Specific Incorporation of Polyunsaturated Fatty Acids into the sn-2 Position of Phosphatidylglycerol Accelerates Photodamage to Photosystem II under Strong Light

Free fatty acids (FFAs) are generated by the reaction of lipases with membrane lipids. Generated polyunsaturated fatty acids (PUFAs) containing more than two double bonds have toxic effects in photosynthetic organisms. In the present study, we examined the effect of exogenous FFAs in the growth medium on the activity of photosystem II (PSII) under strong light in the cyanobacterium Synechocystis sp. PCC 6803 (Synechocystis). PUFAs but not monounsaturated fatty acids accelerated the rate of photodamage to PSII by inactivating electron transfer at the oxygen-evolving complex. Moreover, supplemented PUFAs were specifically incorporated into the sn-2 position of phosphatidylglycerol (PG), which usually contains C16 fatty acids at the sn-2 position in Synechocystis cells. The disruption of the gene for an acyl-ACP synthetase reduced the effect of PUFAs on the photoinhibition of PSII. Thus, the specific incorporation of PUFAs into PG molecules requires acyl-ACP synthetase and leads to an unstable PSII, thereby accelerating photodamage to PSII. Our results are a breakthrough into elucidating the molecular mechanism of the toxicity of PUFAs to photosynthetic organisms.


Introduction
Fatty acids (FAs) are essential components consisting of membrane lipids in living cells. Free fatty acids (FFAs) are produced by lipases that hydrolyze acyl chains from membrane lipids and become substrates used for the biosynthesis of storage compounds and plant hormones as well as acylation of proteins. FFAs have a broad diversity in terms of the chain length, number, and position of double bonds with cis/trans configuration. Such a wide spectrum has different effects on photosynthetic activities, as reported in Synechocystis sp. PCC 6803 (hereafter Synechocystis) [1]. A polyunsaturated FA (PUFA), α-linolenic acid (18:3 ∆9, 12,15 ), increased the sensitivity of photosystem II (PSII) to lightinduced damage to photosynthesis (photoinhibition), which had toxic effects on cellular growth [1,2]. It also inhibited photosynthetic activity, and the inhibition was mediated by its incorporation into membrane lipids via the action of the acyl-acyl carrier protein

PUFAs Accelerate Photodamage by Inhibiting Electron Transfer at the Oxygen-Evolving Complex
Unsaturated FFAs induce light-inducible photodamage in PSII, as we reported previously [1]. To further investigate the effect of unsaturated FAs on photosynthesis, we examined the effect of several unsaturated FFAs containing double bonds of different numbers, at different positions and with different configurations (cis/trans) on the photoinhibition of PSII (Figure 1a). Photodamage to PSII was observed in the presence of lincomycin, an inhibitor of protein synthesis that is required for the repair of PSII ( Figure 1b). As we reported previously, 18:3 ∆9, 12,15 , which has three cis double bonds at the ∆9, 12, and 15 positions, accelerated photoinhibition and photodamage to PSII, but oleic acid (18:1 ∆9 ), with a cis double bond at the ∆9 position, did not ( Figure 1). Other monounsaturated FAs (MUFAs), such as cis-12-octadecenoic acid (18:1 ∆12 ), containing a cis double bond at ∆12, and elaidic acid (18:1 ∆9t ), containing a trans double bond at ∆9, did not accelerate the photodamage to PSII (Figure 1b). γ-linolenic acid (18:3 ∆6, 9,12 ) and linoleic acid (18:2 ∆9,12 ) induced photoinhibition and photodamage to PSII to a similar extent to that induced by 18:3 ∆9,12,15 (Figure 1), thus PUFAs but not cisor trans-MUFAs induced photodamage to PSII. By contrast, under weak light, 18:3 ∆9,12,15 did not affect the PSII activity. We chose 18:3 ∆9,12,15 as a model PUFA and 18:1 ∆12 for MUFA for the further experiments as described below. The content of these unsaturated FAs was very low in Synechocystis cells grown without FA supplementation, and thus, it was easy to trace them after incorporation into the cells. According to the two-step model, photodamage to PSII can occur in two steps: (1) corruption of the OEC and (2) inactivation of electron transfer in the reaction center of PSII [14,15]. We examined the effect of PUFAs on the electron transfer activity of PSII by using diphenylcarbazide (DPC) and 2,6-dichlorophenolindophenol (DCIP) as an electron donor and acceptor, respectively [17]. The activity of electron transfer from H2O or DPC to DCIP in the thylakoid membranes from wild-type (WT) cells incubated with or without 18:3 Δ9,12,15 decreased under exposure to strong light for 20 min ( Figure 2). Moreover, 18:3 Δ9,12,15 inhibited the electron transfer activity from H2O to DCIP (Figure 2a), which corresponds to the whole PSII activity and is consistent with the results obtained in Figure  1a. The electron transfer activity from DPC to DCIP, which corresponds to the electron transfer within D1/D2, decreased at a similar rate to that in the thylakoid membranes from cells incubated without 18:3 Δ9,12,15 (Figure 2b). When incubated with 18:1 Δ12 , the electron transfer activity from H2O or DPC to DCIP decreased at a similar rate to that in thylakoid membranes from cells incubated without 18:1 Δ12 (Figure 2). Therefore, 18:3 Δ9,12,15 accelerated the corruption of the OEC but not inactivation of electron transfer within the reaction center of PSII.  According to the two-step model, photodamage to PSII can occur in two steps: (1) corruption of the OEC and (2) inactivation of electron transfer in the reaction center of PSII [14,15]. We examined the effect of PUFAs on the electron transfer activity of PSII by using diphenylcarbazide (DPC) and 2,6-dichlorophenolindophenol (DCIP) as an electron donor and acceptor, respectively [17]. The activity of electron transfer from H 2 O or DPC to DCIP in the thylakoid membranes from wild-type (WT) cells incubated with or without 18:3 ∆9,12,15 decreased under exposure to strong light for 20 min ( Figure 2). Moreover, 18:3 ∆9,12,15 inhibited the electron transfer activity from H 2 O to DCIP (Figure 2a), which corresponds to the whole PSII activity and is consistent with the results obtained in Figure 1a. The electron transfer activity from DPC to DCIP, which corresponds to the electron transfer within D1/D2, decreased at a similar rate to that in the thylakoid membranes from cells incubated without 18:3 ∆9,12,15 (Figure 2b). When incubated with 18:1 ∆12 , the electron transfer activity from H 2 O or DPC to DCIP decreased at a similar rate to that in thylakoid membranes from cells incubated without 18:1 ∆12 (Figure 2). Therefore, 18:3 ∆9,12,15 accelerated the corruption of the OEC but not inactivation of electron transfer within the reaction center of PSII.

18:3 Δ9,12,15 Destabilizes Photosynthetic Complexes under Strong Light
To investigate the effect of 18:3 Δ9,12,15 on photosynthetic complexes under strong light, we analyzed protein complexes in thylakoid membranes from cells treated with 50 μmol L −1 18:3 Δ9,12,15 under strong light ( Figure 3). In cells treated without 18:3 Δ9,12,15 , exposure to strong light slightly decreased the amount of PSII dimer and PSI trimer, and increased PSI and PSII monomers ( Figure 3). By contrast, in cells incubated with 18:3 Δ9,12,15 under strong light, mega-complexes and the amount of PSI trimer and PSII dimer and monomer complexes decreased greatly, and the amount of PSI monomer and CP43-less PSII monomer complexes (RC47) increased ( Figure 3). Therefore, 18:3 Δ9,12,15 destabilized and dissociated photosystem complexes. However, 18:1 Δ12 did not have a significant effect on the photosystem complexes under strong light ( Figure 3). Several lipid molecules located between protein subunits of the photosystems function as glue to stabilize the tertiary structures [18]. Thus, the incorporation of PUFAs into membrane lipids might affect the stability of protein complexes in the thylakoid membrane, which would lead to disassembly and inactivation of the PSII complexes. Values are the means ± SD (bars) of three independent experiments. Asterisks indicate statistically significant differences (* p < 0.05; Student's t-test).

18:3 ∆9,12,15 Destabilizes Photosynthetic Complexes under Strong Light
To investigate the effect of 18:3 ∆9,12,15 on photosynthetic complexes under strong light, we analyzed protein complexes in thylakoid membranes from cells treated with 50 µmol L −1 18:3 ∆9,12,15 under strong light ( Figure 3). In cells treated without 18:3 ∆9,12,15 , exposure to strong light slightly decreased the amount of PSII dimer and PSI trimer, and increased PSI and PSII monomers ( Figure 3). By contrast, in cells incubated with 18:3 ∆9,12,15 under strong light, mega-complexes and the amount of PSI trimer and PSII dimer and monomer complexes decreased greatly, and the amount of PSI monomer and CP43-less PSII monomer complexes (RC47) increased ( Figure 3). Therefore, 18:3 ∆9,12,15 destabilized and dissociated photosystem complexes. However, 18:1 ∆12 did not have a significant effect on the photosystem complexes under strong light ( Figure 3). Several lipid molecules located between protein subunits of the photosystems function as glue to stabilize the tertiary structures [18]. Thus, the incorporation of PUFAs into membrane lipids might affect the stability of protein complexes in the thylakoid membrane, which would lead to disassembly and inactivation of the PSII complexes.

FFAs Are Incorporated into the sn-2 Position of PG Molecules by the Action of AAS under Strong Light
To investigate the incorporation of 18:3 ∆9,12,15 into membrane lipids, we analyzed the complexes (RC47) increased ( Figure 3). Therefore, 18:3 Δ9,12,15 destabilized and dissociated photosystem complexes. However, 18:1 Δ12 did not have a significant effect on the photosystem complexes under strong light ( Figure 3). Several lipid molecules located between protein subunits of the photosystems function as glue to stabilize the tertiary structures [18]. Thus, the incorporation of PUFAs into membrane lipids might affect the stability of protein complexes in the thylakoid membrane, which would lead to disassembly and inactivation of the PSII complexes. for 20 min were solubilized by 1% (w/v) n-dodecyl-β-D-maltoside and separated by Blue-Native polyacrylamide gel electrophoresis. Bands on the gels were visualized by washing out excess Coomassie Brilliant Blue G-250 with distilled water. Bands for the photosynthetic complexes are identified as described previously [19]. PSI-T, Photosystem I trimer; PSII-D, Photosystem II dimer; PSI-M, Photosystem I monomer; PSII-M, Photosystem II monomer; RC47-M, CP43-less Photosystem II monomer.

FFAs Are Incorporated into the sn-2 Position of PG Molecules by the Action of AAS under Strong Light
To investigate the incorporation of 18:3   To check whether AAS is involved in the incorporation of 18:3 ∆9,12,15 into PG under strong light, we determined the content of 18:3 ∆9,12,15 in membrane lipids in aas mutant cells generated by insertion of the kanamycin-resistance gene cassette in the middle of the gene ( Figure 5). When the mutant cells were incubated with 18:3 ∆9,12,15 under strong light for 20 min, the specific incorporation of 18:3 ∆9,12,15 into PG was much reduced (Figure 4c). We also checked the incorporation of 18:1 ∆12 , which did not have any effects on PSII photoinhibition (Figure 1a). Synechocystis cells do not contain 18:1 ∆12 because a double bond at the ∆12 position is not introduced by the ∆12 desaturase before its introduction to the ∆9 position by ∆9 desaturase [21]. Therefore, its presence in membrane lipids indicates the incorporation of exogeneous 18:1 ∆12 into membrane lipids. The 18:1 ∆12 was mainly detected in PG (Figure 4d), so 18:1 ∆12 is also specifically incorporated into PG. generated by insertion of the kanamycin-resistance gene cassette in the middle of the gene ( Figure 5). When the mutant cells were incubated with 18:3 Δ9,12,15 under strong light for 20 min, the specific incorporation of 18:3 Δ9,12,15 into PG was much reduced (Figure 4c). We also checked the incorporation of 18:1 Δ12 , which did not have any effects on PSII photoinhibition (Figure 1a). Synechocystis cells do not contain 18:1 Δ12 because a double bond at the Δ12 position is not introduced by the Δ12 desaturase before its introduction to the Δ9 position by Δ9 desaturase [21]. Therefore, its presence in membrane lipids indicates the incorporation of exogeneous 18:1 Δ12 into membrane lipids. The 18:1 Δ12 was mainly detected in PG (Figure 4d), so 18:1 Δ12 is also specifically incorporated into PG. Synechocystis PG molecules usually contain C18 FAs at the sn-1 position and C16 FAs at the sn-2 position (Figure 6a) [20,22]. We checked whether 18:1 Δ12 and 18:3 Δ9,12,15 were incorporated into the sn-1 or 2 position of PG molecules by treating purified PG from FAtreated Synechocystis cells with honey bee phospholipase A2, which specifically cleaves FAs at the sn-2 position of PG molecules. Most of the 18:1 Δ12 and 18:3 Δ9,12,15 was detected at the sn-2 position but not the sn-1 position of PG molecules (Figure 6b,c). Therefore, 18:1 Δ12 and 18:3 Δ9,12,15 were specifically incorporated into the sn-2 position of PG by the action of AAS. Our results are consistent with the finding that FAs in PG molecules are remodeled in Synechocystis cells [23]. These results suggest that the incorporation of PUFAs into the sn-2 position of PG destabilizes photosystems, as shown in Figure 3, and enhances the photodamage to PSII. Synechocystis PG molecules usually contain C18 FAs at the sn-1 position and C16 FAs at the sn-2 position (Figure 6a) [20,22]. We checked whether 18:1 ∆12 and 18:3 ∆9,12,15 were incorporated into the sn-1 or 2 position of PG molecules by treating purified PG from FA-treated Synechocystis cells with honey bee phospholipase A 2 , which specifically cleaves FAs at the sn-2 position of PG molecules. Most of the 18:1 ∆12 and 18:3 ∆9,12,15 was detected at the sn-2 position but not the sn-1 position of PG molecules (Figure 6b,c). Therefore, 18:1 ∆12 and 18:3 ∆9,12,15 were specifically incorporated into the sn-2 position of PG by the action of AAS. Our results are consistent with the finding that FAs in PG molecules are remodeled in Synechocystis cells [23]. These results suggest that the incorporation of PUFAs into the sn-2 position of PG destabilizes photosystems, as shown in Figure 3, and enhances the photodamage to PSII.

Disruption of the Gene for AAS Protects PSII against PUFA-Induced Photoinhibition
Disruption of the gene for AAS induces tolerance in cells to the PUFA-induced inhibition of cellular growth [3,4]. To examine the role of AAS in PUFA-induced PSII photoinhibition, we performed similar experiments with a mutant lacking AAS. AAS was required for the incorporation of 18:3 ∆9,12,15 into membrane lipids, especially PG (Figure 4c). In wild-type cells, PSII activity dropped to 8% of the initial activity after the 80 min exposure to strong light in the presence of 18:3 ∆9,12,15 (Figure 1a). AAS mutant cells had 46% PSII activity in the same time (Figure 7a), indicating that the lack of AAS reduced the effect of 18:3 ∆9,12,15 on PSII photoinhibition or photodamage (Figure 7). Therefore, the incorporation of PUFAs into PG mediated by the action of AAS accelerates photodamage to PSII. In Syne-chococcus elongatus PCC 7942, the lack of AAS accelerated the photoinhibition of PSII [24]. A mutant lacking AAS had 42% of the initial PSII activity after 80 min of exposure to strong light, whereas wild-type cells retained 50% under the same conditions (Figures 1a and 7a). Thus, the lack of AAS in Synechocystis slightly accelerated the photoinhibition of PSII as well. However, the effect was smaller than that observed in Synechococcus elongatus PCC 7942 under these conditions.

Disruption of the Gene for AAS Protects PSII against PUFA-Induced Photoinhibition
Disruption of the gene for AAS induces tolerance in cells to the PUFA-induced inhibition of cellular growth [3,4]. To examine the role of AAS in PUFA-induced PSII photoinhibition, we performed similar experiments with a mutant lacking AAS. AAS was required for the incorporation of 18:3 Δ9,12,15 into membrane lipids, especially PG (Figure 4c). In wild-type cells, PSII activity dropped to 8% of the initial activity after the 80 min exposure to strong light in the presence of 18:3 Δ9,12,15 (Figure 1a). AAS mutant cells had 46% PSII activity in the same time (Figure 7a), indicating that the lack of AAS reduced the effect of 18:3 Δ9,12,15 on PSII photoinhibition or photodamage (Figure 7). Therefore, the incorporation of PUFAs into PG mediated by the action of AAS accelerates photodamage to PSII. In Synechococcus elongatus PCC 7942, the lack of AAS accelerated the photoinhibition of PSII [24]. A mutant lacking AAS had 42% of the initial PSII activity after 80 min of exposure to strong light, whereas wild-type cells retained 50% under the same conditions (Figures 1a  and 7a). Thus, the lack of AAS in Synechocystis slightly accelerated the photoinhibition of PSII as well. However, the effect was smaller than that observed in Synechococcus elongatus PCC 7942 under these conditions.

Disruption of the Gene for AAS Protects PSII against PUFA-Induced Photoinhibition
Disruption of the gene for AAS induces tolerance in cells to the PUFA-induced inhibition of cellular growth [3,4]. To examine the role of AAS in PUFA-induced PSII photoinhibition, we performed similar experiments with a mutant lacking AAS. AAS was required for the incorporation of 18:3 Δ9,12,15 into membrane lipids, especially PG (Figure 4c). In wild-type cells, PSII activity dropped to 8% of the initial activity after the 80 min exposure to strong light in the presence of 18:3 Δ9,12,15 (Figure 1a). AAS mutant cells had 46% PSII activity in the same time (Figure 7a), indicating that the lack of AAS reduced the effect of 18:3 Δ9,12,15 on PSII photoinhibition or photodamage (Figure 7). Therefore, the incorporation of PUFAs into PG mediated by the action of AAS accelerates photodamage to PSII. In Synechococcus elongatus PCC 7942, the lack of AAS accelerated the photoinhibition of PSII [24]. A mutant lacking AAS had 42% of the initial PSII activity after 80 min of exposure to strong light, whereas wild-type cells retained 50% under the same conditions (Figures 1a  and 7a). Thus, the lack of AAS in Synechocystis slightly accelerated the photoinhibition of PSII as well. However, the effect was smaller than that observed in Synechococcus elongatus PCC 7942 under these conditions.

Effect of PUFAs on Photodamage to PSII
In the present study, the specific incorporation of 18:3 ∆9,12,15 into the sn-2 position of PG molecules increased the sensitivity of PSII to photodamage by destabilizing the PSII complexes. A mutant of Synechocystis lacking two lysophosphatidic acid acyltransferases (LPAATs) contained a high amount of C18 FAs at the sn-2 position in all membrane lipids and shows increased PSII sensitivity to strong light as compared with WT cells [25]. Hence, C16 FAs at the sn-2 position of membrane lipids are important. Our results also indicated the importance of C16 FAs at the sn-2 position of PG. A marine cyanobacterium, Synechococcus sp. PCC 7002 (hereafter Synechococcus), that lacked a ∆6 desaturase showed extreme sensitivity to PUFAs [2]. A transgenic strain of Synechococcus overexpressing desD for ∆6 desaturase (DesD) from Synechocystis generated 18:3 ∆6, 9,12 and 18:4 ∆6,9,12,15 , and grew better than the WT [26]. Synechocystis DesD introduces a double bond into fatty acids bound to galactolipids rather than those bound to SQDG and PG [27,28]; thus, the severe toxicity of PUFAs for Synechococcus might be due to the incorporation of PUFAs into PG molecules but not galactolipids. PG has crucial roles in the function and maintenance of PSII [29][30][31][32]. PG is also required for the oligomerization of both the PSI and PSII complexes [33,34] and for the binding of extrinsic proteins stabilizing the OEC in PSII [35]. More than 30 mol% of the total PG in the thylakoid membranes is located in the photosystems [36]. Therefore, unusual PG molecules containing PUFAs at the sn-2 position might specifically disorder the photosystems, especially the PSII complexes, with the resulting inactivation of PSII.
In bacteria, PG is used as a substrate for the lipidation of lipoproteins [37]. In cyanobacteria, several lumenal proteins such as CyanoP and CyanoQ, homologs of PsbP and PsbQ in land plants, are modified by diacylglycerol and palmitate [38]. CyanoP is required for stabilizing the OEC [39], and a lack of CyanoQ protein destabilizes PSII [40]. These observations suggest that unusual PG containing PUFAs at the sn-2 position becomes a substrate for lipidation of these lipoproteins, which might affect the stability of the OEC. A recent study by Knoppová et al. revealed that a PSII assembly factor, Ycf48, was lipidated by PG [41]. Therefore, the unstable PSII complex in the presence of PUFAs shown in Figure 3 might be due to the lipidation of Ycf48 by the unusual PG molecules.

Incorporation of PUFAs into PG
We showed that the effect of PUFAs on photosynthesis depends on the specific incorporation of PUFAs into the sn-2 position of PG molecules, which is mediated by AAS. Synechocystis AAS reacts nonspecifically with FFAs to synthesize acyl-ACPs in vitro [5]. These findings suggest that Synechocystis has a phospholipase A 2 that synthesizes 1-acyl-lysoPG (lysoPG) by digesting an acyl group at the sn-2 position of PG, and that a lysoPG acyltransferase (LPGAT) transfers PUFAs into the sn-2 position of 1-acyl-lysoPG from ACP-attached PUFAs. Weier et al. (2005) identified an LPGAT that transferred both the C16-and C18-acyl groups from acyl-ACP but not acyl-CoA to lysoPG [42]. Therefore, this lysoPG acyltransferase might be responsible for the specific incorporation of PUFAs into the sn-2 position of PG molecules. However, tolerance to PUFAs in the mutant lacking LP-GAT has not been investigated. Synechocystis has a phospholipase encoded by sll1969 [43], which suggests that lysoPG is synthesized by the phospholipase in Synechocystis cells. Further biochemical analysis of the phospholipase is needed to investigate whether this phospholipase is involved in the synthesis of 1-acyl-lysoPG. Synechococcus elongatus PCC 7942, a mutant lacking AAS, produced more palmitic acid (16:0) and lyso-lipids under strong light [24]. This finding supports our result that the specific incorporation of PUFAs into PG molecules was accelerated under strong light.
AAS is broadly conserved among oxygenic photosynthetic organisms such as cyanobacteria and land plants [5]. PUFAs are toxic to other photosynthetic organisms as well [44], so the uptake of PUFAs into membrane lipids mediated by the action of AAS might cause a universal effect of PUFAs in photosynthetic organisms.
In plants, PG in the chloroplasts is exclusively synthesized inside the chloroplasts by the plastid pathway, which results in PG molecules containing 16:0 or ∆3-trans-hexadecenoic acid (16:1 ∆3t ) at the sn-2 position owing to the substrate specificity of LPAAT [45,46] and the desaturation catalyzed by desaturase (FAD4) [47] in the chloroplasts. However, glycolipids in the chloroplasts, such as MGDG, DGDG, and SQDG, are synthesized not only by the plastid pathway but also the endoplasmic reticulum (ER) pathway [48,49]. In the ER pathway, FAs synthesized in the chloroplasts are exported to the cytosol and used for the synthesis of phospholipids in the ER, then some synthesized phospholipids in the ER are transferred to the chloroplasts and used for the synthesis of glycolipids. The glycolipids synthesized through the ER pathway contain C18 PUFAs at the sn-2 position owing to the substrate specificity of LPAAT and the action of desaturases (FAD2 and FAD3) in the ER [50][51][52]. Our results suggest that PG needs to be synthesized by the plastid pathway to avoid the synthesis of PG molecules containing PUFAs at the sn-2 position because such PG molecules destabilize the PSII complex and accelerate the photoinhibition of PSII.

Strains and Culture Conditions
Cells of glucose-tolerant Synechocystis GT-I as WT cells were grown photoautotrophically in BG-11 medium at 32 • C, as described in [1]. Cells in the culture with an optical density at 730 nm of 1.0 ± 0.1 (about 3.6 µg mL −1 Chl a) were used for the assays. Generation of a mutant lacking AAS was as described in [4].

Targeted Inactivation of aas in Synechocystis
The aas gene (slr1609) of Synechocystis was amplified by PCR using TaKaRa LA Taq DNA polymerase (TAKARA BIO INC., Japan) and the primer pair a1 (5 -ACGCTTTGGTGA TGAACACTGG-3 )/a2 (5 -TTGGCGTAGGGGAATGGCT-3 ). The amplified DNA fragment was cloned into the pGEM-T easy vector (Promega, Madison, WI, USA). A 1.0-kb DNA fragment carrying the nptI kanamycin resistance gene was excised from the plasmid pRL250 [53] and ligated into the SmaI recognition site in the cloned aas gene. The resultant plasmid was used to transform the wild-type Synechocystis cells through homologous recombination into the kanamycin resistance type. After three rounds of streak purification of single colonies, the absence of the wild-type aas copy in selected colonies was confirmed by PCR. The resulting mutant, which carried an insertionally inactivated aas gene, was named dAS13 ( Figure S1).

Photosynthetic Activity of PSII
PSII activity was measured by oxygen evolution from PSII in the presence of 1 mM 1,4benzoquinone (BQ) and 1 mM K 3 Fe(CN) 6 , as described in [1]. For the analysis of photoinhibition of PSII, cells in the culture were exposed to strong light at 1500 µmol photons m −2 s −1 at 32 • C under ambient aeration. For assaying photodamage, the cell culture was supplemented with 200 µg mL −1 lincomycin before illumination. FFA compounds purchased from FujiFilm Wako (Tokyo, Japan) and Tokyo Chemical Industry (Tokyo, Japan) were added to the cell culture just before illumination.
To determine the electron transfer in PSII, we measured the rate of the photoreduction of DCIP, which accepts electrons from Q B , with isolated thylakoid membranes in the absence or presence of diphenylcarbazide (DPC), an electron donor to a tyrosine Z in PSII. Electron transfer rates from H 2 O or DPC to 2,6-dichlorophenolindophenol (DCIP) were measured as described in [17] with some modifications. Thylakoid membranes were isolated from cells incubated under strong light at 1500 µmol photons m −2 s −1 for 20 min with or without 50 µmol L −1 18:3 ∆9,12,15 or 18:1 ∆12 , then the cells were disrupted by bead-beating in Solution 8 (0.4 mol L −1 sucrose, 2 mmol L −1 MgCl 2 , 5 mmol L −1 NaCl, 40 mmol L −1 Mes-NaOH, pH 6.5). To measure the electron transfer rate from DPC to DCIP, thylakoid membranes were treated with 5 µmol L −1 NH 2 OH for 5 min on ice. NH 2 OHtreated thylakoid membranes were washed twice, then resuspended in Solution 8. The light-induced decrease in the absorbance by DCIP at 580 nm in the thylakoid membrane suspension (5 µg Chl mL −1 ) in the presence of 80 mmol L −1 DCIP and 80 µmol L −1 DPC under strong light at 1500 µmol photons m −2 s −1 at 32 • C for 5 min was measured by spectrophotometry (V-730Bio, JASCO, Tokyo, Japan).

Lipid Analysis
To analyze the incorporation of 18:3 ∆9,12,15 into membrane lipids, cells were incubated with 50 µmol L −1 18:3 ∆9,12,15 (Wako, Japan) under the same conditions as the analysis of PSII photoinhibition described above. A 5 mL aliquot of the cell culture was removed at the designated time and kept on ice until lipid extraction. Lipids in Synechocystis cells were extracted as described in [31]. Each membrane lipid was separated and isolated on a thin-layer chromatography (TLC) plate (TLC silica gel 60; Merck, Kenilworth, NJ, USA) developed with chloroform:methanol: 28% NH 4 OH = 65:35:5 (v/v/v), and FAs binding to each sugar lipid were methyl-esterified with HCl-methanol at 85 • C for more than 3 h. Since FFA is located at the same position as PG in TLC (Supplementary Figure S1), PG was methyl-esterified with 2 M KOH-methanol at room temperature for 2 min, which only methyl-esterified glycerolipids but not FFA, as described previously [55]. The obtained FA methylesters were analyzed by gas chromatography (GC) (GC-2014, Shimazu, Tokyo, Japan). To determine the content of FAs at the sn-1 and -2 positions of PG, PG isolated from cells treated with 50 µmol L −1 18:3 ∆9,12,15 for 20 min under strong light was treated with 0.5 µg honey bee phospholipase A 2 (Sigma-Aldrich, St. Louis, MO, USA) in a lipase buffer (50 mmol L −1 Tris-HCl pH 7.5 with 0.05% (w/v) Triton X-100) for 30 min at 37 • C. The resulting FFAs and 1-acyl-lyso-phosphatidylglycerol (LysoPG) were separated on a TLC plate developed with chloroform:acetone:methanol:acetic acid:water = 50:20:10:15:5 (v/v/v/v/v), and FAs in 1-acyl-lysoPG was methyl-esterified with HCl-methanol at 85 • C for 3 h. The obtained FA methylesters were analyzed by GC (GC-2014, Shimadzu). FA content at the sn-2 position of the PG molecules was calculated by subtracting the FA content at the sn-1 position from the total fatty acid composition; 15:0 (5 nmol) was used as the internal control.