Genetically Programmed Changes in Photosynthetic Cofactor Metabolism in Copper-deficient Chlamydomonas*

Genetic and genomic studies indicate that copper deficiency triggers changes in the expression of genes encoding key enzymes in various chloroplast-localized lipid/pigment biosynthetic pathways. Among these are CGL78 involved in chlorophyll biosynthesis and HPPD1, encoding 4-hydroxyphenylpyruvate dioxygenase catalyzing the committed step of plastoquinone and tocopherol biosyntheses. Copper deficiency in wild-type cells does not change the chlorophyll content, but a survey of chlorophyll protein accumulation in this situation revealed increased accumulation of LHCSR3, which is blocked at the level of mRNA accumulation when either CGL78 expression is reduced or in the crd1 mutant, which has a copper-nutrition conditional defect at the same step in chlorophyll biosynthesis. Again, like copper-deficient crd1 strains, cgl78 knock-down lines also have reduced chlorophyll content concomitant with loss of PSI-LHCI super-complexes and reduced abundance of a chlorophyll binding subunit of PSI, PSAK, which connects LHCI to PSI. For HPPD1, increased mRNA results in increased abundance of the corresponding protein in copper-deficient cells concomitant with CRR1-dependent increased accumulation of γ-tocopherols, but not plastoquinone-9 nor total tocopherols. In crr1 mutants, where increased HPPD1 expression is blocked, plastochromanol-8, derived from plastoquinone-9 and purported to also have an antioxidant function, is found instead. Although not previously found in algae, this metabolite may occur only in stress conditions.

Copper is an essential cofactor for most forms of life because of its function as a catalyst of oxygen chemistry and redox reactions (1). In photosynthetic organs of land plants or in unicellular phototrophs like algae and cyanobacteria, plastocyanin is a major copper-requiring protein (2). Many algae and cyanobac-teria reduce their dependence on copper (the copper quota) by replacing plastocyanin with a heme protein called cytochrome (Cyt) 4 c 6 (reviewed in Ref. 3). In Chlamydomonas, this switch is controlled by copper response elements associated with the CYC6 gene encoding Cyt c 6 and a transcription factor, copper response regulator 1 (CRR1) (4 -6). CRR1 has a SQUAMOSA promoter binding protein (SBP) domain that is conserved in the green lineage (6). Its homologue in Arabidopsis, SPL7, controls nutritional copper homeostasis as well (7,8). Transcriptome experiments revealed multiple targets of this transcription factor in both organisms but besides the genes encoding assimilatory copper transporters (COPT/CTR), which are dramatically up-regulated in both organisms under copper deficiency (9,10), the target genes are distinct (8,11). This may be because modification of the photosynthetic apparatus is a key response in Chlamydomonas but not in Arabidopsis.
We noted previously that redox proteins, especially O 2 -dependent metabolic enzymes, are enriched among the CRR1 targets in Chlamydomonas (11). Among these are CRD1 and CPX1, encoding enzymes regulating two rate-limiting steps in tetrapyrrole biosynthesis: crd1 mutants are chlorotic in copperdeficient cells because of reduced chlorophyll (Chl) content accompanied by reduced abundance of PSI and LHCI (12,13). The abundance of CAO1 mRNA encoding another O 2 -dependent enzyme in Chl biosynthesis is also increased in copper deficiency but this may occur independently of CRR1 (11). The motivation for regulation of tetrapyrrole biosynthesis is not clear, but one hypothesis is that copper deficiency necessitates modification of photosystem biogenesis pathways. FAB2, another CRR1 target, encodes an acyl-ACP desaturase, whose activity is required for the synthesis of unsaturated fatty acids in the chloroplast. Indeed, the galactolipids of the thylakoid membrane from copper-deficient cells are more unsaturated relative to those from copper-replete cells (11). Again, the underlying significance is not known, but because of the importance of unsaturated fatty acids for thylakoid membrane function, we hypothesized that replacement of plastocyanin with Cyt c 6 may require structural modifications of the membrane.
In this work, we probe the function of two other CRR1 target genes encoding enzymes in pathways affecting pigment and lipid metabolism in the chloroplast, CGL78 (also called LCAA and Ycf54 in Arabidopsis/tobacco and cyanobacteria), a conserved protein that has been proposed to function with CRD1 in tetrapyrrole biosynthesis at the Mg Protoporphyrin IX monomethyl cyclase step (14 -16), and HPPD1, encoding one of two isoforms of 4-hydroxyphenylpyruvate dioxygenase (HPPD), which catalyzes the conversion of 4-hydroxyphenylpyruvate to homogentisate (17). Although HPPD is critical for tyrosine catabolism in animals, in plants, this enzyme catalyzes the rate-limiting step of an anabolic branch that forms plastoquinone-9 (PQ-9) and tocopherols. PQ-9 is the mobile carrier of 2 reducing eq between PSII and the Cyt b 6 f complex, whereas tocopherols have antioxidant activity by scavenging singlet oxygen produced by excess excitation energy input into PSII and protecting membrane lipids from peroxidation (18 -21). There are four tocopherol types, ␣-, ␤-, ␥-, and ␦-tocopherols, based on the number and position of methyl groups on the chromanol ring. The increase in abundance of ␣-tocopherols in response to high light is well documented for both algae and plants (22)(23)(24), and there is also a role for tocopherols in signaling pathways, although the mechanistic aspects are not yet fully elucidated (25).
For CGL78, we used an artificial micro-RNA-mediated reverse genetics approach to dissect its function, whereas for HPPD, we monitored pathway end products under conditions affecting its expression to validate its contribution to plastoquinone and tocopherol biosynthesis. We discovered: 1) that LHCSR3 accumulation is increased in copper-deficient cells and 2) that LHCSR3 accumulation was especially sensitive to loss of CGL78 as was the accumulation of PSAK, a connector between PSI and LHCI. This phenotype is similar to that of the crd1 mutant, which is blocked at the same step in Chl biosynthesis. For HPPD, we noted 3) that, an increase of HPPD1 is, surprisingly, not accompanied by an increase in all end products of the tocopherol/plastoquinone biosynthesis pathway. Rather, copper-deficient cells accumulate specifically ␥-tocopherols. Moreover, when HPPD1 expression is blocked, as in the crr1 mutant, ␥-tocopherol accumulation is abolished as well, supportive of a causal connection between HPPD abundance and ␥-tocopherol accumulation. We speculate that operation of the light reactions is less effective in copper-deficient cells, requiring subtle modifications of mechanisms for handling excess excitation energy and stress.

Results
Reduced CGL78 Expression Impacts PSI/II-LHC Super-complex Formation-To assess the function of CGL78 in Chlamydomonas, we used an inducible amiRNA system to generate strains with conditionally reduced abundance of CGL78 mRNA. Because the amiRNA is driven by the NIT1 promoter, it is ammonium repressible (26). Therefore, strains cgl78-ami11 and cgl78-ami13 carrying the knock-down constructs accumulate CGL78 mRNA at normal abundance in medium with ammonium but have reduced abundance in medium lacking ammonium (8 and 5% as compared with control strains, CGL78, respectively (Fig. 1A). Accordingly, the cgl78-ami strains are chlorotic on ammonium-free medium (Fig. 1B) with significantly reduced Chl content relative to CGL78 (36%), although the ratio of Chl a to b is unaffected (Fig. 1C). There is no effect of cgl78-ami knock-down on growth (Fig. 1D), making it unlikely that loss of CGL78 affects any other pathway. The chlorotic phenotype indicated a direct role of CGL78 in chlorophyll-binding protein accumulation, but the photosystem II capacity was unaffected in cgl78-ami lines (Fig. 1E). To distinguish which chlorophyll-binding proteins might be reduced in cgl78-ami strains, we analyzed thylakoid membrane proteins after separation by BN-PAGE, which resolves various photosystem-containing complexes (27). We noticed a striking loss of PSI/II-LHC super-complexes (marked with a red arrow) with a corresponding increase in PSII core monomers (Fig. 2, compare unfilled with filled arrows).
Because CGL78 is a target of CRR1 and both mRNA and protein abundances are increased 7-and 4-fold, respectively, in copper-deficient medium (Refs. 11 and 28 and Fig. 3A), we tested whether the cgl78-ami phenotype was affected by copper nutrition status. The copper status was verified by expression of the previously characterized sentinel gene CYC6 and accumulation of the corresponding protein (29) (Fig. 3, A and C). Nev-FIGURE 1. Loss of CGL78 expression results in a chlorotic phenotype. A, abundance of CGL78 mRNA of independent cgl78-amiRNA lines (11,13) was determined by quantitative RT-PCR. RNA was isolated from cells that were either grown on TAP media containing ammonia or nitrate as the sole nitrogen source. B, representative picture of the pale green phenotype of cgl78-ami11 at a density of 8 ϫ 10 6 cells/ml. C, chlorophyll content and chlorophyll a/b ratio. The averages and S.D. of different cgl78-ami lines are shown. D, representative growth curves of cgl78-ami11 and an empty vector control line (CGL78) that were grown in media containing either ammonia or nitrate as sole nitrogen source. Cells were grown in TAP medium containing ammonium (NH 4 ϩ ) to repress the amiRNA construct or nitrate (NO 3 ) to induce the amiRNA construct. Experiments were done at least twice. E, the maximum PSII capacity was measured in independent CGL78 or cgl78-ami lines. Shown are the averages of three independent cultures Ϯ S.D. ertheless, copper neither exacerbated nor ameliorated the phenotype (Fig. 3B). Separation of chlorophyll-binding proteins by BN-PAGE revealed that copper deficiency did not affect the PSI/II-LHC super-complex formation in wild-type lines (Fig.  3D, red arrow), but copper-deficient cgl78-amiRNA lines showed reduced amounts of PSI/II-LHC super-complexes ( Fig.  3D, marked by an asterisk). The pattern of PSI/II-LHC supercomplexes of thylakoids isolated from cgl78-ami lines in copper-depleted versus copper-replete grown cells look similar except that PS core subunits are more enriched in the former (Fig. 3D, circle). On the other hand, accumulation of PSII was the same in cgl78-ami strains and control lines based on the abundance of D1 (Fig. 4).
In previous work, we observed that PSI-LHCI complexes are destabilized in iron-deficient cells, coincident with a loss of PSAK, which is the subunit connecting PSI and LHCI. The loss of PSI-LHCI complexes prompted us to monitor PSAK accumulation in cgl78-ami lines, where we found it to be decreased to 50% (Figs. 4 and 6C). This molecular phenotype is reminiscent of the crd1 mutant (30). CRD1 is one of two isoforms of a subunit of Mg Protoporphyrin IX monomethyl cyclase. Its expression is up-regulated by poor copper nutrition via CRR1 (12). The other isoform, named CTH1, is reciprocally regulated, also by CRR1 (13). Because CGL78 functions in concert with CRD1 and CTH1, we tested whether CRD1 and CTH1 expression might be affected in the cgl78-ami lines (Fig. 4C). We did not see any significant change in CRD1/CTH1 abundances in cgl78-ami lines (Fig.  4, A and B) nor in the pattern of mRNA accumulation as assessed by qRT-PCR (Fig. 4C). Likewise, CGL78 mRNA abundance is also not changed in crd1 as compared with CRD1 cells (Fig. 4D). Therefore, the expression of CGL78 seems not to be interconnected with that of CTH1/CRD1 and we can attribute the cgl78 molecular phenotype in copper deficiency directly to loss of CGL78 rather than indirectly to loss of its partner protein(s) in the complex.
In previous work, we showed that the presence of PSAK was related to functional association of the LHCI antenna with PSI (30). For instance, in iron-deficient cells, which are impaired in chlorophyll biosynthesis, PSAK is reduced, resulting in less energy transfer from LHCI to PSI. Therefore, we collected fluorescence emission spectra to assess PSI-LHCI interactions in wild-type versus knock-down lines as a function of copper nutrition. The copper-replete control cells (CGL78) showed characteristic peaks at 685 and 711 nm, resulting from fluorescence of LHCII attached to PSII and of LHCI to PSI, respectively (Ref. 31 and Fig. 7A). Copper-depleted control cells showed the same changes of the emission spectra that we had observed in copper-depleted wild-type cells in previous work (13), namely an increase in intensity of the peak at 710 nm relative to the peak at 685 nm, indicating modifications of the PSI-LHCI interaction in copper-deficient cells (Fig. 7B). In cgl78-ami lines, the PSI-LHCI peak at 710 nm was blue-shifted (32) and reduced in intensity under both, copper-depleted and copper-replete conditions (Fig. 7, A and B). The latter is consistent with the observation that PSAK is reduced in cgl78-ami lines. When wild-type strains are grown in high light, we note reduced intensity and a blue-shift of the peak corresponding to PSI-LHCI (Fig. 7C), whereas cgl78-ami lines subject to high light show a blue-shift of the PSII-LHCII peak to 682 nm and loss of the PSI-LHCI peak (Fig. 7C). The latter situation is reminiscent of iron-starved cells (Ref. 30 and Fig. 7D), and is consistent with a complete loss in both cases of PSAK (Ref. 30 and Fig. 6C).
Copper-deficient Cells Accumulate a Stress-related Antenna Protein, LHCSR3-Besides the change in PSI-LHCI interactions noted above, a survey of chlorophyll-protein accumulation revealed a previously unrecognized difference between copper-depleted and copper-replete cells. Specifically, LHCSR3 is increased in copper-deficient wild-type (CGL78) cells severalfold, but not in copper-deficient cgl78-ami strains (Fig. 4). Quantification by a dilution series indicates that LHCSR3 abundance in cgl78-ami strains is about 12.5% of the level in the control strains (Fig. 4, A and B).
When we surveyed the copper-deficiency transcriptome relative to the iron-deficiency transcriptome, where up-regulation of LHCSR3 has already been described (33), we noted a small but consistent increase in LHCSR3.1 and LHCSR3.2 mRNA abundance in copper-deficient relative to copper-replete cells ( Table 1). The change is smaller than that documented for ironstarved cells (33) and substantially less than that in cells exposed to high light, but it is statistically significant. When we compared the abundance of LHCSR3 in copper-deficient cells to the abundance in iron-limited cells, again the increase was correspondingly lower in the copper-deficient situation (Fig. 5). A simple explanation might be that copper-deficient cells are also secondarily iron-deficient, but measurements of the iron content in copper-deficient cells indicate that the iron content in copper-replete cells (10.6 Ϯ 1.4 ϫ 10 7 atoms/cell) does not differ from the iron content in copper-depleted cells (10.4 Ϯ 1.1 ϫ 10 7 atoms/cell). Therefore, we conclude a direct consequence of copper nutrition on the performance of the PSI-LHCI complex.
Because LHCSR3 seemed to be the most affected chlorophyll-binding protein in copper-deficient cgl78-ami lines, and LHCSR3 has a known role in photo-protection (34 -38) we wondered if cgl78-ami strains are more sensitive to high light relative to the control. Indeed, if we grew cgl78-ami lines in high light, we observed that the cells became bleached (Fig. 6A). The chlorophyll content of cgl78-ami strains that were grown for 20 h in high light is reduced to 5.5%, as compared with CGL78 wild-type lines (Fig. 6B), concomitant with a strong reduction of LHCSR3 abundance and nearly complete absence of PSAK (Figs. 6C and 7, A and D). Reduced abundance of LHCSR3 likely occurs at the level of mRNA accumulation because the increased LHCSR3 expression in copper deficiency is blocked in the cgl78-ami lines. In fact, the high light responsive increase in LHCSR3 mRNA is also blocked in the cgl78-ami lines (Fig.  6C). This prompted us to test LHCSR3 expression in the crd1 mutant, and we noted that this molecular phenotype is recapitulated in crd1. The latter suggests that the cyclase in Chl bio-synthesis might signal to the LHCSR3 gene. Another possibility is that a functional electron transport chain in PSI is required for increased LHCSR3 expression (39).
Tocopherol Metabolism Is Altered in Copper-deficient Chlamydomonas-Besides Chl biosynthetic enzymes, previous transcriptome analyses had identified other changes in cofactor metabolizing enzymes in copper-deficient cells. One of these is HPPD1, a likely target of CRR1, and one of two genes encoding HPPD, which catalyzes the first step in a branched pathway leading to the terpenoid cofactors, plastoquinone and tocopherols (Fig. 8). Sequence analysis indicates that both genes are more highly related to homologues in plants versus animals. The two isoforms are paralogues; they are most closely related to algal homologues but even more closely related to each other (90% identity at the nucleotide level and 93.5% similarity at the protein level) consistent with a recent duplication event. Previous transcriptome and proteome profiling indicated that HPPD1 mRNA and the corresponding protein are increased in copper-deficient cells and in cells that experienced hypoxia (11,28) (Fig. 9A), which was validated in this work by real time PCR with paralog-specific primers and immunoblotting (Fig. 9, B and C). Antibodies raised against the carrot protein recognized an approximate 46-kDa protein whose abundance is increased in copper-deficient relative to copper-replete Chlamydomonas cells (Fig. 9C). The impact of CRR1-dependent regulation is an ϳ2-fold increase in mRNA templates encoding HPPD in copper-deficient growth conditions (Fig. 10B). Given the function FIGURE 3. Copper deficiency does not impact the phenotype of cgl78-ami strains. A, CYC6 and CGL78 mRNA abundance was estimated by qRT-PCR. RNA was isolated from cells that were grown in medium containing nitrate to induce the amiRNA targeting CGL78, the presence of Cu(II)-EDTA in the medium is indicated. Control strains (black) or cgl78-ami lines (white) were grown in TAP medium to a density of 5-8 ϫ 10 6 cells/ml. Each symbol represents an independent experiment analyzed in technical duplicates. B, chlorophyll content in cells grown in medium supplemented either with ammonia or nitrate as nitrogen source and supplemented with or without copper as indicted. Error bars represent S.D. of at least three experiments. C, soluble fractions corresponding to 10 g of protein were separated on sodium dodecyl sulfate-containing denaturing gels (15% acrylamide) and immuno-detection with anti-Cyt c 6 or OEE1 as a loading control. D, PSI/II-LHC super-complex formation is also diminished in copper-depleted cgl78-ami lines. Blue native PAGE of thylakoid membranes from CGL78 or cgl78-ami lines grown either in copper-replete or copper-depleted TAP medium with nitrate as nitrogen source. Experiments were performed at least twice.
of these molecules in electron transfer and photoprotection, we measured the abundance of the end products of the pathway to assess the impact of increased expression.
Surprisingly, despite the increase in mRNA and protein at a rate-limiting step, the total plastoquinone or tocopherol content was not changed in copper-deficient versus copper-replete FIGURE 4. PSAK and LHCSR3 accumulation is dependent on CGL78. A, the abundance of the ␣ and ␤ subunits of CF 1 , appearing as a single band, were visualized as a loading control (bottom panel). Membrane fractions, equivalent to 10 g of protein, were loaded on sodium dodecyl sulfate-containing denaturing gels. B, samples used for panel A were re-evaluated for estimating relative abundances. An asterisk marks the copper-replete sample, whereas the circle represents the copper-depleted sample from either cgl78-ami11 cells that were grown in medium containing nitrate as nitrogen source. C, shown are the relative mRNA abundances of CRD1, CTH1 (2.1 kb functional transcript), CTH1 (3.1 kb non-functional transcript) in either CGL78 (black squares) or cgl78-ami11 (white apex-down triangle or cgl78-ami13 (white apex-up triangle) that were grown in copper-depleted or copper-replete medium as indicated. D, CGL78 mRNA abundance in crd1 (white circles) or CRD1 (gray circles) was determined by qRT-PCR. Each symbol represents the average of a duplicate qPCR of an independent experiment.

TABLE 1 LHCSR3 expression is up-regulated in copper-depleted and iron-limiting conditions
Shown is the expression in fpkm (fragments per kilobase of exon per million fragments mapped) of LHCSR3.1 (Cre08.g367500), LHCSR3.2 (Cre08.g367400), and the total amount of LHCSR3-encoding mRNA in the condition indicated. For the light experiments cells were grown for 12 h in the dark and then shifted for 4 h to light. Data were obtained from Refs. 11 and 33.  FIGURE 5. LHCSR3 increase in copper-deficient cells compared with iron limited and high light-exposed cells. Change in abundance of LHCSR3 in response to growth in copper-deficient medium, iron-deficient medium or high light conditions. The abundance of CF 1 was visualized as a loading control (bottom panel). Membrane fractions, equivalent to 10 g of protein, were loaded on sodium dodecyl sulfate-containing denaturing gels.

Condition
Chlamydomonas (Fig. 9, D and E). Nevertheless, we noted a consistently higher proportion of ␥-tocopherol (42% increase on average) in copper-deficient cells of strain 4532 (wild-type for CRR1) (Fig. 9E). To evaluate the contribution of HPPD1 to ␥-tocopherol content, we took advantage of the crr1 mutant (because HPPD1 is a CRR1 target) and two independent com-plemented lines, CRR1 (in which the wild-type gene rescued crr1) and crr1-⌬Cys (in which a gene encoding a C terminally truncated version of CRR1 complements the crr1 mutation). The copper status of the cultures was validated on the basis of sentinel protein accumulation (Fig. 10A). Both complemented lines, but not the crr1 mutant, show normal CRR1-and copperresponsive expression of HPPD1 and the sentinel CYC6 gene (Fig. 10, A-C). The increase in ␥-tocopherol mirrors CRR1 function, namely observed in both complemented lines but not in the crr1 mutant (Fig. 10D). Therefore, we conclude that the change in the ␥-tocopherol to ␣-tocopherol ratio is dependent on copper nutrition and CRR1 and is an intrinsic response to poor copper nutrition. When we surveyed plastoquinone species in crr1 versus CRR1, we found that the total PQ-9 content was unaffected (Fig. 11A), consistent with a lack of effect of copper nutrition on plastoquinone content, but we did observe a new metabolite, plastochromanol-8 (PC-8), previously not observed in algae (Fig. 11B). PC-8 is a stress metabolite generated from plastoquinol-9 by a cyclization reaction catalyzed by VTE1. We hypothesize that PC-8 accumulation in copper-deficient crr1 is part of a general stress acclimation response in this situation because it is independent of an increase in HPPD expression. Its abundance relative to PQ-9 is small (0.35% of total) and is found only in copper-deficient crr1 cells, a situation in which HPPD1 is not up-regulated and ␥-tocopherol increase is not observed. One possibility is that PC-8 is produced for a photo/ROS-protective function in copper-deficient crr1 strains, a hypothesis that is consistent with the up-regulation of LHCSR3 and reduced PSAK in this situation (Fig. 11C).
HPPD1 expression can be increased also in hypoxia in a CRR1-dependent pathway (40) (Fig. 12, A and B). Nevertheless, in this situation the enzymatic activity should actually be compromised because of low O 2 , a substrate of the reaction. Indeed, wild-type hypoxic cells have less ␥-tocopherol than do normoxic cells (Fig. 12C, 66% reduced), which confirms that an oxygen-dependent step is required for the change in tocopherol composition beyond simply increased mRNA accumulation. When we tested crr1 mutants versus complemented strains, again the hypoxic increase in HPPD1 mRNA abundance is dependent on CRR1 (Fig. 13A), but in the absence of oxygen, ␥-tocopherol content is decreased rather than increased relative to normoxic cells (Fig. 13B). When we looked at the plastoquinone profile in the hypoxic cells, we noted that the total PQ-9 content is unchanged in crr1 mutants or complemented strains (Fig. 13C), but the ratio of oxidized to reduced is dramatically reduced in hypoxic cells of all genotypes, which validates the physiology of the treatment.

Acclimation to Copper Deficiency
Cyt c 6 and Acyl-ACP Desaturase-Plastocyanin is the most abundant copper protein in photosynthetic cells and accordingly photosynthesis is dependent on copper in most organisms, except where there is a genetically programmed pathway to accommodate copper deficiency (41). The copper deficiency response occurs in many cyanobacteria and algae (42). The . cgl78-ami strains are more sensitive to high light. A, colonies of CGL78 or cgl78-ami lines were either grown on TAP agar medium containing ammonia or nitrate as the sole nitrogen source for 1 week in low light (40 mol m Ϫ2 s Ϫ1 ) and then exposed for 24 h to high light (600 mol m Ϫ2 s Ϫ1 ). B, chlorophyll content of whole cells. Cells were either grown in TAP liquid medium containing nitrate (NO 3 ) or ammonium (NH 4 ) and transferred at a density of 2 ϫ 10 6 cells/ml to high light (HL, 600 mol m Ϫ2 s Ϫ1 ). Shown are the averages and S.E. of four independent experiments. Asterisks indicate a significant difference compared with CGL78 according to one-way analysis of variance, post hoc Holm-Sidak (*, p value Ͻ 0.05; **, p value Ͻ 0.001). C, protein was extracted from cells that were exposed to high light and grown in TAP medium containing nitrate. Membrane fractions, equivalent to 10 g of protein, were loaded on sodium dodecyl sulfate-containing denaturing gels. The abundance of CF 1 was used as a loading control (bottom panel). Shown is one of two independent experiments. D, LHCSR31 and LHCSR3.2 mRNA abundance in CGL78 (black squares), cgl78-ami11 (white apex down), cgl78-ami13 (white apex up), crd1 (white circles), or CRD1 (gray circles) was determined by qRT-PCR. Shown are the relative mRNA abundances of CRD1, CTH1 (2.1 kb fragment), and CTH1 (3.1 kb fragment). Cells were grown in copper-replete (ϩ) or copper-depleted (ϪϪ) medium in continuous light at a PFD of 100 mol m Ϫ2 s Ϫ1 or exposed to 24 h high light (600 mol m Ϫ2 s Ϫ1 ) as indicated. Each symbol represents the average of a duplicate qRT-PCR of an independent experiment. most well recognized change in the pathway of the light reactions in copper-deficient cells is the replacement of plastocyanin with a copper-independent substitute, a soluble c-type cytochrome or Cyt c 6 (43)(44)(45). In this and previous work, we show that there are, in addition, subtler changes of the photosynthetic apparatus. These changes may be required to accommodate the use of structurally distinct mobile electron donors to PSI or to accommodate Cyt c 6 as a less effective catalyst. Although it is not yet evident in the laboratory setting, the absence of Cyt c 6 in land plants, which makes photosynthesis strictly copper dependent, suggests that plastocyanin offers a selective advantage.
Whole transcriptome analyses of copper-replete versus copper-deficient and CRR1 versus crr1 strains revealed dozens of genes with candidate copper-response elements that are likely targets of nutritional copper signaling (11). Many of these encode key enzymes in plastid lipid or lipid cofactor metabolism, such as coprogen oxidase (CPX1), CRD1 and CGL78 in tetrapyrrole biosynthesis, acyl-ACP desaturase (FAB2), and HPPD (discussed below). There are two possible rationales for increasing the expression of metabolic enzymes in copper deficiency: 1) increased expression is a compensatory mechanism that allows maintenance of end products because the pathway is compromised in copper-deficient cells or 2) increased expression allows increased accumulation of end products, more of which are required in copper-deficient cells. The two models are distinguished by end product analyses in wild-type versus crr1 mutants, and previous analyses showed that the latter is the situation for FAB2 and CRD1. Indeed, there are more unsaturated fatty acids in copper-deficient Chlamydomonas cells, whereas in the crr1 mutant the level of desaturation is unchanged (11). Lipid-profiling indicated that the enrichment was restricted to the galactolipids in the thylakoid membrane, one of the first indications that acclimation to copper deficiency requires modification of the photosynthetic membranes.
␥-Tocopherol and Plastochromanol-8 -In this work, we document other subtle but measurable modifications to the photosynthetic apparatus. Increased expression of HPPD1 is (like for FAB2) causally connected with increased ␥-tocopherol content. The proportion of ␥-tocopherol is small (17% of the total tocopherol pool) and the increase therefore is also small, but it is reproducible and statistically significant, comparable with the magnitude of the change in the galactolipid desaturation. If Cyt c 6 is a less effective catalyst relative to plastocyanin, the change in proportion of ␥-tocopherol may serve to fine tune photosynthetic physiology by adjusting antioxidant or stress capacity. We conclude that like the situation for FAB2 and CRD1, increased expression of HPPD is required for increased accumulation of an end product (in this case ␥-tocopherol).
Plastochromanol-8 has not previously been found in algae, but was revealed here in the crr1 mutant and only in copper deficiency where crr1 grows poorly. Because it is implicated in stress protection (46), its production may be a downstream consequence of stress in crr1 mutants. It would be interesting to test whether other stress situations also allow PC-8 accumulation and to distinguish the relevant regulatory mechanisms. In Arabidopsis, PC-8 levels are regulated by the activity of a type II NADPH dehydrogenase C1 (ortholog is NDA5 in Chlamydomonas) (47,48) and a plastid ABC1-like kinase (49).
We had expected increased HPPD1 expression to yield increased total tocopherol, but this is not the case. It is possible that the biosynthetic pathways for ␣versus ␥-tocopherol in Chlamydomonas are branched at the level of suborganellar organization of enzyme complexes rather than at the level of the VTE3 enzyme. Increased expression of HPPD1 is not by itself sufficient to change the proportion of ␥-tocopherol. In hypoxic cells, expression of HPPD1 is also increased but in the absence of the substrate O 2 , ␥-tocopherol is not increased (Fig.  12C).

Chlorophyll-binding Proteins
We describe two other examples of adjustment of stress protection mechanisms in Chlamydomonas. First, copper-defi-cient cells have increased LHCSR3 content (Fig. 4). Although the increase is not as dramatic as in iron-deficient Chlamydomonas cells (50), it is notable, and more importantly, it is blocked in cgl78-ami lines and in crd1 mutants (Fig. 6C), suggestive of a signaling function of the cyclase with respect to expression of particular chlorophyll proteins. The operation of the tetrapyrrole pathway has been linked to nuclear gene expression in many works; this work documents a new regulatory connection (51).
LHCSR3 functions in photoprotection (35) involving photosystem I in Physcomitrella patens (52); its adjustment in copper-deficient cells may serve as pre-emptive fine tuning to accommodate Cyt c 6 as a less effective donor to PSI relative to plastocyanin. The up-regulation of the tetrapyrrole pathway, documented for coprogen oxidase and the aerobic oxidative cyclase (CPX1, CRD1, CGL78) at the level of RNA and protein (28,29) may allow adjustment of the abundance of specific chlorophyll-binding proteins to optimize the operation of the light reactions in a copper-deficient cell.
Second is de-stabilization of the LHCI-PSI interaction. Survey of photosystem I in the copper-replete situation showed a specific loss of PSAK, leading to de-stabilization of PSI-LHCI super-complexes. The cgl78-ami strains are also deficient in PSAK (Fig. 4), perhaps contributing to the phenotype noted by blue native gel separation of Chl-protein complexes and the observed blue-shift of the PSI-LHCI emission at 77 K. The loss of PSAK is even more dramatic in cgl78-ami lines subject to a high light treatment (Fig. 6). We had suggested previously that PSAK may modulate the connection between LHCI and PSI, consistent with its location in the complex (53), and hence serve a role in photo-protection.
We conclude that replacement of plastocyanin with Cyt c 6 in response to copper deficiency requires considerable adjustments of the photosynthetic apparatus and photoprotective mechanisms. Although CGL78 is 5-fold up-regulated in copper-deficient cells, the cgl78-ami mutants are chlorotic independent of copper nutrition (Fig. 3). This contrasts with the copper nutrition-conditional crd1 phenotype, but this might be because CTH1, the reciprocally expressed paralog of CRD1, covers the loss of function in copper-replete situations.
CGL78-amiRNA strains and Chlamydomonas reinhardtii strains CC4532, crr1-2 (CC5068), CRR1 (CC5071), and crr-⌬cys (CC5073) (56) were grown in Tris acetate/phosphate (TAP) with constant agitation in an Innova incubator (180 rpm, New Brunswick Scientific, Edison, NJ) at 24°C in continuous light (90 mol m Ϫ2 s Ϫ1 ), provided by cool white fluorescent bulbs (4100 K) and warm white fluorescent bulbs (3000 K) in the ratio of 2:1, unless stated otherwise. High light (600 mol m Ϫ2 s Ϫ1 ) was provided by a white fluorescent bulb (3800 K). TAP medium with or without copper was used with revised trace elements (Special K) instead of Hutner's trace elements according to Ref. 57. CGL78-amiRNA strains were either inoculated from a plate into standard TAP(NH 4 ) to repress the artificial micro-RNA or a modified TAP medium, where nitrate was substituted instead of ammonium as the sole nitrogen source, TAP(NO 3 ). Dark hypoxic cells were grown in TAP medium for 24 h in 1% air, 2% CO 2 , and 97% N 2 by bubbling before collection as described in Ref. 56. Cell density (number of cells per milliliter of culture) was determined with a hemocytometer.
Chlorophyll Measurements-Chlorophyll was extracted from whole cells using an 80/20 (v/v) acetone/methanol mixture. Chlorophyll content was estimated according to Ref. 58 from the absorbances at 646.6 and 663.6 nm measured on a PerkinElmer LAMBDA 25 UV/visual spectrometer. The absorption at 750 nm was used as a reference.
RNA Extraction and cDNA Synthesis-2-4 ϫ 10 7 cells were collected by centrifugation for 5 min at 2450 ϫ g, 4°C. RNA was extracted using the TRIzol reagent as described previously (59), treated with Turbo DNase (Ambion), concentrated, and cleaned with the Zymo Research RNA Clean & Concentrator TM -5 FIGURE 9. HPPD accumulation correlates with HPPD1 abundance and increased ␥-tocopherol in copper-deficient Chlamydomonas. A-D, change in HPPD-encoding transcripts in response to growth in copper-deficient medium or dark anoxic conditions. A, data from RNA-seq experiments in Chlamydomonas strain CC4532. The sizes of the circles are proportional to the relative abundance (RPKM) of the total HPPD transcript pool in each condition. B, abundance of CYC6, HPPD1, and HPPD2 transcripts was also estimated in independent samples by qRT-PCR. CC4532 was grown in TAP medium in the condition indicated: ϪCu(II) (black squares) or ϩCu(II) (white squares). Cells were collected after reaching a density of 5-8 ϫ 10 6 cells/ml and analyzed for RNA abundance by real-time PCR. Each symbol represents an independent experiment analyzed in technical duplicates. C, proteins were extracted from copper-depleted or copper-replet conditions and 10 g of soluble protein was separated by 10% PAGE, followed by immuno-detection with an antibody against Arabidopsis HPPD. D, plastoquinone, and E, ␣and ␥-tocopherol contents were measured in extracts of Chlamydomonas cells. Data are the averages of three experimental replicates Ϯ S.D.
Kit according to the manufacturer's instructions. Reverse transcription was primed with oligo(dT) 18  Thylakoid Membrane Preparation and Blue Native Gel Electrophoresis-Thylakoid membranes were prepared as described in Ref. 27 with the following modification: cells were collected by centrifugation at 3100 ϫ g at 4°C and re-sus- FIGURE 10. The increase of ␥-tocopherol content is dependent on CRR1. A, for immuno-detection of plastocyanin and cytochrome c 6 , soluble proteins were separated by SDS-polyacrylamide gel electrophoresis (15%) and transferred to a nitrocellulose membrane (0.1 N pore size). The abundance of OEE1 was used as a loading control (bottom panel). B and C, abundance of CYC6 (B) and HPPD1 (C) transcripts. Independent cultures corresponding to CRR1 (black circles), crr1 (triangles apex up), or crr1-⌬Cys (triangles apex down) were grown in copper-replete (black symbols) or copper-depleted (white symbols) TAP medium. Cells were collected after reaching a density of 2-6 ϫ 10 6 cells/ml and analyzed for RNA abundance. Each symbol represents an independent experiment analyzed in technical triplicates. D, the tocopherol con-tent in Chlamydomonas crr1 and corresponding strains that have been complemented with full-length CRR1 (crr1(CRR1)) or with a truncated version of CRR1 lacking a cysteine-rich domain (crr1-⌬Cys) was measured as described under "Experimental Procedures." Data are the averages of three biological replicates Ϯ S.D. FIGURE 11. Copper-deficient crr1 accumulates PC-8. A, plastoquinone or, B, plastochromanol content in Chlamydomonas crr1 and a strain that has been complemented with full-length CRR1 (crr1(CRR1)). Respective strains were either grown in medium with (ϩ) or without copper (Ϫ) as indicated. Data are the averages of three biological replicates Ϯ S.D. C, LHCSR3.1 (black) and LHCSR3.2 (white) transcript abundance was determined by RNA-seq in a copper-depleted crr1 and CRR1 (crr1(CRR1)) (11).
Protein Analyses-SDS-PAGE was performed using 20 g of total protein for each lane and transferred in a semi-dry blotter to nitrocellulose membranes (Amersham Biosciences Protran 0.1 NC). The membrane was blocked 30 min with 3% dried milk in PBS (137 mM NaCl, 2.7 mM KCl, 10 mM Na 2 HPO 4 , 1.8 mM KH 2 PO 4 ) containing 0.1% (w/v) Tween 20 and incubated in primary antiserum; this solution was used as the diluent for both primary and secondary antibodies, for 1 h, respectively. PBS containing 0.1% (w/v) Tween 20 was used for washing membranes twice for 15 min each time. The secondary antibody, used at 1:10,000, was goat anti-rabbit conjugated to alkaline phosphatase. Antibodies directed against CF 1 1,000), and Cyt c 6 (1:2,000) and HPPD (affinity purified against Arabidopsis HPPD, kindly provided by M. Matringe) (1:1,000) were used.
Quantitative Metal, Phosphorus and Sulfur Content Analysis-1 ϫ 10 8 Chlamydomonas cells of a culture at a density of 3-5 ϫ 10 6 cells/ml were collected by centrifugation at 2450 ϫ g for 3 min in a 50-ml Falcon tube. The cells were washed three times in 1 mM Na 2 -EDTA, pH 8 (to remove cell surface-associated metals), and once in Milli-Q water. The cell pellet, after removing the water, was overlaid with 286 l of 70% nitric acid and digested at room temperature for 24 h at 65°C for about 2 h before being diluted to a final nitric acid concen- FIGURE 12. HPPD1 is also regulated by dark hypoxia but ␥-tocopherol content is decreased in dark hypoxic Chlamydomonas. A, abundance of CYC6, HPPD1, and HPPD2 transcripts. Independent cultures corresponding to Chlamydomonas wild-type strain CC4532 were grown in dark hypoxia. Cells were collected after reaching a density of 2-6 ϫ 10 6 cells/ml and analyzed for RNA abundance. Each symbol represents an independent experiment analyzed in technical triplicates. B, transcript abundance of genes encoding enzymes of the tocopherol biosynthesis pathway were determined by RNA sequencing in dark grown and dark anoxic Chlamydomonas strain CC4532 (38). The sizes of the circles are proportional to the relative abundance (RPKM) of the total transcript pool of genes encoding enzymes regulating the plastoquinone/tocopherol pathway as shown in Fig. 8. C, the tocopherol content in Chlamydomonas CC4532 that was either grown in the dark, normoxic (high) or in the dark, hypoxic (low) as indicated. FIGURE 13. The decrease of ␥-tocopherol in dark hypoxia is independent of CRR1. A, abundance of HPPD1 transcripts as determined by quantitative real time PCR. Independent cultures corresponding to CRR1 (black circles), crr1 (triangles apex up), and crr1-⌬Cys (triangles apex down) were grown in TAP medium in the conditions indicated: dark hypoxia (white symbols) relative to dark (black symbols). Cells were collected after reaching a density of 2-6 ϫ 10 6 cells/ml and analyzed for RNA abundance. Each symbol represents an independent experiment analyzed in technical triplicates. Tocopherol (B) and plastoquinone (C) content in Chlamydomonas crr1 and corresponding strains that have been complemented with full-length CRR1 (crr1(CRR1)) or with a truncated version of CRR1 lacking a cysteine-rich domain (crr1-⌬Cys). Respective strains were either grown in the dark, normoxic (high) or hypoxic (low) as indicated. Data are the averages of three biological replicates Ϯ S.D.
tration of 2% (v/v) with Milli-Q water. Aliquots of fresh or spent culture medium were treated with nitric acid to a final concentration of 2% (v/v). Metal, sulfur, and phosphorous contents were determined by inductively coupled plasma mass spectrometry on an Agilent 8800 Triple Quadrupole ICP-MS instrument, in comparison to an environmental calibration standard (Agilent 5183-4688), a sulfur (Inorganic Ventures CGS1), and phosphorus (Inorganic Ventures CGP1) standard, using 89 Y (the 89 isotope of the chemical element Yttrium) as an internal standard (Inorganic Ventures MSY-100PPM). The levels of all analytes were determined in MS/MS mode, where 63 Cu were measured directly using helium in the collision reaction cell, whereas 56 Fe was directly determined using H 2 as a cell gas. The average of 4 technical replicate measurements was used for each individual biological sample, the average variation between the technical replicate measurements was 1.1% for all analytes and never exceeded 5% for an individual sample. Triplicate samples (independent cultures) were used to determine the variation in between cultures, average, and standard deviation between these replicates are depicted in figures.
Fluorescence Measurements-Chlorophyll fluorescence emission spectra at 77 K were recorded with a TECANXP-HP8000 using intact cells diluted to a concentration of 5 g of chlorophyll/ml in 30% glycerol, 5 mM Hepes, pH 7.5, and 10 mM EDTA and snap frozen in liquid nitrogen in a customized 96-well plate. The samples were excited at 435 nm (bandwidth 5 nm), and emission was monitored between 640 and 760 nm (bandwidth 5 nm). The gain was manually adjusted to 150, the Z-position was 22,000 m. The signals obtained were processed with the Adjacent-Average function (OriginLab) and normalized to the value obtained at a wavelength of 685 nm.
Plastoquinone-9 and Tocochromanol Analyses-Chlamydomonas cells were grown to a density of 4 -8 ϫ 10 6 cells ϫ ml Ϫ1 and 50 -100 ml were collected by centrifugation (5 min at 2450 ϫ g, 4°C). For normalization, 1 ml of each culture was used for chlorophyll determination according to Ref. 58. For plastoquinone-9 and tocochromanol analyses, frozen cell pellets were resuspended in 500 l of 95% (v/v) ethanol, spiked with 50 l of 121 M ubiquinone-10 (6.05 nmol) as an internal standard, and homogenized in a 5-ml Pyrex tissue grinder. The grinder was then rinsed with 500 l of 95% (v/v) ethanol, and the wash was combined to the initial extract. Insoluble material was removed by centrifugation (5 min at 18,000 ϫ g) and the supernatants were immediately analyzed by HPLC as previously described (60).