contents 19 1. Supplementary Table S1 and Explanatory Notes 20 2. Supplementary Text S1 21 3. Supplementary Table S2 and Supplementary Text S2 22 4. Supplementary Table S3 and Supplementary Text S3 23 24 **1.** **Supplementary Table S1 and Explanatory Notes.** 25 We summarize the results of PSI photoinhibition studies. Experimental conditions and results are shown as follows. 26  27 Samples: Intact leaves, detached leaves, leaf discs, or isolated thylakoids are categorized.  28 Species: Species names (common names) are listed. Experiments using the _pgr5_ mutant of _Arabidopsis_ 29 _thaliana_ are shown in bold. 30  Photoinhibition temperatures: _T_ (°C).  31 Photoinhibition light regimes: Light treatments used are categorized into continuous (C), fluctuating (F) and 32 intermittent (I) lighting regimes.  33 Photoinhibition irradiances: The irradiance level (photosynthetic photon flux density) in μmol m [-2] s [-1] is 34 shown. For fluctuating light treatments, all irradiance levels employed are shown like A/B/C.  35 The photoinhibition photon-exposure, or cumulative irradiance, in mol m [-2] is listed.  36 The residual PSI activity after the treatment is shown as %PSI. 37  The rate constant of PSI photoinhibition divided by irradiance, _K_ pi / _I_ ( (mol photons) [-1] m [2] ) is shown. For _K_ pi 38 calculation, see main text. 2 39 **Supplementary Table S1.** Summary of PSI photoinhibition studies |Paper|Samples|Species|T|Light|Irradiance|Exposure|PSI%|K pi/I| |---|---|---|---|---|---|---|---|---| |Satoh 1970a|thylakoids|_Spinacia oleracea_ L. (Spinach)|25|C|240|0.86|42|1.004| |Satoh 1970a|thylakoids|Spinach|4|C|240|0.86|88|0.148| |Sonoike 1995|thylakoids|_Cucumis sativus_ L. (Cucumber)|5|C|300|5.4|50|0.128| |Sonoike 1995|thylakoids|Cucumber|5|C|100|1.8|63|0.257| |Sonoike 1995|thylakoids|Cucumber|5|C|200|3.6|40|0.255| |Sonoike 1995|thylakoids|Cucumber|25|C|200|1.08|70|0.330| |Sonoike 1995|thylakoids|Spinach|25|C|300|1.62|28|0.797| |Sonoike 1995|thylakoids|Spinach|5|C|100|1.8|28|0.717| |Sonoike 1995|thylakoids|Spinach|25|C|100|0.54|33|2.053| |Sonoike 1999|thylakoids|Cucumber|0|C|110|1.98|38|0.489| |Sonoike 1999|thylakoids|Cucumber|4|C|110|1.98|42|0.438| |Sonoike 1999|thylakoids|Cucumber|0|C|60|1.08|45|0.739| |Sonoike 1999|thylakoids|Cucumber|4|C|60|1.08|48|0.680| |Terashima et al. 1994|Leaf|Cucumber|4|C|220|5.94|25|0.233| |Havaux and Davaud 1994|Leaf disc|_Solanum tuberosum_ L. (Potato)|3|C|3500|4.2|75|0.068| |Havaux and Davaud 1994|Leaf disc|Potato|2|C|900|8.1|57|0.069| |Havaux and Davaud 1994|Leaf disc|Potato|2|C|3700|6.66|64|0.067| |Terashima et al. 1998|Leaf intact|Cucumber|5|C|200|2.16|60|0.236| |Terashima et al. 1998|Leaf intact|Cucumber|5|C|200|4.32|40|0.212| |Tjus et al. 1998|Detached leaf|_Hordeum vulgare_ L. (Barley)|4|C|100|2.88|80|0.077| |Ivanov et al. 1998|Detached leaf|_Secale cereale_ L. (Winter rye)|5|C|1600|23.0|66|0.018| |Ivanov et al. 1998|Detached leaf|Winter rye|5|C|1600|23.0|69|0.016| |Ivanov et al. 1998|Detached leaf|Winter rye|5|C|1600|23.0|92|0.004| 3 |Ivanovetal 1998<br>.|Detachedleaf|Winterrye|5|C|1600|230<br>.|59|0023<br>.| |---|---|---|---|---|---|---|---|---| |Ivanov et al. 1998|Detached leaf|Winter rye|5|C|1600|23.0|108|-0.003| |Teicher et al. 2000|Leaf|Barley|Chilling|C|150|30.2|88|0.004| |Teicher et al. 2000|Leaf|Barley|Chilling|C|150|30.2|90|0.004| |Kornyeyev et al. 2003|Leaf disc|_Gossypium hirsutum_ L. (Cotton)|10|C|500|7.2|50|0.096| |Kornyeyev et al. 2003|Leaf disc|Cotton|10|C|500|7.2|67|0.056| |Zhang and Scheller 2004|Leaf|_Arabidopsis thaliana_ (L.)<br>Heynh. WT.<br>(_A. thaliana_ WT)|4|C|150|2.7|71|0.126| |Ivanov et al. 2012|Leaf|_A_. _thaliana_ WT|5|**C **|600|4.32|33|0.260| |Ivanov et al. 2012|Leaf|_A_. _thaliana_ WT|5|C|600|4.32|63|0.107| |Cazzaniga et al. 2012|Leaf intact|_A_. _thaliana_ WT|8|C|550|5.94|65|0.073| |Zhang et al. 2014|Leaf disc|Cucumber|6|C|400|13.0|51|0.052| |Zhang et al. 2014|Leaf disc|Cucumber|6|C|400|13.0|36|0.079| |Zhang et al. 2016|Leaf intact|Cucumber|6|C|200|4.32|33|0.257| |Zhang et al. 2016|Leaf intact|Cucumber|6|C|200|4.32|63|0.107| |Havaux and Davaud 1994|Leaf disc|Potato|23|C|3500|4.2|96|0.010| |Havaux and Davaud 1994|Leaf disc|Potato|25|C|3700|6.66|82|0.030| |Sonoike 1999|Detached leaf|Cucumber|20|C|200|3.6|100|0.000| |Tikkanen et al. 2014|Leaf|_A_. _thaliana_ WT|RT|C|800|5.76|100|0.000| |**Tikkanen et al. 2014**|**Leaf**|**_A_. ****_thaliana_ ****_pgr5_**|**RT**|**C **|**800**|**5.76**|**80**|**0.039**| |Tikkanen et al. 2014|Leaf|_A_. _thaliana_ WT|RT|C|800|5.76|100|0.000| |**Tikkanen et al. 2014**|**Leaf**|**_A_. ****_thaliana_ ****_pgr5_**|**RT**|**C **|**800**|**5.76**|**50**|**0.120**| |**Lima-Melo et al. 2019**|**Leaf intact**|**_A_. ****_thaliana_ ****_pgr5_**|**23**|**C **|**1000**|**0.9**|**60**|**0.568**| |Huang et al. 2016|Leaf intact|_Psychotria henryi_H. Lév.|RT|C|2258|13.5|75|0.021| 4 |Oguchietal 2021<br>.|Leafintact|CapsicumannuumL<br>.|RT|C|950|456<br>.|96|0009<br>.| |---|---|---|---|---|---|---|---|---| |Oguchi et al. 2021|Leaf intact|_C. annuum_|RT|C|950|4.56|92|0.018| |**Tiwari et al. 2024**|**Detached leaf**|**_A_. ****_thaliana_ ****_pgr5_**|**RT**|**C **|**850**|**6.12**|**34**|**0.177**| |Tiwari et al. 2024|Detached leaf|_A_. _thaliana_ WT|RT|C|850|6.12|100|0.000| |Tsuyama and Kobayashi<br>2009|Leaf|_A_. _thaliana_ WT|RT|F|8000/60|12.3|72|0.027| |Tsuyama and Kobayashi<br>2009|Leaf|_A_. _thaliana_ _crr2-2_|RT|F|8000/60|12.3|33|0.090| |Kono et al. 2014|Leaf|_A_. _thaliana_ WT|RT|F|240/30|0.340|89|0.343| |**Kono et al. 2014**|**Leaf**|**_A_. ****_thaliana_ ****_pgr5_**|**RT**|**F **|**240/30**|**0.340**|**62**|**1.405**| |Huang et al. 2016|Leaf intact|_P. henryi_|RT|F|1294/100|2.51|95|0.019| |Huang et al. 2016|Leaf intact|_P. henryi_|RT|F|2258/94|7.08|93|0.010| |Kono et al. 2017|Leaf|_A_. _thaliana_ WT|RT|F|1200/30|0.84|72|0.399| |Kono et al. 2017|Leaf|_A_. _thaliana_ WT|RT|F|1200/135|1.54|85|0.106| |Kono et al. 2017|Leaf|_A_. _thaliana_ WT|RT|F|1200+FR(25.6 W<br>m-2)|0.64|97|0.048| |Kono et al. 2017|Leaf|_A_. _thaliana_ WT|RT|F|1200+FR(12.1 W<br>m-2)|0.64|95|0.080| |Kono et al. 2017|Leaf|_A_. _thaliana_ WT|RT|F|1200+FR(3.93 W<br>m-2)|0.64|86|0.236| |Tikkanen and Grebe 2018|Leaf|_A_. _thaliana_ WT|RT|F|5000/100/80|3.73|30|0.323| |Yang et al. 2019|Leaf|_Prunus cerasoides_Buch.-Ham.<br>ex D. Don|RT|F|1809/59|1.12|95|0.046| |Yang et al. 2019|Leaf|_P. cerasoides_|RT|F|1809/59|1.12|84|0.156| |Lempiäinen et al. 2022|Leaf intact|_A_. _thaliana_ WT|25|F|3200/100/35|2.17|39|0.434| 5 |Lempiäinenetal 2022<br>.|Leafintact|A thalianaWT<br>.|25|F|3200/100/35|434<br>.|15|0437<br>.| |---|---|---|---|---|---|---|---|---| |Kono et al. 2017|Leaf|_A_. _thaliana_ WT|RT|I|1200|0.64|63|0.722| |Takagi et al. 2017b|Leaf|_Helianthus annuus_ L.<br>(Sunflower)|25|I|5000|1.80|35|0.583| |Takagi et al. 2017b|Leaf|_Zea mays_L.|25|I|5000|1.80|38|0.538| |Takagi et al. 2017b|Leaf|_Rhipsalis baccifera_(J.S.Muell.)<br>Stearn|25|I|5000|1.80|57|0.312| |Takagi et al. 2017b|Leaf|_Eucalyptus globulus_Labill.|25|I|5000|1.80|44|0.456| |Takagi et al. 2017b|Leaf|_Olea europaea_L.|25|I|5000|1.80|46|0.431| |Takagi et al. 2017b|Leaf|_Quercus glauca_Thunb.|25|I|5000|1.80|37|0.552| |Takagi et al. 2017b|Leaf|_Camellia japonica_L.|25|I|5000|1.80|75|0.160| |Takagi et al. 2017b|Leaf|_Osmanthus fragrans_(Thunb.)<br>Lour.|25|I|5000|1.80|41|0.495| |Takagi et al. 2017b|Leaf|_Acer palmatum_Thunb.|25|I|5000|1.80|60|0.284| |Takagi et al. 2017b|Leaf|_Populus nigra L._|25|I|5000|1.80|59|0.293| |Takagi et al. 2017b|Leaf|_Prunus × yedoensis_Matsum.<br>‘Somei-yoshino’|25|I|5000|1.80|51|0.374| |Takagi et al. 2017b|Leaf|_Sequoia sempervirens_(Lamb.)<br>Endl.|25|I|5000|1.80|89|0.065| |Takagi et al. 2017b|Leaf|_Pinus thunbergii_Parl.|25|I|5000|1.80|96|0.023| |Takagi et al. 2017b|Leaf|_Cedrus deodara_(Roxb. ex<br>D.Don) G.Don|25|I|5000|1.80|92|0.046| |Takagi et al. 2017b|Leaf|_Juniperus chinensis_L.|25|I|5000|1.80|94|0.034| |Takagi et al. 2017b|Leaf|_Podocarpus macrophyllus_<br>(Thunb.) Sweet|25|I|5000|1.80|90|0.059| 6 |Takagietal 2017b<br>.|Leaf|CycasrevolutaThunb<br>.|25|I|5000|180<br>.|95|0028<br>.| |---|---|---|---|---|---|---|---|---| |Takagi et al. 2017b|Leaf|_Ginkgo biloba_L.|25|I|5000|1.80|79|0.131| |Takagi et al. 2017b|Leaf|_Matteuccia struthiopteris_(L.)<br>Tod.|25|I|5000|1.80|73|0.179| |Takagi et al. 2017b|Leaf|_Selaginella moellendorffii_<br>Hieron.|25|I|5000|1.80|70|0.198| |Takagi et al. 2017b|Leaf|_Phlebodium aureum_(L.) J.Sm.|25|I|5000|1.80|79|0.131| |Takagi et al. 2017b|Leaf|_Nephrolepis cordifolia_(L.)<br>C.Presl|25|I|5000|1.80|80|0.124| |Takagi et al. 2017b|Leaf|_Equisetum arvense_L.|25|I|5000|1.80|87|0.077| |Takagi et al. 2017b|Leaf|_Pyrrosia lingua_(Thunb.) Farw.|25|I|5000|1.80|76|0.152| |Takagi et al. 2017b|Leaf|_Dryopteris erythrosora_<br>(D.C.Eaton) Kuntze|25|I|5000|1.80|78|0.138| |Takagi et al. 2017b|Leaf|_Botrychium ternatum_(Thunb.)<br>Sw.|25|I|5000|1.80|90|0.059| |Takagi et al. 2017b|Leaf|_Marchantia polymorpha L._|25|I|5000|1.80|70|0.198| |Takagi et al. 2017b|Leaf|_Conocephalum conicum_(L.)<br>Dum.|25|I|5000|1.80|48|0.414| |Takagi et al. 2017b|Leaf|_Polytrichum juniperinum_Willd.<br>ex Hedw.|25|I|5000|1.80|96|0.023| |Takagi et al. 2017b|Leaf|_Hypopterygium fauriei_Besch.|25|I|5000|1.80|62|0.266| |Sejima et al. 2014|Leaf|Sunflower|RT|F|21400/1400|28.80|85|0.006| |Sejima et al. 2014|Leaf|Sunflower|RT|F|20580/580|16.99|65|0.025| |Sejima et al. 2014|Leaf|Sunflower|RT|F|20200/200|11.52|21|0.135| |Sejima et al. 2014|Leaf|Sunflower|RT|I|20000|8.64|10|0.267| 7 |Zivcaketal 2015<br>.|Leaf|TriticumaestivumL (Wheat)<br>.|RT|I|15000|041<br>.|45|1972<br>.| |---|---|---|---|---|---|---|---|---| |Takagi et al. 2016|Isolated chloroplasts|Spinach|25|I|20000|2.16|15|0.878| |Takagi et al. 2019|Leaf|Wheat|25|I|20000|2.16|35|0.486| |Takagi et al. 2019|Leaf|Wheat|25|I|20000|2.16|58|0.256| |Furutani et al. 2023|Leaf intact|_A_. _thaliana_ WT|RT|I|15000|0.81|28|1.554| |**Furutani et al. 2023**|**Leaf intact**|**_A_. ****_thaliana_ ****_pgr5_**|**RT**|**I **|**15000**|**0.81**|**28**|**1.556**| 40 Satoh (1970a) irradiated the thylakoid samples with light from tungsten lamps at 12000 lux, which may be converted to PPFD at 240 µmol m [-2] s [-1] . 41 8 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.02.763831; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.02.763831) preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 42 **References** 43 44 Cazzaniga S, Li Z, Niyogi KK, Bassi R, Dall'Osto L (2012) The _Arabidopsis_ _szl1_ mutant reveals a critical role of β45 carotene in photosystem I photoprotection. Plant Physiol 159:1745–1758. 46 [https://doi.org/10.1104/pp.112.201137](https://doi.org/10.1104/pp.112.201137) 47 Furutani R, Wada S, Ifuku K, Maekawa S, Miyake C (2023) Higher reduced state of Fe/S-signals, with the 48 suppressed oxidation of P700, causes PSI inactivation in _Arabidopsis thaliana_ . Antioxidants 12:21. 49 https://doi.org/10.3390/antiox12010021 50 Havaux M, Davaud A (1994) Photoinhibition of photosynthesis in chilled potato leaves is not correlated with a loss 51 of photosystem-II activity: preferential inactivation of photosystem I. Photosynth Res 40:75–92. 52 Huang W, Yang YJ, Hu H, Zhang SB (2016) Responses of photosystem I compared with photosystem II to 53 fluctuating light in the shade-establishing tropical tree species _Psychotria henryi_ . Front Plant Sci 7:1549. 54 https://doi.org/10.3389/fpls.2016.01549 55 Ivanov AG, Morgan RM, Gray GR, Velitchkova MY, Huner NPA (1998) Temperature/light dependent development 56 of selective resistance to photoinhibition of photosystem I. FEBS Lett 430:288–292. 57 [https://doi.org/10.1016/S0014-5793(98)00681-4](https://doi.org/10.1016/S0014-5793(98)00681-4) 58 Ivanov AG, Rosso D, Savitch LV, Stachula P, Rosembert M, Oquist G, Hurry V, Hüner NPA (2012) Implications of 59 alternative electron sinks in increased resistance of PSII and PSI photochemistry to high light stress in cold60 acclimated Arabidopsis thaliana. Photosynthesis Res 113:191-206 61 Kono M, Noguchi K, Terashima I (2014) Roles of the cyclic electron flow around PSI (CEF-PSI) and O 2 -dependent 62 alternative pathways in regulation of the photosynthetic electron flow in short-term fluctuating light in 63 _Arabidopsis thaliana_ [. Plant Cell Physiol 55:990–1004. https://doi.org/10.1093/pcp/pcu033](https://doi.org/10.1093/pcp/pcu033) 64 Kono M, Yamori W, Suzuki Y, Terashima I (2017) Photoprotection of PSI by far-red light against the fluctuating 65 light-induced photoinhibition in _Arabidopsis thaliana_ and field-grown plants. Plant Cell Physiol 58:35–45. 66 https://doi.org/10.1093/pcp/pcw215 67 Kornyeyev D, Logan BA, Allen RD, Holaday AS (2003) Effect of chloroplastic overproduction of ascorbate 68 peroxidase on photosynthesis and photoprotection in cotton leaves subjected to low temperature 69 [photoinhibition. Plant Sci 165:1033–1041. https://doi.org/10.1016/S0168-9452(03)00294-2](https://doi.org/10.1093/pcp/pcg129) 70 Lempiäinen T, Rintamäki E, Aro EM, Tikkanen M (2022) Plants acclimate to photosystem I photoinhibition by 71 [readjusting the photosynthetic machinery. Plant Cell Environ 45:2954–2971. https://doi.org/10.1111/pce.14400](https://doi.org/10.1093/pcp/pcg129) 72 Lima-Melo Y, Alencar VTCB, Lobo AKM, Sousa RHV, Tikkanen M, Aro EM, Silveira JAG, Gollan PJ (2019) 73 Photoinhibition of photosystem I provides oxidative protection during imbalanced photosynthetic electron 74 transport in _Arabidopsis thaliana_ [. Front Plant Sci 10:916. https://doi.org/10.3389/fpls.2019.00916](https://doi.org/10.3389/fpls.2019.00916) 75 Oguchi R, Terashima I, Chow WS (2021) The effect of different spectral light quality on the photoinhibition of 76 [Photosystem I in intact leaves. Photosynth Res 149:83–92. https://doi.org/10.1007/s11120-020-00805-z](https://doi.org/10.3389/fpls.2019.00916) 77 Satoh K (1970a) Mechanism of photoinactivation in photosynthetic systems. I. The dark reaction in photoinactivation. 78 Plant Cell Physiol 11:15–27. https://doi.org/10.1093/oxfordjournals.pcp.a074487 79 Sejima T, Takagi D, Fukayama H, Makino A, Miyake C (2014) Repetitive short-pulse light mainly inactivates 80 [Photosystem I in sunflower leaves. Plant Cell Physiol 55:1184–1193. https://doi.org/10.1093/pcp/pcu061](https://doi.org/10.3389/fpls.2019.00916) 81 Sonoike K (1995) Selective photoinhibition of Photosystem I in isolated thylakoid membranes from cucumber and 82 [spinach. Plant Cell Physiol 36:825–830. https://doi.org/10.1093/oxfordjournals.pcp.a078827](https://doi.org/10.1093/oxfordjournals.pcp.a078827) 83 Sonoike K (1999) The different roles of chilling temperatures in the photoinhibition of photosystem I and 84 [photosystem II. J Photochem Photobiol B Biol 48:136–141. https://doi.org/10.1016/S1011-1344(99)00030-5](https://doi.org/10.1016/S1011-1344(99)00030-5) 85 Takagi D, Takumi S, Hashiguchi M, Sejima T, Miyake C (2016) Superoxide and singlet oxygen produced within the 86 thylakoid membranes both cause photosystem I photoinhibition. Plant Physiol 171:1626–1634. 87 https://doi.org/10.1104/pp.16.00246 88 Takagi D, Ishizaki K, Hanawa H, Mabuchi T, Shimakawa G, Yamamoto H, Miyake C (2017b) Diversity of 89 strategies for escaping reactive oxygen species production within photosystem I among land plants: P700 9 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.02.763831; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.02.763831) preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 90 oxidation system is prerequisite for alleviating photoinhibition in photosystem I. Physiol Plant 161:56–74. 91 https://doi.org/10.1111/ppl.12562 92 Takagi D, Ihara H, Takumi S, Miyake C (2019) Growth light environment changes the sensitivity of Photosystem I 93 photoinhibition depending on common wheat cultivars. Front Plant Sci 10:686. 94 https://doi.org/10.3389/fpls.2019.00686 95 Teicher HB, Møller BL, Scheller HV (2000) Photoinhibition of photosystem I in field-grown barley ( _Hordeum_ 96 _vulgare_ L.): Induction, recovery and acclimation. Photosynth Res 64:53–61. 97 [https://doi.org/10.1023/A:1026524302191](https://doi.org/10.1023/a:1026524302191) 98 Terashima I, Funayama S, Sonoike K (1994) The site of photoinhibition in leaves of _Cucumis sativus_ L at low99 temperatures is photosystem I, not photosystem II. Planta 193:300–306. https://doi.org/10.1007/BF00192544 100 Terashima I, Noguchi K, Itoh-Nemoto T, Park YM, Kubo A, Tanaka K (1998) The cause of PSI photoinhibition at 101 low temperatures in leaves of _Cucumis sativus_, a chilling-sensitive plant. Physiol Plant 103:295–303. 102 [https://doi.org/10.1034/j.1399-3054.1998.1030301.x](https://doi.org/10.1034/j.1399-3054.1998.1030301.x) 103 Tikkanen M, Grebe S (2018) Switching off photoprotection of photosystem I-a novel tool for gradual PSI 104 [photoinhibition. Physiol Plant 162:156–161. https://doi.org/10.1111/ppl.12618](https://doi.org/10.1111/ppl.12618) 105 Tikkanen M, Mekala NR, Aro EM (2014) Photosystem II photoinhibition-repair cycle protects Photosystem I from 106 irreversible damage. Biochim Biophys Acta Bioenerg 1837:210–215. 107 [https://doi.org/10.1016/j.bbabio.2013.10.001](https://doi.org/10.1016/j.bbabio.2013.10.001) 108 Tiwari A, Mamedov F, Fitzpatrick D, Gunell S, Tikkanen M, Aro EM (2024) Differential FeS cluster photodamage 109 plays a critical role in regulating excess electron flow through photosystem I. Nat Plants 10:1592–1603. 110 [https://doi.org/10.1038/s41477-024-01780-2](https://doi.org/10.1038/s41477-024-01780-2) 111 Tjus SE, Møller BL, Scheller HV (1998) Photosystem I is an early target of photoinhibition in barley illuminated at 112 [chilling temperatures. Plant Physiol 116:755–764. https://doi.org/10.1104/pp.116.2.755](https://doi.org/10.1104/pp.116.2.755) 113 Tsuyama M, Kobayashi Y (2009) Reduction of the primary donor P700 of photosystem I during steady-state 114 photosynthesis under low light in _Arabidopsis_ [. Photosynth Res 99:37–47. https://doi.org/10.1007/s11120-008-](https://doi.org/10.1007/s11120-008-9379-x) 115 [9379-x](https://doi.org/10.1007/s11120-008-9379-x) 116 Yang YJ, Zhang SB, Wang JH, Huang W (2019) Photosynthetic regulation under fluctuating light in field-grown 117 _Cerasus cerasoides_ : A comparison of young and mature leaves. Biochim Biophys Acta Bioenerg 1860:469– 118 [477. https://doi.org/10.1016/j.bbabio.2019.04.006](https://doi.org/10.1016/j.bbabio.2019.04.006) 119 Zhang S, Scheller HV (2004) Photoinhibition of photosystem I at chilling temperature and subsequent recovery in 120 _Arabidopsis thaliana_ [. Plant Cell Physiol 45:1595–1602. https://doi.org/10.1093/pcp/pch180](https://doi.org/10.1093/pcp/pch180) 121 Zhang ZS, Yang C, Gao HY, Zhang LT, Fan XL, Liu MJ (2014) The higher sensitivity of PSI to ROS results in 122 lower chilling-light tolerance of photosystems in young leaves of cucumber. J Photochem Photobiol B Biol 123 [137:127–134. https://doi.org/10.1016/j.jphotobiol.2013.12.012](https://doi.org/10.1016/j.jphotobiol.2013.12.012) 124 Zhang ZS, Jin LQ, Li YT, Tikkanen M, Li QM, Ai XZ, Gao HY (2016) Ultraviolet-B radiation (UV-B) relieves 125 chilling-light-induced PSI photoinhibition and accelerates the recovery of CO 2 assimilation in cucumber 126 ( _Cucumis sativus_ L.) leaves. Sci Rep 6:34455. https://doi.org/10.1038/srep34455 127 Zivcak M, Brestic M, Kunderlikova K, Sytar O, Allakhverdiev SI (2015) Repetitive light pulse-induced 128 photoinhibition of photosystem I severely affects CO 2 assimilation and photoprotection in wheat leaves. 129 [Photosynth Res 126:449–463. https://doi.org/10.1007/s11120-015-0121-1](https://doi.org/10.1007/s11120-015-0121-1) 130 131 10 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.02.763831; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.02.763831) preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 132 133 **2.** **Supplementary Text S1.** 134 **Affinity of CEF-PSI pathways to Fd.** 135 This text expands the paragraph "Interaction with Fd" in § 5.1. It gives the sources and conditions of the values 136 quoted there, compares the two CEF-PSI pathways with the other Fd-dependent enzymes of the stroma, summarizes 137 what is known about the amount and redox state of Fd _in vivo_, and lists the open questions, which are taken up in 138 Table S3, prediction 5. 139 **How the two pathways bind Fd.** For NDH, the Fd-binding subunit is NdhS (CRR31), a protein with an SH3-like 140 fold (Yamamoto et al. 2011). In ruptured _Arabidopsis_ chloroplasts, NDH-dependent PQ reduction is almost saturated 141 at 3 µM Fd in the wild type, whereas in _ndhs_ it remains low even at 20 µM Fd; NdhS is thus required for high-affinity 142 binding, although it is not essential for NDH activity (Yamamoto et al. 2011). The binding is electrostatic, as it is for 143 all known Fd complexes (Knaff and Hirasawa 1991): of nine conserved charged residues in the SH3-like domain, 144 only Arg-193 is required, and replacing it by a neutral, acidic or hydrophobic residue lowers the efficiency of Fd145 dependent PQ reduction (Yamamoto and Shikanai 2013). The structure of cyanobacterial NDH-1 shows two 146 candidate Fd-binding sites (Laughlin et al. 2019). For the PGR5-dependent pathway, the picture is less complete. In 147 yeast two-hybrid assays, PGRL1 interacts with Fd, PGR5, the PSI subunit PsaD, Cyt _b_ ₆ and both FNR isoforms, and 148 its interaction with Fd depends on a C-terminal loop (DalCorso et al. 2008). Recombinant PGRL1 binds iron through 149 six conserved cysteines, accepts electrons from Fd reduced by FNR and NADPH, and reduces quinones in a reaction 150 that antimycin A inhibits; on this basis, Hertle et al. (2013) proposed that PGRL1 is the Fd–PQ reductase (FQR). 151 They also quantified the proteins: about 0.70 mmol mol⁻¹ Chl for PGRL1 and about 0.09 for both PGR5 and the 152 NDH subunit NdhH. An involvement of FNR in the antimycin A-sensitive pathway was suggested earlier, because 153 inhibitors of FNR (heparin, dimaleimide) and Fd modified so that it cannot donate electrons to FNR both lowered 154 the FQR activity of pea thylakoids (Cleland and Bendall 1992). Two findings argue against a direct catalytic role of 155 PGR5: it has no cofactor-binding motif (Shikanai 2007), and in leaves of _pgr5_, Fd is more reduced and the Fd156 dependent CEF-PSI measured from the oxidation kinetics of Fd is higher than in the wild type (Maekawa et al. 2024). 157 The same _in vivo_ measurement shows that this Fd-dependent activity does not require NDH, because it is present in 158 _crr4_ (Ohnishi et al. 2023). The identity of the enzyme that oxidizes Fd in the antimycin A-sensitive pathway therefore 159 remains open (Table S2, row 8). 160 **What the affinity values mean.** The only study that measured both pathways in the same system is that of Fisher et 161 al. (2019), who followed the post-illumination fluorescence rise in uncoupled thylakoids supplied with Fd and 162 NADPH. In _A. hybridus_ thylakoids with antimycin A, which isolates NDH, the apparent _K_ m for Fd was 2.1 ± 0.6 µM; 163 the rate was still rising at the highest concentration tested, 5.5 µM, which was about 70% of the fitted maximum, so 164 this value is a lower estimate. In spinach thylakoids without antimycin A, in which most of the activity is antimycin 165 A-sensitive, the apparent _K_ m was 1.3 ± 0.4 µM and the rate saturated near 2 µM. The two values thus differ by less 166 than their uncertainty, and they were obtained in different species. Earlier studies of the antimycin A-sensitive 167 pathway reported larger Fd requirements, but under different conditions. Tagawa et al. (1963) found that cyclic 168 photophosphorylation by broken spinach chloroplasts under anaerobic conditions needed up to ten times more Fd 169 than non-cyclic photophosphorylation. Arnon and Chain (1975) measured cyclic photophosphorylation under 715170 nm light in air and found an optimum at about 10 µM Fd, with about ten times more needed under anaerobic 171 conditions; they also showed that the cycle required a small supply of electrons from PSII, because DCMU abolished 172 it. Cleland and Bendall (1992) measured the oxidation of Fd by FQR directly, from the absorbance of Fd at 463 nm, 173 in uncoupled pea thylakoids under anaerobic conditions with DCMU. The rate increased with the concentration of 174 reduced Fd and saturated at about 75 µM. These assays differ from that of Fisher et al. (2019) in three ways. First, 175 they measured ATP synthesis or the oxidation of Fd rather than the reduction of PQ. Second, the redox poise of the 176 system was set differently. In the assays in air, O₂ also accepts electrons from reduced Fd; Arnon and Chain (1975) 177 attributed the lower activity at Fd concentrations above the optimum to this re-oxidation of reduced Fd by O₂ rather 178 than by the cycle. In the anaerobic assay of Cleland and Bendall (1992), Fd was reduced photochemically with 11 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.02.763831; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.02.763831) preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 179 deazaflavin before the thylakoids were added. Third, none of them measured NDH. The older values therefore 180 describe the dependence of the antimycin A-sensitive pathway on reduced Fd rather than its affinity for Fd, and they 181 cannot be compared directly with the _K_ m values of Fisher et al. (2019). What the old and new studies agree on is that 182 the antimycin A-sensitive pathway is not saturated at 1 µM Fd, the concentration used in many assays (e.g., Fisher et 183 al. 2019). 184 **Comparison with FNR and the assimilatory enzymes.** The affinities of the stromal Fd-dependent enzymes for Fd 185 were reviewed by Knaff and Hirasawa (1991). For FNR, the dissociation constant of the Fd–FNR complex at low 186 ionic strength is below 50 nM in the absence of NADP⁺ and rises to about 0.5 µM when the NADP⁺ site is occupied, 187 a negative cooperativity that helps the oxidized Fd to leave the enzyme after each turnover (Batie and Kamin 1984a, 188 b); at physiological ionic strength, the dissociation constant is 1–10 µM (Knaff and Hirasawa 1991). The dissociation 189 constants for Fd are 0.63–1.0 µM for nitrite reductase, rising to 11 µM when nitrite is bound; 9–14.5 µM for Fd190 dependent glutamate synthase, whose _K_ m for Fd is about 2 µM; and below 100 nM for Fd–thioredoxin reductase 191 (Knaff and Hirasawa 1991). All these complexes are electrostatic, and FNR, nitrite reductase and glutamate synthase 192 may share a common binding surface on Fd, whereas Fd–thioredoxin reductase may use a different one (Knaff and 193 Hirasawa 1991). Two conclusions follow for CEF-PSI. First, the _K_ m values of the two pathways (1–2 µM) are of the 194 same order as the affinities of FNR and the assimilatory enzymes, so neither pathway has a special advantage in the 195 competition for reduced Fd. Second, FNR is the most abundant competitor: in spinach chloroplasts, the molar ratio 196 of Fd to FNR is about 5 : 3 (Böhme 1978) and about 2 : 1 (Terashima and Inoue 1985). As long as NADP⁺ is available, 197 FNR with its high affinity and high turnover will take most of the reduced Fd, and the cyclic pathways receive 198 electrons mainly when NADP⁺ runs short. This is exactly what Tagawa et al. (1963) observed in broken chloroplasts, 199 where cyclic photophosphorylation started at the moment when NADP⁺ was exhausted, and what is seen in leaves, 200 where the Fd-dependent CEF-PSI appears when linear electron flow is saturated (Ohnishi et al. 2023; see below). 201 **Fd content, its redox state and the supply to CEF-PSI.** The amount of Fd in spinach chloroplasts has been 202 measured directly. By quantitative immunoelectrophoresis, Böhme (1978) found Fd : FNR : plastocyanin ≈ 5 : 3 : 4 203 per P700, with about 80% of the Fd soluble. In chloroplasts isolated from palisade and spongy tissue, Terashima and 204 Inoue (1985) found 6.1 and 2.9 mmol Fd mol⁻¹ Chl against 2.55 and 2.43 mmol P700 mol⁻¹ Chl, that is, about 2.4 205 and 1.2 Fd per P700, with FNR at about half the amount of Fd; both Fd and FNR were thus lower in the spongy 206 chloroplasts, which receive the weaker and FR-enriched light inside the leaf. Values of two to six Fd per PSI have 207 been compiled for higher plants (Antal et al. 2013). Combined with a chlorophyll concentration of 30–60 mM in the 208 chloroplast (Antal et al. 2013), the Fd contents of 3–6 mmol mol⁻¹ Chl correspond to about 0.1–0.4 mM. Antal et al. 209 (2013) list a somewhat higher value, about 0.6 mM Fd together with 0.3 mM FNR and 0.1 mM PSI, whereas Hanke 210 and Hase (2008) quote about 50 µM; the spread reflects the reference volume and the source of each estimate, but 211 even the lowest value is more than ten times the _K_ m values of the two pathways. Experiments that changed the amount 212 of Fd point to the same conclusion. In transplastomic tobacco with about four times the wild-type Fd content, CEF213 PSI and NPQ were higher, whereas CO₂ assimilation was not (Yamamoto et al. 2006). In antisense potato with 40– 214 80% of the wild-type Fd, CEF-PSI was also higher (the ratio of the quantum yields of PSI and PSII rose from about 215 1.3 to above 3), Q A and P700 were more reduced, and plants with less than 40% Fd were not viable (Holtgrefe et al. 216 2003). That both an increase and a decrease in Fd stimulate CEF-PSI and over-reduce the PQ pool (Hanke and Hase 217 2008) shows that the amount of Fd acts on the redox poise of the chain rather than as a simple substrate. The redox 218 state of Fd can now be followed in leaves with near-infrared absorbance. At steady state, part of the Fd pool remains 219 reduced, and the reduced fraction rises with light intensity (Satoh et al. 2025). The oxidation rate of Fd that is not 220 accounted for by linear electron flow, taken as the rate of Fd-dependent CEF-PSI, appears when linear electron flow 221 is saturated and the PQ pool is oxidized (Ohnishi et al. 2023), increases with the reduced fraction of Fd and falls 222 when the PQ pool becomes reduced (Satoh et al. 2025), and is higher in _pgr5_, in which Fd is more reduced (Maekawa 223 et al. 2024). These results fit a rate proportional to the product of reduced Fd and oxidized PQ, and they identify the 224 reduced fraction of Fd, not its total amount, as the quantity that limits the supply to CEF-PSI _in vivo_ . The isoforms 225 of Fd may add a further level of control. In _Arabidopsis_, Fd2 accounts for 90–99% of the leaf Fd; lowering Fd2 slows 226 linear electron flow, whereas lowering the minor isoform Fd1 increases it and oxidizes the PQ pool, which suggests 227 that Fd1 tends to keep electrons within the chain (Hanke and Hase 2008). In maize, the bundle-sheath isoform FdII 12 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.02.763831; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.02.763831) preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 228 reacts more slowly with FNR than the mesophyll isoform FdI and donates more electrons to the PQ pool (Kimata229 Ariga et al. 2000). Which isoform feeds each CEF-PSI pathway in C₃ leaves is not known. 230 **Open questions (Table S3, prediction 5).** (a) The apparent _K_ m values of the two pathways have not been measured 231 in the same species with the same Fd preparation; the comparison of Fisher et al. (2019) is between _A. hybridus_ and 232 spinach. (b) All _K_ m values come from uncoupled thylakoids. Whether the affinity for Fd changes with Δp has not 233 been tested, and Cleland and Bendall (1992) noted that a regulatory effect of adenylates might appear only in coupled 234 thylakoids. (c) The reduced fraction of Fd under FR, and the share of it taken by FNR, have not been measured; this 235 is the quantity that decides how much substrate the cyclic pathways receive when PSII supplies few electrons. (d) 236 Fisher et al. (2019) found inhibition of both pathways by ATP, whereas Cleland and Bendall (1992) found no effect 237 of ATP or ADP up to 1 mM on the antimycin A-sensitive pathway; the reason for the difference is unknown (Table 238 S2, row 3). (e) Whether PGRL1 is the FQR, and what the role of PGR5 is, remain unsettled (Table S2, row 8). (f) 239 The Fd isoform used by each pathway is not known, and the absolute stromal concentration of Fd is known only to 240 within an order of magnitude. 241 13 242 **3.** **Supplementary Table S2 and Supplementary Text S2** 243 **Supplementary Table S2.** Points of tension in the proposed division of roles between the CEF-PSI pathways **Issue** **Evidence that complicates the model** **Our position** **What remains open** 1 Does the pmf hold NDH back _in vivo_ ? 2 Redox-side constraint 3 ATP feedback as an alternative 4 Two orders of activation 5 Donor side or acceptor side? A light-induced pmf did not inhibit NDH as added ATP did (Fisher et al. 2019); NDH-dependent CEFPSI active in GO5 despite a pmf ~2× the wild type (Strand et al. 2015) Under FR alone in air, the chain is too oxidized for cyclic flow; fast NDH-dependent cycling needs partly reduced PQ and NADP pools (Joët et al. 2002); in isolated chloroplasts, cyclic photophosphorylation under 715-nm light needs a small supply of electrons from PSII (Arnon and Chain 1975) A moderate fall in stromal ATP activates NDH, a severe fall the PGR5 pathway; _K_ i ~670 vs ~240 µM (Fisher et al. 2019); but ATP and ADP up to 1 mM had no effect on the antimycin A-sensitive pathway in an earlier assay (Cleland and Bendall 1992) NADPH → PGR5 pathway → H 2 O 2 → NDH over minutes to hours (Strand et al. 2017) vs ATP fall → NDH first, within seconds (Fisher et al. 2019) The primary requirement for protection is an oxidized acceptor side (Sonoike 2025); P700 can be oxidized Gating defined by the net driving force (Δ _E_ h − _n_ - Δp) is compatible with GO5 (strongly reduced stroma) and with Fisher et al. if their pmf lay below the limiting range, as the same authors expect at higher pmf; supported by the increase in the NDH-dependent post-illumination fluorescence rise on uncoupling in spinach, _A. hybridus_ and _A. thaliana_, smaller in _A. thaliana_ owing to its larger antimycin A-sensitive share (Strand et al. 2017), by NDH acting on the PSI acceptor side in the low-light but not the high-light phases of red FL, possibly owing to its reverse reaction at large ΔpH (Zhou et al. 2023), and by its largest contribution under low light in _hope2_ (Degen et al. 2023) An oxidized PQ pool favors NDH only while reduced donors remain; beyond that point, substrate rather than back-pressure limits the flux Points the same way as gating under FR and low light, but is a different mechanism Different triggers and time scales; both can be true; the fast sequence is the relevant one for 10-s cycles Y(ND) is protective only together with low Y(NA); FR meets both conditions 14 Sign of the net driving force never measured in a leaf (Table S3, 1–3) Whether the 10-s low-light phase stays within the operating window of NDH (Table S3, 1 and 4) Stromal ATP under FR never measured, and the two ATP results conflict (Table S3, 5); separated from gating by Table S3, 2 Which route activates NDH under FR (basal content, carry-over from red pulses, induction after PSI damage) is untested FR raises Y(ND) and lowers Y(NA) together, so the two cannot be separated when the stroma is over-reduced (Rantala et al. 2020) with present data (Table S3, 8) 6 Pulse-type regime 7 How many protons does NDH pump? 8 What does PGR5 do? 9 NDH–PSI supercomplex 10 What cannot yet be measured 1-s pulses of ~3,000 µmol photons m [−2] s [−1] : _pgr5_ and _ndho_ lose P m to a similar extent as the wild type (Rantala et al. 2020); the mechanism may differ from that under FL (Sonoike 2025) 2 H [+] /e [−] inferred from homology and uncoupler sensitivity; _in vivo_ slopes give 4.6–8.8 H [+] /e [−] for NDHdependent CEF-PSI (Strand et al. 2017); NDH may poise the chain rather than feed the pmf (Yamori et al. 2011) No redox cofactor (Shikanai 2007); high PGR5dependent CEF-PSI cannot impose photosynthetic control when _g_ H [+] is high ( _hope2_ ; Degen et al. 2023); FLV expression restores the pmf and normalizes _g_ H [+] in _pgr5_ (Degen et al. 2023; Yamamoto et al. 2016); a steady-state cycle cannot oxidize P700 by itself (Vilyanen and Kozuleva 2025) _lhca6 pgr5_ phenotype lies between _pgr5_ and _crr pgr5_ although NDH is retained (Shikanai et al. 2025); NDH works without a supercomplex in _Marchantia_ (Ueda et al. 2012); two candidate Fd-binding sites (Laughlin et al. 2019) NDH flux never measured in illuminated leaves; PIFR is a dark assay of capacity; Fd-dependent PQ reduction uses ruptured chloroplasts; P700 reoxidation by FR and ETR(I) − ETR(II) are indirect; proton-flux slopes assume the stoichiometry of linear flow Agrees with our FR-free controls (~70% in all genotypes); genotype dependence appears only with FR and then runs through NDH The value of ~2 in Table 1 is supported but not measured; the poising function does not require a large contribution to the pmf Our argument needs only a pathway that builds the pmf quickly without its own pump; if PGR5 is a regulator, the "capacity" entries of Table 1 describe the pathway, not the protein Not used to explain the wavelength dependence, because both pathways draw on one pool of reduced Fd Our FR data rely on Y(ND), which reports P700 oxidation but not the engagement of photosynthetic control Whether the 800-ms/1,200 µmol photons m [−2] s [−1] regime is of the pulse or the FL type; mutants compensate after damage (NDH-1 up in _pgr5_, PGR5 up in _ndho_ ) Direct measurement of the stoichiometry (Table S3, 3) Cause of the high _g_ H [+] in _pgr5_ unknown Role in Fd supply under FR (Table S3, 9) Simultaneous Δ _E_ h and Δp in leaves; rate constant of P700 [+] re-reduction (Degen and Johnson 2024) (Table S3, 1 and 8) 244 Details and references for rows 1–3 and 5–10 are given in Text S2; the Fd data of rows 2, 3 and 8 are discussed in Text S1. Numbers in the last column refer to the predictions in Table S3. 245 ETR(I), ETR(II), electron transport rates through PSI and PSII; PIFR, post-illumination fluorescence rise. 246 247 248 249 15 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.02.763831; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.02.763831) preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 250 **Points of tension and open questions** 251 This text expands Table S2. The rows of Table S2 discussed under each heading are given in parentheses; row 4 (two 252 orders of activation) is summarized in the table only. The question of whether P700⁺ is protective in itself, raised in 253 § 5.3, is discussed under row 5. 254 **Is NDH really held back by pmf** _**in vivo**_ **? (Table S2, rows 1–3).** The redox–Δp gating hypothesis rests on the 255 thermodynamics of a proton pump, and the evidence that this thermodynamic process operates in a leaf is still indirect. 256 Two observations weigh against it. First, in a later study from the same laboratory, the authors tried to impose a pmf 257 on thylakoids by adding ATP, so that the ATP synthase ran in reverse and pumped protons into the lumen. ATP strongly 258 inhibits NDH, but a light-induced pmf of a similar size does not (Fisher et al. 2019). The authors concluded that ATP 259 directly inhibits NDH. For our purpose, the important point is the other half of the result: a pmf of this size, generated 260 by light, did not hold back NDH. The same authors nevertheless suggested that NDH may become 261 thermodynamically limited at higher pmf (Fisher et al. 2019); therefore, the two observations are not in conflict if 262 the pmf reached in their light experiment was below the range in which back-pressure becomes limiting. Second, 263 NDH-dependent CEF-PSI is clearly active in the H₂O₂-accumulating line GO5, in which the pmf is approximately 264 twice that of the wild type (Strand et al. 2015); we have argued in § 5.3 that this is compatible with gating defined 265 by the net driving force (Δ _E_ h − _n_ - Δp), but it is not evidence for it. The uncoupler experiments of Strand et al. (2017), 266 by contrast, support gating rather than weaken it. Collapsing the pmf with uncouplers increased the NDH-dependent 267 post-illumination fluorescence rise in thylakoids from spinach, _Amaranthus hybridus_ and _Arabidopsis thaliana_ . The 268 increase was about twofold in spinach and _A. hybridus_ and smaller in _A. thaliana_, which the authors attributed to the 269 larger share of the antimycin A-sensitive pathway, which is not limited by back-pressure, in the fluorescence rise of 270 this species. A further constraint arises from the redox side. Under FR alone, the electron transport chain of tobacco 271 leaves in air is too oxidized for significant cyclic flow, including flow through NDH, and rapid cycling appears only 272 when the PQ and NADP pools are partly reduced (Joët et al. 2002); in isolated chloroplasts, cyclic 273 photophosphorylation under 715-nm light likewise requires a small supply of electrons from PSII, and DCMU 274 abolishes it (Arnon and Chain 1975). Therefore, an oxidized PQ pool thermodynamically favors NDH only as long 275 as enough reduced donors remain; beyond that point, the lack of substrate, not back-pressure, limits the flux. Thus, 276 we present gating as a hypothesis that adds to, and does not replace, the established reasons why NDH contributes 277 little under red high light: its low abundance of approximately 0.02 per PSI, its slow turnover, and its competition 278 with PSII for oxidized PQ (Ermakova et al. 2024a; Table 1). Gating adds a reason why the balance should shift with 279 light quality as well as with irradiance. Two mutant studies point the same way. The first is the _hope2_ mutant, in 280 which a defect in the regulation of the ATP synthase is compensated by a high rate of PGR5-dependent CEF: the 281 NDH pathway made its largest contribution under low light, where it is less limited by back-pressure from the pmf, 282 and the authors proposed that competition for electrons at the acceptor side together with thermodynamic constraints 283 on turnover decides which pathway is favored (Degen et al. 2023). The second is the _crr2-2_ mutant under red FL 284 (1616/40 | 1 min/5 min): Y(NA) was elevated in the low-light phases but not in the high-light phases, and the authors 285 suggested that a large ΔpH and a reduced PQ pool may favor the reverse reaction of NDH in high light (Zhou et al. 286 2023). However, no study has simultaneously determined the pmf and redox states of PQ and Fd in a leaf under red 287 and FR; therefore, the sign of the net driving force _in vivo_ is unknown. It should also be said that ATP feedback could 288 produce a similar division of roles. In the scheme of Fisher et al. (2019), a moderate fall in stromal ATP activates 289 NDH and a severe fall activates the PGR5-dependent pathway; under FR and low light, the demand for ATP persists 290 while its supply is low. Thermodynamic gating and ATP feedback point in the same direction, but they are different 291 mechanisms that can be separated experimentally. 292 **Donor side or acceptor side? (Table S2, row 5).** Two recent papers question the direct protective role of P700⁺. 293 Sonoike (2025) argues that the primary requirement for PSI protection is an oxidized acceptor side, that the 294 dissipation of excitation by P700⁺ should have a smaller effect on photoinhibition than the reduction state of the Fe– 295 S clusters, and that FR would aggravate rather than prevent photoinhibition if it merely moved electrons from the 296 donor side to the acceptor side. In his view, the protective agent is the CEF-PSI that FR induces, not FR itself. Rantala 297 et al. (2020) argued that P700 can also become oxidized when the stromal acceptors are over-reduced, so that a rapid 16 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.02.763831; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.02.763831) preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 298 oxidation of P700 does not prove that PSI is protected. Taking account of these claims, we propose the roles of P700⁺ 299 in the following way. Y(ND) is an indicator of donor-side limitation, and it is protective only when it increases 300 together with an oxidation of the acceptor side. When FR is added, these two conditions are met: because the electron 301 flux through PSI is small, the sinks downstream of PSI can accept the electrons released by P700 oxidation, so that 302 P700⁺ accumulates while Fd and NADP⁺ remain oxidized. This is exactly the combination of high Y(ND) and low 303 Y(NA) seen under FR (§ 5.3). The protective functions of P700⁺ in this state are twofold. First, excitation that reaches 304 an oxidized reaction center is dissipated without transferring an electron downstream. Photoacoustic measurements 305 on intact leaves show that the extent of P700 oxidation induced by FR is strictly proportional to a positive variable 306 component of heat emission (Bukhov and Carpentier 2003, their Fig. 3c), and that the fraction of reduced P700 307 matches the fraction of energy stored photochemically (their Fig. 4c); open reaction centers are strong photochemical 308 quenchers, and heat emission rises in proportion to the oxidized P700. Second, charge recombination between P700⁺ 309 and reduced Fe–S clusters can re-oxidize the clusters that would otherwise react with H₂O₂ (Shimakawa and Miyake 310 2018). Both functions depend on P700 being kept oxidized. When the electron sink downstream of PSI becomes very 311 small, P700 can no longer be kept oxidized even though the pmf is high, because the oxidation of excited P700 312 becomes rate-limiting (Miyake 2020). The protective effect of FR is therefore conditional: it holds as long as FR 313 keeps the whole chain, and not only P700, in an oxidized state. 314 The difficulty, which Sonoike (2025) also acknowledges, is that our FR data cannot separate the two. Under 315 red FL, Kono and Terashima (2016) argued that the relief of the acceptor side must come before the rise in Y(ND) (§ 316 5.2); this is a logical requirement rather than a measured order, and in our FR experiments both parameters were read 317 at the same time points, and at each of them FR had already raised Y(ND) and lowered Y(NA) (Kono et al. 2017); 318 thus, the protection could be attributed to either change. Separating them will require an experiment, for example a 319 series of FR intensities in which Y(ND) rises before Y(NA) falls, or a condition in which P700 is kept oxidized while 320 the acceptor side is deliberately reduced. 321 **The pulse-type regime (Table S2, row 6).** Rantala et al. (2020) exposed _Arabidopsis_ leaves to repeated 1-s pulses 322 of approximately 3,000 µmol photons m⁻² s⁻¹ on a weak background for 2 or 4 h and measured P m after 24 h of 323 recovery. P m fell to approximately 60%, 50%, and 50% after 2 h and to approximately 30%, 15%, and 25% after 4 h 324 in the wild type, _pgr5_, and _ndho_, respectively. Because the relative loss was similar, the authors concluded that neither 325 pathway protects PSI against this type of burst. This conclusion agrees with our own data: in the 800-ms pulse regime 326 of Kono et al. (2017), P m fell to approximately 70% in all genotypes when FR was absent, with no significant 327 difference between them. In pulse regimes, genotype dependence appears only when FR is added, and it then runs 328 through NDH; in the 2-min regime, it runs through PGR5. Two points follow. First, the statement that CEF-PSI does 329 not protect PSI is specific to the pulse regime, and Rantala et al. (2020) accept that PGR5 is required under FL and 330 NDH-1 under rapidly fluctuating light. Sonoike (2025) has argued that photoinhibition by repetitive saturating pulses 331 proceeds by a mechanism different from that under slower FL, with the Fe–S clusters left intact. Second, the fact that 332 FR protects PSI in a pulse regime in which CEF-PSI alone does not is what the masking hypothesis and redox–Δp 333 gating together predict: FR removes the need for a fast, pmf-driven oxidation of P700, and NDH supplies the low334 flux recycling that remains necessary. A further complication is that the mutants compensate: after the pulse treatment, 335 NDH-1 increased in _pgr5_ and PGR5 in _ndho_ (Rantala et al. 2020), so that a mutant measured after photoinhibition is 336 not the same plant as before it. 337 **How many protons does NDH pump? (Table S2, row 7).** The value of 2 H⁺ pumped by NDH per electron used 338 throughout § 5 is well supported but not measured. This is based on the homology of the membrane arm of NDH 339 with the proton-pumping subunits of respiratory complex I (Laughlin et al. 2019), on the sensitivity of the post340 illumination fluorescence rise to uncouplers, which requires a stoichiometry of at least 2 (Strand et al. 2017), and on 341 the slopes of proton flux against PSI electron flux in leaves of the wild-type and the high-CEF mutant _hcef1_ . These 342 slopes were 1.8 and 2.75 when fitted freely and 1.6 and 4.7 when forced through the origin (Strand et al. 2017); with 343 3 H⁺ per electron for linear flow, they correspond to about 4.6 to 8.8 H⁺ per electron for NDH-dependent CEF, that 344 is, to between about 2.6 and 6.8 protons pumped by NDH itself. The _in vivo_ data thus show that NDH pumps protons 345 but do not fix the number. Yamori et al. (2011) proposed the alternative view that the flux through NDH is too small 346 to contribute to the pmf and that NDH acts instead by poising the redox state of the intersystem carriers. In our 17 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.02.763831; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.02.763831) preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 347 framework, the poising function (§ 5.3) does not depend on a large contribution to the pmf. 348 **What does PGR5 do? (Table S2, row 8).** The molecular role of PGR5 remains uncertain. PGR5 carries no motif for 349 a redox cofactor and is unlikely to transfer electrons (Shikanai 2007), Fd-dependent PQ reduction is reduced but not 350 abolished in _pgr5_ (Munekage et al. 2002), and the mutant's failure to hold the pmf at a light increase has been 351 interpreted as a loss of control over linear electron transport and over the ATP synthase (Suorsa et al. 2012), an 352 interpretation supported by the elevated proton conductivity of _pgr5_ under high light (Rantala et al. 2020; Wang et 353 al. 2015). Two recent studies have sharpened this picture. In _hope2_, a mutant of the ATP synthase γ-subunit with a 354 constitutively high proton conductivity, the pmf and qE are kept at wild-type levels by an increased PGR5-dependent 355 CEF, yet steady-state P700 oxidation is virtually absent and PSI remains susceptible to photoinhibition: a high rate 356 of PGR5-dependent CEF cannot impose photosynthetic control while the proton conductivity remains high (Degen 357 and Johnson 2024; Degen et al. 2023). On the other hand, the expression of the flavodiiron proteins of _Physcomitrium_ 358 _patens_ (Hedw.) Mitt. (syn. _Physcomitrella patens_ (Hedw.) Bruch & Schimp.) in _pgr5_ restores the pmf and P700 359 oxidation, relieves the acceptor-side limitation (Yamamoto et al. 2016), and normalizes the elevated proton 360 conductivity (Degen et al. 2023), which argues against a direct regulation of the ATP synthase by PGR5; the cause 361 of the high conductivity in _pgr5_ is still unknown (Degen et al. 2023). Vilyanen and Kozuleva (2025) add a general 362 point: a cycle at steady state returns exactly as many electrons to PSI as it removes and therefore cannot keep P700 363 oxidized by itself; its protective effect must come from the protons it moves, either through extra ATP that speeds the 364 Calvin–Benson cycle and relieves the acceptor side, or through lumen acidification and photosynthetic control. Our 365 account in § 5.2 uses both routes and does not require a decision on the molecular role of PGR5. What it requires is 366 only that the PGR5-dependent pathway builds the pmf quickly from a large electron flux without a proton pump of 367 its own; if PGR5 turns out to act mainly as a regulator, the "capacity" entries of Table 1 describe the pathway rather 368 than the protein, and the rescue of _pgr5_ by FR is read as FR removing the need for that regulation—the conclusion 369 is unchanged. 370 **The NDH–PSI supercomplex (Table S2, row 9).** We have not used the supercomplex to explain the wavelength 371 dependence (§ 5.1), but its functional significance remains unresolved. In _lhca6_, which cannot form the supercomplex, 372 the NDH subunits are largely retained, yet the _lhca6 pgr5_ double mutant shows a phenotype between _pgr5_ and _crr_ 373 _pgr5_, a discrepancy that Shikanai et al. (2025) could not explain; in _Marchantia_, NDH functions without any 374 supercomplex (Ueda et al. 2012). In chilled cucumber leaves, the partial disassembly of the supercomplex 375 accompanies the loss of NDH activity and PSI photoinhibition (Takeuchi et al. 2026; § 1.3). Cyanobacterial NDH-1 376 has two candidate Fd-binding sites (Laughlin et al. 2019). Whether the supercomplex changes the supply of reduced 377 Fd to NDH or its local redox environment under FR is a legitimate question (Table S3, prediction 9). 378 **What cannot yet be measured (Table S2, row 10).** Finally, the flux through NDH has never been directly measured 379 in illuminated leaves. Every available method carries an assumption: the post-illumination fluorescence rise is 380 measured in the dark, when Δp has collapsed and NDH is least constrained, so that it reports capacity rather than the 381 flux in the light (Strand et al. 2017); Fd-dependent PQ reduction is measured in ruptured chloroplasts; the re-oxidation 382 of P700 by FR (Shikanai et al. 2025) and the difference between the electron transport rates of PSI and PSII are 383 indirect; and the proton-flux slopes of Strand et al. (2017) assume fixed stoichiometries for linear flow. The two 384 variables of the gating hypothesis, Δ _E_ h and Δp, have not been measured together under red and FR light, and the ATP 385 level under FR has not been measured at all. Our FR data rely on Y(ND), which reports the extent of P700 oxidation 386 but not the engagement of photosynthetic control; the rate constant of P700⁺ re-reduction is a more direct indicator 387 (Degen and Johnson 2024). These gaps are addressed in Text S3. 388 18 389 **4.** **Supplementary Table S3 and Supplementary Text S3** 390 391 **Supplementary Table S3.** Testable predictions of the masking and redox–Δp gating hypotheses 1 The sign of the net driving force (Δ _E_ h − _n_ - Δp) differs between red and FR light 2 The NDH-dependent rate falls as the imposed Δp rises; the antimycin A-sensitive rate does not 3 NDH carries a larger share of cyclic flux under FR in leaves 4 NDH runs forward only within an operating window: upper limit set by the pmf, lower limit by the supply of reduced Fd and NADPH 5 Fd supply, Fd affinity and the ATP level favor NDH under FR Leaves: pmf from ECS and redox states of P700, PC and PQ under red vs red + FR; driving force calculated as in Strand et al. (2017) Coupled thylakoids under FR only; Δp set without ATP (acid–base transition, K [+] diffusion potential, ATP synthase inhibition); PIFR or P700 [+] re-reduction, ± antimycin A or _pgr5_ ; PQ redox state set by the FR:red ratio Proton flux (ECS) vs PSI electron flux (Strand et al. 2017) under red vs FR, ± antimycin A, and in _crr_ ; KEA3 effects on DIRK taken into account NDH share (as in 3) across irradiances under FRsupplemented light in plants of known growth irradiance Apparent _K_ m for Fd (uncoupled, different species: ~2 µM NDH, ~1 µM antimycin A-sensitive; Fisher et al. 2019) in the same species and in coupled thylakoids ± antimycin A, _crr_, _pgr5_ ; reduced fraction of Fd and Fd-dependent CEF PSI from Fd oxidation kinetics in leaves under FR (Ohnishi et al. 2023; Satoh et al. 2025); stromal ATP under FR-supplemented low light 19 Net driving force more positive under FR, mainly through PQ oxidation; the low-light phase stays within the operating window of NDH Monotonic decline with Δp; increase as PQ is oxidized H [+] /e [−] ratio rises with FR, survives antimycin A, and is absent in _crr_ Share rises while the pmf builds and falls once it saturates (~140 µmol photons m [−2] s [−1] for plants grown at 50; Wang et al. 2015); share falls again when the red component is lowered until the PQ pool is almost fully oxidized _K_ m changes with Δp; reduced Fd small but not zero under FR, and Fd-dependent CEFPSI rises with FR in the wild type and _pgr5_ but not in _crr_ ; ATP in the range where NDH is active but the PGR5 pathway is inhibited Net driving force equally positive in both regimes No Δp dependence No change, or a change lost with antimycin A Monotonic increase with irradiance; no decline at the oxidized end _K_ m independent of Δp; Fddependent CEF-PSI rises with FR in _crr_ ; ATP outside that range _Testing the masking hypothesis and the role of P700_ _[+]_ 6 The division of roles is set by the light regime 7 FR protection bypasses PGR5 and requires NDH 8 P700 is oxidized by different means in the two regimes _Structure_ 9 The supercomplex matters only if it supplies Fd to NDH _Ecological and applied questions_ 10 FR and a low cytochrome _b_ 6 _f_ content protect PSI by the same route 11 The NDH growth defect under canopy-like FL depends on FR Full matrix: 2-min regime (Kono and Terashima 2016) and pulse regime (Kono et al. 2017) × _pgr5_, _crr_ × ± FR _pgr5 crr4-3_ double mutant under FR-supplemented FL (young leaves, low growth light) Series of FR intensities or FR:red ratios; P m loss vs residual Y(NA) and Y(ND); rate constant of P700 [+] rereduction measured alongside Y(ND) (Degen and Johnson 2024) _lhca5 lhca6_ double mutant and Lhca6 domain-swap lines (Otani et al. 2017) under FR-supplemented FL Intermediate-Rieske tobacco (Kono et al. 2025) ± FRsupplemented FL; growth under canopy-like light _crr_, _pgr5_ and wild type under canopy-spectrum rapid FL with FR present or filtered out; PSI photoinhibition and biomass measured together 2-min regime: FR rescues _pgr5_ only partially; pulse regime: removing FR abolishes the genotype dependence P m loss tracks Y(NA) (or both); FR raises Y(ND) without slowing re-reduction, red high light raises it by slowing re-reduction Wild-type behavior (if the NDH content is adequate) Combined treatment protects PSI with a smaller loss of capacity than either alone; intermediate lines show intermediate protection Defect larger with FR and then accompanied by P m loss; removing FR shifts the dependence toward PGR5 FR rescues _pgr5_ fully in both regimes FR rescues _pgr5_ but not the double mutant FR rescues the double mutant FR slows P700 [+] re-reduction as red high light does _crr_ -like behavior No additivity Defect independent of FR 392 Predictions are grouped by the question they address; each is described in full in Text S3. DIRK, dark-interval relaxation kinetics; ECS, electrochromic shift; PC, plastocyanin; PIFR, post393 illumination fluorescence rise. 394 395 396 20 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.02.763831; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.02.763831) preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 397 **Experiments to test the masking and redox–Δp gating hypotheses** 398 The hypotheses presented in § 5.3 make predictions that can be tested using existing methods. We list them here, 399 grouped by the question they address, and state for each what result would support or refute the hypothesis. The 400 numbers correspond to those in Table S3. 401 402 **Testing redox–Δp gating directly (Table S3, predictions 1–5)** 403 (1) _The sign of the net driving force (ΔE_ _h_ _− n·Δp) under red and FR light._ The central prediction is that FR shifts the 404 thermodynamic balance of NDH toward the forward reaction. This can be tested in leaves by measuring the pmf from 405 the electrochromic shift and the redox states of P700, plastocyanin, and the PQ pool simultaneously under red light 406 and under red light supplemented with FR, and by calculating the driving force from the measured redox states, as 407 Strand et al. (2017) did for isolated thylakoids. We predict that the driving force for the reduction of PQ by reduced 408 Fd exceeds _n_ - Δp by a larger margin under FR than under the high-light phase of red FL, and that the difference arises 409 mainly from the oxidation of the PQ pool rather than from a change in total pmf (Kono et al. 2020). The same 410 measurements should show whether, in the 10-s low-light phase, the PQ pool stays within the operating window of 411 NDH (§ 5.3): oxidized enough to keep Δ _E_ h large, but not so oxidized that the supply of reduced donors limits cycling 412 (Joët et al. 2002). If the net driving force is found to be equally positive in both regimes, gating cannot explain the 413 wavelength dependence, and the explanation must be sought in capacity, ATP feedback, or activation. 414 (2) _The Δp dependence of NDH in coupled thylakoids without competition from linear flow._ The uncoupler 415 experiments of Strand et al. (2017) showed that NDH is sensitive to Δp, but uncouplers abolish the pmf rather than 416 set it to graded values. In coupled thylakoids illuminated with FR only, so that linear flow is absent, the pmf can be 417 set to defined values without ATP. This can be done by an acid–base transition in which a permeant organic acid 418 carries protons into the lumen (Jagendorf and Uribe 1966; Uribe and Jagendorf 1967a, b), by a K⁺ diffusion potential, 419 or by inhibition of the ATP synthase. The NDH-dependent reduction of PQ can then be followed by the post420 illumination fluorescence rise or by the re-reduction of P700⁺, with the PGR5-dependent component removed by 421 antimycin A or using _pgr5_ . We predict that the NDH-dependent rate falls monotonically as the imposed Δp increases, 422 whereas the antimycin A-sensitive rate does not. Imposing the pmf without ATP is essential because ATP inhibits 423 both pathways directly (Fisher et al. 2019). In the same system, the PQ redox state can be set by the ratio of FR to 424 red light, and the NDH-dependent rate should increase as the pool is oxidized. 425 (3) _The effective flux in leaves under competition from linear flow._ The proton-flux method of Strand et al. (2017), 426 which compares the proton flux from the electrochromic shift with the electron flux through PSI, can be applied to 427 leaves under red and FR-supplemented light, with and without antimycin A, and in the _crr_ mutants. If NDH carries 428 a larger share of the cyclic flux under FR, the ratio of proton to electron flux should increase with FR in the wild type, 429 this increase should survive antimycin A, and it should be absent in _crr_ mutants. The dark-interval relaxation kinetics 430 on which this method rests must be interpreted with care: the thylakoid K⁺/H⁺ antiporter KEA3, which exports protons 431 from the lumen in exchange for K⁺ and is activated when the light decreases and the stromal energy state falls 432 (Armbruster et al. 2014, 2016; Uflewski et al. 2024), can alter the relaxation of the electrochromic shift during the 433 dark interval and thus the proton flux and the ΔpH/ΔΨ partitioning derived from it. The difference between the 434 maximal NDH activity measured in experiment (2) and the effective flux measured in experiment (3) would quantify, 435 for the first time, how much of the capacity of NDH is actually used in a leaf under a given light quality. 436 (4) _The operating window of NDH._ § 5.3 proposed that NDH runs forward only within an operating window, whose 437 upper limit is set by the pmf and whose lower limit is set by the supply of reduced Fd and NADPH. The upper limit 438 can be tested by measuring the NDH-dependent share of cyclic flux, as in experiment (3), across a series of irradiances 439 under FR-supplemented light in plants with known growth irradiance. We predict that the NDH share increases with 440 irradiance while the pmf is still building and falls once the pmf saturates, which in plants grown at 50 µmol photons 441 m⁻² s⁻¹ occurs at approximately 140 µmol photons m⁻² s⁻¹ (Wang et al. 2015). A monotonic increase would argue 21 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.02.763831; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.02.763831) preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 442 against an upper limit set by the pmf. The lower limit can be tested in the same series by lowering the red component 443 of the light until the PQ pool is almost fully oxidized: the NDH share should then fall again, as cycling did in tobacco 444 leaves under FR alone (Joët et al. 2002). 445 (5) _Fd supply, Fd affinity and ATP under FR._ The apparent _K_ m values for Fd of the two pathways (about 2 µM for 446 NDH and about 1 µM for the antimycin A-sensitive pathway) were obtained in uncoupled thylakoids and in different 447 species (Fisher et al. 2019; Text S1). Measuring them in the same species and in coupled thylakoids, with and without 448 antimycin A and in _crr_ and _pgr5_, would show whether the affinity for Fd itself changes with Δp, which would be a 449 second route by which the redox poise could favor one pathway. Because both pathways compete with FNR and the 450 other Fd-dependent enzymes for reduced Fd on roughly equal terms (Text S1), the quantity that limits the supply to 451 CEF-PSI _in vivo_ is the reduced fraction of Fd rather than its total amount. This fraction can now be measured in 452 leaves from the near-infrared absorbance of Fd (Ohnishi et al. 2023; Satoh et al. 2025). Ideally, the decay of reduced 453 Fd would be resolved into the rate constants of its individual sinks, so that the share of each pathway could be read 454 directly; at the present signal-to-noise ratio of the Fd signal this is not feasible, and the comparison has to rely on 455 mutants and inhibitors. We predict that under FR-supplemented low light it is small but not zero, and that the Fd456 dependent CEF-PSI measured from the oxidation kinetics of Fd rises with FR in the wild type and in _pgr5_ but not in 457 _crr_ mutants. In parallel, the stromal ATP level under FR-supplemented low light has never been measured; if it falls 458 into the range in which NDH is active but the PGR5-dependent pathway is inhibited (Fisher et al. 2019), the ATP 459 feedback and the thermodynamic gating would be acting together, and their contributions would have to be separated 460 by experiment (2). The discrepancy between the inhibition by ATP found by Fisher et al. (2019) and its absence in 461 the assay of Cleland and Bendall (1992) should be resolved in the same system (Table S2, row 3). 462 463 **Testing the masking hypothesis and the role of P700⁺ (Table S3, predictions 6–8)** 464 (6) _Regime × FR._ Our two FL regimes have never been compared under the same conditions. We predict that in the 465 2-min regime of Kono and Terashima (2016), in which the high-light phase is long and the PQ pool is reduced 466 throughout it, FR will rescue _pgr5_ only partially, because the pmf-driven oxidation of P700 during the high-light 467 phase still depends on the PGR5-dependent pathway; and that in the pulse regime of Kono et al. (2017), removing 468 FR abolishes the genotype dependence, as observed. The full matrix of two regimes, two mutants, and FR present or 469 absent would test whether the division of roles is set by the light regime, as the model claims. 470 (7) _The pgr5 crr double mutant under FR._ If FR protection runs through NDH and bypasses PGR5, FR should fail to 471 rescue the _pgr5 crr4-3_ double mutant, whereas it should rescue _pgr5_ . The double mutant grows poorly (Munekage 472 et al. 2004); therefore, the experiment must use young leaves under low growth light, but the prediction is 473 unambiguous. 474 (8) _Separating Y(ND) from Y(NA)._ Under FR, the two parameters change together (Text S2). A series of FR intensities 475 or FR-to-red ratios in which Y(ND) rises before Y(NA) falls would separate them; the loss of P m should then correlate 476 with the residual Y(NA) rather than with Y(ND) if the acceptor side is decisive, and with both if P700⁺ contributes 477 in its own right. In the same experiments, the rate constant of P700⁺ re-reduction should be measured alongside 478 Y(ND) (Degen and Johnson 2024). We predict that FR raises Y(ND) in the low-light phase without slowing P700⁺ 479 re-reduction, that is, without engaging photosynthetic control, whereas red high light raises Y(ND) by slowing it. 480 This would provide the first direct evidence that P700 is oxidized by different means in the two regimes (Table 2). 481 482 **Structure (Table S3, prediction 9)** 483 (9) _NDH–PSI supercomplex._ Two kinds of plants lack the supercomplex but retain NDH: the _lhca5 lhca6_ double 484 mutant and the domain-swap lines of Otani et al. (2017), in which the stromal loop of Lhca6 is replaced. These should 485 be exposed to FR-supplemented FL. If the supercomplex is needed only for the stability of NDH, these lines should 486 behave like the wild type, as long as their NDH content is adequate. If the supercomplex is needed for the supply of 22 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.02.763831; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.02.763831) preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 487 reduced Fd to NDH, they should behave like _crr_ mutants under FR. 488 489 **Ecological and applied questions (Table S3, predictions 10 and 11)** 490 (10) _Trading capacity for protection._ Tobacco with a low Cyt _b_ ₆ _f_ content maintains PSI integrity under extreme pulses, 491 but at the cost of photosynthetic capacity (Kono et al. 2025). If FR and a low Cyt _b_ ₆ _f_ content protect PSI by the same 492 route (§ 5.4), a moderate reduction in Rieske protein combined with FR-supplemented FL should protect PSI with a 493 smaller loss of capacity than either treatment alone, and plants with intermediate Rieske content, which have not yet 494 been tested for photoinhibition, should show intermediate protection. The growth of these lines under a canopy-like 495 light regime would show whether the trade-off can be shifted in practice or not. 496 (11) _Light quality as an independent variable in the field._ The growth defect of NDH mutants under FL that 497 reproduced the irradiance, but not the spectrum, of a maize canopy (Theeuwen et al. 2022, preprint) and the protection 498 of PSI by FR in the laboratory (Kono et al. 2017) have not been connected. Why NDH matters under such rapid 499 fluctuations even without FR is itself open, and our framework suggests a hypothesis. A canopy regime spends most 500 of its time at low irradiance, and its bright spots are too short for the pmf to reach the level that stops NDH; the role 501 of NDH under this regime would then be the one described in § 5.3 and in Table S3, prediction 4, namely to keep the 502 intersystem chain poised and to supply ATP during the low-light periods, rather than to protect PSI directly. This 503 agrees with the observation that the Burren _ndhG_ variant affects PSII efficiency only below about 350 µmol photons 504 m⁻² s⁻¹ (Theeuwen et al. 2022, preprint), and it predicts that the growth defect of the NDH mutants under such a 505 regime is accompanied by a more reduced PQ pool and a lower pmf in the low-light periods, but not necessarily by 506 a loss of P m . The experiment that connects them is to grow _crr_, _pgr5_, and wild-type plants under a rapidly fluctuating 507 regime that reproduces a canopy spectrum, with FR either present or removed by a filter, and to measure PSI 508 photoinhibition and biomass together. We predict that the growth defect of the NDH mutants is larger with FR than 509 without, that it is accompanied by a loss of P m when FR is present, and that removing FR shifts the dependence from 510 NDH toward PGR5. This experiment would also test the suggestion of § 9.2 that red LED regimes have overestimated 511 the role of PGR5 in nature. 512 If experiments (1) to (5) fail to show a wavelength-dependent thermodynamic constraint on NDH, the 513 redox–Δp gating hypothesis should be abandoned, and the division of roles would have to be explained by capacity, 514 ATP feedback, and activation alone. The masking hypothesis would survive that outcome, and so would the 515 observations that rest on it: that FR oxidizes P700 by the excitation balance, that this removes the specific need for 516 PGR5, and that NDH provides the low-flux recycling under FR. They do not depend on how NDH is gated, but only 517 on the fact that NDH is the pathway left when a fast, pmf-driven oxidation of P700 is no longer required. 518 23 [bioRxiv preprint doi: https://doi.org/10.64898/2026.11.02.763831; this version posted November 3, 2026. The copyright holder for this](https://doi.org/10.64898/2026.11.02.763831) preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 519 **References cited only in Supplementary Texts S1–S3.** 520 For other references, see the reference list of the main text. 521 Armbruster U, Carrillo LR, Venema K, Pavlovic L, Schmidtmann E, Kornfeld A, Jahns P, Berry JA, Kramer DM, 522 Jonikas MC (2014) Ion antiport accelerates photosynthetic acclimation in fluctuating light environments. Nat 523 [Commun 5:5439. https://doi.org/10.1038/ncomms6439](https://doi.org/10.1038/ncomms6439) 524 Armbruster U, Leonelli L, Correa Galvis V, Strand D, Quinn EH, Jonikas MC, Niyogi KK (2016) Regulation and 525 levels of the thylakoid K+/H+ antiporter KEA3 shape the dynamic response of photosynthesis in fluctuating 526 [light. Plant Cell Physiol 57:1557–1567. https://doi.org/10.1093/pcp/pcw085](https://doi.org/10.1093/pcp/pcw085) 527 Batie CJ, Kamin H (1984a) Ferredoxin:NADP+ oxidoreductase. Equilibria in binary and ternary complexes with 528 NADP+ and ferredoxin. J Biol Chem 259:8832–8839. [https://doi.org/10.1016/S0021-9258(17)47229-2](https://doi.org/10.1016/S0021-9258(17)47229-2) 529 Batie CJ, Kamin H (1984b) Electron transfer by ferredoxin:NADP+ reductase. Rapid-reaction evidence for 530 [participation of a ternary complex. J Biol Chem 259:11976–11985. https://doi.org/10.1016/S0021-](https://doi.org/10.1016/S0021-9258(20)71306-2) 531 [9258(20)71306-2](https://doi.org/10.1016/S0021-9258(20)71306-2) 532 Bukhov NG, Carpentier R (2003) Measurement of photochemical quenching of absorbed quanta in photosystem I 533 of intact leaves using simultaneous measurements of absorbance changes at 830 nm and thermal dissipation. 534 Planta 216:630–638. [https://doi.org/10.1007/s00425-002-0886-2](https://doi.org/10.1007/s00425-002-0886-2) 535 Hanke GT, Hase T (2008) Variable photosynthetic roles of two leaf-type ferredoxins in _Arabidopsis_, as revealed by 536 RNA interference. Photochem Photobiol 84:1302–1309. [https://doi.org/10.1111/j.1751-1097.2008.00411.x](https://doi.org/10.1111/j.1751-1097.2008.00411.x) 537 Holtgrefe S, Bader KP, Horton P, Scheibe R, von Schaewen A, Backhausen JE (2003) Decreased content of leaf 538 ferredoxin changes electron distribution and limits photosynthesis in transgenic potato plants. Plant Physiol 539 133:1768–1778. [https://doi.org/10.1104/pp.103.026013](https://doi.org/10.1104/pp.103.026013) 540 Kimata-Ariga Y, Matsumura T, Kada S, Fujimoto H, Fujita Y, Endo T, Mano J, Sato F, Hase T (2000) Differential 541 electron flow around photosystem I by two C4-photosynthetic-cell-specific ferredoxins. EMBO J 19:5041– 542 [5050. https://doi.org/10.1038/sj.emboj.7593319](https://doi.org/10.1038/sj.emboj.7593319) 543 Uflewski M, Rindfleisch T, Korkmaz K, Tietz E, Mielke S, Correa Galvis V, Dünschede B, Luzarowski M, Skirycz 544 A, Schwarzländer M, Strand DD, Hertle AP, Schünemann D, Walther D, Thalhammer A, Wolff M, Armbruster 545 U (2024) The thylakoid proton antiporter KEA3 regulates photosynthesis in response to the chloroplast energy 546 [status. Nat Commun 15:2792. https://doi.org/10.1038/s41467-024-47151-5](https://doi.org/10.1038/s41467-024-47151-5) 547 Uribe EG, Jagendorf AT (1967a) On the localization of organic acids in acid-induced ATP synthesis. Plant Physiol 548 42:697–705. [https://doi.org/10.1104/pp.42.5.697](https://doi.org/10.1104/pp.42.5.697) 549 Uribe EG, Jagendorf AT (1967b) Organic acid specificity for acid-induced ATP synthesis by isolated chloroplasts. 550 Plant Physiol 42:706–711. [https://doi.org/10.1104/pp.42.5.706](https://doi.org/10.1104/pp.42.5.706) 551 Yamamoto H, Kato H, Shinzaki Y, Horiguchi S, Shikanai T, Hase T, Endo T, Nishioka M, Makino A, Tomizawa K, 552 Miyake C (2006) Ferredoxin limits cyclic electron flow around PSI (CEF-PSI) in higher plants—stimulation of 553 CEF-PSI enhances non-photochemical quenching of Chl fluorescence in transplastomic tobacco. Plant Cell 554 Physiol 47:1355–1371. [https://doi.org/10.1093/pcp/pcl005](https://doi.org/10.1093/pcp/pcl005) 555 24